| Title: | Package for Community Ecology and Suitability Analysis |
|---|---|
| Description: | Graphical User Interface (via the R-Commander) and utility functions (often based on the vegan package) for statistical analysis of biodiversity and ecological communities, including species accumulation curves, diversity indices, Renyi profiles, GLMs for analysis of species abundance and presence-absence, distance matrices, Mantel tests, and cluster, constrained and unconstrained ordination analysis. A book on biodiversity and community ecology analysis is available for free download from the website. In 2012, methods for (ensemble) suitability modelling and mapping were expanded in the package. |
| Authors: | Roeland Kindt [cre, aut]
|
| Maintainer: | Roeland Kindt <[email protected]> |
| License: | GPL-3 |
| Version: | 2.16-1 |
| Built: | 2024-11-12 06:59:20 UTC |
| Source: | CRAN |
This package provides a GUI (Graphical User Interface, via the R-Commander; BiodiversityRGUI) and some utility functions (often based on the vegan package) for statistical analysis of biodiversity and ecological communities, including species accumulation curves, diversity indices, Renyi profiles, GLMs for analysis of species abundance and presence-absence, distance matrices, Mantel tests, and cluster, constrained and unconstrained ordination analysis. A book on biodiversity and community ecology analysis is available for free download from the website.
We warmly thank all that provided inputs that lead to improvement of the Tree Diversity Analysis manual that describes common methods for biodiversity and community ecology analysis and its accompanying software. We especially appreciate the comments received during training sessions with draft versions of this manual and the accompanying software in Kenya, Uganda and Mali. We are equally grateful to the thoughtful reviews by Dr Simoneta Negrete-Yankelevich (Instituto de Ecologia, Mexico) and Dr Robert Burn (Reading University, UK) of the draft version of this manual, and to Hillary Kipruto for help in editing of this manual. We also want to specifically thank Mikkel Grum, Jane Poole and Paulo van Breugel for helping in testing the packaged version of the software. We also want to give special thanks for all the support that was given by Jan Beniest, Tony Simons and Kris Vanhoutte in realizing the book and software.
We highly appreciate the support of the Programme for Cooperation with International Institutes (SII), Education and Development Division of the Netherlands Ministry of Foreign Affairs, and VVOB (The Flemish Association for Development Cooperation and Technical Assistance, Flanders, Belgium) for funding the development for this manual. We also thank VVOB for seconding Roeland Kindt to the World Agroforestry Centre (ICRAF). The tree diversity analysis manual was inspired by research, development and extension activities that were initiated by ICRAF on tree and landscape diversification. We want to acknowledge the various donor agencies that have funded these activities, especially VVOB, DFID, USAID and EU.
We are grateful for the developers of the R Software for providing a free and powerful statistical package that allowed development of BiodiversityR. We also want to give special thanks to Jari Oksanen for developing the vegan package and John Fox for developing the Rcmdr package, which are key packages that are used by BiodiversityR.
Maintainer: Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
We suggest to use this citation for this software as well (together with citations of all other packages that were used)
Provides alternative methods of obtaining species accumulation results than provided by functions specaccum and plot.specaccum (vegan).
accumresult(x, y="", factor="", level, scale="", method="exact", permutations=100, conditioned=T, gamma="boot", ...) accumplot(xr, addit=F, labels="", col=1, ci=2, pch=1, type="p", cex=1, xlim=c(1, xmax), ylim=c(1, rich), xlab="sites", ylab="species richness", cex.lab=1, cex.axis=1, ...) accumcomp(x, y="", factor, scale="", method="exact", permutations=100, conditioned=T, gamma="boot", plotit=T, labelit=T, legend=T, rainbow=T, xlim=c(1, max), ylim=c(0, rich),type="p", xlab="sites", ylab="species richness", cex.lab=1, cex.axis=1, ...)accumresult(x, y="", factor="", level, scale="", method="exact", permutations=100, conditioned=T, gamma="boot", ...) accumplot(xr, addit=F, labels="", col=1, ci=2, pch=1, type="p", cex=1, xlim=c(1, xmax), ylim=c(1, rich), xlab="sites", ylab="species richness", cex.lab=1, cex.axis=1, ...) accumcomp(x, y="", factor, scale="", method="exact", permutations=100, conditioned=T, gamma="boot", plotit=T, labelit=T, legend=T, rainbow=T, xlim=c(1, max), ylim=c(0, rich),type="p", xlab="sites", ylab="species richness", cex.lab=1, cex.axis=1, ...)
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
y |
Environmental data frame. |
factor |
Variable of the environmental data frame that defines subsets to calculate species accumulation curves for. |
level |
Level of the variable to create the subset to calculate species accumulation curves. |
scale |
Continuous variable of the environmental data frame that defines the variable that scales the horizontal axis of the species accumulation curves. |
method |
Method of calculating the species accumulation curve (as in function |
permutations |
Number of permutations to calculate the species accumulation curve (as in function |
conditioned |
Estimation of standard deviation is conditional on the empirical dataset for the exact SAC (as in function |
gamma |
Method for estimating the total extrapolated number of species in the survey area (as in |
addit |
Add species accumulation curve to an existing graph. |
xr |
Result from |
col |
Colour for drawing lines of the species accumulation curve (as in function |
labels |
Labels to plot at left and right of the species accumulation curves. |
ci |
Multiplier used to get confidence intervals from standard deviatione (as in function |
pch |
Symbol used for drawing the species accumulation curve (as in function |
type |
Type of plot (as in function |
cex |
Character expansion factor (as in function |
xlim |
Limits for the X = horizontal axis. |
ylim |
Limits for the Y = vertical axis. |
xlab |
Label for the X = horizontal axis (as in function |
ylab |
Label for the Y = vertical axis (as in function |
cex.lab |
The magnification to be used for X and Y labels relative to the current setting of |
cex.axis |
The magnification to be used for axis annotation relative to the current setting of |
plotit |
Plot the results. |
labelit |
Label the species accumulation curves with the levels of the categorical variable. |
legend |
Add the legend (you need to click in the graph where the legend needs to be plotted). |
rainbow |
Use rainbow colouring for the different curves. |
... |
Other items passed to function |
These functions provide some alternative methods of obtaining species accumulation results, although function specaccum is called by these functions to calculate the actual species accumulation curve.
Functions accumresult and accumcomp allow to calculate species accumulation curves for subsets of the community and environmental data sets. Function accumresult calculates the species accumulation curve for the specified level of a selected environmental variable. Method accumcomp calculates the species accumulation curve for all levels of a selected environmental variable separatedly. Both methods allow to scale the horizontal axis by multiples of the average of a selected continuous variable from the environmental dataset (hint: add the abundance of each site to the environmental data frame to scale accumulation results by mean abundance).
Functions accumcomp and accumplot provide alternative methods of plotting species accumulation curve results, although function plot.specaccum is called by these functions. When you choose to add a legend, make sure that you click in the graph on the spot where you want to put the legend.
The functions provide alternative methods of obtaining species accumulation curve results, although results are similar as obtained by functions specaccum and plot.specaccum.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
https://rpubs.com/Roeland-KINDT
library(vegan) data(dune.env) data(dune) dune.env$site.totals <- apply(dune,1,sum) Accum.1 <- accumresult(dune, y=dune.env, scale='site.totals', method='exact', conditioned=TRUE) Accum.1 accumplot(Accum.1) Accum.2 <- accumcomp(dune, y=dune.env, factor='Management', method='exact', legend=FALSE, conditioned=TRUE, scale='site.totals') ## CLICK IN THE GRAPH TO INDICATE WHERE THE LEGEND NEEDS TO BE PLACED FOR ## OPTION WHERE LEGEND=TRUE (DEFAULT). ## Not run: # ggplot2 plotting method data(warcom) data(warenv) Accum.3 <- accumcomp(warcom, y=warenv, factor='population', method='exact', conditioned=F, plotit=F) library(ggplot2) # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) accum.long3 <- accumcomp.long(Accum.3, ci=NA, label.freq=5) plotgg1 <- ggplot(data=accum.long3, aes(x = Sites, y = Richness, ymax = UPR, ymin= LWR)) + scale_x_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(aes(colour=Grouping), size=2) + geom_point(data=subset(accum.long3, labelit==TRUE), aes(colour=Grouping, shape=Grouping), size=5) + geom_ribbon(aes(colour=Grouping), alpha=0.2, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x = "Trees", y = "Loci", colour = "Population", shape = "Population") plotgg1 ## End(Not run) # dontrunlibrary(vegan) data(dune.env) data(dune) dune.env$site.totals <- apply(dune,1,sum) Accum.1 <- accumresult(dune, y=dune.env, scale='site.totals', method='exact', conditioned=TRUE) Accum.1 accumplot(Accum.1) Accum.2 <- accumcomp(dune, y=dune.env, factor='Management', method='exact', legend=FALSE, conditioned=TRUE, scale='site.totals') ## CLICK IN THE GRAPH TO INDICATE WHERE THE LEGEND NEEDS TO BE PLACED FOR ## OPTION WHERE LEGEND=TRUE (DEFAULT). ## Not run: # ggplot2 plotting method data(warcom) data(warenv) Accum.3 <- accumcomp(warcom, y=warenv, factor='population', method='exact', conditioned=F, plotit=F) library(ggplot2) # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) accum.long3 <- accumcomp.long(Accum.3, ci=NA, label.freq=5) plotgg1 <- ggplot(data=accum.long3, aes(x = Sites, y = Richness, ymax = UPR, ymin= LWR)) + scale_x_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(aes(colour=Grouping), size=2) + geom_point(data=subset(accum.long3, labelit==TRUE), aes(colour=Grouping, shape=Grouping), size=5) + geom_ribbon(aes(colour=Grouping), alpha=0.2, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x = "Trees", y = "Loci", colour = "Population", shape = "Population") plotgg1 ## End(Not run) # dontrun
Calculates scores (coordinates) to plot species for PCoA or NMS results that do not naturally provide species scores. The function can also rescale PCA results to use the choice of rescaling used in vegan for the rda function (after calculating PCA results via PCoA with the euclidean distance first).
add.spec.scores(ordi, comm, method="cor.scores", multi=1, Rscale=F, scaling="1")add.spec.scores(ordi, comm, method="cor.scores", multi=1, Rscale=F, scaling="1")
ordi |
Ordination result as calculated by |
comm |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
method |
Method for calculating species scores. Method "cor.scores" calculates the scores by the correlation between site scores and species vectors (via function |
multi |
Multiplier for the species scores. |
Rscale |
Use the same scaling method used by vegan for |
scaling |
Scaling method as used by |
The function returns a new ordination result with new information on species scores. For PCoA results, the function calculates eigenvalues (not sums-of-squares as provided in results from function cmdscale), the percentage of explained variance per axis and the sum of all eigenvalues. PCA results (obtained by PCoA obtained by function cmdscale with the Euclidean distance) can be scaled as in function rda, or be left at the original scale.
Roeland Kindt
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune) distmatrix <- vegdist(dune, method="euc") # Principal coordinates analysis with 19 axes to estimate total variance Ordination.model1 <- cmdscale (distmatrix, k=19, eig=TRUE, add=FALSE) # Change scores for second axis Ordination.model1$points[,2] <- -1.0 * Ordination.model1$points[,2] Ordination.model1 <- add.spec.scores(Ordination.model1, dune, method='pcoa.scores', Rscale=TRUE, scaling=1, multi=1) # Compare Ordination.model1 with PCA Ordination.model2 <- rda(dune, scale=FALSE) # par(mfrow=c(1,2)) ordiplot(Ordination.model1, type="text") abline(h = 0, lty = 3) abline(v = 0, lty = 3) plot(Ordination.model2, type="text", scaling=1)library(vegan) data(dune) distmatrix <- vegdist(dune, method="euc") # Principal coordinates analysis with 19 axes to estimate total variance Ordination.model1 <- cmdscale (distmatrix, k=19, eig=TRUE, add=FALSE) # Change scores for second axis Ordination.model1$points[,2] <- -1.0 * Ordination.model1$points[,2] Ordination.model1 <- add.spec.scores(Ordination.model1, dune, method='pcoa.scores', Rscale=TRUE, scaling=1, multi=1) # Compare Ordination.model1 with PCA Ordination.model2 <- rda(dune, scale=FALSE) # par(mfrow=c(1,2)) ordiplot(Ordination.model1, type="text") abline(h = 0, lty = 3) abline(v = 0, lty = 3) plot(Ordination.model2, type="text", scaling=1)
Provides species accumulation results calculated from balanced (equal subsample sizes) subsampling from each stratum. Sites can be accumulated in a randomized way, or alternatively sites belonging to the same stratum can be kept together Results are in the same format as specaccum and can be plotted with plot.specaccum (vegan).
balanced.specaccum(comm, permutations=100, strata=strata, grouped=TRUE, reps=0, scale=NULL)balanced.specaccum(comm, permutations=100, strata=strata, grouped=TRUE, reps=0, scale=NULL)
comm |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
permutations |
Number of permutations to calculate the species accumulation curve. |
strata |
Categorical variable used to specify strata. |
grouped |
Should sites from the same stratum be kept together (TRUE) or not. |
reps |
Number of subsamples to be taken from each stratum (see details). |
scale |
Quantitative variable used to scale the sampling effort (see details). |
This function provides an alternative method of obtaining species accumulation results as provided by specaccum and accumresult.
Balanced sampling is achieved by randomly selecting the same number of sites from each stratum. The number of sites selected from each stratum is determined by reps. Sites are selected from strata with sample sizes larger or equal than reps. In case that reps is smaller than 1 (default: 0), then the number of sites selected from each stratum is equal to the smallest sample size of all strata. Sites from the same stratum can be kept together (grouped=TRUE) or the order of sites can be randomized (grouped=FALSE).
The results can be scaled by the average accumulation of a quantitative variable (default is number of sites), as in accumresult (hint: add the abundance of each site to the environmental data frame to scale accumulation results by mean abundance). When sites are not selected from all strata, then the average is calculated only for the strata that provided sites.
The functions provide alternative methods of obtaining species accumulation curve results, although results are similar as obtained by functions specaccum and accumresult.
Roeland Kindt (World Agroforestry Centre)
Kindt, R., Kalinganire, A., Larwanou, M., Belem, M., Dakouo, J.M., Bayala, J. & Kaire, M. (2008) Species accumulation within landuse and tree diameter categories in Burkina Faso, Mali, Niger and Senegal. Biodiversity and Conservation. 17: 1883-1905.
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune.env) data(dune) # not balancing species accumulation Accum.orig <- specaccum(dune) Accum.orig # randomly sample 3 quadrats from each stratum of Management Accum.1 <- balanced.specaccum(dune, strata=dune.env$Management, reps=3) Accum.1 # scale results by number of trees per quadrat dune.env$site.totals <- apply(dune,1,sum) Accum.2 <- balanced.specaccum(dune, strata=dune.env$Management, reps=3, scale=dune.env$site.totals) Accum.2library(vegan) data(dune.env) data(dune) # not balancing species accumulation Accum.orig <- specaccum(dune) Accum.orig # randomly sample 3 quadrats from each stratum of Management Accum.1 <- balanced.specaccum(dune, strata=dune.env$Management, reps=3) Accum.1 # scale results by number of trees per quadrat dune.env$site.totals <- apply(dune,1,sum) Accum.2 <- balanced.specaccum(dune, strata=dune.env$Management, reps=3, scale=dune.env$site.totals) Accum.2
Topography-derived variables and UTM coordinates and UTM coordinates of a 50 ha sample plot (consisting of 50 1-ha quadrats) from Barro Colorado Island of Panama. Dataset BCI provides the tree species composition (trees with diameter at breast height equal or larger than 10 cm) of the same plots.
data(BCI.env)data(BCI.env)
A data frame with 50 observations on the following 6 variables.
UTM.EWUTM easting
UTM.NSUTM northing
elevationmean of the elevation values of the four cell corners
convexmean elevation of the target cell minus the mean elevation of the eight surrounding cells
slopemean angular deviation from horizontal of each of the four triangular planes formed by connecting three of its corners
aspectEWthe sine of aspect
aspectNSthe cosine of aspect
Pyke C.R., Condit R., Aguilar S. and Lao S. (2001). Floristic composition across a climatic gradient in a neotropical lowland forest. Journal of Vegetation Science 12: 553-566.
Condit R., Pitman N., Leigh E.G., Chave J., Terborgh J., Foster R.B., Nunez P., Aguilar S., Valencia R., Villa G., Muller-Landau H.C., Losos E. and Hubbell, S.P. (2002). Beta-diversity in tropical forest trees. Science 295: 666-669.
De Caceres M., P. Legendre, R. Valencia, M. Cao, L.-W. Chang, G. Chuyong, R. Condit, Z. Hao, C.-F. Hsieh, S. Hubbell, D. Kenfack, K. Ma, X. Mi, N. Supardi Noor, A. R. Kassim, H. Ren, S.-H. Su, I-F. Sun, D. Thomas, W. Ye and F. He. (2012). The variation of tree beta diversity across a global network of forest plots. Global Ecology and Biogeography 21: 1191-1202
data(BCI.env)data(BCI.env)
ChangeLog file
BiodiversityR.changeLog()BiodiversityR.changeLog()
This function provides a GUI (Graphical User Interface) for some of the functions of vegan, some other packages and some new functions to run biodiversity analysis, including species accumulation curves, diversity indices, Renyi profiles, rank-abundance curves, GLMs for analysis of species abundance and presence-absence, distance matrices, Mantel tests, cluster and ordination analysis (including constrained ordination methods such as RDA, CCA, db-RDA and CAP). In 2012 methods for ensemble suitability The function depends and builds on Rcmdr, performing all analyses on the community and environmental datasets that the user selects. A thorough description of the package and the biodiversity and ecological methods that it accomodates (including examples) is provided in the freely available Tree Diversity Analysis manual (Kindt and Coe, 2005) that is accessible via the help menu.
BiodiversityRGUI(changeLog = FALSE, backward.compatibility.messages = FALSE)BiodiversityRGUI(changeLog = FALSE, backward.compatibility.messages = FALSE)
changeLog |
Show the changeLog file |
backward.compatibility.messages |
Some notes on backward compatiblity |
The function launches the R-Commander GUI with an extra menu for common statistical methods for biodiversity and community ecology analysis as described in the Tree Diversity Analysis manual of Roeland Kindt and Richard Coe (available via https://www.worldagroforestry.org/output/tree-diversity-analysis]) and expanded systematically with new functions that became available from the vegan community ecology package.
Since 2012, functions for ensemble suitability modelling were included in BiodiversityR. In 2016, a GUI was created for ensemble suitabilty modelling.
The R-Commander is launched by changing the location of the Rcmdr "etc" folder to the "etc" folder of BiodiversityR. As the files of the "etc" folder of BiodiversityR are copied from the Rcmdr, it is possible that newest versions of the R-Commander will not be launched properly. In such situations, it is possible that copying all files from the Rcmdr "etc" folder again and adding the BiodiversityR menu options to the Rcmdr-menus.txt is all that is needed to launch the R-Commander again. However, please alert Roeland Kindt about the issue.
BiodiversityR uses two data sets for biodiversity and community ecology analysis: the community dataset (or community matrix or species matrix) and the environmental dataset (or environmental matrix). The environmental dataset is the same dataset that is used as the "active dataset" of The R-Commander. (Note that you could sometimes use the same dataset as both the community and environmental dataset. For example, you could use the community dataset as environmental dataset as well to add information about specific species to ordination diagrams. As another example, you could use the environmental dataset as community dataset if you first calculated species richness of each site, saved this information in the environmental dataset, and then use species richness as response variable in a regression analysis.) Some options of analysis of ecological distance allow the community matrix to be a distance matrix (the community data set will be interpreted as distance matrix via as.dist prior to further analysis).
For ensemble suitability modelling, different data sets should be created and declared such as the calibration stack, the presence data set and the absence data set. The ensemble suitability modelling menu gives some guidelines on getting started with ensemble suitability modelling.
Besides launching the graphical user interface, the function gives some notes on backward compatibility.
Roeland Kindt (with some help from Jari Oksanen)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
This function provides a method for CAP that follows the procedure as described by the authors of the ordination method (Anderson & Willis 2003). The CAP method implemented in vegan through capscale conforms more to distance-based Redundancy Analysis (Legendre & Anderson, 1999) than to the original description for CAP (Anderson & Willis, 2003 ).
CAPdiscrim(formula, data, dist="bray", axes=4, m=0, mmax=10, add=FALSE, permutations=0, aitchison_pseudocount=1)CAPdiscrim(formula, data, dist="bray", axes=4, m=0, mmax=10, add=FALSE, permutations=0, aitchison_pseudocount=1)
formula |
Formula with a community data frame (with sites as rows, species as columns and species abundance as cell values) or distance matrix on the left-hand side and a categorical variable on the right-hand side (only the first explanatory variable will be used). |
data |
Environmental data set. |
dist |
Method for calculating ecological distance with function |
axes |
Number of PCoA axes ( |
m |
Number of PCoA axes to be investigated by discriminant analysis ( |
mmax |
The maximum number of PCoA axes considered when searching (m=0) for the number of axes that provide the best classification success. |
add |
Add a constant to the non-diagonal dissimilarities such that the modified dissimilarities are Euclidean; see also |
permutations |
The number of permutations for significance testing. |
aitchison_pseudocount |
Pseudocount setting as in |
This function provides a method of Constrained Analysis of Principal Coordinates (CAP) that follows the description of the method by the developers of the method, Anderson and Willis. The method investigates the results of a Principal Coordinates Analysis (function cmdscale) with linear discriminant analysis (lda). Anderson and Willis advocate to use the number of principal coordinate axes that result in the best prediction of group identities of the sites.
Results may be different than those obtained in the PRIMER-e package because PRIMER-e does not consider prior probabilities, does not standardize PCOA axes by their eigenvalues and applies an additional spherical standardization to a common within-group variance/covariance matrix.
For permutations > 0, the analysis is repeated by randomising the observations of the environmental data set. The significance is estimated by dividing the number of times the randomisation generated a larger percentage of correct predictions.
The function returns an object with information on CAP based on discriminant analysis. The object contains following elements:
PCoA |
the positions of the sites as fitted by PCoA |
m |
the number of axes analysed by discriminant analysis |
tot |
the total variance (sum of all eigenvalues of PCoA) |
varm |
the variance of the m axes that were investigated |
group |
the original group of the sites |
CV |
the predicted group for the sites by discriminant analysis |
percent |
the percentage of correct predictions |
percent.level |
the percentage of correct predictions for different factor levels |
x |
the positions of the sites provided by the discriminant analysis |
F |
the squares of the singulare values of the discriminant analysis |
manova |
the results for MANOVA with the same grouping variable |
signi |
the significance of the percentage of correct predictions |
manova |
a summary of the observed randomised prediction percentages |
The object can be plotted with ordiplot, and species scores can be added by add.spec.scores .
Roeland Kindt (World Agroforestry Centre)
Legendre, P. & Anderson, M.J. (1999). Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecological Monographs 69: 1-24.
Anderson, M.J. & Willis, T.J. (2003). Canonical analysis of principal coordinates: a useful method of constrained ordination for ecology. Ecology 84: 511-525.
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
## Not run: library(vegan) library(MASS) data(dune) data(dune.env) # categorical variables should not be ordered dune$Management <- factor(dune$Management, ordered=FALSE) Ordination.model1 <- CAPdiscrim(dune~Management, data=dune.env, dist="bray", axes=2, m=0, add=FALSE) Ordination.model1 plot1 <- ordiplot(Ordination.model1, type="none") ordisymbol(plot1, dune.env, "Management", legend=TRUE) # plot change in classification success against m plot(seq(1:14), rep(-1000, 14), xlim=c(1, 14), ylim=c(0, 100), xlab="m", ylab="classification success (percent)", type="n") for (mseq in 1:14) { CAPdiscrim.result <- CAPdiscrim(dune~Management, data=dune.env, dist="bray", axes=2, m=mseq) points(mseq, CAPdiscrim.result$percent) } ## End(Not run)## Not run: library(vegan) library(MASS) data(dune) data(dune.env) # categorical variables should not be ordered dune$Management <- factor(dune$Management, ordered=FALSE) Ordination.model1 <- CAPdiscrim(dune~Management, data=dune.env, dist="bray", axes=2, m=0, add=FALSE) Ordination.model1 plot1 <- ordiplot(Ordination.model1, type="none") ordisymbol(plot1, dune.env, "Management", legend=TRUE) # plot change in classification success against m plot(seq(1:14), rep(-1000, 14), xlim=c(1, 14), ylim=c(0, 100), xlab="m", ylab="classification success (percent)", type="n") for (mseq in 1:14) { CAPdiscrim.result <- CAPdiscrim(dune~Management, data=dune.env, dist="bray", axes=2, m=mseq) points(mseq, CAPdiscrim.result$percent) } ## End(Not run)
This is a simple function that rescales the ordination coordinates obtained from the distance-based redundancy analysis method implemented in vegan through capscale. The rescaling of the ordination coordinates results in the distances between fitted site scores in ordination results (scaling=1 obtained via ordiplot to be equal to the distances between sites on the axes corresponding to positive eigenvalues obtained from principal coordinates analysis (cmdscale).
caprescale(x,verbose=FALSE)caprescale(x,verbose=FALSE)
x |
Ordination result obtained with |
verbose |
Give some information on the pairwise distances among sites (TRUE) or not. |
The first step of distance-based redundancy analysis involves principal coordinates analysis whereby the distances among sites from a distance matrix are approximated by distances among sites in a multidimensional configuration (ordination). In case that the principal coordinates analysis does not result in negative eigenvalues, then the distances from the distance matrix are the same as the distances among the sites in the ordination. In case that the principal coordinates analysis results in negative eigenvalues, then the distances among the sites on all ordination axes are related to the sum of positive eigenvalues, a sum which is larger than the sum of squared distances of the distance matrix.
The distance-based redundancy analysis method implemented in vegan through capscale uses a specific rescaling method for ordination results. Function caprescale modifies the results of capscale so that an ordination with scaling=1 (a distance biplot) obtained viaordiplot preserves the distances reflected in the principal coordinates analysis implemented as the first step of the analysis. See Legendre and Legendre (1998) about the relationship between fitted site scores and eigenvalues.
The function modifies and returns an object obtained via capscale.
Roeland Kindt (World Agroforestry Centre)
Legendre, P. & Legendre, L. (1998). Numerical Ecology. Amsterdam: Elsevier. 853 pp.
Legendre, P. & Anderson, M.J. (1999). Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecological Monographs 69: 1-24.
library(vegan) library(MASS) data(dune) data(dune.env) Distmatrix.1 <- vegdist(dune,method='bray') Ordination.model1 <- cmdscale(Distmatrix.1, k=19, eig=TRUE, add=FALSE) # Sum of all eigenvalues sum(Ordination.model1$eig) # [1] 4.395807541512926 sum(Ordination.model1$eig[1:14]) # [1] 4.593946896588808 Distmatrix.2 <- as.matrix(vegdist(Ordination.model1$points[,1:14],method='euc')) totalsumsquares1 <- sum(Distmatrix.2^2)/(2*20) # Sum of distances among sites in principal coordinates analysis on axes # corresponding to positive eigenvalues totalsumsquares1 # [1] 4.593946896588808 Ordination.model2 <- capscale(dune ~ Management,dune.env,dist='bray', add=FALSE) # Total sums of positive eigenvalues of the distance-based redundancy analysis Ordination.model2$CA$tot.chi+Ordination.model2$CCA$tot.chi # [1] 4.593946896588808 Ordination.model3 <- caprescale(Ordination.model2, verbose=TRUE) sum1 <- summary(Ordination.model3,axes=17,scaling=1)$constraints Distmatrix.3 <- as.matrix(vegdist(sum1 ,method='euc')) totalsumsquares2 <- sum((Distmatrix.3)^2)/(2*20)/19 totalsumsquares2 # [1] 4.593946896588808library(vegan) library(MASS) data(dune) data(dune.env) Distmatrix.1 <- vegdist(dune,method='bray') Ordination.model1 <- cmdscale(Distmatrix.1, k=19, eig=TRUE, add=FALSE) # Sum of all eigenvalues sum(Ordination.model1$eig) # [1] 4.395807541512926 sum(Ordination.model1$eig[1:14]) # [1] 4.593946896588808 Distmatrix.2 <- as.matrix(vegdist(Ordination.model1$points[,1:14],method='euc')) totalsumsquares1 <- sum(Distmatrix.2^2)/(2*20) # Sum of distances among sites in principal coordinates analysis on axes # corresponding to positive eigenvalues totalsumsquares1 # [1] 4.593946896588808 Ordination.model2 <- capscale(dune ~ Management,dune.env,dist='bray', add=FALSE) # Total sums of positive eigenvalues of the distance-based redundancy analysis Ordination.model2$CA$tot.chi+Ordination.model2$CCA$tot.chi # [1] 4.593946896588808 Ordination.model3 <- caprescale(Ordination.model2, verbose=TRUE) sum1 <- summary(Ordination.model3,axes=17,scaling=1)$constraints Distmatrix.3 <- as.matrix(vegdist(sum1 ,method='euc')) totalsumsquares2 <- sum((Distmatrix.3)^2)/(2*20)/19 totalsumsquares2 # [1] 4.593946896588808
This function makes a cross-tabulation of two variables after transforming the first variable to presence-absence and then returns results of chisq.test.
crosstabanalysis(x,variable,factor)crosstabanalysis(x,variable,factor)
x |
Data set that contains the variables "variable" and "factor". |
variable |
Variable to be transformed in presence-absence in the resulting cross-tabulation. |
factor |
Variable to be used for the cross-tabulation together with the transformed variable. |
The function returns the results of chisq.test on a crosstabulation of two variables, after transforming the first variable to presence-absence first.
Roeland Kindt
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune.env) crosstabanalysis(dune.env,"Manure","Management")library(vegan) data(dune.env) crosstabanalysis(dune.env,"Manure","Management")
This data set provides WorldClim 2.1 bioclimatic data extracted for the baseline (WorldClim 2.1 at 2.5 minutes resolution) and one future (wc2.1_2.5m_bioc_EC-Earth3-Veg_ssp245_2041-2060) climate for presence observations of Cucurbita cordata, C. digitata and C. palmata provided via the CucurbitaData data set of the GapAnalysis package. This data set is used for examples of the ensemble.concave.hull and ensemble.concave.venn functions.
data(CucurbitaClim)data(CucurbitaClim)
Fick SE and Hijmans RJ. 2017. WorldClim 2: new 1km spatial resolution climate surfaces for global land areas. International Journal of Climatology 37: 4302-4315.
Carver et al. 2021. GapAnalysis: an R package to calculate conservation indicators using spatial information. Ecography 44: 1000-1009.
data(CucurbitaClim)data(CucurbitaClim)
This function calculates the percentage of deviance explained by a GLM model and calculates the significance of the model.
deviancepercentage(x,data,test="F",digits=2)deviancepercentage(x,data,test="F",digits=2)
x |
|
data |
Data set to be used for the null model (preferably the same data set used by the 'full' model). |
test |
Test statistic to be used for the comparison between the null model and the 'full' model as estimated by |
digits |
Number of digits in the calculation of the percentage. |
The function calculates the percentage of explained deviance and the significance of the 'full' model by contrasting it with the null model.
For the null model, the data is subjected to na.omit. You should check whether the same data are used for the null and 'full' models.
The function calculates the percentage of explained deviance and the significance of the 'full' model by contrasting it with the null model by ANOVA. The results of the ANOVA are also provided.
Roeland Kindt
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune) data(dune.env) dune.env$Agrostol <- dune$Agrostol Count.model1 <- glm(Agrostol ~ Management + A1, family=quasipoisson(link=log), data=dune.env, na.action=na.omit) summary(Count.model1) deviancepercentage(Count.model1, dune.env, digits=3)library(vegan) data(dune) data(dune.env) dune.env$Agrostol <- dune$Agrostol Count.model1 <- glm(Agrostol ~ Management + A1, family=quasipoisson(link=log), data=dune.env, na.action=na.omit) summary(Count.model1) deviancepercentage(Count.model1, dune.env, digits=3)
Function dist.eval provides one test of a distance matrix, and then continues with distconnected (vegan). Function prepare.bioenv converts selected variables to numeric variables and then excludes all categorical variables in preparation of applying bioenv (vegan).
dist.eval(x, dist) prepare.bioenv(env, as.numeric = c())dist.eval(x, dist) prepare.bioenv(env, as.numeric = c())
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
env |
Environmental data frame with sites as rows and variables as columns. |
dist |
Method for calculating ecological distance with function |
as.numeric |
Vector with names of variables in the environmental data set to be converted to numeric variables. |
Function dist.eval provides two tests of a distance matrix:
(i) The first test checks whether any pair of sites that share some species have a larger distance than any other pair of sites that do not share any species. In case that cases are found, then a warning message is given.
(ii) The second test is the one implemented by the distconnected function (vegan). The distconnected test is only calculated for distances that calculate a value of 1 if sites share no species (i.e. not manhattan or euclidean), using the threshold of 1 as an indication that the sites do not share any species. Interpretation of analysis of distance matrices that provided these warnings should be cautious.
Function prepare.bioenv provides some simple methods of dealing with categorical variables prior to applying bioenv.
The function tests whether distance matrices have some desirable properties and provide warnings if this is not the case.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune) dist.eval(dune,"euclidean") dist.eval(dune,"bray") ## Not run: data(dune.env) dune.env2 <- dune.env[,c('A1', 'Moisture', 'Manure')] dune.env2$Moisture <- as.numeric(dune.env2$Moisture) dune.env2$Manure <- as.numeric(dune.env2$Manure) sol <- bioenv(dune ~ A1 + Moisture + Manure, dune.env2) sol summary(sol) dune.env3 <- prepare.bioenv(dune.env, as.numeric=c('Moisture', 'Manure')) bioenv(dune, dune.env3) ## End(Not run)library(vegan) data(dune) dist.eval(dune,"euclidean") dist.eval(dune,"bray") ## Not run: data(dune.env) dune.env2 <- dune.env[,c('A1', 'Moisture', 'Manure')] dune.env2$Moisture <- as.numeric(dune.env2$Moisture) dune.env2$Manure <- as.numeric(dune.env2$Manure) sol <- bioenv(dune ~ A1 + Moisture + Manure, dune.env2) sol summary(sol) dune.env3 <- prepare.bioenv(dune.env, as.numeric=c('Moisture', 'Manure')) bioenv(dune, dune.env3) ## End(Not run)
Sample units without any species result in "NaN" values in the distance matrix for some of the methods of vegdist (vegan). The function replaces "NA" by "0" if both sample units do not contain any species and "NA" by "1" if only one sample unit does not have any species.
dist.zeroes(comm, dist)dist.zeroes(comm, dist)
comm |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
dist |
Distance matrix as calculated with function |
This functions changes a distance matrix by replacing "NaN" values by "0" if both sample units do not contain any species and by "1" if only one sample unit does not contain any species.
Please note that there is a valid reason (deliberate removal of zero abundance values from calculations) that the original distance matrix contains "NaN", so you may not wish to do this transformation and remove sample units with zero abundances from further analysis.
The function provides a new distance matrix where "NaN" values have been replaced by "0" or "1".
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) matrix <- array(0, dim=c(5,3)) matrix[4,] <- c(1, 2, 3) matrix[5,] <- c(1, 0, 0) dist1 <- vegdist(matrix, method="kulc") dist1 dist2 <- dist.zeroes(matrix, dist1) dist2library(vegan) matrix <- array(0, dim=c(5,3)) matrix[4,] <- c(1, 2, 3) matrix[5,] <- c(1, 0, 0) dist1 <- vegdist(matrix, method="kulc") dist1 dist2 <- dist.zeroes(matrix, dist1) dist2
This function compares the distance among sites as displayed in an ordination diagram (generated by ordiplot) with the actual distances among sites as available from a distance matrix (as generated by vegdist).
distdisplayed(x, ordiplot, distx = "bray", plotit = T, addit = F, method = "spearman", permutations = 100, abline = F, gam = T, ...)distdisplayed(x, ordiplot, distx = "bray", plotit = T, addit = F, method = "spearman", permutations = 100, abline = F, gam = T, ...)
x |
Community data frame (with sites as rows, species as columns and species abundance as cell values) or distance matrix. |
ordiplot |
Ordination diagram generated by |
distx |
Ecological distance used to calculated the distance matrix (theoretically the same distance as displayed in the ordination diagram); passed to |
plotit |
Should a plot comparing the distance in ordination diagram (or the distance matrix) with the distance from the distance matrix be generated (or not). |
addit |
Should the GAM regression result be added to an existing plot (or not). |
method |
Method for calculating the correlation between the ordination distance and the complete distance; from function |
permutations |
Number of permutations to assess the significance of the Mantel test; passed to |
abline |
Should a reference line (y=x) be added to the graph (or not). |
gam |
Evaluate the correspondence between the original distance and the distance from the ordination diagram with GAMas estimated by |
... |
Other arguments passed to |
This function compares the Euclidean distances (between sites) displayed in an ordination diagram with the distances of a distance matrix. Alternatively, the distances of one distance matrix are compared against the distances of another distance matrix.
These distances are compared by a Mantel test (mantel) and (optionally) a GAM regression (gam). Optionally, a graph is provided compairing the distances and adding GAM results.
.
The function returns the results of a Mantel test and (optionally) the results of a GAM analysis.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) library(mgcv) data(dune) # pseudocount used by aitchison distance distmatrix <- vegdist(dune, method="kulc", pseudocount=1) ordination.model1 <- cmdscale(distmatrix,k=2) ordiplot1 <- ordiplot(ordination.model1) distdisplayed(dune, ordiplot=ordiplot1, distx="kulc", plotit=TRUE, method="spearman", permutations=100, gam=TRUE)library(vegan) library(mgcv) data(dune) # pseudocount used by aitchison distance distmatrix <- vegdist(dune, method="kulc", pseudocount=1) ordination.model1 <- cmdscale(distmatrix,k=2) ordiplot1 <- ordiplot(ordination.model1) distdisplayed(dune, ordiplot=ordiplot1, distx="kulc", plotit=TRUE, method="spearman", permutations=100, gam=TRUE)
Transforms a community matrix. Some transformation methods are described by distances for the original community matrix that result in the same distance matrix as calculated with the euclidean distance from the transformed community matrix.
In several cases (methods of "hellinger", "chord", "profiles" and "chi.square), the method makes use of function decostand. In several other cases ("Braun.Blanquet", "Domin", "Hult", "Hill", "fix" and "coverscale.log"), the method makes use of function coverscale. For method "dispweight" a call is made to function dispweight.
disttransform(x, method="hellinger")disttransform(x, method="hellinger")
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
method |
Distance measure for the original community matrix that the euclidean distance will calculate for the transformed community matrix: partial match to "hellinger", "chord", "profiles", "chi.square", "log", "square", "pa", "Braun.Blanquet", "Domin", "Hult", "Hill", "fix", "coverscale.log" and "dispweight". |
This functions transforms a community matrix.
Some transformation methods ("hellinger", "chord", "profiles" and "chi.square") have the behaviour that the euclidean distance from the transformed matrix will equal a distance of choice for the original matrix. For example, using method "hellinger" and calculating the euclidean distance will result in the same distance matrix as by calculating the Hellinger distance from the original community matrix.
Transformation methods ("Braun.Blanquet", "Domin", "Hult", "Hill", "fix" and "coverscale.log") call function coverscale.
Method "dispweight" uses function dispweight without specifying a grouping structure.
The function returns a transformed community matrix.
Roeland Kindt (World Agroforestry Centre)
Legendre, P. & Gallagher, E.D. (2001). Ecologically meaningful transformations for ordination of species data. Oecologia 129: 271-280.
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
## Not run: library(vegan) data(dune) Community.1 <- disttransform(dune, method='hellinger') Distmatrix.1 <- vegdist(Community.1, method='euclidean') Distmatrix.1 ## End(Not run)## Not run: library(vegan) data(dune) Community.1 <- disttransform(dune, method='hellinger') Distmatrix.1 <- vegdist(Community.1, method='euclidean') Distmatrix.1 ## End(Not run)
Provides alternative methods of obtaining results on diversity statistics than provided directly by functions diversity, fisher.alpha, specpool and specnumber (all from vegan), although these same functions are called. Some other statistics are also calculated such as the reciprocal Berger-Parker diversity index and abundance (not a diversity statistic). The statistics can be calculated for the entire community, for each site separately, the mean of the sites can be calculated or a jackknife estimate can be calculated for the community.
diversityresult(x, y = NULL, factor = NULL, level = NULL, index=c("Shannon", "Simpson", "inverseSimpson", "Logalpha", "Berger", "simpson.unb", "simpson.unb.inverse", "richness", "abundance", "Jevenness", "Eevenness", "jack1", "jack2", "chao", "boot"), method=c("pooled", "each site", "mean", "sd", "max", "jackknife"), sortit = FALSE, digits = 8) diversityvariables(x, y, digits=8) diversitycomp(x, y = NULL, factor1 = NULL ,factor2 = NULL, index=c("Shannon", "Simpson", "inverseSimpson", "Logalpha", "Berger", "simpson.unb", "simpson.unb.inverse", "richness", "abundance", "Jevenness", "Eevenness", "jack1", "jack2", "chao", "boot"), method=c("pooled", "mean", "sd", "max", "jackknife"), sortit=FALSE, digits=8)diversityresult(x, y = NULL, factor = NULL, level = NULL, index=c("Shannon", "Simpson", "inverseSimpson", "Logalpha", "Berger", "simpson.unb", "simpson.unb.inverse", "richness", "abundance", "Jevenness", "Eevenness", "jack1", "jack2", "chao", "boot"), method=c("pooled", "each site", "mean", "sd", "max", "jackknife"), sortit = FALSE, digits = 8) diversityvariables(x, y, digits=8) diversitycomp(x, y = NULL, factor1 = NULL ,factor2 = NULL, index=c("Shannon", "Simpson", "inverseSimpson", "Logalpha", "Berger", "simpson.unb", "simpson.unb.inverse", "richness", "abundance", "Jevenness", "Eevenness", "jack1", "jack2", "chao", "boot"), method=c("pooled", "mean", "sd", "max", "jackknife"), sortit=FALSE, digits=8)
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
y |
Environmental data frame. |
factor |
Variable of the environmental data frame that defines subsets to calculate diversity statistics for. |
level |
Level of the variable to create the subset to calculate diversity statistics. |
index |
Type of diversity statistic with "richness" to calculate species richness, "abundance" to calculate abundance, "Shannon" to calculate the Shannon diversity index, "Simpson" to calculate 1-Simpson concentration index, "inverseSimpson" to calculate the reciprocal Simpson diversity index, "simpson.unb" to calculate the unbiased Simpson index, "simpson.unb.inverse" to calculate the unbiased inverse simpson index, "Logalpha" to calculate the log series alpha diversity index, "Berger" to calculate the reciprocal Berger-Parker diversity index, "Jevenness" to calculate one Shannon evenness index, "Eevenness" to calculate another Shannon evenness index, "jack1" to calculate the first-order jackknife gamma diversity estimator, "jack2" to calculate the second-order jackknife gamma diversity estimator, "chao" to calculate the Chao gamma diversity estimator and "boot" to calculate the bootstrap gamma diversity estimator. |
method |
Method of calculating the diversity statistics: "pooled" calculates the diversity of the entire community (all sites pooled), "each site" calculates diversity for each site separetly, "mean" calculates the average diversity of the sites, "sd" calculates the standard deviation of the diversity of the sites, "max" calculates the maximum diversity of the sites, whereas "jackknife" calculates the jackknifed diversity for the entire data frame. |
sortit |
Sort the sites by increasing values of the diversity statistic. |
digits |
Number of digits in the results. |
factor1 |
Variable of the environmental data frame that defines subsets to calculate diversity statistics for. |
factor2 |
Optional second variable of the environmental data frame that defines subsets to calculate diversity statistics for in a crosstabulation with the other variable of the environmental data frame. |
These functions provide some alternative methods of obtaining results with diversity statistics, although functions diversity, fisher.alpha, specpool, estimateR and specnumber (all from vegan) are called to calculate the various statistics.
Function diversityvariables adds variables to the environmental dataset (richness, Shannon, Simpson, inverseSimpson, Logalpha, Berger, Jevenness, Eevenness).
The reciprocal Berger-Parker diversity index is the reciprocal of the proportional abundance of the most dominant species.
J-evenness is calculated as: H / ln(S) where H is the Shannon diversity index and S the species richness.
E-evenness is calculated as: exp(H) / S where H is the Shannon diversity index and S the species richness.
The method of calculating the diversity statistics include following options: "all" calculates the diversity of the entire community (all sites pooled together), "s" calculates the diversity of each site separatedly, "mean" calculates the average diversity of the sites, whereas "Jackknife" calculates the jackknifed diversity for the entire data frame. Methods "s" and "mean" are not available for function diversitycomp. Gamma diversity estimators assume that the method is "all".
Functions diversityresult and diversitycomp allow to calculate diversity statistics for subsets of the community and environmental data sets. Function diversityresult calculates the diversity statistics for the specified level of a selected environmental variable. Function diversitycomp calculates the diversity statistics for all levels of a selected environmental variable separatedly. When a second environmental variable is provided, function diversitycomp calculates diversity statistics as a crosstabulation of both variables.
The functions provide alternative methods of obtaining diversity results. For function diversitycomp, the number of sites is provided as "n".
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
## Not run: library(vegan) data(dune.env) data(dune) diversityresult(dune, y=NULL, index="Shannon", method="each site", sortit=TRUE, digits=5) diversityresult(dune, y=dune.env, factor="Management", level="NM", index="Shannon", method="each site", sortit=TRUE, digits=5) diversityresult(dune, y=NULL, index="Shannon", method="pooled", digits=5) diversityresult(dune, y=dune.env, factor="Management", level="NM", index="Shannon", method="pooled", digits=5) diversityresult(dune, y=NULL, index="Shannon", method="mean", digits=5) diversityresult(dune, y=NULL, index="Shannon", method="sd", digits=5) diversityresult(dune, y=NULL, index="Shannon", method="jackknife", digits=5) diversityresult(dune, y=dune.env, factor="Management", level="NM", index="Shannon", method="jackknife", digits=5) diversitycomp(dune, y=dune.env, factor1="Moisture", index="Shannon", method="pooled", sortit=TRUE) diversitycomp(dune, y=dune.env, factor1="Moisture", index="Shannon", method="mean", sortit=TRUE) diversitycomp(dune, y=dune.env, factor1="Management", index="Shannon", method="jackknife", sortit=TRUE) diversitycomp(dune, y=dune.env, factor1="Management", factor2="Moisture", index="Shannon", method="pooled", digits=6) diversitycomp(dune, y=dune.env, factor1="Management", factor2="Moisture", index="Shannon", method="mean", digits=6) ## End(Not run)## Not run: library(vegan) data(dune.env) data(dune) diversityresult(dune, y=NULL, index="Shannon", method="each site", sortit=TRUE, digits=5) diversityresult(dune, y=dune.env, factor="Management", level="NM", index="Shannon", method="each site", sortit=TRUE, digits=5) diversityresult(dune, y=NULL, index="Shannon", method="pooled", digits=5) diversityresult(dune, y=dune.env, factor="Management", level="NM", index="Shannon", method="pooled", digits=5) diversityresult(dune, y=NULL, index="Shannon", method="mean", digits=5) diversityresult(dune, y=NULL, index="Shannon", method="sd", digits=5) diversityresult(dune, y=NULL, index="Shannon", method="jackknife", digits=5) diversityresult(dune, y=dune.env, factor="Management", level="NM", index="Shannon", method="jackknife", digits=5) diversitycomp(dune, y=dune.env, factor1="Moisture", index="Shannon", method="pooled", sortit=TRUE) diversitycomp(dune, y=dune.env, factor1="Moisture", index="Shannon", method="mean", sortit=TRUE) diversitycomp(dune, y=dune.env, factor1="Management", index="Shannon", method="jackknife", sortit=TRUE) diversitycomp(dune, y=dune.env, factor1="Management", factor2="Moisture", index="Shannon", method="pooled", digits=6) diversitycomp(dune, y=dune.env, factor1="Management", factor2="Moisture", index="Shannon", method="mean", digits=6) ## End(Not run)
Function ensemble.analogue creates the map with climatic distance and provides the locations of the climate analogues (defined as locations with smallest climatic distance to a reference climate). Function ensemble.analogue.object provides the reference values used by the prediction function used by predict .
ensemble.analogue(x = NULL, analogue.object = NULL, analogues = 1, RASTER.object.name = analogue.object$name, RASTER.stack.name = x@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, limits = c(1, 5, 20, 50), limit.colours = c('red', 'orange', 'blue', 'grey'), CATCH.OFF = FALSE) ensemble.analogue.object(ref.location, future.stack, current.stack, name = "reference1", method = "mahal", an = 10000, probs = c(0.025, 0.975), weights = NULL, z = 2)ensemble.analogue(x = NULL, analogue.object = NULL, analogues = 1, RASTER.object.name = analogue.object$name, RASTER.stack.name = x@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, limits = c(1, 5, 20, 50), limit.colours = c('red', 'orange', 'blue', 'grey'), CATCH.OFF = FALSE) ensemble.analogue.object(ref.location, future.stack, current.stack, name = "reference1", method = "mahal", an = 10000, probs = c(0.025, 0.975), weights = NULL, z = 2)
x |
RasterStack object ( |
analogue.object |
Object listing reference values for the environmental layers and additional parameters (covariance matrix for |
analogues |
Number of analogue locations to be provided |
RASTER.object.name |
First part of the names of the raster file that will be generated, expected to identify the area and time period for which ranges were calculated |
RASTER.stack.name |
Last part of the names of the raster file that will be generated, expected to identify the predictor stack used |
RASTER.format |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
limits |
Limits indicating the accumulated number of closest analogue sites. These limits will correspond to different colours in the KML map. In the default setting, the closest analogue will be coloured red and the second to fifth closest analogues will be coloured orange. |
limit.colours |
Colours for the different limits based on number of analogues. |
CATCH.OFF |
Disable calls to function |
ref.location |
Location of the reference location for which analogues are searched for and from which climatic distance will be calculated, typically available in 2-column (lon, lat) dataframe; see also |
future.stack |
RasterStack object ( |
current.stack |
RasterStack object ( |
name |
Name of the object, expect to expected to identify the area and time period for which ranges were calculated and where no novel conditions will be detected |
method |
Method used to calculate climatic distance: |
an |
Number of randomly selected locations points to calculate the covariance matrix ( |
probs |
Numeric vector of probabilities [0,1] as used by |
weights |
Numeric vector of weights by which each variable (difference) should be multiplied by (can be used to give equal weight to 12 monthly rainfall values and 24 minimum and maximum monthly temperature values). Not used for |
z |
Parameter used as exponent for differences calculated between reference climatic variables and variables in the 'current' raster and reciprocal exponent for the sum of all differences. Default value of 2 corresponds to the Euclidean distance. Not used for |
Function ensemble.analogues maps the climatic distance from reference values determined by ensemble.analogues.object and provides the locations of the analogues closest analogues.
The method = "mahal" uses the Mahalanobis distance as environmental (climatic) distance: mahalanobis.
Other methods use a normalization method to handle scale differences between environmental (climatic) variables:
where are the target values for environmental (climatic) variable i,
are the values in the current environmental layers where analogues are searched for,
are the weights for environmental variable i, and
are the normalization parameters for environmental variable i
Function ensemble.analogue.object returns a list with following objects:
name |
name for the reference location |
ref.location |
coordinates of the reference location |
stack.name |
name for time period for which values are extracted from the |
method |
method used for calculating climatic distance |
target.values |
target environmental values to select analogues for through minimum climatic distance |
cov.mahal |
covariance matrix |
norm.values |
parameters by which each difference between target and 'current' value will be divided |
weight.values |
weights by which each difference between target and 'current' value will be multiplied |
z |
parameter to be used as exponent for differences between target and 'current' values |
Roeland Kindt (World Agroforestry Centre) and Eike Luedeling (World Agroforestry Centre)
Bos, Swen PM, et al. "Climate analogs for agricultural impact projection and adaptation-a reliability test." Frontiers in Environmental Science 3 (2015): 65. Luedeling, Eike, and Henry Neufeldt. "Carbon sequestration potential of parkland agroforestry in the Sahel." Climatic Change 115.3-4 (2012): 443-461.
## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio7", "bio8", "bio12", "bio16", "bio17")) predictors predictors@title <- "base" # instead of searching for current analogue of future climate conditions, # search for analogue in southern hemisphere future.stack <- stack(crop(predictors, y=extent(-125, -32, 0, 40))) future.stack@title <- "north" current.stack <- stack(crop(predictors, y=extent(-125, -32, -56, 0))) current.stack@title <- "south" # reference location in Florida # in this case future.stack and current.stack are both current ref.loc <- data.frame(t(c(-80.19, 25.76))) names(ref.loc) <- c("lon", "lat") # climate analogue analysis based on the Mahalanobis distance Florida.object.mahal <- ensemble.analogue.object(ref.location=ref.loc, future.stack=future.stack, current.stack=current.stack, name="FloridaMahal", method="mahal", an=10000) Florida.object.mahal Florida.analogue.mahal <- ensemble.analogue(x=current.stack, analogue.object=Florida.object.mahal, analogues=50) Florida.analogue.mahal # climate analogue analysis based on the Euclidean distance and dividing each variable by the sd Florida.object.sd <- ensemble.analogue.object(ref.location=ref.loc, future.stack=future.stack, current.stack=current.stack, name="FloridaSD", method="sd", z=2) Florida.object.sd Florida.analogue.sd <- ensemble.analogue(x=current.stack, analogue.object=Florida.object.sd, analogues=50) Florida.analogue.sd # plot analogues on climatic distance maps par(mfrow=c(1,2)) analogue.file <- paste(getwd(), "//ensembles//analogue//FloridaMahal_south_analogue.tif", sep="") plot(raster(analogue.file), main="Mahalanobis climatic distance") points(Florida.analogue.sd[3:50, "lat"] ~ Florida.analogue.sd[3:50, "lon"], pch=1, col="red", cex=1) points(Florida.analogue.mahal[3:50, "lat"] ~ Florida.analogue.mahal[3:50, "lon"], pch=3, col="black", cex=1) points(Florida.analogue.mahal[2, "lat"] ~ Florida.analogue.mahal[2, "lon"], pch=22, col="blue", cex=2) legend(x="topright", legend=c("closest", "Mahalanobis", "SD"), pch=c(22, 3 , 1), col=c("blue" , "black", "red")) analogue.file <- paste(getwd(), "//ensembles//analogue//FloridaSD_south_analogue.tif", sep="") plot(raster(analogue.file), main="Climatic distance normalized by standard deviation") points(Florida.analogue.mahal[3:50, "lat"] ~ Florida.analogue.mahal[3:50, "lon"], pch=3, col="black", cex=1) points(Florida.analogue.sd[3:50, "lat"] ~ Florida.analogue.sd[3:50, "lon"], pch=1, col="red", cex=1) points(Florida.analogue.sd[2, "lat"] ~ Florida.analogue.sd[2, "lon"], pch=22, col="blue", cex=2) legend(x="topright", legend=c("closest", "Mahalanobis", "SD"), pch=c(22, 3 , 1), col=c("blue" , "black", "red")) par(mfrow=c(1,1)) ## End(Not run)## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio7", "bio8", "bio12", "bio16", "bio17")) predictors predictors@title <- "base" # instead of searching for current analogue of future climate conditions, # search for analogue in southern hemisphere future.stack <- stack(crop(predictors, y=extent(-125, -32, 0, 40))) future.stack@title <- "north" current.stack <- stack(crop(predictors, y=extent(-125, -32, -56, 0))) current.stack@title <- "south" # reference location in Florida # in this case future.stack and current.stack are both current ref.loc <- data.frame(t(c(-80.19, 25.76))) names(ref.loc) <- c("lon", "lat") # climate analogue analysis based on the Mahalanobis distance Florida.object.mahal <- ensemble.analogue.object(ref.location=ref.loc, future.stack=future.stack, current.stack=current.stack, name="FloridaMahal", method="mahal", an=10000) Florida.object.mahal Florida.analogue.mahal <- ensemble.analogue(x=current.stack, analogue.object=Florida.object.mahal, analogues=50) Florida.analogue.mahal # climate analogue analysis based on the Euclidean distance and dividing each variable by the sd Florida.object.sd <- ensemble.analogue.object(ref.location=ref.loc, future.stack=future.stack, current.stack=current.stack, name="FloridaSD", method="sd", z=2) Florida.object.sd Florida.analogue.sd <- ensemble.analogue(x=current.stack, analogue.object=Florida.object.sd, analogues=50) Florida.analogue.sd # plot analogues on climatic distance maps par(mfrow=c(1,2)) analogue.file <- paste(getwd(), "//ensembles//analogue//FloridaMahal_south_analogue.tif", sep="") plot(raster(analogue.file), main="Mahalanobis climatic distance") points(Florida.analogue.sd[3:50, "lat"] ~ Florida.analogue.sd[3:50, "lon"], pch=1, col="red", cex=1) points(Florida.analogue.mahal[3:50, "lat"] ~ Florida.analogue.mahal[3:50, "lon"], pch=3, col="black", cex=1) points(Florida.analogue.mahal[2, "lat"] ~ Florida.analogue.mahal[2, "lon"], pch=22, col="blue", cex=2) legend(x="topright", legend=c("closest", "Mahalanobis", "SD"), pch=c(22, 3 , 1), col=c("blue" , "black", "red")) analogue.file <- paste(getwd(), "//ensembles//analogue//FloridaSD_south_analogue.tif", sep="") plot(raster(analogue.file), main="Climatic distance normalized by standard deviation") points(Florida.analogue.mahal[3:50, "lat"] ~ Florida.analogue.mahal[3:50, "lon"], pch=3, col="black", cex=1) points(Florida.analogue.sd[3:50, "lat"] ~ Florida.analogue.sd[3:50, "lon"], pch=1, col="red", cex=1) points(Florida.analogue.sd[2, "lat"] ~ Florida.analogue.sd[2, "lon"], pch=22, col="blue", cex=2) legend(x="topright", legend=c("closest", "Mahalanobis", "SD"), pch=c(22, 3 , 1), col=c("blue" , "black", "red")) par(mfrow=c(1,1)) ## End(Not run)
The main function allows for batch processing of different species and different environmental RasterStacks. The function makes internal calls to ensemble.calibrate.weights, ensemble.calibrate.models and ensemble.raster.
ensemble.batch(x = NULL, xn = c(x), species.presence = NULL, species.absence = NULL, presence.min = 20, thin.km = 0.1, an = 1000, excludep = FALSE, target.groups = FALSE, get.block = FALSE, block.default = runif(1) > 0.5, get.subblocks = FALSE, SSB.reduce = FALSE, CIRCLES.d = 250000, k.splits = 4, k.test = 0, n.ensembles = 1, VIF.max = 10, VIF.keep = NULL, SINK = FALSE, CATCH.OFF = FALSE, RASTER.datatype = "INT2S", RASTER.NAflag = -32767, models.save = FALSE, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, ENSEMBLE.weight.min = 0.05, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 0, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1 , SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, formulae.defaults = TRUE, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent", sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.formula = NULL, MAXLIKE.method = "BFGS", GBM.formula = NULL, GBM.n.trees = 2001, GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.formula = NULL, RF.ntree = 751, RF.mtry = floor(sqrt(raster::nlayers(x))), CF.formula = NULL, CF.ntree = 751, CF.mtry = floor(sqrt(raster::nlayers(x))), GLM.formula = NULL, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, STEP.formula = NULL, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.formula = NULL, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.formula = NULL, MGCV.select = FALSE, MGCVFIX.formula = NULL, EARTH.formula = NULL, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.formula = NULL, RPART.xval = 50, NNET.formula = NULL, NNET.size = 8, NNET.decay = 0.01, FDA.formula = NULL, SVM.formula = NULL, SVME.formula = NULL, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.mean(RASTER.species.name = "Species001", RASTER.stack.name = "base", positive.filters = c("tif", "_ENSEMBLE_"), negative.filters = c("xml"), RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, abs.breaks = 6, pres.breaks = 6, sd.breaks = 9, p = NULL, a = NULL, pt = NULL, at = NULL, threshold = -1, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE) ensemble.plot(RASTER.species.name = "Species001", RASTER.stack.name = "base", plot.method=c("suitability", "presence", "count", "consensussuitability", "consensuspresence", "consensuscount", "consensussd"), dev.new.width = 7, dev.new.height = 7, main = paste(RASTER.species.name, " ", plot.method, " for ", RASTER.stack.name, sep=""), positive.filters = c("tif"), negative.filters = c("xml"), p=NULL, a=NULL, threshold = -1, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, abs.breaks = 6, abs.col = NULL, pres.breaks = 6, pres.col = NULL, sd.breaks = 9, sd.col = NULL, absencePresence.col = NULL, count.col = NULL, ...)ensemble.batch(x = NULL, xn = c(x), species.presence = NULL, species.absence = NULL, presence.min = 20, thin.km = 0.1, an = 1000, excludep = FALSE, target.groups = FALSE, get.block = FALSE, block.default = runif(1) > 0.5, get.subblocks = FALSE, SSB.reduce = FALSE, CIRCLES.d = 250000, k.splits = 4, k.test = 0, n.ensembles = 1, VIF.max = 10, VIF.keep = NULL, SINK = FALSE, CATCH.OFF = FALSE, RASTER.datatype = "INT2S", RASTER.NAflag = -32767, models.save = FALSE, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, ENSEMBLE.weight.min = 0.05, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 0, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1 , SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, formulae.defaults = TRUE, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent", sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.formula = NULL, MAXLIKE.method = "BFGS", GBM.formula = NULL, GBM.n.trees = 2001, GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.formula = NULL, RF.ntree = 751, RF.mtry = floor(sqrt(raster::nlayers(x))), CF.formula = NULL, CF.ntree = 751, CF.mtry = floor(sqrt(raster::nlayers(x))), GLM.formula = NULL, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, STEP.formula = NULL, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.formula = NULL, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.formula = NULL, MGCV.select = FALSE, MGCVFIX.formula = NULL, EARTH.formula = NULL, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.formula = NULL, RPART.xval = 50, NNET.formula = NULL, NNET.size = 8, NNET.decay = 0.01, FDA.formula = NULL, SVM.formula = NULL, SVME.formula = NULL, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.mean(RASTER.species.name = "Species001", RASTER.stack.name = "base", positive.filters = c("tif", "_ENSEMBLE_"), negative.filters = c("xml"), RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, abs.breaks = 6, pres.breaks = 6, sd.breaks = 9, p = NULL, a = NULL, pt = NULL, at = NULL, threshold = -1, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE) ensemble.plot(RASTER.species.name = "Species001", RASTER.stack.name = "base", plot.method=c("suitability", "presence", "count", "consensussuitability", "consensuspresence", "consensuscount", "consensussd"), dev.new.width = 7, dev.new.height = 7, main = paste(RASTER.species.name, " ", plot.method, " for ", RASTER.stack.name, sep=""), positive.filters = c("tif"), negative.filters = c("xml"), p=NULL, a=NULL, threshold = -1, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, abs.breaks = 6, abs.col = NULL, pres.breaks = 6, pres.col = NULL, sd.breaks = 9, sd.col = NULL, absencePresence.col = NULL, count.col = NULL, ...)
x |
RasterStack object ( |
xn |
RasterStack object ( |
species.presence |
presence points used for calibrating the suitability models, available in 3-column (species, x, y) or (species, lon, lat) dataframe |
species.absence |
background points used for calibrating the suitability models, either available in a 3-column (species, x, y) or (species, lon, lat), or available in a 2-column (x, y) or (lon, lat) dataframe. In case of a 2-column dataframe, the same background locations will be used for all species. |
presence.min |
minimum number of presence locations for the organism (if smaller, no models are fitted). |
thin.km |
Threshold for minimum distance (km) between presence point locations for focal species for model calibrations in each run. A new data set is randomly selected via |
an |
number of background points for calibration to be selected with |
excludep |
parameter that indicates (if |
target.groups |
Parameter that indicates (if |
get.block |
if |
block.default |
if |
get.subblocks |
if |
SSB.reduce |
If |
CIRCLES.d |
Radius in m of circular neighbourhoods (created with |
k |
If larger than 1, the mumber of groups to split between calibration (k-1) and evaluation (1) data sets (for example, |
k.splits |
If larger than 1, the number of splits for the |
k.test |
If larger than 1, the mumber of groups to split between calibration (k-1) and evaluation (1) data sets when calibrating the final models (for example, |
n.ensembles |
If larger than 1, the number of different ensembles generated per species in batch processing. |
VIF.max |
Maximum Variance Inflation Factor of variables; see |
VIF.keep |
character vector with names of the variables to be kept; see |
SINK |
Append the results to a text file in subfolder 'outputs' (if |
CATCH.OFF |
Disable calls to function |
RASTER.format |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
models.save |
Save the list with model details to a file (if |
threshold.method |
Method to calculate the threshold between predicted absence and presence; possibilities include |
threshold.sensitivity |
Sensitivity value for |
threshold.PresenceAbsence |
If |
ENSEMBLE.best |
The number of individual suitability models to be used in the consensus suitability map (based on a weighted average). In case this parameter is smaller than 1 or larger than the number of positive input weights of individual models, then all individual suitability models with positive input weights are included in the consensus suitability map. In case a vector is provided, |
ENSEMBLE.min |
The minimum input weight (typically corresponding to AUC values) for a model to be included in the ensemble. In case a vector is provided, function |
ENSEMBLE.exponent |
Exponent applied to AUC values to convert AUC values into weights (for example, an exponent of 2 converts input weights of 0.7, 0.8 and 0.9 into 0.7^2=0.49, 0.8^2=0.64 and 0.9^2=0.81). See details. |
ENSEMBLE.weight.min |
The minimum output weight for models included in the ensemble, applying to weights that sum to one. Note that |
input.weights |
array with numeric values for the different modelling algorithms; if |
MAXENT |
Input weight for a maximum entropy model ( |
MAXNET |
number: if larger than 0, then a maximum entropy model ( |
MAXLIKE |
Input weight for a maxlike model ( |
GBM |
Input weight for a boosted regression trees model ( |
GBMSTEP |
Input weight for a stepwise boosted regression trees model ( |
RF |
Input weight for a random forest model ( |
CF |
number: if larger than 0, then a random forest model ( |
GLM |
Input weight for a generalized linear model ( |
GLMSTEP |
Input weight for a stepwise generalized linear model ( |
GAM |
Input weight for a generalized additive model ( |
GAMSTEP |
Input weight for a stepwise generalized additive model ( |
MGCV |
Input weight for a generalized additive model ( |
MGCVFIX |
number: if larger than 0, then a generalized additive model with fixed d.f. regression splines ( |
EARTH |
Input weight for a multivariate adaptive regression spline model ( |
RPART |
Input weight for a recursive partioning and regression tree model ( |
NNET |
Input weight for an artificial neural network model ( |
FDA |
Input weight for a flexible discriminant analysis model ( |
SVM |
Input weight for a support vector machine model ( |
SVME |
Input weight for a support vector machine model ( |
GLMNET |
Input weight for a GLM with lasso or elasticnet regularization ( |
BIOCLIM.O |
Input weight for the original BIOCLIM algorithm ( |
BIOCLIM |
Input weight for the BIOCLIM algorithm ( |
DOMAIN |
Input weight for the DOMAIN algorithm ( |
MAHAL |
Input weight for the Mahalonobis algorithm ( |
MAHAL01 |
Input weight for the Mahalanobis algorithm ( |
PROBIT |
If |
Yweights |
chooses how cases of presence and background (absence) are weighted; |
layer.drops |
vector that indicates which layers should be removed from RasterStack |
factors |
vector that indicates which variables are factors; see also |
dummy.vars |
vector that indicates which variables are dummy variables (influences formulae suggestions) |
formulae.defaults |
Suggest formulae for most of the models (if |
maxit |
Maximum number of iterations for some of the models. See also |
MAXENT.a |
background points used for calibrating the maximum entropy model ( |
MAXENT.an |
number of background points for calibration to be selected with |
MAXENT.path |
path to the directory where output files of the maximum entropy model are stored; see also |
MAXNET.classes |
continuous feature classes, either "default" or any subset of "lqpht" (linear, quadratic, product, hinge, threshold). Note that the "default" option chooses feature classes based on the number of presence locations as "l" (< 10 locations), "lq" (10 - 14 locations), "lqh" (15 - 79 locations) or "lqph" (> 79 locations). See also |
MAXNET.clamp |
restrict predictors and features to the range seen during model training; see also |
MAXNET.type |
type of response required; see also |
MAXLIKE.formula |
formula for the maxlike algorithm; see also |
MAXLIKE.method |
method for the maxlike algorithm; see also |
GBM.formula |
formula for the boosted regression trees algorithm; see also |
GBM.n.trees |
total number of trees to fit for the boosted regression trees model; see also |
GBMSTEP.tree.complexity |
complexity of individual trees for stepwise boosted regression trees; see also |
GBMSTEP.learning.rate |
weight applied to individual trees for stepwise boosted regression trees; see also |
GBMSTEP.bag.fraction |
proportion of observations used in selecting variables for stepwise boosted regression trees; see also |
GBMSTEP.step.size |
number of trees to add at each cycle for stepwise boosted regression trees (should be small enough to result in a smaller holdout deviance than the initial number of trees [50]); see also |
RF.formula |
formula for the random forest algorithm; see also |
RF.ntree |
number of trees to grow for random forest algorithm; see also |
RF.mtry |
number of variables randomly sampled as candidates at each split for random forest algorithm; see also |
CF.formula |
formula for random forest algorithm; see also |
CF.ntree |
number of trees to grow in a forest; see also |
CF.mtry |
number of input variables randomly sampled as candidates at each node for random forest like algorithms; see also |
GLM.formula |
formula for the generalized linear model; see also |
GLM.family |
description of the error distribution and link function for the generalized linear model; see also |
GLMSTEP.steps |
maximum number of steps to be considered for stepwise generalized linear model; see also |
STEP.formula |
formula for the "starting model" to be considered for stepwise generalized linear model; see also |
GLMSTEP.scope |
range of models examined in the stepwise search; see also |
GLMSTEP.k |
multiple of the number of degrees of freedom used for the penalty (only k = 2 gives the genuine AIC); see also |
GAM.formula |
formula for the generalized additive model; see also |
GAM.family |
description of the error distribution and link function for the generalized additive model; see also |
GAMSTEP.steps |
maximum number of steps to be considered in the stepwise generalized additive model; see also |
GAMSTEP.scope |
range of models examined in the step-wise search n the stepwise generalized additive model; see also |
GAMSTEP.pos |
parameter expected to be set to 1 to allow for fitting of the stepwise generalized additive model |
MGCV.formula |
formula for the generalized additive model; see also |
MGCV.select |
if |
MGCVFIX.formula |
formula for the generalized additive model with fixed d.f. regression splines; see also |
EARTH.formula |
formula for the multivariate adaptive regression spline model; see also |
EARTH.glm |
|
RPART.formula |
formula for the recursive partioning and regression tree model; see also |
RPART.xval |
number of cross-validations for the recursive partioning and regression tree model; see also |
NNET.formula |
formula for the artificial neural network model; see also |
NNET.size |
number of units in the hidden layer for the artificial neural network model; see also |
NNET.decay |
parameter of weight decay for the artificial neural network model; see also |
FDA.formula |
formula for the flexible discriminant analysis model; see also |
SVM.formula |
formula for the support vector machine model; see also |
SVME.formula |
formula for the support vector machine model; see also |
GLMNET.nlambda |
The number of |
GLMNET.class |
Use the predicted class to calculate the mean predictions of GLMNET; see also |
BIOCLIM.O.fraction |
Fraction of range representing the optimal limits, default value of 0.9 as in the original BIOCLIM software ( |
MAHAL.shape |
parameter that influences the transformation of output values of |
RASTER.species.name |
First part of the names of the raster files, expected to identify the modelled species (or organism). |
RASTER.stack.name |
Last part of the names of the raster files, expected to identify the predictor stack used. |
positive.filters |
vector that indicates parts of filenames for files that will be included in the calculation of the mean probability values |
negative.filters |
vector that indicates parts of filenames for files that will not be included in the calculation of the mean probability values |
abs.breaks |
Number of breaks in the colouring scheme for absence (only applies to |
pres.breaks |
Number of breaks in the colouring scheme for presence (only applies to |
sd.breaks |
Number of breaks in the colouring scheme for standard deviation (only applies to |
p |
presence points used for calibrating the suitability models, typically available in 2-column (x, y) or (lon, lat) dataframe; see also |
a |
background points used for calibrating the suitability models, typically available in 2-column (x, y) or (lon, lat) dataframe; see also |
pt |
presence points used for evaluating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
at |
background points used for calibrating the suitability models, typicall available in 2-column (lon, lat) dataframe; see also |
threshold |
Threshold value that will be used to distinguish between presence and absence. If < 0, then a threshold value will be calculated from the provided presence |
plot.method |
Choice of maps to be plotted: |
dev.new.width |
Width for new graphics device ( |
dev.new.height |
Heigth for new graphics device ( |
main |
main title for the plots. |
abs.col |
specify colours for absence (see examples on how not to plot areas where the species is predicted absent) |
pres.col |
specify colours for presence |
sd.col |
specify colours for standard deviation |
absencePresence.col |
specify colours for absence - presence maps (see examples on how not to plot areas where the species is predicted absent) |
count.col |
specify colours for number of algorithms or ensembles (see examples on how not to plot areas where the species is predicted absent) |
... |
Other items passed to function |
This function allows for batch processing of different species and different environmental RasterStacks. The function makes internal calls to ensemble.spatialThin, ensemble.VIF, ensemble.calibrate.weights, ensemble.calibrate.models and ensemble.raster.
Different ensemble runs allow for different random selection of k-fold subsets, background locations or spatial thinning of presence locations.
ensemble.calibrate.weights results in a cross-validation procedure whereby the data set is split in calibration and testing subsets and the best weights for the ensemble model are determined (including the possibility for weights = 0).
ensemble.calibrate.models is the step whereby models are calibrated using all the available presence data.
ensemble.raster is the final step whereby raster layers are produced for the ensemble model.
Function ensemble.mean results in raster layers that are based on the summary of several ensemble layers: the new ensemble has probability values that are the mean of the probabilities of the different raster layers, the presence-absence threshold is derived for this new ensemble layer, whereas the count reflects the number of ensemble layers where presence was predicted. Note the assumption that input probabilities are scaled between 0 and 1000 (as the output from ensemble.raster), whereas thresholds are based on actual probabilities (scaled between 0 and 1). After the mean probability has been calculated, the niche overlap (nicheOverlap) with the different input layers is calculated.
Function ensemble.plot plots suitability, presence-absence or count maps. In the case of suitability maps, the presence-absence threshold needs to be provide as suitabilities smaller than the threshold will be coloured red to orange, whereas suitabilities larger than the threshold will be coloured light blue to dark blue.
The function finally results in ensemble raster layers for each species, including the fitted values for the ensemble model, the estimated presence-absence and the count of the number of submodels prediction presence and absence.
Roeland Kindt (World Agroforestry Centre), Eike Luedeling (World Agroforestry Centre) and Evert Thomas (Bioversity International)
Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009
Buisson L, Thuiller W, Casajus N, Lek S and Grenouillet G. 2010. Uncertainty in ensemble forecasting of species distribution. Global Change Biology 16: 1145-1157
Phillips SJ, Dudik M, Elith J et al. 2009. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecological Applications 19: 181-197.
ensemble.calibrate.weights, ensemble.calibrate.models, ensemble.raster
## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',') pres[,1] <- rep("Bradypus", nrow(pres)) # choose background points background <- randomPoints(predictors, n=1000, extf = 1.00) # north and south for new predictions (as if new climates) ext2 <- extent(-90, -32, 0, 23) predictors2 <- crop(predictors, y=ext2) predictors2 <- stack(predictors2) predictors2@title <- "north" ext3 <- extent(-90, -32, -33, 0) predictors3 <- crop(predictors, y=ext3) predictors3 <- stack(predictors3) predictors3@title <- "south" # fit 3 ensembles with batch processing, choosing the best ensemble model based on the # average weights of 4-fold split of calibration and testing data # final models use all available presence data and average weights determined by the # ensemble.calibrate.weights function (called internally) # batch processing can handle several species by using 3-column species.presence and # species.absence data sets # note that these calculations can take a while ensemble.nofactors <- ensemble.batch(x=predictors, xn=c(predictors, predictors2, predictors3), species.presence=pres, species.absence=background, k.splits=4, k.test=0, n.ensembles=3, SINK=TRUE, layer.drops=c("biome"), ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, PROBIT=TRUE, Yweights="BIOMOD", formulae.defaults=TRUE) # summaries for the 3 ensembles for the species # summaries are based on files in folders ensemble/suitability, # ensemble/presence and ensemble/count # ensemble.mean is used internally in ensemble.batch ensemble.mean(RASTER.species.name="Bradypus", RASTER.stack.name="base", p=pres, a=background) # plot mean suitability without specifying colours plot1 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensussuitability", p=pres, a=background, abs.breaks=4, pres.breaks=9) plot1 # only colour the areas where species is predicted to be present # option is invoked by having no absence breaks # same colourscheme as \url{http://www.worldagroforestry.org/atlas-central-america} LAatlascols <- grDevices::colorRampPalette(c("#FFFF80", "#38E009","#1A93AB", "#0C1078")) plot2 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensussuitability", p=pres, a=background, abs.breaks=0, pres.breaks=9, pres.col=LAatlascols(8)) plot2 # only colour the areas where species is predicted to be present # option is invoked by only setting one colour for absence-presence plot3 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensuspresence", absencePresence.col=c("#90EE90")) # only colour presence area by specifying colours > 0 plot4 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensuscount", count.col=LAatlascols(3)) ## End(Not run)## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',') pres[,1] <- rep("Bradypus", nrow(pres)) # choose background points background <- randomPoints(predictors, n=1000, extf = 1.00) # north and south for new predictions (as if new climates) ext2 <- extent(-90, -32, 0, 23) predictors2 <- crop(predictors, y=ext2) predictors2 <- stack(predictors2) predictors2@title <- "north" ext3 <- extent(-90, -32, -33, 0) predictors3 <- crop(predictors, y=ext3) predictors3 <- stack(predictors3) predictors3@title <- "south" # fit 3 ensembles with batch processing, choosing the best ensemble model based on the # average weights of 4-fold split of calibration and testing data # final models use all available presence data and average weights determined by the # ensemble.calibrate.weights function (called internally) # batch processing can handle several species by using 3-column species.presence and # species.absence data sets # note that these calculations can take a while ensemble.nofactors <- ensemble.batch(x=predictors, xn=c(predictors, predictors2, predictors3), species.presence=pres, species.absence=background, k.splits=4, k.test=0, n.ensembles=3, SINK=TRUE, layer.drops=c("biome"), ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, PROBIT=TRUE, Yweights="BIOMOD", formulae.defaults=TRUE) # summaries for the 3 ensembles for the species # summaries are based on files in folders ensemble/suitability, # ensemble/presence and ensemble/count # ensemble.mean is used internally in ensemble.batch ensemble.mean(RASTER.species.name="Bradypus", RASTER.stack.name="base", p=pres, a=background) # plot mean suitability without specifying colours plot1 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensussuitability", p=pres, a=background, abs.breaks=4, pres.breaks=9) plot1 # only colour the areas where species is predicted to be present # option is invoked by having no absence breaks # same colourscheme as \url{http://www.worldagroforestry.org/atlas-central-america} LAatlascols <- grDevices::colorRampPalette(c("#FFFF80", "#38E009","#1A93AB", "#0C1078")) plot2 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensussuitability", p=pres, a=background, abs.breaks=0, pres.breaks=9, pres.col=LAatlascols(8)) plot2 # only colour the areas where species is predicted to be present # option is invoked by only setting one colour for absence-presence plot3 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensuspresence", absencePresence.col=c("#90EE90")) # only colour presence area by specifying colours > 0 plot4 <- ensemble.plot(RASTER.species.name="Bradypus", RASTER.stack.name="base", plot.method="consensuscount", count.col=LAatlascols(3)) ## End(Not run)
Implementation of the BIOCLIM algorithm more similar to the original BIOCLIM algorithm and software than the implementation through bioclim. Function ensemble.bioclim creates the suitability map. Function ensemble.bioclim.object provides the reference values used by the prediction function used by predict .
ensemble.bioclim(x = NULL, bioclim.object = NULL, RASTER.object.name = bioclim.object$species.name, RASTER.stack.name = x@title, RASTER.format = "GTiff", CATCH.OFF = FALSE) ensemble.bioclim.object(x = NULL, p = NULL, fraction = 0.9, quantiles = TRUE, species.name = "Species001", factors = NULL)ensemble.bioclim(x = NULL, bioclim.object = NULL, RASTER.object.name = bioclim.object$species.name, RASTER.stack.name = x@title, RASTER.format = "GTiff", CATCH.OFF = FALSE) ensemble.bioclim.object(x = NULL, p = NULL, fraction = 0.9, quantiles = TRUE, species.name = "Species001", factors = NULL)
x |
RasterStack object ( |
bioclim.object |
Object listing optimal and absolute minima and maxima for bioclimatic variables, used by the prediction function that is used internally by |
RASTER.object.name |
First part of the names of the raster file that will be generated, expected to identify the species or crop for which ranges were calculated |
RASTER.stack.name |
Last part of the names of the raster file that will be generated, expected to identify the predictor stack used |
RASTER.format |
Format of the raster files that will be generated. See |
CATCH.OFF |
Disable calls to function |
p |
presence points used for calibrating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
fraction |
Fraction of range representing the optimal limits, default value of 0.9 as in the original BIOCLIM software. |
quantiles |
If |
species.name |
Name by which the model results will be saved. |
factors |
vector that indicates which variables are factors; these variables will be ignored by the BIOCLIM algorithm |
Function ensemble.bioclim maps suitability for a species based on optimal (percentiles, typically 5 and 95 percent) and absolute (minimum to maximum) limits for bioclimatic variables. If all values at a given location are within the optimal limits, suitability values are mapped as 1 (suitable). If not all values are within the optimal limits, but all values are within the absolute limits, suitability values are mapped as 0.5 (marginal). If not all values are within the absolute limits, suitability values are mapped as 0 (unsuitable).
Function ensemble.bioclim.object calculates the optimal and absolute limits. Optimal limits are calculated based on the parameter fraction, resulting in optimal limits that correspond to 0.5-fraction/2 and 0.5+fraction/2 (the default value of 0.9 therefore gives a lower limit of 0.05 and a upper limit of 0.95). Two methods are implemented to obtain optimal limits for the lower and upper limits. One method (quantiles = FALSE) uses mean, standard deviation and a cutoff parameter calculated with qnorm. The other method (quantiles = TRUE) calculates optimal limits via the quantile function. To handle possible asymmetrical distributions better, the second method is used as default.
When x is a RasterStack and point locations are provided, then optimal and absolute limits correspond to the bioclimatic values observed for the locations. When x is RasterStack and point locations are not provided, then optimal and absolute limits correspond to the bioclimatic values of the RasterStack.
Applying to algorithm without providing point locations will provide results that are similar to the ensemble.novel function, whereby areas plotted as not suitable will be the same areas that are novel.
Function ensemble.bioclim.object returns a list with following objects:
lower.limits |
vector with lower limits for each bioclimatic variable |
upper.limits |
vector with upper limits for each bioclimatic variable |
minima |
vector with minima for each bioclimatic variable |
maxima |
vector with maxima for each bioclimatic variable |
means |
vector with mean values for each bioclimatic variable |
medians |
vector with median values for each bioclimatic variable |
sds |
vector with standard deviation values for each bioclimatic variable |
cutoff |
cutoff value for the normal distribution |
fraction |
fraction of values within the optimal limits |
species.name |
name for the species |
Roeland Kindt (World Agroforestry Centre) with inputs from Trevor Booth (CSIRO)
Nix HA. 1986. A biogeographic analysis of Australian elapid snakes. In: Atlas of Elapid Snakes of Australia. (Ed.) R. Longmore, pp. 4-15. Australian Flora and Fauna Series Number 7. Australian Government Publishing Service: Canberra.
Booth TH, Nix HA, Busby JR and Hutchinson MF. 2014. BIOCLIM: the first species distribution modelling package, its early applications and relevance to most current MAXENT studies. Diversity and Distributions 20: 1-9
bioclim, ensemble.bioclim.graph and ensemble.novel
## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] background <- dismo::randomPoints(predictors, n=100) colnames(background)=c('lon', 'lat') pres.dataset <- data.frame(extract(predictors, y=pres)) names(pres.dataset) <- names(predictors) pres.dataset$biome <- as.factor(pres.dataset$biome) Bradypus.bioclim <- ensemble.bioclim.object(predictors, quantiles=T, p=pres, factors="biome", species.name="Bradypus") Bradypus.bioclim # obtain the same results with a data.frame Bradypus.bioclim2 <- ensemble.bioclim.object(pres.dataset, quantiles=T, species.name="Bradypus") Bradypus.bioclim2 # obtain results for entire rasterStack Bradypus.bioclim3 <- ensemble.bioclim.object(predictors, p=NULL, quantiles=T, factors="biome", species.name="America") Bradypus.bioclim3 ensemble.bioclim(x=predictors, bioclim.object=Bradypus.bioclim) ensemble.bioclim(x=predictors, bioclim.object=Bradypus.bioclim3) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) rasterfull1 <- paste("ensembles//Bradypus_base_BIOCLIM_orig.tif", sep="") raster::plot(raster(rasterfull1), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="original method") rasterfull2 <- paste("ensembles//America_base_BIOCLIM_orig.tif", sep="") raster::plot(raster(rasterfull2), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="America") graphics::par(par.old) # compare with implementation bioclim in dismo bioclim.dismo <- bioclim(predictors, p=pres) rasterfull2 <- paste("ensembles//Bradypus_base_BIOCLIM_dismo.tif", sep="") raster::predict(object=predictors, model=bioclim.dismo, na.rm=TRUE, filename=rasterfull2, progress='text', overwrite=TRUE) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) raster::plot(raster(rasterfull1), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="original method") raster::plot(raster(rasterfull2), main="dismo method") graphics::par(par.old) # use dummy variables to deal with factors predictors <- stack(predictor.files) biome.layer <- predictors[["biome"]] biome.layer ensemble.dummy.variables(xcat=biome.layer, most.frequent=0, freq.min=1, overwrite=TRUE) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors.dummy <- subset(predictors, subset=c("biome_1", "biome_2", "biome_3", "biome_4", "biome_5", "biome_7", "biome_8", "biome_9", "biome_10", "biome_12", "biome_13", "biome_14")) predictors.dummy predictors.dummy@title <- "base_dummy" Bradypus.dummy <- ensemble.bioclim.object(predictors.dummy, quantiles=T, p=pres, species.name="Bradypus") Bradypus.dummy ensemble.bioclim(x=predictors.dummy, bioclim.object=Bradypus.dummy) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) rasterfull3 <- paste("ensembles//Bradypus_base_dummy_BIOCLIM_orig.tif", sep="") raster::plot(raster(rasterfull1), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="numeric predictors") raster::plot(raster(rasterfull3), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="dummy predictors") graphics::par(par.old) ## End(Not run)## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] background <- dismo::randomPoints(predictors, n=100) colnames(background)=c('lon', 'lat') pres.dataset <- data.frame(extract(predictors, y=pres)) names(pres.dataset) <- names(predictors) pres.dataset$biome <- as.factor(pres.dataset$biome) Bradypus.bioclim <- ensemble.bioclim.object(predictors, quantiles=T, p=pres, factors="biome", species.name="Bradypus") Bradypus.bioclim # obtain the same results with a data.frame Bradypus.bioclim2 <- ensemble.bioclim.object(pres.dataset, quantiles=T, species.name="Bradypus") Bradypus.bioclim2 # obtain results for entire rasterStack Bradypus.bioclim3 <- ensemble.bioclim.object(predictors, p=NULL, quantiles=T, factors="biome", species.name="America") Bradypus.bioclim3 ensemble.bioclim(x=predictors, bioclim.object=Bradypus.bioclim) ensemble.bioclim(x=predictors, bioclim.object=Bradypus.bioclim3) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) rasterfull1 <- paste("ensembles//Bradypus_base_BIOCLIM_orig.tif", sep="") raster::plot(raster(rasterfull1), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="original method") rasterfull2 <- paste("ensembles//America_base_BIOCLIM_orig.tif", sep="") raster::plot(raster(rasterfull2), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="America") graphics::par(par.old) # compare with implementation bioclim in dismo bioclim.dismo <- bioclim(predictors, p=pres) rasterfull2 <- paste("ensembles//Bradypus_base_BIOCLIM_dismo.tif", sep="") raster::predict(object=predictors, model=bioclim.dismo, na.rm=TRUE, filename=rasterfull2, progress='text', overwrite=TRUE) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) raster::plot(raster(rasterfull1), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="original method") raster::plot(raster(rasterfull2), main="dismo method") graphics::par(par.old) # use dummy variables to deal with factors predictors <- stack(predictor.files) biome.layer <- predictors[["biome"]] biome.layer ensemble.dummy.variables(xcat=biome.layer, most.frequent=0, freq.min=1, overwrite=TRUE) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors.dummy <- subset(predictors, subset=c("biome_1", "biome_2", "biome_3", "biome_4", "biome_5", "biome_7", "biome_8", "biome_9", "biome_10", "biome_12", "biome_13", "biome_14")) predictors.dummy predictors.dummy@title <- "base_dummy" Bradypus.dummy <- ensemble.bioclim.object(predictors.dummy, quantiles=T, p=pres, species.name="Bradypus") Bradypus.dummy ensemble.bioclim(x=predictors.dummy, bioclim.object=Bradypus.dummy) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) rasterfull3 <- paste("ensembles//Bradypus_base_dummy_BIOCLIM_orig.tif", sep="") raster::plot(raster(rasterfull1), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="numeric predictors") raster::plot(raster(rasterfull3), breaks=c(-0.1, 0, 0.5, 1), col=c("grey", "blue", "green"), main="dummy predictors") graphics::par(par.old) ## End(Not run)
The main graph function makes graphs that show mean, median, minimum, maximum and lower and upper limits for species or climates. The ensemble.bioclim.graph.data function creates input data, using ensemble.bioclim.object internally.
ensemble.bioclim.graph(graph.data = NULL, focal.var = NULL, species.climates.subset = NULL, cols = NULL, var.multiply = 1.0, ref.lines = TRUE) ensemble.bioclim.graph.data( x=NULL, p=NULL, fraction = 0.9, species.climate.name="Species001_base", factors = NULL)ensemble.bioclim.graph(graph.data = NULL, focal.var = NULL, species.climates.subset = NULL, cols = NULL, var.multiply = 1.0, ref.lines = TRUE) ensemble.bioclim.graph.data( x=NULL, p=NULL, fraction = 0.9, species.climate.name="Species001_base", factors = NULL)
graph.data |
Input data with same variables as created by |
focal.var |
Bioclimatic variable to be plotted in the graph |
species.climates.subset |
Character vector with subset of names of species and climates to be plotted in the graph (if not provided, then all species and climates will be plotted). |
cols |
colours for the different species and climates |
var.multiply |
multiplier for the values to be plotted; 0.1 should be used if the bioclimatic variable was multiplied by 10 in the raster layers as in WorldClim and AFRICLIM |
ref.lines |
If |
x |
RasterStack object ( |
p |
presence points used for calibrating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
fraction |
Fraction of range representing the optimal limits, default value of 0.9 as in the original BIOCLIM software; see also |
species.climate.name |
Name for the species or climate that will be used as label in the graph. |
factors |
vector that indicates which variables are factors; these variables will be ignored by the BIOCLIM algorithm; see also |
The function creates a graph that shows mean, median, minimum, maximum and upper and lower limits for a range of species and climates. The graph can be useful in interpreting results of ensemble.bioclim or ensemble.novel.
In the graphs, means are indicated by an asterisk (pch=8 and medians as larger circles (pch=1)).
function ensemble.bioclim.graph.data creates a data frame, function ensemble.bioclim.graph allows for plotting.
Roeland Kindt (World Agroforestry Centre)
## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # climates for north and south (use same process for future climates) ext2 <- extent(-90, -32, 0, 23) predictors2 <- crop(predictors, y=ext2) predictors2 <- stack(predictors2) predictors2@title <- "north" ext3 <- extent(-90, -32, -33, 0) predictors3 <- crop(predictors, y=ext3) predictors3 <- stack(predictors3) predictors3@title <- "south" graph.data1 <- ensemble.bioclim.graph.data(predictors, p=pres, factors="biome", species.climate.name="Bradypus") graph.data2 <- ensemble.bioclim.graph.data(predictors, p=NULL, factors="biome", species.climate.name="baseline") graph.data3 <- ensemble.bioclim.graph.data(predictors2, p=NULL, factors="biome", species.climate.name="north") graph.data4 <- ensemble.bioclim.graph.data(predictors3, p=NULL, factors="biome", species.climate.name="south") graph.data.all <- rbind(graph.data1, graph.data2, graph.data3, graph.data4) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(2, 2)) ensemble.bioclim.graph(graph.data.all, focal.var="bio5", var.multiply=0.1, cols=c("black", rep("blue", 3))) ensemble.bioclim.graph(graph.data.all, focal.var="bio6", var.multiply=0.1, cols=c("black", rep("blue", 3))) ensemble.bioclim.graph(graph.data.all, focal.var="bio16", var.multiply=1.0, cols=c("black", rep("blue", 3))) ensemble.bioclim.graph(graph.data.all, focal.var="bio17", var.multiply=1.0, cols=c("black", rep("blue", 3))) graphics::par(par.old) ## End(Not run)## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # climates for north and south (use same process for future climates) ext2 <- extent(-90, -32, 0, 23) predictors2 <- crop(predictors, y=ext2) predictors2 <- stack(predictors2) predictors2@title <- "north" ext3 <- extent(-90, -32, -33, 0) predictors3 <- crop(predictors, y=ext3) predictors3 <- stack(predictors3) predictors3@title <- "south" graph.data1 <- ensemble.bioclim.graph.data(predictors, p=pres, factors="biome", species.climate.name="Bradypus") graph.data2 <- ensemble.bioclim.graph.data(predictors, p=NULL, factors="biome", species.climate.name="baseline") graph.data3 <- ensemble.bioclim.graph.data(predictors2, p=NULL, factors="biome", species.climate.name="north") graph.data4 <- ensemble.bioclim.graph.data(predictors3, p=NULL, factors="biome", species.climate.name="south") graph.data.all <- rbind(graph.data1, graph.data2, graph.data3, graph.data4) par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(2, 2)) ensemble.bioclim.graph(graph.data.all, focal.var="bio5", var.multiply=0.1, cols=c("black", rep("blue", 3))) ensemble.bioclim.graph(graph.data.all, focal.var="bio6", var.multiply=0.1, cols=c("black", rep("blue", 3))) ensemble.bioclim.graph(graph.data.all, focal.var="bio16", var.multiply=1.0, cols=c("black", rep("blue", 3))) ensemble.bioclim.graph(graph.data.all, focal.var="bio17", var.multiply=1.0, cols=c("black", rep("blue", 3))) graphics::par(par.old) ## End(Not run)
The basic function ensemble.calibrate.models allows to evaluate different algorithms for (species) suitability modelling, including maximum entropy (MAXENT), boosted regression trees, random forests, generalized linear models (including stepwise selection of explanatory variables), generalized additive models (including stepwise selection of explanatory variables), multivariate adaptive regression splines, regression trees, artificial neural networks, flexible discriminant analysis, support vector machines, the BIOCLIM algorithm, the DOMAIN algorithm and the Mahalanobis algorithm. These sets of functions were developed in parallel with the biomod2 package, especially for inclusion of the maximum entropy algorithm, but also to allow for a more direct integration with the BiodiversityR package, more direct handling of model formulae and greater focus on mapping. Researchers and students of species distribution are strongly encouraged to familiarize themselves with all the options of the BIOMOD and dismo packages.
ensemble.calibrate.models(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, target.groups = FALSE, k = 0, pt = NULL, at = NULL, SSB.reduce = FALSE, CIRCLES.d = 250000, TrainData = NULL, TestData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, CATCH.OFF = FALSE, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, evaluations.keep = FALSE, models.list = NULL, models.keep = FALSE, models.save = FALSE, species.name = "Species001", ENSEMBLE.tune = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, ENSEMBLE.weight.min = 0.05, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 1, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1 , SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, formulae.defaults = TRUE, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent_", species.name, sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.formula = NULL, MAXLIKE.method = "BFGS", GBM.formula = NULL, GBM.n.trees = 2001, GBMSTEP.gbm.x = 2:(ncol(TrainData.orig)), GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.formula = NULL, RF.ntree = 751, RF.mtry = floor(sqrt(ncol(TrainData.vars))), CF.formula = NULL, CF.ntree = 751, CF.mtry = floor(sqrt(ncol(TrainData.vars))), GLM.formula = NULL, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, STEP.formula = NULL, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.formula = NULL, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.formula = NULL, MGCV.select = FALSE, MGCVFIX.formula = NULL, EARTH.formula = NULL, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.formula = NULL, RPART.xval = 50, NNET.formula = NULL, NNET.size = 8, NNET.decay = 0.01, FDA.formula = NULL, SVM.formula = NULL, SVME.formula = NULL, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.calibrate.weights(x = NULL, p = NULL, TrainTestData=NULL, a = NULL, an = 1000, get.block = FALSE, block.default = TRUE, get.subblocks = FALSE, SSB.reduce = FALSE, CIRCLES.d = 100000, excludep = FALSE, target.groups = FALSE, k = 4, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, CATCH.OFF = FALSE, data.keep = FALSE, species.name = "Species001", threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, ENSEMBLE.tune = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, ENSEMBLE.weight.min = 0.05, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 1, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1 , SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, formulae.defaults = TRUE, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent_", species.name, sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.formula = NULL, MAXLIKE.method = "BFGS", GBM.formula = NULL, GBM.n.trees = 2001, GBMSTEP.gbm.x = 2:(length(var.names)+1), GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.formula = NULL, RF.ntree = 751, RF.mtry = floor(sqrt(length(var.names))), CF.formula = NULL, CF.ntree = 751, CF.mtry = floor(sqrt(length(var.names))), GLM.formula = NULL, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, STEP.formula = NULL, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.formula = NULL, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.formula = NULL, MGCV.select = FALSE, MGCVFIX.formula = NULL, EARTH.formula = NULL, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.formula = NULL, RPART.xval = 50, NNET.formula = NULL, NNET.size = 8, NNET.decay = 0.01, FDA.formula = NULL, SVM.formula = NULL, SVME.formula = NULL, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.calibrate.models.gbm(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, k = 4, TrainData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, species.name = "Species001", Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, GBMSTEP.gbm.x = 2:(ncol(TrainData.orig)), complexity = c(3:6), learning = c(0.005, 0.002, 0.001), GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100) ensemble.calibrate.models.nnet(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, k = 4, TrainData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, species.name = "Species001", Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, formulae.defaults = TRUE, maxit = 100, NNET.formula = NULL, sizes = c(2, 4, 6, 8), decays = c(0.1, 0.05, 0.01, 0.001) ) ensemble.drop1(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, target.groups = FALSE, k = 0, pt = NULL, at = NULL, SSB.reduce = FALSE, CIRCLES.d = 100000, TrainData = NULL, TestData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, species.name = "Species001", difference = FALSE, variables.alone = FALSE, ENSEMBLE.tune = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 0, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1, SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent_", species.name, sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.method = "BFGS", GBM.n.trees = 2001, GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.ntree = 751, CF.ntree = 751, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.select = FALSE, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.xval = 50, NNET.size = 8, NNET.decay = 0.01, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.weights(weights = c(0.9, 0.8, 0.7, 0.5), best = 0, min.weight = 0, exponent = 1, digits = 6) ensemble.strategy(TrainData = NULL, TestData = NULL, verbose = FALSE, ENSEMBLE.best = c(4:10), ENSEMBLE.min = c(0.7), ENSEMBLE.exponent = c(1, 2, 3) ) ensemble.formulae(x, layer.drops = NULL, factors = NULL, dummy.vars = NULL, weights = NULL) ensemble.threshold(eval, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, Pres, Abs) ensemble.VIF(x = NULL, a = NULL, an = 10000, VIF.max = 10, keep = NULL, layer.drops = NULL, factors = NULL, dummy.vars = NULL) ensemble.VIF.dataframe(x=NULL, VIF.max=10, keep=NULL, car=TRUE, silent=F) ensemble.pairs(x = NULL, a = NULL, an = 10000)ensemble.calibrate.models(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, target.groups = FALSE, k = 0, pt = NULL, at = NULL, SSB.reduce = FALSE, CIRCLES.d = 250000, TrainData = NULL, TestData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, CATCH.OFF = FALSE, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, evaluations.keep = FALSE, models.list = NULL, models.keep = FALSE, models.save = FALSE, species.name = "Species001", ENSEMBLE.tune = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, ENSEMBLE.weight.min = 0.05, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 1, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1 , SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, formulae.defaults = TRUE, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent_", species.name, sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.formula = NULL, MAXLIKE.method = "BFGS", GBM.formula = NULL, GBM.n.trees = 2001, GBMSTEP.gbm.x = 2:(ncol(TrainData.orig)), GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.formula = NULL, RF.ntree = 751, RF.mtry = floor(sqrt(ncol(TrainData.vars))), CF.formula = NULL, CF.ntree = 751, CF.mtry = floor(sqrt(ncol(TrainData.vars))), GLM.formula = NULL, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, STEP.formula = NULL, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.formula = NULL, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.formula = NULL, MGCV.select = FALSE, MGCVFIX.formula = NULL, EARTH.formula = NULL, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.formula = NULL, RPART.xval = 50, NNET.formula = NULL, NNET.size = 8, NNET.decay = 0.01, FDA.formula = NULL, SVM.formula = NULL, SVME.formula = NULL, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.calibrate.weights(x = NULL, p = NULL, TrainTestData=NULL, a = NULL, an = 1000, get.block = FALSE, block.default = TRUE, get.subblocks = FALSE, SSB.reduce = FALSE, CIRCLES.d = 100000, excludep = FALSE, target.groups = FALSE, k = 4, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, CATCH.OFF = FALSE, data.keep = FALSE, species.name = "Species001", threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, ENSEMBLE.tune = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, ENSEMBLE.weight.min = 0.05, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 1, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1 , SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, formulae.defaults = TRUE, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent_", species.name, sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.formula = NULL, MAXLIKE.method = "BFGS", GBM.formula = NULL, GBM.n.trees = 2001, GBMSTEP.gbm.x = 2:(length(var.names)+1), GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.formula = NULL, RF.ntree = 751, RF.mtry = floor(sqrt(length(var.names))), CF.formula = NULL, CF.ntree = 751, CF.mtry = floor(sqrt(length(var.names))), GLM.formula = NULL, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, STEP.formula = NULL, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.formula = NULL, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.formula = NULL, MGCV.select = FALSE, MGCVFIX.formula = NULL, EARTH.formula = NULL, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.formula = NULL, RPART.xval = 50, NNET.formula = NULL, NNET.size = 8, NNET.decay = 0.01, FDA.formula = NULL, SVM.formula = NULL, SVME.formula = NULL, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.calibrate.models.gbm(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, k = 4, TrainData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, species.name = "Species001", Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, GBMSTEP.gbm.x = 2:(ncol(TrainData.orig)), complexity = c(3:6), learning = c(0.005, 0.002, 0.001), GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100) ensemble.calibrate.models.nnet(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, k = 4, TrainData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, PLOTS = FALSE, species.name = "Species001", Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, formulae.defaults = TRUE, maxit = 100, NNET.formula = NULL, sizes = c(2, 4, 6, 8), decays = c(0.1, 0.05, 0.01, 0.001) ) ensemble.drop1(x = NULL, p = NULL, a = NULL, an = 1000, excludep = FALSE, target.groups = FALSE, k = 0, pt = NULL, at = NULL, SSB.reduce = FALSE, CIRCLES.d = 100000, TrainData = NULL, TestData = NULL, VIF = FALSE, COR = FALSE, SINK = FALSE, species.name = "Species001", difference = FALSE, variables.alone = FALSE, ENSEMBLE.tune = FALSE, ENSEMBLE.best = 0, ENSEMBLE.min = 0.7, ENSEMBLE.exponent = 1, input.weights = NULL, MAXENT = 1, MAXNET = 1, MAXLIKE = 1, GBM = 1, GBMSTEP = 0, RF = 1, CF = 1, GLM = 1, GLMSTEP = 1, GAM = 1, GAMSTEP = 1, MGCV = 1, MGCVFIX = 0, EARTH = 1, RPART = 1, NNET = 1, FDA = 1, SVM = 1, SVME = 1, GLMNET = 1, BIOCLIM.O = 0, BIOCLIM = 1, DOMAIN = 1, MAHAL = 1, MAHAL01 = 1, PROBIT = FALSE, Yweights = "BIOMOD", layer.drops = NULL, factors = NULL, dummy.vars = NULL, maxit = 100, MAXENT.a = NULL, MAXENT.an = 10000, MAXENT.path = paste(getwd(), "/models/maxent_", species.name, sep=""), MAXNET.classes = "default", MAXNET.clamp = FALSE, MAXNET.type = "cloglog", MAXLIKE.method = "BFGS", GBM.n.trees = 2001, GBMSTEP.tree.complexity = 5, GBMSTEP.learning.rate = 0.005, GBMSTEP.bag.fraction = 0.5, GBMSTEP.step.size = 100, RF.ntree = 751, CF.ntree = 751, GLM.family = binomial(link = "logit"), GLMSTEP.steps = 1000, GLMSTEP.scope = NULL, GLMSTEP.k = 2, GAM.family = binomial(link = "logit"), GAMSTEP.steps = 1000, GAMSTEP.scope = NULL, GAMSTEP.pos = 1, MGCV.select = FALSE, EARTH.glm = list(family = binomial(link = "logit"), maxit = maxit), RPART.xval = 50, NNET.size = 8, NNET.decay = 0.01, GLMNET.nlambda = 100, GLMNET.class = FALSE, BIOCLIM.O.fraction = 0.9, MAHAL.shape = 1) ensemble.weights(weights = c(0.9, 0.8, 0.7, 0.5), best = 0, min.weight = 0, exponent = 1, digits = 6) ensemble.strategy(TrainData = NULL, TestData = NULL, verbose = FALSE, ENSEMBLE.best = c(4:10), ENSEMBLE.min = c(0.7), ENSEMBLE.exponent = c(1, 2, 3) ) ensemble.formulae(x, layer.drops = NULL, factors = NULL, dummy.vars = NULL, weights = NULL) ensemble.threshold(eval, threshold.method = "spec_sens", threshold.sensitivity = 0.9, threshold.PresenceAbsence = FALSE, Pres, Abs) ensemble.VIF(x = NULL, a = NULL, an = 10000, VIF.max = 10, keep = NULL, layer.drops = NULL, factors = NULL, dummy.vars = NULL) ensemble.VIF.dataframe(x=NULL, VIF.max=10, keep=NULL, car=TRUE, silent=F) ensemble.pairs(x = NULL, a = NULL, an = 10000)
x |
RasterStack object ( |
p |
presence points used for calibrating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
a |
background points used for calibrating the suitability models (except for |
an |
number of background points for calibration to be selected with |
excludep |
parameter that indicates (if |
target.groups |
Parameter that indicates (if |
k |
If larger than 1, the number of groups to split between calibration (k-1) and evaluation (1) data sets (for example, |
pt |
presence points used for evaluating the suitability models, available in 2-column (lon, lat) dataframe; see also |
at |
background points used for evaluating the suitability models, available in 2-column (lon, lat) dataframe; see also |
SSB.reduce |
If |
CIRCLES.d |
Radius in m of circular neighbourhoods (created with |
TrainData |
dataframe with first column 'pb' describing presence (1) and absence (0) and other columns containing explanatory variables; see also |
TestData |
dataframe with first column 'pb' describing presence (1) and absence (0) and other columns containing explanatory variables; see also |
VIF |
Estimate the variance inflation factors based on a linear model calibrated on the training data (if |
COR |
Provide information on the correlation between the numeric explanatory variables (if |
SINK |
Append the results to a text file in subfolder 'outputs' (if |
PLOTS |
Disabled option of plotting evaluation results(BiodiversityR version 2.9-1) - see examples how to plot results afterwards and also |
CATCH.OFF |
Disable calls to function |
threshold.method |
Method to calculate the threshold between predicted absence and presence; possibilities include |
threshold.sensitivity |
Sensitivity value for |
threshold.PresenceAbsence |
If |
evaluations.keep |
Keep the results of evaluations (if |
models.list |
list with 'old' model objects such as |
models.keep |
store the details for each suitability modelling algorithm (if |
models.save |
Save the list with model details to a file (if |
species.name |
Name by which the model details will be saved to a file; see also argument |
data.keep |
Keep the data for each k-fold cross-validation run (if |
ENSEMBLE.tune |
Determine weights for the ensemble model based on AUC values (if |
ENSEMBLE.best |
The number of individual suitability models to be used in the consensus suitability map (based on a weighted average). In case this parameter is smaller than 1 or larger than the number of positive input weights of individual models, then all individual suitability models with positive input weights are included in the consensus suitability map. In case a vector is provided, |
ENSEMBLE.min |
The minimum input weight (typically corresponding to AUC values) for a model to be included in the ensemble. In case a vector is provided, function |
ENSEMBLE.exponent |
Exponent applied to AUC values to convert AUC values into weights (for example, an exponent of 2 converts input weights of 0.7, 0.8 and 0.9 into 0.7^2=0.49, 0.8^2=0.64 and 0.9^2=0.81). See details. |
ENSEMBLE.weight.min |
The minimum output weight for models included in the ensemble, applying to weights that sum to one. Note that |
input.weights |
array with numeric values for the different modelling algorithms; if |
MAXENT |
number: if larger than 0, then a maximum entropy model ( |
MAXNET |
number: if larger than 0, then a maximum entropy model ( |
MAXLIKE |
number: if larger than 0, then a maxlike model ( |
GBM |
number: if larger than 0, then a boosted regression trees model ( |
GBMSTEP |
number: if larger than 0, then a stepwise boosted regression trees model ( |
RF |
number: if larger than 0, then a random forest model ( |
CF |
number: if larger than 0, then a random forest model ( |
GLM |
number: if larger than 0, then a generalized linear model ( |
GLMSTEP |
number: if larger than 0, then a stepwise generalized linear model ( |
GAM |
number: if larger than 0, then a generalized additive model ( |
GAMSTEP |
number: if larger than 0, then a stepwise generalized additive model ( |
MGCV |
number: if larger than 0, then a generalized additive model ( |
MGCVFIX |
number: if larger than 0, then a generalized additive model with fixed d.f. regression splines ( |
EARTH |
number: if larger than 0, then a multivariate adaptive regression spline model ( |
RPART |
number: if larger than 0, then a recursive partioning and regression tree model ( |
NNET |
number: if larger than 0, then an artificial neural network model ( |
FDA |
number: if larger than 0, then a flexible discriminant analysis model ( |
SVM |
number: if larger than 0, then a support vector machine model ( |
SVME |
number: if larger than 0, then a support vector machine model ( |
GLMNET |
number: if larger than 0, then a GLM with lasso or elasticnet regularization ( |
BIOCLIM.O |
number: if larger than 0, then the original BIOCLIM algorithm ( |
BIOCLIM |
number: if larger than 0, then the BIOCLIM algorithm ( |
DOMAIN |
number: if larger than 0, then the DOMAIN algorithm ( |
MAHAL |
number: if larger than 0, then the Mahalanobis algorithm ( |
MAHAL01 |
number: if larger than 0, then the Mahalanobis algorithm ( |
PROBIT |
If |
Yweights |
chooses how cases of presence and background (absence) are weighted; |
layer.drops |
vector that indicates which layers should be removed from RasterStack |
factors |
vector that indicates which variables are factors; see also |
dummy.vars |
vector that indicates which variables are dummy variables (influences formulae suggestions) |
formulae.defaults |
Suggest formulae for most of the models (if |
maxit |
Maximum number of iterations for some of the models. See also |
MAXENT.a |
background points used for calibrating the maximum entropy model ( |
MAXENT.an |
number of background points for calibration to be selected with |
MAXENT.path |
path to the directory where output files of the maximum entropy model are stored; see also |
MAXNET.classes |
continuous feature classes, either "default" or any subset of "lqpht" (linear, quadratic, product, hinge, threshold). Note that the "default" option chooses feature classes based on the number of presence locations as "l" (< 10 locations), "lq" (10 - 14 locations), "lqh" (15 - 79 locations) or "lqph" (> 79 locations). See also |
MAXNET.clamp |
restrict predictors and features to the range seen during model training; see also |
MAXNET.type |
type of response required; see also |
MAXLIKE.formula |
formula for the maxlike algorithm; see also |
MAXLIKE.method |
method for the maxlike algorithm; see also |
GBM.formula |
formula for the boosted regression trees algorithm; see also |
GBM.n.trees |
total number of trees to fit for the boosted regression trees model; see also |
GBMSTEP.gbm.x |
indices of column numbers with explanatory variables for stepwise boosted regression trees; see also |
GBMSTEP.tree.complexity |
complexity of individual trees for stepwise boosted regression trees; see also |
GBMSTEP.learning.rate |
weight applied to individual trees for stepwise boosted regression trees; see also |
GBMSTEP.bag.fraction |
proportion of observations used in selecting variables for stepwise boosted regression trees; see also |
GBMSTEP.step.size |
number of trees to add at each cycle for stepwise boosted regression trees (should be small enough to result in a smaller holdout deviance than the initial number of trees [50]); see also |
RF.formula |
formula for random forest algorithm; see also |
RF.ntree |
number of trees to grow for random forest algorithm; see also |
RF.mtry |
number of variables randomly sampled as candidates at each split for random forest algorithm; see also |
CF.formula |
formula for random forest algorithm; see also |
CF.ntree |
number of trees to grow in a forest; see also |
CF.mtry |
number of input variables randomly sampled as candidates at each node for random forest like algorithms; see also |
GLM.formula |
formula for the generalized linear model; see also |
GLM.family |
description of the error distribution and link function for the generalized linear model; see also |
GLMSTEP.steps |
maximum number of steps to be considered for stepwise generalized linear model; see also |
STEP.formula |
formula for the "starting model" to be considered for stepwise generalized linear model; see also |
GLMSTEP.scope |
range of models examined in the stepwise search; see also |
GLMSTEP.k |
multiple of the number of degrees of freedom used for the penalty (only k = 2 gives the genuine AIC); see also |
GAM.formula |
formula for the generalized additive model; see also |
GAM.family |
description of the error distribution and link function for the generalized additive model; see also |
GAMSTEP.steps |
maximum number of steps to be considered in the stepwise generalized additive model; see also |
GAMSTEP.scope |
range of models examined in the step-wise search n the stepwise generalized additive model; see also |
GAMSTEP.pos |
parameter expected to be set to 1 to allow for fitting of the stepwise generalized additive model |
MGCV.formula |
formula for the generalized additive model; see also |
MGCV.select |
if |
MGCVFIX.formula |
formula for the generalized additive model with fixed d.f. regression splines; see also |
EARTH.formula |
formula for the multivariate adaptive regression spline model; see also |
EARTH.glm |
|
RPART.formula |
formula for the recursive partioning and regression tree model; see also |
RPART.xval |
number of cross-validations for the recursive partioning and regression tree model; see also |
NNET.formula |
formula for the artificial neural network model; see also |
NNET.size |
number of units in the hidden layer for the artificial neural network model; see also |
NNET.decay |
parameter of weight decay for the artificial neural network model; see also |
FDA.formula |
formula for the flexible discriminant analysis model; see also |
SVM.formula |
formula for the support vector machine model; see also |
SVME.formula |
formula for the support vector machine model; see also |
GLMNET.nlambda |
The number of |
GLMNET.class |
Use the predicted class to calculate the mean predictions of GLMNET; see |
BIOCLIM.O.fraction |
Fraction of range representing the optimal limits, default value of 0.9 as in the original BIOCLIM software ( |
MAHAL.shape |
parameter that influences the transformation of output values of |
TrainTestData |
dataframe with first column 'pb' describing presence (1) and absence (0) and other columns containing explanatory variables; see also |
get.block |
if |
block.default |
if |
get.subblocks |
if |
complexity |
vector with values of complexity of individual trees ( |
learning |
vector with values of weights applied to individual trees ( |
sizes |
vector with values of number of units in the hidden layer for the artificial neural network model; see also |
decays |
vector with values of weight decay for the artificial neural network model; see also |
difference |
if |
variables.alone |
if |
weights |
input weights for the |
best |
The number of final weights. In case this parameter is smaller than 1 or larger than the number of positive input weights of individual models, then this parameter is ignored. |
min.weight |
The minimum input weight to be included in the output. |
exponent |
Exponent applied to AUC values to convert AUC values into weights (for example, an exponent of 2 converts input weights of 0.7, 0.8 and 0.9 into 0.7^2=0.49, 0.8^2=0.64 and 0.9^2=0.81). See details. |
digits |
Number of number of decimal places in the output weights; see also |
verbose |
If |
eval |
ModelEvaluation object obtained by |
Pres |
Suitabilities (probabilities) at presence locations |
Abs |
Suitabilities (probabilities) at background locations |
VIF.max |
Maximum Variance Inflation Factor of the selected subset of variables. In case that at least one of the variables has VIF larger than VIF.max, then the variable with the highest VIF will be removed in the next step. |
keep |
character vector with names of the variables to be kept. |
car |
Also provide results from |
silent |
Limit textual output. |
The basic function ensemble.calibrate.models first calibrates individual suitability models based on presence locations p and background locations a, then evaluates these suitability models based on presence locations pt and background locations at. While calibrating and testing individual models, results obtained via the evaluate function can be saved (evaluations.keep).
As an alternative to providing presence locations p, models can be calibrated with data provided in TrainData. In case that both p and TrainData are provided, then models will be calibrated with TrainData.
Calibration of the maximum entropy (MAXENT) algorithm is not based on background locations a, but based on background locations MAXENT.a instead. However, to compare evaluations with evaluations of other algorithms, during evaluations of the MAXENT algorithm, presence locations p and background locations a are used (and not background locations MAXENT.a).
Output from the GLMNET algorithm is calculated as the mean of the output from predict.glmnet. With option GLMNET.class = TRUE, the mean output is the mean prediction of class 1. With option GLMNET.class = FALSE, the mean output is the mean of the responses.
As the Mahalanobis function (mahal) does not always provide values within the range of 0 - 1, the output values are rescaled with option MAHAL01 by first subtracting the value of 1 - MAHAL.shape from each prediction, followed by calculating the absolute value, followed by calculating the reciprocal value and finally multiplying this reciprocal value with MAHAL.shape. As this rescaling method does not estimate probabilities, inclusion in the calculation of a (weighted) average of ensemble probabilities may be problematic and the PROBIT transformation may help here (the same applies to other distance-based methods).
With parameter ENSEMBLE.best, the subset of best models (evaluated by the individual AUC values) can be selected and only those models will be used for calculating the ensemble model (in other words, weights for models not included in the ensemble will be set to zero). It is possible to further increase the contribution to the ensemble model for models with higher AUC values through parameter ENSEMBLE.exponent. With ENSEMBLE.exponent = 2, AUC values of 0.7, 0.8 and 0.9 are converted into weights of 0.7^2=0.49, 0.8^2=0.64 and 0.9^2=0.81). With ENSEMBLE.exponent = 4, AUC values of 0.7, 0.8 and 0.9 are converted into weights of 0.7^4=0.2401, 0.8^4=0.4096 and 0.9^2=0.6561).
ENSEMBLE.tune will result in an internal procedure whereby the best selection of parameter values for ENSEMBLE.min, ENSEMBLE.best or ENSEMBLE.exponent can be identified. Through a factorial procedure, the ensemble model with best AUC for a specific combination of parameter values is identified. The procedure also provides the weights that correspond to the best ensemble. In case that ENSEMBLE.tune is set to FALSE, then the ensemble is calculated based on the input weights.
Function ensemble.calibrate.weights splits the presence and background locations in a user-defined (k) number of subsets (i.e. k-fold cross-validation), then sequentially calibrates individual suitability models with (k-1) combined subsets and evaluates those with the remaining one subset, whereby each subset is used once for evaluation in the user-defined number (k) of runs. For example, k = 4 results in splitting the locations in 4 subsets, then using one of these subsets in turn for evaluations (see also kfold). Note that for the maximum entropy (MAXENT) algorithm, the same background data will be used in each cross-validation run (this is based on the assumption that a large number (~10000) of background locations are used).
Among the output from function ensemble.calibrate.weights are suggested weights for an ensemble model (output.weights and output.weights.AUC), and information on the respective AUC values of the ensemble model with the suggested weights for each of the (k) subsets. Suggested weights output.weights are calculated as the average of the weights of the different algorithms (submodels) of the k ensembles. Suggested weights output.weights.AUC are calculated as the average of the AUC of the different algorithms of the for the k runs.
Function ensemble.calibrate.models.gbm allows to test various combinations of parameters tree.complexity and learning.rate for the gbm.step model.
Function ensemble.calibrate.models.nnet allows to test various combinations of parameters size and decay for the nnet model.
Function ensemble.drop1 allows to test the effects of leaving out each of the explanatory variables, and comparing these results with the "full" model. Note that option of difference = TRUE may result in positive values, indicating that the model without the explanatory variable having larger AUC than the "full" model. A procedure is included to estimate the deviance of a model based on the fitted values, using -2 * (sum(x*log(x)) + sum((1-x)*log(1-x))) where x is a vector of the fitted values for a respective model. (It was checked that this procedure results in similar deviance estimates for the null and 'full' models for glm, but note that it is not certain whether deviance can be calculated in a similar way for other submodels.)
Function ensemble.formulae provides suggestions for formulae that can be used for ensemble.calibrate.models and ensemble.raster. This function is always used internally by the ensemble.drop1 function.
The ensemble.weights function is used internally by the ensemble.calibrate.models and ensemble.raster functions, using the input weights for the different suitability modelling algorithms. Ties between input weights result in the same output weights.
The ensemble.strategy function is used internally by the ensemble.calibrate.models function, using the train and test data sets with predictions of the different suitability modelling algorithms and different combinations of parameters ENSEMBLE.best, ENSEMBLE.min and ENSEMBLE.exponent. The final ensemble model is based on the parameters that generate the best AUC.
The ensemble.threshold function is used internally by the ensemble.calibrate.models, ensemble.mean and ensemble.plot functions. threshold2005.mean results in calculating the mean value of threshold methods that resulted in better results (calculated by optimal.thresholds with methods of ObsPrev, MeanProb, MaxSens+Spec, Sens=Spec and MinROCdist) in a study by Liu et al. (Ecography 28: 385-393. 2005). threshold2005.min results in calculating the mean value of threshold methods that resulted in better results (calculated by optimal.thresholds with methods of ObsPrev, MeanProb and MaxSens+Spec) in a study by Liu et al. (Ecography 28: 385-393. 2005). threshold2013.mean results in calculating the mean value of threshold methods that resulted in better results (calculated by optimal.thresholds with methods of ObsPrev, MeanProb, MaxSens+Spec, Sens=Spec and MinROCdist) in a study by Liu et al. (J. Biogeogr. 40: 778-789. 2013). threshold2013.min results in calculating the minimum value of threshold methods that resulted in better results (calculated by optimal.thresholds with methods of ObsPrev, MeanProb, MaxSens+Spec, Sens=Spec and MinROCdist) in a study by Liu et al. (J. Biogeogr. 40: 778-789. 2013).
Function ensemble.VIF implements a stepwise procedure whereby the explanatory variable with highest Variance Inflation Factor is removed from the list of variables. The procedure ends when no variable has VIF larger than parameter VIF.max.
Function ensemble.pairs provides a matrix of scatterplots similar to the example of pairs for version 3.4.3 of that package.
Function ensemble.calibrate.models (potentially) returns a list with results from evaluations (via evaluate) of calibration and test runs of individual suitability models.
Function ensemble.calibrate.weights returns a matrix with, for each individual suitability model, the AUC of each run and the average AUC over the runs. Models are sorted by the average AUC. The average AUC for each model can be used as input weights for the ensemble.raster function.
Functions ensemble.calibrate.models.gbm and ensemble.calibrate.models.nnet return a matrix with, for each combination of model parameters, the AUC of each run and the average AUC. Models are sorted by the average AUC.
Roeland Kindt (World Agroforestry Centre)
Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009
Buisson L, Thuiller W, Casajus N, Lek S and Grenouillet G. 2010. Uncertainty in ensemble forecasting of species distribution. Global Change Biology 16: 1145-1157
Liu C, Berry PM, Dawson TP and Pearson RC. 2005. Selecting thresholds of occurrence in the prediction of species distributions. Ecography 28: 385-393
Liu C, White M and Newell G. 2013. Selecting thresholds for the prediction of species occurrence with presence-only data. Journal of Biogeography 40: 778-789
Phillips SJ, Dudik M, Elith J et al. 2009. Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecological Applications 19: 181-197.
ensemble.raster, ensemble.batch
## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "predictors" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/4) and training (3/4) data groupp <- kfold(pres, 4) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 4) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # formulae for random forest and generalized linear model # compare with: ensemble.formulae(predictors, factors=c("biome")) rfformula <- as.formula(pb ~ bio5+bio6+bio16+bio17) glmformula <- as.formula(pb ~ bio5 + I(bio5^2) + I(bio5^3) + bio6 + I(bio6^2) + I(bio6^3) + bio16 + I(bio16^2) + I(bio16^3) + bio17 + I(bio17^2) + I(bio17^3) ) # fit four ensemble models (RF, GLM, BIOCLIM, DOMAIN) # factors removed for BIOCLIM, DOMAIN, MAHAL ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", ENSEMBLE.tune=TRUE, ENSEMBLE.min = 0.65, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE, models.keep=TRUE, RF.formula=rfformula, GLM.formula=glmformula) # with option models.keep, all model objects are saved in ensemble object # the same slots can be used to replace model objects with new calibrations ensemble.nofactors$models$RF summary(ensemble.nofactors$models$GLM) ensemble.nofactors$models$BIOCLIM ensemble.nofactors$models$DOMAIN ensemble.nofactors$models$formulae # evaluations are kept in different slot attributes(ensemble.nofactors$evaluations) plot(ensemble.nofactors$evaluations$RF.T, "ROC") # fit four ensemble models (RF, GLM, BIOCLIM, DOMAIN) using default formulae # variable 'biome' is not included as explanatory variable # results are provided in a file in the 'outputs' subfolder of the working # directory ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, layer.drops="biome", species.name="Bradypus", ENSEMBLE.tune=TRUE, ENSEMBLE.min=0.65, SINK=TRUE, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", evaluations.keep=TRUE, formulae.defaults=TRUE) # after fitting the individual algorithms (submodels), # transform predictions with a probit link. ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, layer.drops="biome", species.name="Bradypus", SINK=TRUE, ENSEMBLE.tune=TRUE, ENSEMBLE.min=0.65, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, PROBIT=TRUE, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE, formulae.defaults=TRUE) # Instead of providing presence and background locations, provide data.frames. # Because 'biome' is a factor, RasterStack needs to be provided # to check for levels in the Training and Testing data set. TrainData1 <- prepareData(x=predictors, p=pres_train, b=backg_train, factors=c("biome"), xy=FALSE) TestData1 <- prepareData(x=predictors, p=pres_test, b=backg_test, factors=c("biome"), xy=FALSE) ensemble.factors1 <- ensemble.calibrate.models(x=predictors, TrainData=TrainData1, TestData=TestData1, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", SINK=TRUE, ENSEMBLE.tune=TRUE, ENSEMBLE.min=0.65, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=0, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE) # compare different methods of calculating ensembles ensemble.factors2 <- ensemble.calibrate.models(x=predictors, TrainData=TrainData1, TestData=TestData1, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", SINK=TRUE, ENSEMBLE.tune=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.best=c(4:10), ENSEMBLE.exponent=c(1, 2, 3), Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE) # test performance of different suitability models # data are split in 4 subsets, each used once for evaluation ensemble.nofactors2 <- ensemble.calibrate.weights(x=predictors, p=pres, a=background, k=4, species.name="Bradypus", SINK=TRUE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, Yweights="BIOMOD", formulae.defaults=TRUE) ensemble.nofactors2$AUC.table ensemble.nofactors2$eval.table.all # test the result of leaving out one of the variables from the model # note that positive differences indicate that the model without the variable # has higher AUC than the full model ensemble.variables <- ensemble.drop1(x=predictors, p=pres, a=background, k=4, species.name="Bradypus", SINK=TRUE, difference=TRUE, VIF=TRUE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, Yweights="BIOMOD", factors="biome") ensemble.variables # use function ensemble.VIF to select a subset of variables # factor variables are not handled well by the function # and therefore factors are removed # however, one can check for factors with car::vif through # the ensemble.calibrate.models function # VIF.analysis$var.drops can be used as input for ensemble.calibrate.models or # ensemble.calibrate.weights predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio8", "bio12", "bio16", "bio17", "biome")) ensemble.pairs(predictors) VIF.analysis <- ensemble.VIF(predictors, factors="biome") VIF.analysis # alternative solution where bio1 and bio12 are kept VIF.analysis <- ensemble.VIF(predictors, factors="biome", keep=c("bio1", "bio12")) VIF.analysis ## End(Not run)## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "predictors" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/4) and training (3/4) data groupp <- kfold(pres, 4) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 4) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # formulae for random forest and generalized linear model # compare with: ensemble.formulae(predictors, factors=c("biome")) rfformula <- as.formula(pb ~ bio5+bio6+bio16+bio17) glmformula <- as.formula(pb ~ bio5 + I(bio5^2) + I(bio5^3) + bio6 + I(bio6^2) + I(bio6^3) + bio16 + I(bio16^2) + I(bio16^3) + bio17 + I(bio17^2) + I(bio17^3) ) # fit four ensemble models (RF, GLM, BIOCLIM, DOMAIN) # factors removed for BIOCLIM, DOMAIN, MAHAL ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", ENSEMBLE.tune=TRUE, ENSEMBLE.min = 0.65, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE, models.keep=TRUE, RF.formula=rfformula, GLM.formula=glmformula) # with option models.keep, all model objects are saved in ensemble object # the same slots can be used to replace model objects with new calibrations ensemble.nofactors$models$RF summary(ensemble.nofactors$models$GLM) ensemble.nofactors$models$BIOCLIM ensemble.nofactors$models$DOMAIN ensemble.nofactors$models$formulae # evaluations are kept in different slot attributes(ensemble.nofactors$evaluations) plot(ensemble.nofactors$evaluations$RF.T, "ROC") # fit four ensemble models (RF, GLM, BIOCLIM, DOMAIN) using default formulae # variable 'biome' is not included as explanatory variable # results are provided in a file in the 'outputs' subfolder of the working # directory ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, layer.drops="biome", species.name="Bradypus", ENSEMBLE.tune=TRUE, ENSEMBLE.min=0.65, SINK=TRUE, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", evaluations.keep=TRUE, formulae.defaults=TRUE) # after fitting the individual algorithms (submodels), # transform predictions with a probit link. ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, layer.drops="biome", species.name="Bradypus", SINK=TRUE, ENSEMBLE.tune=TRUE, ENSEMBLE.min=0.65, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, PROBIT=TRUE, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE, formulae.defaults=TRUE) # Instead of providing presence and background locations, provide data.frames. # Because 'biome' is a factor, RasterStack needs to be provided # to check for levels in the Training and Testing data set. TrainData1 <- prepareData(x=predictors, p=pres_train, b=backg_train, factors=c("biome"), xy=FALSE) TestData1 <- prepareData(x=predictors, p=pres_test, b=backg_test, factors=c("biome"), xy=FALSE) ensemble.factors1 <- ensemble.calibrate.models(x=predictors, TrainData=TrainData1, TestData=TestData1, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", SINK=TRUE, ENSEMBLE.tune=TRUE, ENSEMBLE.min=0.65, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=0, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE) # compare different methods of calculating ensembles ensemble.factors2 <- ensemble.calibrate.models(x=predictors, TrainData=TrainData1, TestData=TestData1, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", SINK=TRUE, ENSEMBLE.tune=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.best=c(4:10), ENSEMBLE.exponent=c(1, 2, 3), Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE) # test performance of different suitability models # data are split in 4 subsets, each used once for evaluation ensemble.nofactors2 <- ensemble.calibrate.weights(x=predictors, p=pres, a=background, k=4, species.name="Bradypus", SINK=TRUE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, Yweights="BIOMOD", formulae.defaults=TRUE) ensemble.nofactors2$AUC.table ensemble.nofactors2$eval.table.all # test the result of leaving out one of the variables from the model # note that positive differences indicate that the model without the variable # has higher AUC than the full model ensemble.variables <- ensemble.drop1(x=predictors, p=pres, a=background, k=4, species.name="Bradypus", SINK=TRUE, difference=TRUE, VIF=TRUE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, Yweights="BIOMOD", factors="biome") ensemble.variables # use function ensemble.VIF to select a subset of variables # factor variables are not handled well by the function # and therefore factors are removed # however, one can check for factors with car::vif through # the ensemble.calibrate.models function # VIF.analysis$var.drops can be used as input for ensemble.calibrate.models or # ensemble.calibrate.weights predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio8", "bio12", "bio16", "bio17", "biome")) ensemble.pairs(predictors) VIF.analysis <- ensemble.VIF(predictors, factors="biome") VIF.analysis # alternative solution where bio1 and bio12 are kept VIF.analysis <- ensemble.VIF(predictors, factors="biome", keep=c("bio1", "bio12")) VIF.analysis ## End(Not run)
Building on methodologies described by Pironon et al. (doi:10.1038/s41558-019-0585-7), function ensemble.concave.hull constructs two hulls in environmental space for the baseline and a changed (typically a future climate, but possibly also a historical or paleo-climate) for a focal species. Functions ensemble.concave.venn and ensemble.concave.union create a third hull for candidate accessions that represent different geographies and/or different species. Subsequently overlaps between hulls are investigated. Information is also provided for each accession of the focal species in the novel climate if these are included within the hull of the candidate accessions.
ensemble.concave.hull( baseline.data, change.data, complete.cases = TRUE, VIF = TRUE, VIF.max = 20, VIF.silent = TRUE, method = c("rda", "pca", "prcomp"), ax1 = 1, ax2 = 2, concavity = 2.5, buffer.dist = NA, ggplot = TRUE) ensemble.concave.venn( x, candidate.data, concavity = x$concavity, buffer.dist = x$buffer.dist, ggplot = TRUE, show.candidate.points = TRUE) ensemble.concave.union( x, candidate.venns, buffer.dist = x$buffer.dist, ggplot = TRUE, show.candidate.points = TRUE) ensemble.outliers( x, ID.var = NULL, bioc.vars = NULL, fence.k = 2.5, n_min = 5)ensemble.concave.hull( baseline.data, change.data, complete.cases = TRUE, VIF = TRUE, VIF.max = 20, VIF.silent = TRUE, method = c("rda", "pca", "prcomp"), ax1 = 1, ax2 = 2, concavity = 2.5, buffer.dist = NA, ggplot = TRUE) ensemble.concave.venn( x, candidate.data, concavity = x$concavity, buffer.dist = x$buffer.dist, ggplot = TRUE, show.candidate.points = TRUE) ensemble.concave.union( x, candidate.venns, buffer.dist = x$buffer.dist, ggplot = TRUE, show.candidate.points = TRUE) ensemble.outliers( x, ID.var = NULL, bioc.vars = NULL, fence.k = 2.5, n_min = 5)
baseline.data |
data.frame with climatic variables for the accessions in the baseline climate. |
change.data |
data.frame with climatic variables for the accessions in the changed (potentially future) climate. |
complete.cases |
Reduce cases with those without missing data via |
VIF |
Select a subset of climatic variables via |
VIF.max |
Argument setting for |
VIF.silent |
Argument setting for |
method |
Method of constructing the hulls; see details. |
ax1 |
Idex for the first ordination axis to be analyzed; see also |
ax2 |
Index for second ordination axis to be analyzed; see also |
concavity |
A relative measure of concavity used by |
buffer.dist |
Buffer width used internally by |
ggplot |
Should a ggplot object be included in the output? |
x |
Output similar to those of |
candidate.data |
data.frame with climatic variables for candidate accessions such as accessions from other geographical areas or other species. |
show.candidate.points |
Should the ggplot object show the locations of the candidate accessions? |
candidate.venns |
list with outputs from the |
ID.var |
Variable name used as identifier |
bioc.vars |
Variables included in the analysis of outliers |
fence.k |
Multiplier to calculate distance of observation from Interquartile lower and upper limits as used by Tukey's Fences method to detect outliers |
n_min |
Minimum number of variables for identifying outliers |
Whereas the metholology of Pironon et al. (2019) uses convex hulls, concave hulls can also be used in the methodology provided here. Convex hulls will be obtained by using large values for the concavity argument (see the description for the concaveman function). By using more concave hulls, the influence of outliers on measures of niche overlap can be reduced.
Three methods are available for mapping accessions in environmental space. Methods pca and prcomp use principal components analysis, respectively via the rda and prcomp functions. In both the methods, climatic variables are scaled. As results with pca are also rescaled via caprescale, both methods of pca and prcomp should theoretically result in the same configurations.
Method rda internally uses envfit to select a subset of climatic variables that are significantly correlated (P <= 0.05, R2 >= 0.50) with the first two axes of a redundancy analysis that uses the climate (baseline vs. changed) as predictor variable.
Candidate accessions are mapped in the environmental space created by ensemble.concave.hull via prediction methods available from predict.cca and predict.prcomp.
Function ensemble.concave.union combines candidate hulls obtained from ensemble.concave.venn, using st_union internally.
Both ensemble.concave.venn and ensemble.concave.union return measures of niche overlap based on areas of overlap between the candidate hull and the part of hull for the changed climate that is not covered by the hull for the baseline climate. These functions also indicate for each of the accessions of the focal species in the changed climate whether they occur in a novel climate (novel == TRUE; this information was obtained by ensemble.concave.hull) and whether they are inside the hull of the candidate accessions (candidate.in == TRUE).
The optional plot shows the locations of accessions for the changed climate. For ensemble.concave.hull, colouring is based on having novel climates (not occurring in the overlap between the two hulls) or not. For the other functions, locations are only shown for accessions with novel climates. Colouring is based on being inside the hull for the candidate accessions or not.
Function ensemble.outliers generalizes Tukey's fences method to require that a multivariate outlier is a univariate outlier for a minimum number of n_min variables (see )
Function ensemble.concave.hull returns a list with following elements:
- rda.object: result of the ordination method used; - method: method used in the function; - baseline.hull: polygon for the hull for the baseline climate; - baseline.area: area of the baseline hull; - change.hull: polygon for the hull for the changed climate; - change.area: area of the hull for the changed climate; - overlap.hull: polygon for the overlap (intersection) of the baseline and changed hull; - overlap.area: area of the overlap hull; - novel.hull: polygon for the part of the changed hull that does not cover the baseline hull; - change.area: area of the novel hull; - buffer.dist: distance used in checking whether accessions are in novel conditions; - change.points: plotting coordinates and details on novel conditions for accessions of the changed climate; - baseline.points: plotting coordinates for accessions of the baseline climate
Roeland Kindt (World Agroforestry Centre) and Maarten van Zonneveld (World Vegetable Center)
Pironon et al. (2019). Potential adaptive strategies for 29 sub-Saharan crops under future climate change. Nat. Clim. Chang. 9: 758-736. doi:10.1038/s41558-019-0585-7
van Zonneveld et al. (2018). Tree genetic resources at risk in South America: a spatial threat assessment to prioritize populations for conservation. Diversity and Distributions 24: 718-729
van Zonneveld et al. (2023). Forgotten food crops in sub-Saharan Africa for healthy diets in a changing climate. Proceedings of the National Academy of Sciences (PNAS) 120 (14) e2205794120. doi:10.1073/pnas.2205794120
## Not run: library(ggplot2) library(sf) library(concaveman) data(CucurbitaClim) alata.data <- CucurbitaClim[CucurbitaClim$species == "Cucurbita_palmata", ] bioc.names <- paste0("bioc", 1:19) alata.data2 <- alata.data[alata.data$ADM0_A3 == "USA", ] alata.base <- alata.data2[alata.data2$climate == "baseline", bioc.names] alata.fut <- alata.data2[alata.data2$climate == "future", bioc.names] conc2.res <- ensemble.concave.hull(baseline.data=alata.base, change.data=alata.fut, method="pca", VIF.max=40, concavity=2) plot(conc2.res$ggplot.out) conc2.res$baseline.area conc2.res$change.area conc2.res$novel.area conc2.res$novel.area / conc2.res$change.area # Which accessions have novel climates? summary(conc2.res$change.points) change.points <- conc2.res$change.points rownames(change.points[change.points$novel == TRUE, ]) nrow(change.points[change.points$novel == TRUE, ]) / nrow(change.points) # Analysis via convex hulls conc100.res <- ensemble.concave.hull(baseline.data=alata.base, change.data=alata.fut, method="pca", concavity=100) plot(conc100.res$ggplot.out) conc100.res$baseline.area conc100.res$change.area conc100.res$novel.area conc100.res$novel.area / conc100.res$change.area # Which accessions have novel climates? summary(conc100.res$change.points) change.points <- conc100.res$change.points rownames(change.points[change.points$novel == TRUE, ]) nrow(change.points[change.points$novel == TRUE, ]) / nrow(change.points) # Checking niche overlaps with other accessions # Alternative 1: niche overlap with accessions from Mexico alata.data2 <- alata.data[alata.data$ADM0_A3 == "MEX", ] alata.MEX <- alata.data2[alata.data2$climate == "baseline", bioc.names] venn2.res <- ensemble.concave.venn(conc2.res, candidate.data=alata.MEX, concavity=2) plot(venn2.res$ggplot.out) table(venn2.res$change.points[ , c("novel", "candidate.in")]) # alternative 1 for convex hulls venn100.res <- ensemble.concave.venn(conc100.res, candidate.data=alata.MEX, concavity=100) plot(venn100.res$ggplot.out) table(venn100.res$change.points[ , c("novel", "candidate.in")]) # alternative 2: niche overlap with other species cucurbita2 <- CucurbitaClim[CucurbitaClim$climate == "baseline", ] cordata.data <- cucurbita2[cucurbita2$species == "Cucurbita_cordata", bioc.names] digitata.data <- cucurbita2[cucurbita2$species == "Cucurbita_digitata", bioc.names] venn.cordata <- ensemble.concave.venn(conc2.res, candidate.data=cordata.data, concavity=2) plot(venn.cordata$ggplot.out) venn.digitata <- ensemble.concave.venn(conc2.res, candidate.data=digitata.data, concavity=2) plot(venn.digitata$ggplot.out) # check the union of the two species spec.res <- vector("list", 2) spec.res[[1]] <- venn.cordata spec.res[[2]] <- venn.digitata union2.res <- ensemble.concave.union(conc2.res, candidate.venns=spec.res) table(union2.res$change.points[ , c("novel", "candidate.in")]) # Analysis via convex hulls venn.digitata <- ensemble.concave.venn(conc100.res, candidate.data=digitata.data, concavity=100) venn.cordata <- ensemble.concave.venn(conc100.res, candidate.data=cordata.data, concavity=100) spec.res <- vector("list", 2) spec.res[[1]] <- venn.cordata spec.res[[2]] <- venn.digitata union100.res <- ensemble.concave.union(conc100.res, candidate.venns=spec.res) plot(union100.res$ggplot.out) table(union100.res$change.points[ , c("novel", "candidate.in")]) # Identify outliers baseline.outliers <- ensemble.outliers(alata.base, bioc.vars=paste0("bioc", 1:19)) baseline.outliers[baseline.outliers$outlier == TRUE, ] ## End(Not run)## Not run: library(ggplot2) library(sf) library(concaveman) data(CucurbitaClim) alata.data <- CucurbitaClim[CucurbitaClim$species == "Cucurbita_palmata", ] bioc.names <- paste0("bioc", 1:19) alata.data2 <- alata.data[alata.data$ADM0_A3 == "USA", ] alata.base <- alata.data2[alata.data2$climate == "baseline", bioc.names] alata.fut <- alata.data2[alata.data2$climate == "future", bioc.names] conc2.res <- ensemble.concave.hull(baseline.data=alata.base, change.data=alata.fut, method="pca", VIF.max=40, concavity=2) plot(conc2.res$ggplot.out) conc2.res$baseline.area conc2.res$change.area conc2.res$novel.area conc2.res$novel.area / conc2.res$change.area # Which accessions have novel climates? summary(conc2.res$change.points) change.points <- conc2.res$change.points rownames(change.points[change.points$novel == TRUE, ]) nrow(change.points[change.points$novel == TRUE, ]) / nrow(change.points) # Analysis via convex hulls conc100.res <- ensemble.concave.hull(baseline.data=alata.base, change.data=alata.fut, method="pca", concavity=100) plot(conc100.res$ggplot.out) conc100.res$baseline.area conc100.res$change.area conc100.res$novel.area conc100.res$novel.area / conc100.res$change.area # Which accessions have novel climates? summary(conc100.res$change.points) change.points <- conc100.res$change.points rownames(change.points[change.points$novel == TRUE, ]) nrow(change.points[change.points$novel == TRUE, ]) / nrow(change.points) # Checking niche overlaps with other accessions # Alternative 1: niche overlap with accessions from Mexico alata.data2 <- alata.data[alata.data$ADM0_A3 == "MEX", ] alata.MEX <- alata.data2[alata.data2$climate == "baseline", bioc.names] venn2.res <- ensemble.concave.venn(conc2.res, candidate.data=alata.MEX, concavity=2) plot(venn2.res$ggplot.out) table(venn2.res$change.points[ , c("novel", "candidate.in")]) # alternative 1 for convex hulls venn100.res <- ensemble.concave.venn(conc100.res, candidate.data=alata.MEX, concavity=100) plot(venn100.res$ggplot.out) table(venn100.res$change.points[ , c("novel", "candidate.in")]) # alternative 2: niche overlap with other species cucurbita2 <- CucurbitaClim[CucurbitaClim$climate == "baseline", ] cordata.data <- cucurbita2[cucurbita2$species == "Cucurbita_cordata", bioc.names] digitata.data <- cucurbita2[cucurbita2$species == "Cucurbita_digitata", bioc.names] venn.cordata <- ensemble.concave.venn(conc2.res, candidate.data=cordata.data, concavity=2) plot(venn.cordata$ggplot.out) venn.digitata <- ensemble.concave.venn(conc2.res, candidate.data=digitata.data, concavity=2) plot(venn.digitata$ggplot.out) # check the union of the two species spec.res <- vector("list", 2) spec.res[[1]] <- venn.cordata spec.res[[2]] <- venn.digitata union2.res <- ensemble.concave.union(conc2.res, candidate.venns=spec.res) table(union2.res$change.points[ , c("novel", "candidate.in")]) # Analysis via convex hulls venn.digitata <- ensemble.concave.venn(conc100.res, candidate.data=digitata.data, concavity=100) venn.cordata <- ensemble.concave.venn(conc100.res, candidate.data=cordata.data, concavity=100) spec.res <- vector("list", 2) spec.res[[1]] <- venn.cordata spec.res[[2]] <- venn.digitata union100.res <- ensemble.concave.union(conc100.res, candidate.venns=spec.res) plot(union100.res$ggplot.out) table(union100.res$change.points[ , c("novel", "candidate.in")]) # Identify outliers baseline.outliers <- ensemble.outliers(alata.base, bioc.vars=paste0("bioc", 1:19)) baseline.outliers[baseline.outliers$outlier == TRUE, ] ## End(Not run)
The basic function ensemble.dummy.variables creates new raster layers representing dummy variables (coded 0 or 1) for all or the most frequent levels of a caterogical variable. Sometimes the creation of dummy variables is needed for proper handling of categorical data for some of the suitability modelling algorithms.
ensemble.dummy.variables(xcat = NULL, freq.min = 50, most.frequent = 5, new.levels = NULL, overwrite = TRUE, ...) ensemble.accepted.categories(xcat = NULL, categories = NULL, filename = NULL, overwrite = TRUE, ...) ensemble.simplified.categories(xcat = NULL, p = NULL, filename = NULL, overwrite = TRUE, ...)ensemble.dummy.variables(xcat = NULL, freq.min = 50, most.frequent = 5, new.levels = NULL, overwrite = TRUE, ...) ensemble.accepted.categories(xcat = NULL, categories = NULL, filename = NULL, overwrite = TRUE, ...) ensemble.simplified.categories(xcat = NULL, p = NULL, filename = NULL, overwrite = TRUE, ...)
xcat |
RasterLayer object ( |
freq.min |
Minimum frequency for a dummy raster layer to be created for the corresponding factor level. See also |
most.frequent |
Number of dummy raster layers to be created (if larger than 0), corresponding to the same number of most frequent factor levels See also |
new.levels |
character vector giving factor levels that are not encountered in |
overwrite |
overwrite an existing file name with the same name (if |
... |
additional arguments for |
categories |
numeric vector providing the accepted levels of a categorical raster layer; expected to correspond to the levels encountered during calibration |
filename |
name for the output file. See also |
p |
presence points that will be used for calibrating the suitability models, typically available in 2-column (x, y) or (lon, lat) dataframe; see also |
The basic function ensemble.dummy.variables creates dummy variables from a RasterLayer object (see raster) that represents a categorical variable. With freq.min and most.frequent it is possible to limit the number of dummy variables that will be created. For example, most.frequent = 5 results in five dummy variables to be created.
Function ensemble.accepted.categories modifies the RasterLayer object (see raster) by replacing cell values for categories (levels) that are not accepted with missing values.
Function ensemble.simplified.categories modifies the RasterLayer object (see raster) by replacing cell values for categories (levels) where none of the presence points occur with the same level. This new level is coded by the maximum coding level for these 'outside categories'.
The basic function ensemble.dummy.variables mainly results in the creation of raster layers that correspond to dummy variables.
Roeland Kindt (World Agroforestry Centre) and Evert Thomas (Bioversity International)
ensemble.calibrate.models, ensemble.raster
## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) biome.layer <- predictors[["biome"]] biome.layer # create dummy layers for the 5 most frequent factor levels ensemble.dummy.variables(xcat=biome.layer, most.frequent=5, overwrite=TRUE) # check whether dummy variables were created predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors names(predictors) # once dummy variables were created, avoid using the original categorical data layer predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome_1", "biome_2", "biome_7", "biome_8", "biome_13")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/5) and training (4/5) data groupp <- kfold(pres, 5) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 5) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # note that dummy variables with no variation are not used by DOMAIN # note that dummy variables are not used by MAHAL and MAHAL01 # (neither are categorical variables) ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", VIF=T, MAXENT=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, GLM=1, GLMSTEP=0, GAM=1, GAMSTEP=0, MGCV=1, MGCVFIX=0, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, Yweights="BIOMOD", dummy.vars=c("biome_1", "biome_2", "biome_7", "biome_8", "biome_13"), PLOTS=FALSE, evaluations.keep=TRUE) ## End(Not run)## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) biome.layer <- predictors[["biome"]] biome.layer # create dummy layers for the 5 most frequent factor levels ensemble.dummy.variables(xcat=biome.layer, most.frequent=5, overwrite=TRUE) # check whether dummy variables were created predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors names(predictors) # once dummy variables were created, avoid using the original categorical data layer predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome_1", "biome_2", "biome_7", "biome_8", "biome_13")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/5) and training (4/5) data groupp <- kfold(pres, 5) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 5) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # note that dummy variables with no variation are not used by DOMAIN # note that dummy variables are not used by MAHAL and MAHAL01 # (neither are categorical variables) ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", VIF=T, MAXENT=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, GLM=1, GLMSTEP=0, GAM=1, GAMSTEP=0, MGCV=1, MGCVFIX=0, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, Yweights="BIOMOD", dummy.vars=c("biome_1", "biome_2", "biome_7", "biome_8", "biome_13"), PLOTS=FALSE, evaluations.keep=TRUE) ## End(Not run)
Function ensemble.ecocrop creates the map with novel conditions. Function ensemble.novel.object provides the reference values used by the prediction function used by predict .
ensemble.ecocrop(x = NULL, ecocrop.object = NULL, RASTER.object.name = ecocrop.object$name, RASTER.stack.name = "xTitle", RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, CATCH.OFF = FALSE) ensemble.ecocrop.object(temp.thresholds, rain.thresholds, name = "crop01", temp.multiply = 1, annual.temps = TRUE, transform = 1)ensemble.ecocrop(x = NULL, ecocrop.object = NULL, RASTER.object.name = ecocrop.object$name, RASTER.stack.name = "xTitle", RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, CATCH.OFF = FALSE) ensemble.ecocrop.object(temp.thresholds, rain.thresholds, name = "crop01", temp.multiply = 1, annual.temps = TRUE, transform = 1)
x |
RasterStack object ( |
ecocrop.object |
Object listing optimal and absolute minima and maxima for the rainfall and temperature values, used by the prediction function that is used internally by |
RASTER.object.name |
First part of the names of the raster file that will be generated, expected to identify the species or crop for which ranges were calculated |
RASTER.stack.name |
Last part of the names of the raster file that will be generated, expected to identify the predictor stack used |
RASTER.format |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
CATCH.OFF |
Disable calls to function |
temp.thresholds |
Optimal and absolute thresholds for temperatures. These will be sorted as: absolute minimum temperature, optimal minimum temperature, optimal maximum temperature and absolute maximum temperature. |
rain.thresholds |
Optimal and absolute thresholds for annual rainfall. These will be sorted as: absolute minimum rainfall, optimal minimum rainfall, optimal maximum rainfall and absolute maximum rainfall. |
name |
Name of the object, expect to expected to identify the species or crop |
temp.multiply |
Multiplier for temperature values. The value of 10 is to be used with raster layers where temperature was multiplied by 10 such as Worldclim version 1 or AFRICLIM. |
annual.temps |
If |
transform |
Exponent used to transform probability values obtained from interpolating between optimal and absolute limits. Exponent of 2 results in squaring probabilities, for example input probabilities of 0.5 transformed to 0.5^2 = 0.25. |
Function ensemble.ecocrop maps suitability for a species or crop based on optimal and absolute temperature and rainfall limits. Where both temperature and rainfall are within the optimal limits, suitability of 1000 is calculated. Where both temperature and rainfall are outside the absolute limits, suitability of 0 is calculated. In situations where temperature or rainfall is in between the optimal and absolute limits, then suitability is interpolated between 0 and 1000, and the lowest suitability from temperature and rainfall is calculated. Setting very wide rainfall limits will simulate the effect of irrigation, i.e. where suitability only depends on temperature limits.
For a large range of crop and plant species, optimal and absolute limits are available from the FAO ecocrop database (https://gaez.fao.org/pages/ecocrop-search), hence the name of the function. A different implementation of suitability mapping based on ecocrop limits is available from ecocrop. Ecocrop thresholds for several species are available from: getCrop
Function ensemble.ecocrop.object returns a list with following objects:
name |
name for the crop or species |
temp.thresholds |
optimal and absolute minimum and maximum temperature limits |
rain.thresholds |
optimal and absolute minimum and maximum annual rainfall limits |
annual.temps |
logical indicating whether temperature limits apply to annual temperatures |
transform |
exponent to transform suitability values |
Roeland Kindt (World Agroforestry Centre)
## Not run: # test with Brazil nut (limits from FAO ecocrop) # temperature: (12) 20-36 (40) # annnual rainfall: (1400) 2400-2800 (3500) # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio12")) predictors predictors@title <- "base" # As the raster data correspond to WorldClim version 1, # the temperatures need to be multiplied by 10 Brazil.ecocrop <- ensemble.ecocrop.object(temp.thresholds=c(20, 36, 12, 40), rain.thresholds=c(2400, 2800, 1400, 3500), temp.multiply=10, annual.temps=FALSE, name="Bertholletia_excelsa") Brazil.ecocrop ensemble.ecocrop(predictors, ecocrop.object=Brazil.ecocrop, RASTER.stack.name="base") dev.new() par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) rasterfull1 <- paste("ensembles//ecocrop//Bertholletia_excelsa_base.tif", sep="") rasterfull1 <- raster(rasterfull1) # raster file saved probabilities as integer values between 0 and 1000 rasterfull1 <- rasterfull1/1000 raster::plot(rasterfull1, main="Ecocrop suitability") GBIFloc <- gbif(genus="Bertholletia", species="excelsa", geo=TRUE) GBIFpres <- GBIFloc[, c("lon", "lat")] GBIFpres <- GBIFpres[complete.cases(GBIFpres), ] GBIFpres <- GBIFpres[duplicated(GBIFpres) == FALSE, ] point.suitability <- extract(rasterfull1, y=GBIFpres) point.suitability[is.na(point.suitability)] <- -1 GBIFpres.optimal <- GBIFpres[point.suitability == 1, ] GBIFpres.suboptimal <- GBIFpres[point.suitability < 1 & point.suitability > 0, ] GBIFpres.not <- GBIFpres[point.suitability == 0, ] raster::plot(rasterfull1, main="GBIF locations", sub="blue: optimal, cyan: suboptimal, red: not suitable") bg.legend <- c("blue", "cyan", "red") points(GBIFpres.suboptimal, pch=21, cex=1.2, bg=bg.legend[2]) points(GBIFpres.optimal, pch=21, cex=1.2, bg=bg.legend[1]) points(GBIFpres.not, pch=21, cex=1.2, bg=bg.legend[3]) graphics::par(par.old) ## End(Not run)## Not run: # test with Brazil nut (limits from FAO ecocrop) # temperature: (12) 20-36 (40) # annnual rainfall: (1400) 2400-2800 (3500) # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio12")) predictors predictors@title <- "base" # As the raster data correspond to WorldClim version 1, # the temperatures need to be multiplied by 10 Brazil.ecocrop <- ensemble.ecocrop.object(temp.thresholds=c(20, 36, 12, 40), rain.thresholds=c(2400, 2800, 1400, 3500), temp.multiply=10, annual.temps=FALSE, name="Bertholletia_excelsa") Brazil.ecocrop ensemble.ecocrop(predictors, ecocrop.object=Brazil.ecocrop, RASTER.stack.name="base") dev.new() par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) rasterfull1 <- paste("ensembles//ecocrop//Bertholletia_excelsa_base.tif", sep="") rasterfull1 <- raster(rasterfull1) # raster file saved probabilities as integer values between 0 and 1000 rasterfull1 <- rasterfull1/1000 raster::plot(rasterfull1, main="Ecocrop suitability") GBIFloc <- gbif(genus="Bertholletia", species="excelsa", geo=TRUE) GBIFpres <- GBIFloc[, c("lon", "lat")] GBIFpres <- GBIFpres[complete.cases(GBIFpres), ] GBIFpres <- GBIFpres[duplicated(GBIFpres) == FALSE, ] point.suitability <- extract(rasterfull1, y=GBIFpres) point.suitability[is.na(point.suitability)] <- -1 GBIFpres.optimal <- GBIFpres[point.suitability == 1, ] GBIFpres.suboptimal <- GBIFpres[point.suitability < 1 & point.suitability > 0, ] GBIFpres.not <- GBIFpres[point.suitability == 0, ] raster::plot(rasterfull1, main="GBIF locations", sub="blue: optimal, cyan: suboptimal, red: not suitable") bg.legend <- c("blue", "cyan", "red") points(GBIFpres.suboptimal, pch=21, cex=1.2, bg=bg.legend[2]) points(GBIFpres.optimal, pch=21, cex=1.2, bg=bg.legend[1]) points(GBIFpres.not, pch=21, cex=1.2, bg=bg.legend[3]) graphics::par(par.old) ## End(Not run)
envirem package for data.frames.
Function generateEnvirem uses RasterStack (stack) objects as input and also generates outputs in the same format. The functions described here can be used to generate the bioclimatic variables for data.frames while using envirem functions internally. This feature can be useful in situations where models are calibrated with higher resolution data, but where maps will only be generated in lower resolutions, thus avoiding the need to generate the higher resolution envirem layers first.
ensemble.envirem.masterstack( x, precipstack, tmaxstack, tminstack, tmeanstack = NULL, envirem3 = TRUE) ensemble.envirem.solradstack( x, solrad, envirem3 = TRUE) ensemble.envirem.run( masterstack, solradstack, var = "all", ...)ensemble.envirem.masterstack( x, precipstack, tmaxstack, tminstack, tmeanstack = NULL, envirem3 = TRUE) ensemble.envirem.solradstack( x, solrad, envirem3 = TRUE) ensemble.envirem.run( masterstack, solradstack, var = "all", ...)
x |
Point locations provided in 2-column (eg, LON-LAT) format. |
precipstack |
RasterStack object ( |
tmaxstack |
RasterStack object ( |
tminstack |
RasterStack object ( |
tmeanstack |
RasterStack object ( |
envirem3 |
generate a SpatRaster object ( |
solrad |
RasterStack object ( |
masterstack |
RasterStack object ( |
solradstack |
RasterStack object ( |
var |
Names of bioclimatic variables to be created; see: |
... |
Other arguments for |
The objective of these functions is to expand a data.frame of explanatory variables that will be used for calibrating species distribution models with bioclimatic variables that are generated by the envirem package (See details in generateEnvirem).
It is important that monthly values are sorted sequentially (January - December) as the ensemble.envirem.masterstack and ensemble.envirem.solradstack functions expect the inputs to be sorted in this order.
Function ensemble.envirem.solradstack requires monthly extraterrestrial solar radiation layers at the same resolution as the climatic layers. It is possible, however, to also calculate these values directly for point observation data as shown below in the examples.
Function ensemble.envirem.run returns a data.frame with bioclimatic variables for each point location.
Roeland Kindt (World Agroforestry Centre)
Title P.O., Bemmels J.B. 2018. ENVIREM: An expanded set of bioclimatic and topographic variables increases flexibility and improves performance of ecological niche modeling. Ecography 41: 291-307.
Kindt R. 2023. TreeGOER: A database with globally observed environmental ranges for 48,129 tree species. Global Change Biology. doi:10.1111/gcb.16914
generateEnvirem, ensemble.calibrate.models, ensemble.calibrate.weights
## Not run: # Based on examples in envirem package for envirem::generateEnvirem # Modified Sep-2023 due to change in function name in envirem library(terra) library(envirem) # Find example rasters rasterFiles <- list.files(system.file('extdata', package='envirem'), full.names=TRUE) precip.files <- rasterFiles[grepl(pattern="prec_", x=rasterFiles)] precip.files <- precip.files[c(1, 5:12, 2:4)] precip.stack <- terra::rast(precip.files) precip.stack names(precip.stack) tmin.files <- rasterFiles[grepl(pattern="tmin_", x=rasterFiles)] tmin.files <- tmin.files[c(1, 5:12, 2:4)] tmin.stack <- terra::rast(tmin.files) tmin.stack names(tmin.stack) tmax.files <- rasterFiles[grepl(pattern="tmax_", x=rasterFiles)] tmax.files <- tmax.files[c(1, 5:12, 2:4)] tmax.stack <- terra::rast(tmax.files) tmax.stack names(tmax.stack) tmean.files <- rasterFiles[grepl(pattern="tmean_", x=rasterFiles)] tmean.files <- tmean.files[c(1, 5:12, 2:4)] tmean.stack <- terra::rast(tmean.files) tmean.stack names(tmean.stack) # Random locations locs <- dismo::randomPoints(raster::stack(precip.stack[[1]]), n=50) # Climatic data master.input <- ensemble.envirem.masterstack(x=locs, precipstack=precip.stack, tmaxstack=tmax.stack, tminstack=tmin.stack, tmeanstack=tmean.stack) # Calculate solar radiation for 1975 # (Use other midpoint for the 1970-2000 WorldClim 2.1 baseline) solrad.stack <- ETsolradRasters(precip.stack[[1]], year = 1975-1950, outputDir = NULL) solrad.input <- ensemble.envirem.solradstack(x=locs, solrad=solrad.stack) # Obtain the envirem bioclimatic data envirem.data1 <- ensemble.envirem.run(masterstack=master.input, solradstack=solrad.input, tempScale=10) # Generate all the envirem layers, then extract # See envirem package for envirem::generateEnvirem worldclim <- rast(c(precip.files, tmax.files, tmin.files, tmean.files)) names(worldclim) assignNames(precip = 'prec_##') # generate all possible envirem variables envirem.stack <- generateEnvirem(worldclim, solrad.stack, var='all', tempScale = 10) # set back to defaults assignNames(reset = TRUE) envirem.data2 <- extract(envirem.stack, y=locs) # compare envirem.data1 - envirem.data2 # Calculate extraterrestrial solar radiation for point observations solrad1 <- extract(solrad.stack, y=locs) solrad2 <- array(dim=c(nrow(locs), 12)) for (i in 1:nrow(locs)) { lat.i <- locs[i, 2] for (m in 1:12) { solrad2[i, m] <- envirem:::calcSolRad(year=1975-1950, lat=lat.i, month=m) } } solrad1 - solrad2 ## End(Not run)## Not run: # Based on examples in envirem package for envirem::generateEnvirem # Modified Sep-2023 due to change in function name in envirem library(terra) library(envirem) # Find example rasters rasterFiles <- list.files(system.file('extdata', package='envirem'), full.names=TRUE) precip.files <- rasterFiles[grepl(pattern="prec_", x=rasterFiles)] precip.files <- precip.files[c(1, 5:12, 2:4)] precip.stack <- terra::rast(precip.files) precip.stack names(precip.stack) tmin.files <- rasterFiles[grepl(pattern="tmin_", x=rasterFiles)] tmin.files <- tmin.files[c(1, 5:12, 2:4)] tmin.stack <- terra::rast(tmin.files) tmin.stack names(tmin.stack) tmax.files <- rasterFiles[grepl(pattern="tmax_", x=rasterFiles)] tmax.files <- tmax.files[c(1, 5:12, 2:4)] tmax.stack <- terra::rast(tmax.files) tmax.stack names(tmax.stack) tmean.files <- rasterFiles[grepl(pattern="tmean_", x=rasterFiles)] tmean.files <- tmean.files[c(1, 5:12, 2:4)] tmean.stack <- terra::rast(tmean.files) tmean.stack names(tmean.stack) # Random locations locs <- dismo::randomPoints(raster::stack(precip.stack[[1]]), n=50) # Climatic data master.input <- ensemble.envirem.masterstack(x=locs, precipstack=precip.stack, tmaxstack=tmax.stack, tminstack=tmin.stack, tmeanstack=tmean.stack) # Calculate solar radiation for 1975 # (Use other midpoint for the 1970-2000 WorldClim 2.1 baseline) solrad.stack <- ETsolradRasters(precip.stack[[1]], year = 1975-1950, outputDir = NULL) solrad.input <- ensemble.envirem.solradstack(x=locs, solrad=solrad.stack) # Obtain the envirem bioclimatic data envirem.data1 <- ensemble.envirem.run(masterstack=master.input, solradstack=solrad.input, tempScale=10) # Generate all the envirem layers, then extract # See envirem package for envirem::generateEnvirem worldclim <- rast(c(precip.files, tmax.files, tmin.files, tmean.files)) names(worldclim) assignNames(precip = 'prec_##') # generate all possible envirem variables envirem.stack <- generateEnvirem(worldclim, solrad.stack, var='all', tempScale = 10) # set back to defaults assignNames(reset = TRUE) envirem.data2 <- extract(envirem.stack, y=locs) # compare envirem.data1 - envirem.data2 # Calculate extraterrestrial solar radiation for point observations solrad1 <- extract(solrad.stack, y=locs) solrad2 <- array(dim=c(nrow(locs), 12)) for (i in 1:nrow(locs)) { lat.i <- locs[i, 2] for (m in 1:12) { solrad2[i, m] <- envirem:::calcSolRad(year=1975-1950, lat=lat.i, month=m) } } solrad1 - solrad2 ## End(Not run)
The main function of ensemble.evaluate calculates various model evaluation statistics. Function ENSEMBLE.SEDI calculates the Symmetric Extremal Dependence Index (SEDI) from the True Positive Rate (TPR = Sensitivity = Hit Rate) and the False Positive Rate (FPR = False Alarm Rate = 1 - Specificity).
ensemble.evaluate(eval, fixed.threshold = NULL, eval.train = NULL) ensemble.SEDI(TPR, FPR, small = 1e-9) ensemble.Tjur(eval)ensemble.evaluate(eval, fixed.threshold = NULL, eval.train = NULL) ensemble.SEDI(TPR, FPR, small = 1e-9) ensemble.Tjur(eval)
eval |
ModelEvaluation object ( |
fixed.threshold |
Absence-presence threshold to create the confusion matrix. See also ( |
eval.train |
ModelEvaluation object ( |
TPR |
True Presence Rate, equal to correctly predicted presence observations divided by total number of presence observations. Also known as Sensitivity or Hit Rate. |
FPR |
False Presence Rate, equal to wrongly predicted absence observations divided by total number of absence observations. Also known as False Alarm Rate. |
small |
small amount that replaces zeroes in calculations. |
Details for the True Skill Statistic (TSS = TPR + TNR - 1 = TPR - FPR), Symmetric Extremal Dependence Index (SEDI), False Negative Rate (omission or miss rate) and AUCdiff (AUCtrain - AUCtest) are available from Ferro and Stephenson (2011), Wunderlich et al. (2019) and Castellanos et al. (2019).
Tjur's (2009) coefficient of discrimination is calculated as the differences between the averages of fitted values for successes and failures (see also Erikson & Smith 2023).
Values for TSS and SEDI are given for the fixed absence-presence threshold, as well as their maximal values across the entire range of candidate threshold values calculate by evaluate.
In case that fixed.threshold is not provided, it is calculated from the calibration ModelEvaluation as the threshold that maximizes the sum of TPR (sensitivity) and TNR (specificity) (and thus also maximizes the TSS for the calibration).
A numeric vector with following values.
- AUC: Area Under The Curve for the testing ModelEvaluation
- TSS: maximum value of the True Skill Statistic over range of threshold values
- SEDI: maximum value of the Symmetric Extremal Dependence Index over range of threshold values
- TSS.fixed: True Skill Statistic at the fixed threshold value
- SEDI.fixed: SEDI at the fixed threshold value
- FNR.fixed: False Negative Rate (= omission rate) at the fixed threshold value
- MCR.fixed: Missclassification Rate at the fixed threshold value
- AUCdiff: Difference between AUC for calibration and the testing data
- Tjur: Coefficient of Discrimination proposed by Tjur (2009)
Roeland Kindt (World Agroforestry Centre)
Ferro CA, Stephenson DB. 2011. Extremal Dependence Indices: Improved Verification Measures for Deterministic Forecasts of Rare Binary Events. Wea. Forecasting 26: 699-713.
Wunderlich RF, Lin Y-P, Anthony J, Petway JR. 2019. Two alternative evaluation metrics to replace the true skill statistic in the assessment of species distribution models. Nature Conservation 35: 97-116. doi:10.3897/natureconservation.35.33918
Castellanos AA, Huntley JW, Voelker G, Lawing AM. 2019. Environmental filtering improves ecological niche models across multiple scales. Methods in Ecology and Evolution 10: 481-492.
Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009
Tjur T. 2009. Coefficient of determination in logistic regression models - a new proposal: the coefficient of discrimination. The American Statistician 63: 366-372. doi:10.1198/tast.2009.08210
Erickson KD, Smith AB. 2023. Modelling the rarest of the rare: a comparison between multi-species distribution models, ensembles of small models, and single-species models at extremely low sample sizes. Ecography e06500 doi:10.1111/ecog.06500
## check examples from Ferro and Stephenson (2011) ## see their Tables 2 - 5 TPR.Table2 <- 55/100 FPR.Table2 <- 45/900 ensemble.SEDI(TPR=TPR.Table2, FPR=FPR.Table2) TPR.Table4 <- 195/300 FPR.Table4 <- 105/700 ensemble.SEDI(TPR=TPR.Table4, FPR=FPR.Table4) ## Not run: ## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "predictors" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/4) and training (3/4) data groupp <- kfold(pres, 4) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 4) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # formulae for random forest and generalized linear model # compare with: ensemble.formulae(predictors, factors=c("biome")) rfformula <- as.formula(pb ~ bio5+bio6+bio16+bio17) glmformula <- as.formula(pb ~ bio5 + I(bio5^2) + I(bio5^3) + bio6 + I(bio6^2) + I(bio6^3) + bio16 + I(bio16^2) + I(bio16^3) + bio17 + I(bio17^2) + I(bio17^3) ) # fit four ensemble models (RF, GLM, BIOCLIM, DOMAIN) # factors removed for BIOCLIM, DOMAIN, MAHAL ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", ENSEMBLE.tune=TRUE, ENSEMBLE.min = 0.65, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE, models.keep=FALSE, RF.formula=rfformula, GLM.formula=glmformula) # with option evaluations.keep, all model evaluations are saved in the ensemble object attributes(ensemble.nofactors$evaluations) # Get evaluation statistics for the ENSEMBLE model eval.ENSEMBLE <- ensemble.nofactors$evaluations$ENSEMBLE.T eval.calibrate.ENSEMBLE <- ensemble.nofactors$evaluations$ENSEMBLE.C ensemble.evaluate(eval=eval.ENSEMBLE, eval.train=eval.calibrate.ENSEMBLE) # TSS is maximum where specificity + sensitivity is maximum threshold.specsens <- threshold(eval.ENSEMBLE, stat="spec_sens") ensemble.evaluate(eval=eval.ENSEMBLE, fixed.threshold=threshold.specsens, eval.train=eval.calibrate.ENSEMBLE) # usual practice to calculate threshold from calibration data ensemble.evaluate(eval=eval.ENSEMBLE, eval.train=eval.calibrate.ENSEMBLE) ## End(Not run)## check examples from Ferro and Stephenson (2011) ## see their Tables 2 - 5 TPR.Table2 <- 55/100 FPR.Table2 <- 45/900 ensemble.SEDI(TPR=TPR.Table2, FPR=FPR.Table2) TPR.Table4 <- 195/300 FPR.Table4 <- 105/700 ensemble.SEDI(TPR=TPR.Table4, FPR=FPR.Table4) ## Not run: ## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "predictors" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/4) and training (3/4) data groupp <- kfold(pres, 4) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 4) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # formulae for random forest and generalized linear model # compare with: ensemble.formulae(predictors, factors=c("biome")) rfformula <- as.formula(pb ~ bio5+bio6+bio16+bio17) glmformula <- as.formula(pb ~ bio5 + I(bio5^2) + I(bio5^3) + bio6 + I(bio6^2) + I(bio6^3) + bio16 + I(bio16^2) + I(bio16^3) + bio17 + I(bio17^2) + I(bio17^3) ) # fit four ensemble models (RF, GLM, BIOCLIM, DOMAIN) # factors removed for BIOCLIM, DOMAIN, MAHAL ensemble.nofactors <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, species.name="Bradypus", ENSEMBLE.tune=TRUE, ENSEMBLE.min = 0.65, MAXENT=0, MAXNET=0, MAXLIKE=0, GBM=0, GBMSTEP=0, RF=1, CF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", factors="biome", evaluations.keep=TRUE, models.keep=FALSE, RF.formula=rfformula, GLM.formula=glmformula) # with option evaluations.keep, all model evaluations are saved in the ensemble object attributes(ensemble.nofactors$evaluations) # Get evaluation statistics for the ENSEMBLE model eval.ENSEMBLE <- ensemble.nofactors$evaluations$ENSEMBLE.T eval.calibrate.ENSEMBLE <- ensemble.nofactors$evaluations$ENSEMBLE.C ensemble.evaluate(eval=eval.ENSEMBLE, eval.train=eval.calibrate.ENSEMBLE) # TSS is maximum where specificity + sensitivity is maximum threshold.specsens <- threshold(eval.ENSEMBLE, stat="spec_sens") ensemble.evaluate(eval=eval.ENSEMBLE, fixed.threshold=threshold.specsens, eval.train=eval.calibrate.ENSEMBLE) # usual practice to calculate threshold from calibration data ensemble.evaluate(eval=eval.ENSEMBLE, eval.train=eval.calibrate.ENSEMBLE) ## End(Not run)
Function ensemble.novel creates the map with novel conditions. Function ensemble.novel.object provides the reference values used by the prediction function used by predict .
ensemble.novel(x = NULL, novel.object = NULL, RASTER.object.name = novel.object$name, RASTER.stack.name = x@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, CATCH.OFF = FALSE) ensemble.novel.object(x = NULL, name = "reference1", mask.raster = NULL, quantiles = FALSE, probs = c(0.05, 0.95), factors = NULL)ensemble.novel(x = NULL, novel.object = NULL, RASTER.object.name = novel.object$name, RASTER.stack.name = x@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, CATCH.OFF = FALSE) ensemble.novel.object(x = NULL, name = "reference1", mask.raster = NULL, quantiles = FALSE, probs = c(0.05, 0.95), factors = NULL)
x |
RasterStack object ( |
novel.object |
Object listing minima and maxima for the environmental layers, used by the prediction function that is used internally by |
RASTER.object.name |
First part of the names of the raster file that will be generated, expected to identify the area and time period for which ranges were calculated |
RASTER.stack.name |
Last part of the names of the raster file that will be generated, expected to identify the predictor stack used |
RASTER.format |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
CATCH.OFF |
Disable calls to function |
name |
Name of the object, expect to expected to identify the area and time period for which ranges were calculated and where no novel conditions will be detected |
mask.raster |
RasterLayer object ( |
quantiles |
If |
probs |
Numeric vector of probabilities [0, 1] as used by |
factors |
vector that indicates which variables are factors; these variables will be ignored for novel conditions |
Function ensemble.novel maps zones (coded '1') that are novel (outside the minimum-maximum range) relative to the range provided by function ensemble.novel.object. Values that are not novel (inside the range of minimum-maximum values) are coded '0'. In theory, the maps show the same areas that have negative Multivariate Environmental Similarity Surface (MESS) values ((mess))
Function ensemble.novel.object returns a list with following objects:
minima |
minima of the reference environmental conditions |
maxima |
maxima of the reference environmental conditions |
name |
name for the reference area and time period |
Roeland Kindt (World Agroforestry Centre)
ensemble.raster, ensemble.bioclim and ensemble.bioclim.graph
## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio7", "bio8", "bio12", "bio16", "bio17")) predictors predictors@title <- "base" # reference area to calculate environmental ranges ext <- extent(-70, -50, -10, 10) extent.values2 <- c(-70, -50, -10, 10) predictors.current <- crop(predictors, y=ext) predictors.current <- stack(predictors.current) novel.test <- ensemble.novel.object(predictors.current, name="noveltest") novel.test novel.raster <- ensemble.novel(x=predictors, novel.object=novel.test) novel.raster plot(novel.raster) # no novel conditions within reference area rect(extent.values2[1], extent.values2[3], extent.values2[2], extent.values2[4]) # use novel conditions as a simple species suitability mapping method # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] pres.data <- data.frame(extract(predictors, y=pres)) # ranges and maps Bradypus.ranges1 <- ensemble.novel.object(pres.data, name="Bradypus", quantiles=F) Bradypus.ranges1 Bradypus.novel1 <- ensemble.novel(x=predictors, novel.object=Bradypus.ranges1) Bradypus.novel1 par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) # suitable where there are no novel conditions raster::plot(Bradypus.novel1, breaks=c(-0.1, 0, 1), col=c("green", "grey"), main="Suitability mapping using minimum to maximum range") points(pres[, 2] ~ pres[, 1], pch=1, col="red", cex=0.8) # use 90 percent intervals similar to BIOCLIM methodology Bradypus.ranges2 <- ensemble.novel.object(pres.data, name="BradypusQuantiles", quantiles=T) Bradypus.ranges2 Bradypus.novel2 <- ensemble.novel(x=predictors, novel.object=Bradypus.ranges2) Bradypus.novel2 raster::plot(Bradypus.novel2, breaks=c(-0.1, 0, 1), col=c("green", "grey"), main="Suitability mapping using quantile range") points(pres[, 2] ~ pres[, 1], pch=1, col="red", cex=0.8) graphics::par(par.old) # deal with novel factor levels through dummy variables predictors <- stack(predictor.files) biome.layer <- predictors[["biome"]] biome.layer ensemble.dummy.variables(xcat=biome.layer, most.frequent=0, freq.min=1, overwrite=TRUE) predictors.dummy <- stack(predictor.files) predictors.dummy <- subset(predictors.dummy, subset=c("biome_1", "biome_2", "biome_3", "biome_4", "biome_5", "biome_7", "biome_8", "biome_9", "biome_10", "biome_12", "biome_13", "biome_14")) predictors.dummy predictors.dummy@title <- "base_dummy" predictors.dummy.current <- crop(predictors.dummy, y=ext) predictors.dummy.current <- stack(predictors.dummy.current) novel.levels <- ensemble.novel.object(predictors.dummy.current, name="novellevels") novel.levels novel.levels.raster <- ensemble.novel(x=predictors.dummy, novel.object=novel.levels) novel.levels.raster novel.levels.quantiles <- ensemble.novel.object(predictors.dummy.current, quantiles=TRUE, name="novellevels_quantiles") novel.levels.quantiles novel.levels.quantiles.raster <- ensemble.novel(x=predictors.dummy, novel.object=novel.levels.quantiles) novel.levels.quantiles.raster # difference in ranges for variables with low frequencies background <- dismo::randomPoints(predictors.dummy.current, n=10000, p=NULL, excludep=F) extract.data <- extract(predictors.dummy.current, y=background) colSums(extract.data)/sum(extract.data)*100 novel.levels novel.levels.quantiles par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) raster::plot(novel.levels.raster, breaks=c(-0.1, 0, 1), col=c("grey", "green"), main="novel outside minimum to maximum range") rect(extent.values2[1], extent.values2[3], extent.values2[2], extent.values2[4]) raster::plot(novel.levels.quantiles.raster, breaks=c(-0.1, 0, 1), col=c("grey", "green"), main="novel outside quantile range") rect(extent.values2[1], extent.values2[3], extent.values2[2], extent.values2[4]) graphics::par(par.old) ## End(Not run)## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio7", "bio8", "bio12", "bio16", "bio17")) predictors predictors@title <- "base" # reference area to calculate environmental ranges ext <- extent(-70, -50, -10, 10) extent.values2 <- c(-70, -50, -10, 10) predictors.current <- crop(predictors, y=ext) predictors.current <- stack(predictors.current) novel.test <- ensemble.novel.object(predictors.current, name="noveltest") novel.test novel.raster <- ensemble.novel(x=predictors, novel.object=novel.test) novel.raster plot(novel.raster) # no novel conditions within reference area rect(extent.values2[1], extent.values2[3], extent.values2[2], extent.values2[4]) # use novel conditions as a simple species suitability mapping method # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] pres.data <- data.frame(extract(predictors, y=pres)) # ranges and maps Bradypus.ranges1 <- ensemble.novel.object(pres.data, name="Bradypus", quantiles=F) Bradypus.ranges1 Bradypus.novel1 <- ensemble.novel(x=predictors, novel.object=Bradypus.ranges1) Bradypus.novel1 par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) # suitable where there are no novel conditions raster::plot(Bradypus.novel1, breaks=c(-0.1, 0, 1), col=c("green", "grey"), main="Suitability mapping using minimum to maximum range") points(pres[, 2] ~ pres[, 1], pch=1, col="red", cex=0.8) # use 90 percent intervals similar to BIOCLIM methodology Bradypus.ranges2 <- ensemble.novel.object(pres.data, name="BradypusQuantiles", quantiles=T) Bradypus.ranges2 Bradypus.novel2 <- ensemble.novel(x=predictors, novel.object=Bradypus.ranges2) Bradypus.novel2 raster::plot(Bradypus.novel2, breaks=c(-0.1, 0, 1), col=c("green", "grey"), main="Suitability mapping using quantile range") points(pres[, 2] ~ pres[, 1], pch=1, col="red", cex=0.8) graphics::par(par.old) # deal with novel factor levels through dummy variables predictors <- stack(predictor.files) biome.layer <- predictors[["biome"]] biome.layer ensemble.dummy.variables(xcat=biome.layer, most.frequent=0, freq.min=1, overwrite=TRUE) predictors.dummy <- stack(predictor.files) predictors.dummy <- subset(predictors.dummy, subset=c("biome_1", "biome_2", "biome_3", "biome_4", "biome_5", "biome_7", "biome_8", "biome_9", "biome_10", "biome_12", "biome_13", "biome_14")) predictors.dummy predictors.dummy@title <- "base_dummy" predictors.dummy.current <- crop(predictors.dummy, y=ext) predictors.dummy.current <- stack(predictors.dummy.current) novel.levels <- ensemble.novel.object(predictors.dummy.current, name="novellevels") novel.levels novel.levels.raster <- ensemble.novel(x=predictors.dummy, novel.object=novel.levels) novel.levels.raster novel.levels.quantiles <- ensemble.novel.object(predictors.dummy.current, quantiles=TRUE, name="novellevels_quantiles") novel.levels.quantiles novel.levels.quantiles.raster <- ensemble.novel(x=predictors.dummy, novel.object=novel.levels.quantiles) novel.levels.quantiles.raster # difference in ranges for variables with low frequencies background <- dismo::randomPoints(predictors.dummy.current, n=10000, p=NULL, excludep=F) extract.data <- extract(predictors.dummy.current, y=background) colSums(extract.data)/sum(extract.data)*100 novel.levels novel.levels.quantiles par.old <- graphics::par(no.readonly=T) graphics::par(mfrow=c(1,2)) raster::plot(novel.levels.raster, breaks=c(-0.1, 0, 1), col=c("grey", "green"), main="novel outside minimum to maximum range") rect(extent.values2[1], extent.values2[3], extent.values2[2], extent.values2[4]) raster::plot(novel.levels.quantiles.raster, breaks=c(-0.1, 0, 1), col=c("grey", "green"), main="novel outside quantile range") rect(extent.values2[1], extent.values2[3], extent.values2[2], extent.values2[4]) graphics::par(par.old) ## End(Not run)
Internally, the function first determines different dry seasons, defined by consecutive months where precipitation is smaller than potential evapotranspiration. The function then returns the summation of monthly balances of precipitation minus potential evapotranspiration that is largest (most negative) of the different dry seasons.
ensemble.PET.season(PREC.stack = NULL, PET.stack = NULL, filename = NULL, overwrite = TRUE, CATCH.OFF = FALSE, ...)ensemble.PET.season(PREC.stack = NULL, PET.stack = NULL, filename = NULL, overwrite = TRUE, CATCH.OFF = FALSE, ...)
PREC.stack |
stack object ( |
PET.stack |
stack object ( |
filename |
Name for writing the resulting raster layer (as in |
overwrite |
Replace a previous version of the same file. |
CATCH.OFF |
Disable calls to function |
... |
Additional arguments for |
Unlike the methodology described by Chave et al. (2014), the assumption is not made that there is a single drought season. Internally, the function first identifies dry months as months where the balance of precipitation minus potential evapotranspiration is negative. Then dry seasons are identified as consecutive dry months. For each dry season, the total sum of balances is calculated. The function finally identifies and returns the largest of these balances.
The algorithm of the function should obtain the same values of the Maximum Cumulative Water Deficit as from rules described by Aragao et al. 2007 (section 2.2), when using fixed monthly PET values of 100 mm instead of calculated monthly PET values (calculated, for example, from monthly mean temperatures and extraterrestrial solar radiation through the Hargreaves method).
Note that calculation may take a while for larger raster data sets.
The function returns and writes a raster layer
Roeland Kindt (World Agroforestry Centre)
Chave J et al. 2014. Improved allometric models to estimate the aboveground biomass of tropical trees. Global Change Biology 20: 3177-3190.
Aragao LZ et al. 2007. Spatial patterns and fire response of recent Amazonian droughts. Geophysical Research Letters 34 L07701
## Not run: ## Not run: library(raster) stack1 <- stack(monthly.prec.files) stack2 <- stack(monthly.PET.files) # note that the stacks should be of the same extend and resolution ensemble.PET.season(PREC.stack=stack1, PET.stack=stack2, filename=paste(getwd(), '//Aridity.deficit.tif', sep="")) ## End(Not run)## Not run: ## Not run: library(raster) stack1 <- stack(monthly.prec.files) stack2 <- stack(monthly.PET.files) # note that the stacks should be of the same extend and resolution ensemble.PET.season(PREC.stack=stack1, PET.stack=stack2, filename=paste(getwd(), '//Aridity.deficit.tif', sep="")) ## End(Not run)
The main function of ensemble.PET.seasons calculates the number of growing seasons and their starts and lengths from the dry period criterion of 2 * P < PET (https://www.fao.org/3/w2962e/w2962e-03.htm). Functions ensemble.PREC.season and ensemble.TMEAN.season calculate the total precipitation and average temperature for provided starts and lengths of a growing season. Together with data on optimal and absolute precipitation and temperature limits for a selected crop (as available from FAO's ECOCROP database), these layers enable the calculation of crop suitability using methods detailed in Chapman et al. (2020).
ensemble.PET.seasons(PREC.stack=NULL, PET.stack=NULL, index=c("seasons", "start1", "length1", "start2", "length2", "start3", "length3"), filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...) ensemble.prec.season(PREC.stack=NULL, start.layer=NULL, length.layer=NULL, filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...) ensemble.tmean.season(TMEAN.stack=NULL, start.layer=NULL, length.layer=NULL, filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...) ensemble.season.suitability(season.raster=NULL, thresholds=NULL, filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...)ensemble.PET.seasons(PREC.stack=NULL, PET.stack=NULL, index=c("seasons", "start1", "length1", "start2", "length2", "start3", "length3"), filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...) ensemble.prec.season(PREC.stack=NULL, start.layer=NULL, length.layer=NULL, filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...) ensemble.tmean.season(TMEAN.stack=NULL, start.layer=NULL, length.layer=NULL, filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...) ensemble.season.suitability(season.raster=NULL, thresholds=NULL, filename=NULL, overwrite=TRUE, CATCH.OFF=FALSE, ...)
PREC.stack |
stack object ( |
PET.stack |
stack object ( |
TMEAN.stack |
stack object ( |
index |
selection of type of output - see details. |
start.layer |
raster layer with index of the month of the start of the growing season. |
length.layer |
raster layer with index of the length of the growing season. |
season.raster |
raster layer with seasonal precipitation or mean temperature. |
thresholds |
optimal and absolute thresholds of crop suitability, defined similarly as by ECOCROP. |
filename |
Name for writing the resulting raster layer (as in |
overwrite |
Replace a previous version of the same file. |
CATCH.OFF |
Disable calls to function |
... |
Additional arguments for |
Function ensemble.PET.seasons calculates the number, starts and lengths of growing seasons after first internally determining dry periods from the criterion of 2 * P < PET. The function was developed with data sets with monthly precipitatin and PET values, but probably can also work with data sets of other temporal resolution. Where there are multiple gaps between dry seasons, different growing periods are identified.
The definition of dry periods is less strict than the definition of P < PET used in ensemble.PET.season, following the methodologies for this function.
Argument index determines the contents of the output rasters:
- seasons selects the number of growing periods to be returned;
- start1 selects the index of the start of the first or only growing period to be returned;
- length1 selects the index of the end of the first or only growing period to be returned;
- start2 selects the index of the start of the second growing period to be returned;
- length2 selects the index of the end of the second growing period to be returned;
- start3 selects the index of the start of the third growing period to be returned; and
- length3 selects the index of the end of the third growing period to be returned.
The methodology of calculating crop suitability is directly based on Chapman et al. (2020), following their equations 2 (temperature suitability, based on the mean temperature of the growing season) and 3 (precipitation suitability, based on the total precipitation of the growing season). The methods of Chapman et al. (2020) are based on Ramirez-Villegas et al. (2013), including the calculation of crop suitability as the product of temperature suitability and crop suitability (their respective equations 1 and 3).
Crop thresholds are available from the FAO ECOCROP database, which are also available via function getCrop.
Note that calculations can take a while for larger data sets.
The function returns and writes raster layers.
Roeland Kindt (World Agroforestry Centre)
Ramirez-Villegas J, Jarvis A and Laderach P. 2013. Empirical approaches for assessing impacts of climate change on agriculture: The EcoCrop model and a case study with grain sorghum. Agricultural and Forest Meteorology doi:10.1016/j.agrformet.2011.09.005
Chapman et al. 2020. Impact of climate change on crop suitability in sub-Saharan Africa in parameterized and convection-permitting regional climate models. Environmental Research Letters 15:094086.
## Not run: ## Not run: library(raster) P.stack <- stack(monthly.prec.files) PE.stack <- stack(monthly.PET.files) # Calculate average monthly values similarly as in TMIN.stack <- stack(monthly.tmin.files) TMAX.stack <- stack(monthly.tmax.files) T.stack <- stack(0.5*(TMIN.stack + TMAX.stack)) # step 1: determine number of seasons, start and length of season 1 seasons.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="seasons", filename="seasons.tif", CATCH.OFF=TRUE) start1.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="start1", filename="start1.tif", CATCH.OFF=TRUE) length1.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="length1", filename="length1.tif", CATCH.OFF=TRUE) start2.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="start2", filename="start2.tif", CATCH.OFF=TRUE) length2.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="length2", filename="length2.tif", CATCH.OFF=TRUE) # step 2: calculate total precipitation in first rainy season, # then use this value to calculate precipitation suitability prec.season <- ensemble.prec.season(PREC.stack=P.stack, start.layer=start1.raster, length.layer=length1.raster, filename="precSeason.tif", CATCH.OFF=FALSE) dismo::getCrop("Sorghum (med. altitude)") prec.suit <- ensemble.season.suitability(season.raster=prec.season, thresholds=c(300, 500, 1000, 3000), filename="precSuitability.tif", CATCH.OFF=FALSE) # step 3: calculate average temperature in first rainy season, # then use this value to calculate temperature suitability tmean.season <- ensemble.tmean.season(TMEAN.stack=T.stack, start.layer=start1.raster, length.layer=length1.raster, filename="tmeanSeason.tif", CATCH.OFF=FALSE) temp.suit <- ensemble.season.suitability(season.raster=tmean.season, thresholds=c(10, 24, 35, 40), filename="tempSuitability.tif", CATCH.OFF=FALSE) # step 4: seasonal crop suitability is product of precipitation suitability # and temperature suitability sorghum.suit <- prec.suit * temp.suit plot(sorghum.suit) ## End(Not run)## Not run: ## Not run: library(raster) P.stack <- stack(monthly.prec.files) PE.stack <- stack(monthly.PET.files) # Calculate average monthly values similarly as in TMIN.stack <- stack(monthly.tmin.files) TMAX.stack <- stack(monthly.tmax.files) T.stack <- stack(0.5*(TMIN.stack + TMAX.stack)) # step 1: determine number of seasons, start and length of season 1 seasons.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="seasons", filename="seasons.tif", CATCH.OFF=TRUE) start1.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="start1", filename="start1.tif", CATCH.OFF=TRUE) length1.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="length1", filename="length1.tif", CATCH.OFF=TRUE) start2.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="start2", filename="start2.tif", CATCH.OFF=TRUE) length2.raster <- ensemble.PET.seasons(PREC.stack=P.stack, PET.stack=PE.stack, index="length2", filename="length2.tif", CATCH.OFF=TRUE) # step 2: calculate total precipitation in first rainy season, # then use this value to calculate precipitation suitability prec.season <- ensemble.prec.season(PREC.stack=P.stack, start.layer=start1.raster, length.layer=length1.raster, filename="precSeason.tif", CATCH.OFF=FALSE) dismo::getCrop("Sorghum (med. altitude)") prec.suit <- ensemble.season.suitability(season.raster=prec.season, thresholds=c(300, 500, 1000, 3000), filename="precSuitability.tif", CATCH.OFF=FALSE) # step 3: calculate average temperature in first rainy season, # then use this value to calculate temperature suitability tmean.season <- ensemble.tmean.season(TMEAN.stack=T.stack, start.layer=start1.raster, length.layer=length1.raster, filename="tmeanSeason.tif", CATCH.OFF=FALSE) temp.suit <- ensemble.season.suitability(season.raster=tmean.season, thresholds=c(10, 24, 35, 40), filename="tempSuitability.tif", CATCH.OFF=FALSE) # step 4: seasonal crop suitability is product of precipitation suitability # and temperature suitability sorghum.suit <- prec.suit * temp.suit plot(sorghum.suit) ## End(Not run)
The basic function ensemble.raster creates two consensus raster layers, one based on a (weighted) average of different suitability modelling algorithms, and a second one documenting the number of modelling algorithms that predict presence of the focal organisms. Modelling algorithms include maximum entropy (MAXENT), boosted regression trees, random forests, generalized linear models (including stepwise selection of explanatory variables), generalized additive models (including stepwise selection of explanatory variables), multivariate adaptive regression splines, regression trees, artificial neural networks, flexible discriminant analysis, support vector machines, the BIOCLIM algorithm, the DOMAIN algorithm and the Mahalonobis algorithm. These sets of functions were developed in parallel with the biomod2 package, especially for inclusion of the maximum entropy algorithm, but also to allow for a more direct integration with the BiodiversityR package, more direct handling of model formulae and greater focus on mapping. Researchers and students of species distribution are strongly encouraged to familiarize themselves with all the options of the biomod2 and dismo packages.
ensemble.raster(xn = NULL, models.list = NULL, input.weights = models.list$output.weights, thresholds = models.list$thresholds, RASTER.species.name = models.list$species.name, RASTER.stack.name = xn@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, RASTER.models.overwrite = TRUE, evaluate = FALSE, SINK = FALSE, p = models.list$p, a = models.list$a, pt = models.list$pt, at = models.list$at, CATCH.OFF = FALSE) ensemble.habitat.change(base.map = file.choose(), other.maps = utils::choose.files(), change.folder = "ensembles/change", RASTER.names = "changes", RASTER.format = "GTiff", RASTER.datatype = "INT1U", RASTER.NAflag = 255) ensemble.area(x=NULL, km2=TRUE)ensemble.raster(xn = NULL, models.list = NULL, input.weights = models.list$output.weights, thresholds = models.list$thresholds, RASTER.species.name = models.list$species.name, RASTER.stack.name = xn@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, RASTER.models.overwrite = TRUE, evaluate = FALSE, SINK = FALSE, p = models.list$p, a = models.list$a, pt = models.list$pt, at = models.list$at, CATCH.OFF = FALSE) ensemble.habitat.change(base.map = file.choose(), other.maps = utils::choose.files(), change.folder = "ensembles/change", RASTER.names = "changes", RASTER.format = "GTiff", RASTER.datatype = "INT1U", RASTER.NAflag = 255) ensemble.area(x=NULL, km2=TRUE)
xn |
RasterStack object ( |
models.list |
list with 'old' model objects such as |
input.weights |
array with numeric values for the different modelling algorithms; if |
thresholds |
array with the threshold values separating predicted presence for each of the algorithms. |
RASTER.species.name |
First part of the names of the raster files that will be generated, expected to identify the modelled species (or organism). |
RASTER.stack.name |
Last part of the names of the raster files that will be generated, expected to identify the predictor stack used. |
RASTER.format |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
RASTER.models.overwrite |
Overwrite the raster files that correspond to each suitability modelling algorithm (if |
evaluate |
if |
SINK |
Append the results to a text file in subfolder 'outputs' (if |
p |
presence points used for calibrating the suitability models, typically available in 2-column (x, y) or (lon, lat) dataframe; see also |
a |
background points used for calibrating the suitability models, typically available in 2-column (x, y) or (lon, lat) dataframe; see also |
pt |
presence points used for evaluating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
at |
background points used for calibrating the suitability models, typicall available in 2-column (lon, lat) dataframe; see also |
CATCH.OFF |
Disable calls to function |
base.map |
filename with baseline map used to produce maps that show changes in suitable habitat |
other.maps |
files with other maps used to produce maps that show changes in suitable habitat from a defined base.map |
change.folder |
folder where new maps documenting changes in suitable habitat will be stored. NOTE: please ensure that the base folder (eg: ../ensembles) exists already. |
RASTER.names |
names for the files in the change.folder (previously set as names of the other maps). |
x |
RasterLayer object ( |
km2 |
Provide results in square km rather than square m. See also |
The basic function ensemble.raster fits individual suitability models for all models with positive input weights. In subfolder "models" of the working directory, suitability maps for the individual suitability modelling algorithms are stored. In subfolder "ensembles", a consensus suitability map based on a weighted average of individual suitability models is stored. In subfolder "ensembles/presence", a presence-absence (1-0) map will be provided. In subfolder "ensembles/count", a consensus suitability map based on the number of individual suitability models that predict presence of the focal organism is stored.
Note that values in suitability maps are integer values that were calculated by multiplying probabilities by 1000 (see also trunc).
The ensemble.habitat.change function produces new raster layers that show changes in suitable and not suitable habitat between a base raster and a list of other rasters. The output uses the following coding: 0 = areas that remain unsuitable, 11 = areas that remain suitable, 10 = areas of lost habitat, 1 = areas of new habitat. (Codes are inspired on a binary classification of habitat suitability in base [1- or 0-] and other layer [-1 or -0], eg new habitat is coded 01 = 1).
The ensemble.area function calculates the area of different categories with areaPolygon
The basic function ensemble.raster mainly results in the creation of raster layers that correspond to fitted probabilities of presence of individual suitability models (in folder "models") and consensus models (in folder "ensembles"), and the number of suitability models that predict presence (in folder "ensembles"). Prediction of presence is based on a threshold usually defined by maximizing the sum of the true presence and true absence rates (see threshold.method and also ModelEvaluation).
If desired by the user, the ensemble.raster function also saves details of fitted suitability models or data that can be plotted with the evaluation.strip.plot function.
Roeland Kindt (World Agroforestry Centre), Eike Luedeling (World Agroforestry Centre) and Evert Thomas (Bioversity International)
Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009
Buisson L, Thuiller W, Casajus N, Lek S and Grenouillet G. 2010. Uncertainty in ensemble forecasting of species distribution. Global Change Biology 16: 1145-1157
evaluation.strip.plot, ensemble.calibrate.models, ensemble.calibrate.weights, ensemble.batch
## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "base" # presence points # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # choose background points background <- randomPoints(predictors, n=1000, extf = 1.00) # if desired, change working directory where subfolders of "models" and # "ensembles" will be created # raster layers will be saved in subfolders of /models and /ensembles: getwd() # first calibrate the ensemble # calibration is done in two steps # in step 1, a k-fold procedure is used to determine the weights # in step 2, models are calibrated for all presence and background locations # factor is not used as it is not certain whether correct levels will be used # it may therefore be better to use dummy variables # step 1: determine weights through 4-fold cross-validation ensemble.calibrate.step1 <- ensemble.calibrate.weights( x=predictors, p=pres, a=background, k=4, SINK=TRUE, species.name="Bradypus", MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, PROBIT=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=c(0.65, 0.7), Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 1 generated the weights for each algorithm model.weights <- ensemble.calibrate.step1$output.weights x.batch <- ensemble.calibrate.step1$x p.batch <- ensemble.calibrate.step1$p a.batch <- ensemble.calibrate.step1$a MAXENT.a.batch <- ensemble.calibrate.step1$MAXENT.a factors.batch <- ensemble.calibrate.step1$factors dummy.vars.batch <- ensemble.calibrate.step1$dummy.vars # step 2: calibrate models with all available presence locations # weights determined in step 1 calculate ensemble in step 2 ensemble.calibrate.step2 <- ensemble.calibrate.models( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, SINK=TRUE, species.name="Bradypus", models.keep=TRUE, input.weights=model.weights, ENSEMBLE.tune=FALSE, PROBIT=TRUE, Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 3: use previously calibrated models to create ensemble raster layers # re-evaluate the created maps at presence and background locations # (note that re-evaluation will be different due to truncation of raster layers # as they wered saved as integer values ranged 0 to 1000) ensemble.raster.results <- ensemble.raster(xn=predictors, models.list=ensemble.calibrate.step2$models, input.weights=model.weights, SINK=TRUE, evaluate=TRUE, RASTER.species.name="Bradypus", RASTER.stack.name="base") # use the base map to check for changes in suitable habitat # this type of analysis is typically done with different predictor layers # (for example, predictor layers representing different possible future climates) # In this example, changes from a previous model (ensemble.raster.results) # are contrasted with a newly calibrated model (ensemble.raster.results2) # step 1: 4-fold cross-validation ensemble.calibrate2.step1 <- ensemble.calibrate.weights( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, k=4, SINK=TRUE, species.name="Bradypus", MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, PROBIT=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=c(0.65, 0.7), Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) model.weights2 <- ensemble.calibrate2.step1$output.weights ensemble.calibrate2.step2 <- ensemble.calibrate.models( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, SINK=TRUE, species.name="Bradypus", models.keep=TRUE, input.weights=model.weights2, ENSEMBLE.tune=FALSE, PROBIT=TRUE, Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) ensemble.raster.results2 <- ensemble.raster( xn=predictors, models.list=ensemble.calibrate2.step2$models, input.weights=model.weights2, SINK=TRUE, evaluate=TRUE, RASTER.species.name="Bradypus", RASTER.stack.name="recalibrated") base.file <- paste(getwd(), "/ensembles/presence/Bradypus_base.tif", sep="") other.file <- paste(getwd(), "/ensembles/presence/Bradypus_recalibrated.tif", sep="") changed.habitat <- ensemble.habitat.change(base.map=base.file, other.maps=c(other.file), change.folder="ensembles/change", RASTER.names="Bradypus_recalibrated") change.file <- paste(getwd(), "/ensembles/change/Bradypus_recalibrated.tif", sep="") par.old <- graphics::par(no.readonly=T) dev.new() par(mfrow=c(2,2)) raster::plot(raster(base.file), breaks=c(-1, 0, 1), col=c("grey", "green"), legend.shrink=0.8, main="base presence") raster::plot(raster(other.file), breaks=c(-1, 0, 1), col=c("grey", "green"), legend.shrink=0.8, main="other presence") raster::plot(raster(change.file), breaks=c(-1, 0, 1, 10, 11), col=c("grey", "blue", "red", "green"), legend.shrink=0.8, main="habitat change", sub="11 remaining, 10 lost, 1 new") graphics::par(par.old) areas <- ensemble.area(raster(change.file)) areas ## End(Not run)## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "base" # presence points # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # choose background points background <- randomPoints(predictors, n=1000, extf = 1.00) # if desired, change working directory where subfolders of "models" and # "ensembles" will be created # raster layers will be saved in subfolders of /models and /ensembles: getwd() # first calibrate the ensemble # calibration is done in two steps # in step 1, a k-fold procedure is used to determine the weights # in step 2, models are calibrated for all presence and background locations # factor is not used as it is not certain whether correct levels will be used # it may therefore be better to use dummy variables # step 1: determine weights through 4-fold cross-validation ensemble.calibrate.step1 <- ensemble.calibrate.weights( x=predictors, p=pres, a=background, k=4, SINK=TRUE, species.name="Bradypus", MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, PROBIT=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=c(0.65, 0.7), Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 1 generated the weights for each algorithm model.weights <- ensemble.calibrate.step1$output.weights x.batch <- ensemble.calibrate.step1$x p.batch <- ensemble.calibrate.step1$p a.batch <- ensemble.calibrate.step1$a MAXENT.a.batch <- ensemble.calibrate.step1$MAXENT.a factors.batch <- ensemble.calibrate.step1$factors dummy.vars.batch <- ensemble.calibrate.step1$dummy.vars # step 2: calibrate models with all available presence locations # weights determined in step 1 calculate ensemble in step 2 ensemble.calibrate.step2 <- ensemble.calibrate.models( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, SINK=TRUE, species.name="Bradypus", models.keep=TRUE, input.weights=model.weights, ENSEMBLE.tune=FALSE, PROBIT=TRUE, Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 3: use previously calibrated models to create ensemble raster layers # re-evaluate the created maps at presence and background locations # (note that re-evaluation will be different due to truncation of raster layers # as they wered saved as integer values ranged 0 to 1000) ensemble.raster.results <- ensemble.raster(xn=predictors, models.list=ensemble.calibrate.step2$models, input.weights=model.weights, SINK=TRUE, evaluate=TRUE, RASTER.species.name="Bradypus", RASTER.stack.name="base") # use the base map to check for changes in suitable habitat # this type of analysis is typically done with different predictor layers # (for example, predictor layers representing different possible future climates) # In this example, changes from a previous model (ensemble.raster.results) # are contrasted with a newly calibrated model (ensemble.raster.results2) # step 1: 4-fold cross-validation ensemble.calibrate2.step1 <- ensemble.calibrate.weights( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, k=4, SINK=TRUE, species.name="Bradypus", MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, PROBIT=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=c(0.65, 0.7), Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) model.weights2 <- ensemble.calibrate2.step1$output.weights ensemble.calibrate2.step2 <- ensemble.calibrate.models( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, SINK=TRUE, species.name="Bradypus", models.keep=TRUE, input.weights=model.weights2, ENSEMBLE.tune=FALSE, PROBIT=TRUE, Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) ensemble.raster.results2 <- ensemble.raster( xn=predictors, models.list=ensemble.calibrate2.step2$models, input.weights=model.weights2, SINK=TRUE, evaluate=TRUE, RASTER.species.name="Bradypus", RASTER.stack.name="recalibrated") base.file <- paste(getwd(), "/ensembles/presence/Bradypus_base.tif", sep="") other.file <- paste(getwd(), "/ensembles/presence/Bradypus_recalibrated.tif", sep="") changed.habitat <- ensemble.habitat.change(base.map=base.file, other.maps=c(other.file), change.folder="ensembles/change", RASTER.names="Bradypus_recalibrated") change.file <- paste(getwd(), "/ensembles/change/Bradypus_recalibrated.tif", sep="") par.old <- graphics::par(no.readonly=T) dev.new() par(mfrow=c(2,2)) raster::plot(raster(base.file), breaks=c(-1, 0, 1), col=c("grey", "green"), legend.shrink=0.8, main="base presence") raster::plot(raster(other.file), breaks=c(-1, 0, 1), col=c("grey", "green"), legend.shrink=0.8, main="other presence") raster::plot(raster(change.file), breaks=c(-1, 0, 1, 10, 11), col=c("grey", "blue", "red", "green"), legend.shrink=0.8, main="habitat change", sub="11 remaining, 10 lost, 1 new") graphics::par(par.old) areas <- ensemble.area(raster(change.file)) areas ## End(Not run)
Function ensemble.red is a wrapper function for estimation of AOO and EOO computed for redlisting of species based on IUCN criteria (https://www.iucnredlist.org/about/regional). Function ensemble.chull.create creates a mask layer based on a convex hull around known presence locations, inspired by mcp argument of the map.sdm function.
ensemble.red(x) ensemble.chull.create(x.pres = NULL, p = NULL, buffer.width = 0.2, buffer.maxmins = FALSE, lonlat.dist = FALSE, poly.only = FALSE, RASTER.format = "GTiff", RASTER.datatype = "INT1U", RASTER.NAflag = 255, overwrite = TRUE, ...) ensemble.chull.apply(x.spec = NULL, mask.layer=NULL, keep.old=T, RASTER.format="GTiff", RASTER.datatype="INT1U", RASTER.NAflag = 255, overwrite=TRUE, ...) ensemble.chull.buffer.distances(p = NULL, buffer.maxmins = FALSE, lonlat.dist = FALSE) ensemble.chull.MSDM(p = NULL, a = NULL, species.name = NULL, suit.file = NULL, suit.divide = 1000, MSDM.dir = NULL, method = "BMCP", threshold = "spec_sens", buffer = "species_specific")ensemble.red(x) ensemble.chull.create(x.pres = NULL, p = NULL, buffer.width = 0.2, buffer.maxmins = FALSE, lonlat.dist = FALSE, poly.only = FALSE, RASTER.format = "GTiff", RASTER.datatype = "INT1U", RASTER.NAflag = 255, overwrite = TRUE, ...) ensemble.chull.apply(x.spec = NULL, mask.layer=NULL, keep.old=T, RASTER.format="GTiff", RASTER.datatype="INT1U", RASTER.NAflag = 255, overwrite=TRUE, ...) ensemble.chull.buffer.distances(p = NULL, buffer.maxmins = FALSE, lonlat.dist = FALSE) ensemble.chull.MSDM(p = NULL, a = NULL, species.name = NULL, suit.file = NULL, suit.divide = 1000, MSDM.dir = NULL, method = "BMCP", threshold = "spec_sens", buffer = "species_specific")
x |
RasterLayer object ( |
x.pres |
RasterLayer object ( |
p |
known presence locations, available in 2-column (lon, lat) dataframe; see also |
buffer.width |
multiplier to create buffer (via |
buffer.maxmins |
Calculate the buffer width based on the two neighbouring locations that are furthest apart (maximum of minimum distances from each location). |
lonlat.dist |
Estimate the distance in km for longitude latitude data. |
poly.only |
Only return the polygon with the convex hull, but do not create the mask layer. |
RASTER.format |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
overwrite |
Overwrite existing raster files. See |
... |
Additional arguments for |
x.spec |
RasterLayer object ( |
mask.layer |
RasterLayer object ( |
keep.old |
keep a copy of the RasterLayer before the mask is applied. |
a |
absence of background locations, available in 2-column (lon, lat) dataframe. |
species.name |
name of the species, ideally without spaces. |
suit.file |
file with raster data corresponding to suitability values of the focal species. |
suit.divide |
number by which values in the suitability raster should be divided to result in probabilities (BiodiversityR saves data as 1000 * suitability, hence these values need to be divided by 1000). |
MSDM.dir |
name of the directory where input and processed raster files will be saved. |
method |
method for MSDM_Posteriori function from c("OBR", "PRES", "LQ", "MCP", "BMCP"). |
threshold |
threshold for MSDM_Posteriori function from c("kappa", "spec_sens", "no_omission", "prevalence", "equal_sens_spec", "sensitivty"). |
buffer |
buffer for MSDM_Posteriori function. |
Function ensemble.red calculates AOO (aoo) and EOO (aoo) statistics calculated for areas with different consensus levels on species presence (1 model predicting presence, 2 models predicting presence, ...). In case that these statistics are within IUCN criteria for Critically Endangered (CR), Endangered (EN) or Vulnerable (VU), then this information is added in columns documenting the types of AOO and EOO.
Function ensemble.chull.create first creates a convex hull around known presence locations. Next, a buffer is created around the convex hull where the width of this buffer is calculated as the maximum distance among presence locations (pointDistance) multiplied by argument buffer.width. Finally, the mask is created by including all polygons of predicted species presence that are partially covered by the convex hull and its buffer.
Function ensemble.red returns an array with AOO and EOO
Function ensemble.chull.create creates a mask layer based on a convex hull around known presence locations.
Function ensemble.chull.MSDM prepares the input data and script for the MSDM_Posteriori function of the MSDM package.
Roeland Kindt (World Agroforestry Centre)
Cardoso P. 2017. red - an R package to facilitate species red list assessments according to the IUCN criteria. Biodiversity Data Journal 5:e20530. doi:10.3897/BDJ.5.e20530
Mendes, P.; Velazco S.J.E.; Andrade, A.F.A.; De Marco, P. (2020) Dealing with overprediction in species distribution models: how adding distance constraints can improve model accuracy, Ecological Modelling, in press. doi:10.1016/j.ecolmodel.2020.109180
Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009
## Not run: ## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "red" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',') # fit 4 ensemble models (could take some time!) # (examples for the red package use 100 models) ensembles <- ensemble.batch(x=predictors, xn=c(predictors), species.presence=pres, thin.km=100, k.splits=4, k.test=0, n.ensembles=4, SINK=TRUE, ENSEMBLE.best=10, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.6, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=0, GLM=0, GLMSTEP=1, GAM=1, GAMSTEP=0, MGCV=1, MGCVFIX=0, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=0, MAHAL=0, MAHAL01=0, PROBIT=TRUE, Yweights="BIOMOD", formulae.defaults=TRUE) # first application of ensemble.red before applying the convex hull mask # AOO and EOO are determined for each count level library(red) count.file <- paste(getwd(), "/ensembles/consensuscount/Bradypus variegatus_red.tif", sep="") count.raster <- raster(count.file) ensemble.red(count.raster) # do not predict presence in polygons completely outside convex hull # of known presence locations pres.file <- paste(getwd(), "/ensembles/consensuspresence/Bradypus variegatus_red.tif", sep="") pres.raster <- raster(pres.file) pres1 <- pres[, -1] chull.created <- ensemble.chull.create(x.pres=pres.raster, p=pres1) mask.raster <- chull.created$mask.layer plot(mask.raster, col=c("black", "green")) mask.poly <- chull.created$convex.hull pres.chull <- ensemble.chull.apply(pres.raster, mask=mask.raster, keep.old=T) par.old <- graphics::par(no.readonly=T) par(mfrow=c(1,2)) plot(pres.raster, breaks=c(-1, 0, 1), col=c("grey", "green"), main="before convex hull") points(pres1, col="blue") # load new pres.file.new <- paste(getwd(), "/ensembles/chull/Bradypus variegatus_red.tif", sep="") pres.raster.new <- raster(pres.file.new) plot(pres.raster.new, breaks=c(-1, 0, 1), col=c("grey", "green"), main="after convex hull") plot(mask.poly, add=T, border="blue") # create a smaller hull (0.05 * largest distance) chull.created <- ensemble.chull.create(x.pres=pres.raster, p=pres1, buffer.width=0.05, lonlat.dist=TRUE) mask.raster <- chull.created$mask.layer mask.poly <- chull.created$convex.hull pres.chull <- ensemble.chull.apply(pres.raster, mask=mask.raster, keep.old=T) par(mfrow=c(1,2)) plot(pres.raster, breaks=c(-1, 0, 1), col=c("grey", "green"), main="before convex hull") points(pres1, col="blue") pres.raster.new <- raster(pres.file.new) plot(pres.raster.new, breaks=c(-1, 0, 1), col=c("grey", "green"), main="after convex hull") plot(mask.poly, add=T, border="blue") # create a hull based on the distance to the location with the farthest neighbour chull.created <- ensemble.chull.create(x.pres=pres.raster, p=pres1, buffer.maxmins=TRUE, buffer.width=0.9, lonlat.dist=TRUE) mask.raster <- chull.created$mask.layer mask.poly <- chull.created$convex.hull pres.chull <- ensemble.chull.apply(pres.raster, mask=mask.raster, keep.old=T) par(mfrow=c(1,2)) plot(pres.raster, breaks=c(-1, 0, 1), col=c("grey", "green"), main="before convex hull") points(pres1, col="blue") pres.raster.new <- raster(pres.file.new) plot(pres.raster.new, breaks=c(-1, 0, 1), col=c("grey", "green"), main="after convex hull") plot(mask.poly, add=T, border="blue") par.old <- graphics::par(no.readonly=T) # how distances were derived # maximum distance between observations ensemble.chull.buffer.distances(pres1, lonlat.dist=TRUE) # the closest neigbhour that is farthest away from each observation # this is the distance calculated by MSDM_posteriori for buffer="species_specific" ensemble.chull.buffer.distances(pres1, buffer.maxmins=TRUE, lonlat.dist=TRUE) ## End(Not run)## Not run: ## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "red" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',') # fit 4 ensemble models (could take some time!) # (examples for the red package use 100 models) ensembles <- ensemble.batch(x=predictors, xn=c(predictors), species.presence=pres, thin.km=100, k.splits=4, k.test=0, n.ensembles=4, SINK=TRUE, ENSEMBLE.best=10, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.6, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=0, GLM=0, GLMSTEP=1, GAM=1, GAMSTEP=0, MGCV=1, MGCVFIX=0, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, BIOCLIM.O=0, BIOCLIM=1, DOMAIN=0, MAHAL=0, MAHAL01=0, PROBIT=TRUE, Yweights="BIOMOD", formulae.defaults=TRUE) # first application of ensemble.red before applying the convex hull mask # AOO and EOO are determined for each count level library(red) count.file <- paste(getwd(), "/ensembles/consensuscount/Bradypus variegatus_red.tif", sep="") count.raster <- raster(count.file) ensemble.red(count.raster) # do not predict presence in polygons completely outside convex hull # of known presence locations pres.file <- paste(getwd(), "/ensembles/consensuspresence/Bradypus variegatus_red.tif", sep="") pres.raster <- raster(pres.file) pres1 <- pres[, -1] chull.created <- ensemble.chull.create(x.pres=pres.raster, p=pres1) mask.raster <- chull.created$mask.layer plot(mask.raster, col=c("black", "green")) mask.poly <- chull.created$convex.hull pres.chull <- ensemble.chull.apply(pres.raster, mask=mask.raster, keep.old=T) par.old <- graphics::par(no.readonly=T) par(mfrow=c(1,2)) plot(pres.raster, breaks=c(-1, 0, 1), col=c("grey", "green"), main="before convex hull") points(pres1, col="blue") # load new pres.file.new <- paste(getwd(), "/ensembles/chull/Bradypus variegatus_red.tif", sep="") pres.raster.new <- raster(pres.file.new) plot(pres.raster.new, breaks=c(-1, 0, 1), col=c("grey", "green"), main="after convex hull") plot(mask.poly, add=T, border="blue") # create a smaller hull (0.05 * largest distance) chull.created <- ensemble.chull.create(x.pres=pres.raster, p=pres1, buffer.width=0.05, lonlat.dist=TRUE) mask.raster <- chull.created$mask.layer mask.poly <- chull.created$convex.hull pres.chull <- ensemble.chull.apply(pres.raster, mask=mask.raster, keep.old=T) par(mfrow=c(1,2)) plot(pres.raster, breaks=c(-1, 0, 1), col=c("grey", "green"), main="before convex hull") points(pres1, col="blue") pres.raster.new <- raster(pres.file.new) plot(pres.raster.new, breaks=c(-1, 0, 1), col=c("grey", "green"), main="after convex hull") plot(mask.poly, add=T, border="blue") # create a hull based on the distance to the location with the farthest neighbour chull.created <- ensemble.chull.create(x.pres=pres.raster, p=pres1, buffer.maxmins=TRUE, buffer.width=0.9, lonlat.dist=TRUE) mask.raster <- chull.created$mask.layer mask.poly <- chull.created$convex.hull pres.chull <- ensemble.chull.apply(pres.raster, mask=mask.raster, keep.old=T) par(mfrow=c(1,2)) plot(pres.raster, breaks=c(-1, 0, 1), col=c("grey", "green"), main="before convex hull") points(pres1, col="blue") pres.raster.new <- raster(pres.file.new) plot(pres.raster.new, breaks=c(-1, 0, 1), col=c("grey", "green"), main="after convex hull") plot(mask.poly, add=T, border="blue") par.old <- graphics::par(no.readonly=T) # how distances were derived # maximum distance between observations ensemble.chull.buffer.distances(pres1, lonlat.dist=TRUE) # the closest neigbhour that is farthest away from each observation # this is the distance calculated by MSDM_posteriori for buffer="species_specific" ensemble.chull.buffer.distances(pres1, buffer.maxmins=TRUE, lonlat.dist=TRUE) ## End(Not run)
The functions internally calls blockCV::spatialBlock and blockCV::envBlock. Syntax is very similar to that of BiodiversityR::ensemble.calibrate.weights.
ensemble.spatialBlock(x = NULL, p = NULL, a = NULL, an = 1000, EPSG=NULL, excludep = FALSE, target.groups = FALSE, k = 4, factors = NULL, theRange = NULL, return.object = FALSE, ...) ensemble.envBlock(x = NULL, p = NULL, a = NULL, an = 1000, EPSG=NULL, excludep = FALSE, target.groups = FALSE, k = 4, factors = NULL, return.object = FALSE, ...)ensemble.spatialBlock(x = NULL, p = NULL, a = NULL, an = 1000, EPSG=NULL, excludep = FALSE, target.groups = FALSE, k = 4, factors = NULL, theRange = NULL, return.object = FALSE, ...) ensemble.envBlock(x = NULL, p = NULL, a = NULL, an = 1000, EPSG=NULL, excludep = FALSE, target.groups = FALSE, k = 4, factors = NULL, return.object = FALSE, ...)
x |
RasterStack object ( |
p |
presence points used for calibrating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
a |
background points used for calibrating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
an |
number of background points for calibration to be selected with |
EPSG |
EPSG number (see https://spatialreference.org/) to be assigned internally to the coordinate reference system of the locations via |
excludep |
parameter that indicates (if |
target.groups |
Parameter that indicates (if |
k |
Integer value. The number of desired folds for cross-validation. The default is |
factors |
vector that indicates which variables are factors; see also |
theRange |
Numeric value of the specified range by which blocks are created and training/testing data are separated. This distance should be in metres. See also |
return.object |
If |
... |
Other arguments to pass to |
The functions internally call spatialBlock or envBlock.
The result of the function includes a list (k) with following elements. This list can be directly imported into ensemble.calibrate.weights, but only elements groupp and groupa will be used.
- p : Presence locations, created by ensemble.calibrate.models where points with missing data were excluded and possibly points were added for missing factor levels
- a : Background locations, created by ensemble.calibrate.models where points with missing data were excluded and possibly points were added for missing factor levels
- groupp : k-fold identities for the presence locations
- groupa : k-fold identities for the background locations
Optionally the function also returns elements block.object and speciesData. These can be used to visualize data with foldExplorer.
The function returns a list with the following elements:.
k |
A list with data on folds that can be directly used by |
block.object |
the results of |
speciesData |
a |
Roeland Kindt (World Agroforestry Centre)
Roberts et al., 2017. Cross-validation strategies for data with temporal, spatial, hierarchical, or phylogenetic structure. Ecography. 40: 913-929.
## Not run: library(blockCV) library(sf) # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[, -1] # choose background points background <- randomPoints(predictors, n=1000, p=pres, excludep=T, extf=1.00) background <- data.frame(background) colnames(background)=c('lon', 'lat') # spatial blocking with square blocks of 1000 km and minimum 20 points in each categor # fails if EPSG is not assigned block.data <- ensemble.spatialBlock(x=predictors, p=pres, a=background, EPSG=NULL, showBlocks=F, theRange=1000000, k=4, numLimit=20, iteration=1000, return.object=T) block.data <- ensemble.spatialBlock(x=predictors, p=pres, a=background, EPSG=4326, showBlocks=F, theRange=1000000, k=4, numLimit=20, iteration=1000, return.object=T) # explore the results foldExplorer(blocks=block.data$block.object, rasterLayer=predictors, speciesData=block.data$speciesData) # apply in calibration of ensemble weights # make sure that folds apply to subset of points p.spatial <- block.data$k$p a.spatial <- block.data$k$a k.spatial <- block.data$k ensemble.w1 <- ensemble.calibrate.weights(x=predictors, p=p.spatial, a=a.spatial, k=k.spatial, species.name="Bradypus", SINK=FALSE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=0, CF=1, GLM=1, GLMSTEP=0, GAM=1, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=1, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=0, MAHAL=0, MAHAL01=0, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, Yweights="BIOMOD", formulae.defaults=TRUE) # confirm that correct folds were used all.equal(ensemble.w1$groupp, block.data$k$groupp) all.equal(ensemble.w1$groupa, block.data$k$groupa) # environmental blocking with minimum 5 points in each category block.data2 <- ensemble.envBlock(x=predictors, p=pres, a=background, factors="biome", k=4, numLimit=5, return.object=T) # explore the results foldExplorer(blocks=block.data2$block.object, rasterLayer=predictors, speciesData=block.data2$speciesData) # apply in calibration of ensemble weights # make sure that folds apply to subset of points p.env <- block.data2$k$p a.env <- block.data2$k$a k.env <- block.data2$k ensemble.w2 <- ensemble.calibrate.weights(x=predictors, p=p.env, a=a.env, k=k.env, species.name="Bradypus", SINK=FALSE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=0, CF=1, GLM=1, GLMSTEP=0, GAM=1, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=1, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=0, MAHAL=0, MAHAL01=0, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, factors="biome", Yweights="BIOMOD", formulae.defaults=TRUE) # confirm that correct folds were used all.equal(ensemble.w2$groupp, block.data2$k$groupp) all.equal(ensemble.w2$groupa, block.data2$k$groupa) ## End(Not run)## Not run: library(blockCV) library(sf) # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[, -1] # choose background points background <- randomPoints(predictors, n=1000, p=pres, excludep=T, extf=1.00) background <- data.frame(background) colnames(background)=c('lon', 'lat') # spatial blocking with square blocks of 1000 km and minimum 20 points in each categor # fails if EPSG is not assigned block.data <- ensemble.spatialBlock(x=predictors, p=pres, a=background, EPSG=NULL, showBlocks=F, theRange=1000000, k=4, numLimit=20, iteration=1000, return.object=T) block.data <- ensemble.spatialBlock(x=predictors, p=pres, a=background, EPSG=4326, showBlocks=F, theRange=1000000, k=4, numLimit=20, iteration=1000, return.object=T) # explore the results foldExplorer(blocks=block.data$block.object, rasterLayer=predictors, speciesData=block.data$speciesData) # apply in calibration of ensemble weights # make sure that folds apply to subset of points p.spatial <- block.data$k$p a.spatial <- block.data$k$a k.spatial <- block.data$k ensemble.w1 <- ensemble.calibrate.weights(x=predictors, p=p.spatial, a=a.spatial, k=k.spatial, species.name="Bradypus", SINK=FALSE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=0, CF=1, GLM=1, GLMSTEP=0, GAM=1, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=1, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=0, MAHAL=0, MAHAL01=0, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, Yweights="BIOMOD", formulae.defaults=TRUE) # confirm that correct folds were used all.equal(ensemble.w1$groupp, block.data$k$groupp) all.equal(ensemble.w1$groupa, block.data$k$groupa) # environmental blocking with minimum 5 points in each category block.data2 <- ensemble.envBlock(x=predictors, p=pres, a=background, factors="biome", k=4, numLimit=5, return.object=T) # explore the results foldExplorer(blocks=block.data2$block.object, rasterLayer=predictors, speciesData=block.data2$speciesData) # apply in calibration of ensemble weights # make sure that folds apply to subset of points p.env <- block.data2$k$p a.env <- block.data2$k$a k.env <- block.data2$k ensemble.w2 <- ensemble.calibrate.weights(x=predictors, p=p.env, a=a.env, k=k.env, species.name="Bradypus", SINK=FALSE, PROBIT=TRUE, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=0, CF=1, GLM=1, GLMSTEP=0, GAM=1, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=1, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=0, MAHAL=0, MAHAL01=0, ENSEMBLE.tune=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=0.7, factors="biome", Yweights="BIOMOD", formulae.defaults=TRUE) # confirm that correct folds were used all.equal(ensemble.w2$groupp, block.data2$k$groupp) all.equal(ensemble.w2$groupa, block.data2$k$groupa) ## End(Not run)
Function ensemble.spatialThin creates a randomly selected subset of point coordinates where the shortest distance (geodesic) is above a predefined minimum. The geodesic is calculated more accurately (via distGeo) than in the spThin or red packages.
ensemble.spatialThin(x, thin.km = 0.1, runs = 100, silent = FALSE, verbose = FALSE, return.notRetained = FALSE) ensemble.spatialThin.quant(x, thin.km = 0.1, runs = 100, silent = FALSE, verbose = FALSE, LON.length = 21, LAT.length = 21) ensemble.environmentalThin(x, predictors.stack = NULL, extracted.data=NULL, thin.n = 50, runs = 100, pca.var = 0.95, silent = FALSE, verbose = FALSE, return.notRetained = FALSE) ensemble.environmentalThin.clara(x, predictors.stack = NULL, thin.n = 20, runs = 100, pca.var = 0.95, silent = FALSE, verbose = FALSE, clara.k = 100) ensemble.outlierThin(x, predictors.stack = NULL, k = 10, quant = 0.95, pca.var = 0.95, return.outliers = FALSE)ensemble.spatialThin(x, thin.km = 0.1, runs = 100, silent = FALSE, verbose = FALSE, return.notRetained = FALSE) ensemble.spatialThin.quant(x, thin.km = 0.1, runs = 100, silent = FALSE, verbose = FALSE, LON.length = 21, LAT.length = 21) ensemble.environmentalThin(x, predictors.stack = NULL, extracted.data=NULL, thin.n = 50, runs = 100, pca.var = 0.95, silent = FALSE, verbose = FALSE, return.notRetained = FALSE) ensemble.environmentalThin.clara(x, predictors.stack = NULL, thin.n = 20, runs = 100, pca.var = 0.95, silent = FALSE, verbose = FALSE, clara.k = 100) ensemble.outlierThin(x, predictors.stack = NULL, k = 10, quant = 0.95, pca.var = 0.95, return.outliers = FALSE)
x |
Point locations provided in 2-column (lon, lat) format. |
thin.km |
Threshold for minimum distance (km) in final point location data set. |
runs |
Number of runs to maximize the retained number of point coordinates. |
silent |
Do not provide any details on the process. |
verbose |
Provide some details on each run. |
return.notRetained |
Return in an additional data set the point coordinates that were thinned out. |
LON.length |
Number of quantile limits to be calculated from longitudes; see also |
LAT.length |
Number of quantile limits to be calculated from latitudes; see also |
predictors.stack |
RasterStack object ( |
extracted.data |
Data set with the environmental data at the point locations. If this data is provided, then this data will be used in the analysis and data will not be extracted from the predictors.stack. |
thin.n |
Target number of environmentally thinned points. |
pca.var |
Minimum number of axes based on the fraction of variance explained (default value of 0.95 indicates that at least 95 percent of variance will be explained on the selected number of axes). Axes and coordinates are obtained from Principal Components Analysis ( |
clara.k |
The number of clusters in which the point coordinates will be divided by |
k |
The number of neighbours for the Local Outlier Factor analysis; see |
quant |
The quantile probability above with local outlier factors are classified as outliers; see also |
return.outliers |
Return in an additional data set the point coordinates that were flagged as outliers. |
Locations with distances smaller than the threshold distance are randomly removed from the data set until no distance is smaller than the threshold. The function uses a similar algorithm as functions in the spThin or red packages, but the geodesic is more accurately calculated via distGeo.
With several runs (default of 100 as in the red package or some spThin examples), the (first) data set with the maximum number of records is retained.
Function ensemble.spatialThin.quant was designed to be used with large data sets where the size of the object with pairwise geographical distances could create memory problems. With this function, spatial thinning is only done within geographical areas defined by quantile limits of geographical coordinates.
Function ensemble.environmentalThin performs an analysis in environmental space similar to the analysis in geographical space by ensemble.spatialThin. However, the target number of retained point coordinates needs to be defined by the user. Coordinates are obtained in environmental space by a principal components analysis (function rda). Internally, first points are randomly selected from the pair with the smallest environmental distance until the selected target number of retained point coordinates is reached. From the retained point coordinates, the minimum environmental distance is determined. In a second step (more similar to spatial thinning), locations are randomly removed from all pairs that have a distance larger than the minimum distance calculated in step 1.
Function ensemble.environmentalThin.clara was designed to be used with large data sets where the size of the object with pairwise environmental distances could create memory problems. With this function, environmental thinning is done sequentially for each of the clusters defined by clara. Environmental space is obtained by by a principal components analysis (function rda). Environmental distances are calculated as the pairwise Euclidean distances between the point locations in the environmental space.
Function ensemble.outlierThin selects point coordinates that are less likely to be local outliers based on a Local Outlier Factor analysis (lof). Since LOF does not result in strict classification of outliers, a user-defined quantile probability is used to identify outliers.
The function returns a spatially or environmentally thinned point location data set.
Roeland Kindt (World Agroforestry Centre)
Aiello-Lammens ME, Boria RA, Radosavljevic A, Vilela B and Anderson RP. 2015. spThin: an R package for spatial thinning of species occurrence records for use in ecological niche models. Ecography 38: 541-545
## Not run: # get predictor variables, only needed for plotting library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[, -1] # number of locations nrow(pres) par.old <- graphics::par(no.readonly=T) par(mfrow=c(2,2)) pres.thin1 <- ensemble.spatialThin(pres, thin.km=100, runs=10, verbose=T) plot(predictors[[1]], main="5 runs", ext=extent(SpatialPoints(pres.thin1))) points(pres, pch=20, col="black") points(pres.thin1, pch=20, col="red") pres.thin2 <- ensemble.spatialThin(pres, thin.km=100, runs=10, verbose=T) plot(predictors[[1]], main="5 runs (after fresh start)", ext=extent(SpatialPoints(pres.thin2))) points(pres, pch=20, col="black") points(pres.thin2, pch=20, col="red") pres.thin3 <- ensemble.spatialThin(pres, thin.km=100, runs=100, verbose=T) plot(predictors[[1]], main="100 runs", ext=extent(SpatialPoints(pres.thin3))) points(pres, pch=20, col="black") points(pres.thin3, pch=20, col="red") pres.thin4 <- ensemble.spatialThin(pres, thin.km=100, runs=100, verbose=T) plot(predictors[[1]], main="100 runs (after fresh start)", ext=extent(SpatialPoints(pres.thin4))) points(pres, pch=20, col="black") points(pres.thin4, pch=20, col="red") graphics::par(par.old) ## thinning in environmental space env.thin <- ensemble.environmentalThin(pres, predictors.stack=predictors, thin.n=60, return.notRetained=T) pres.env1 <- env.thin$retained pres.env2 <- env.thin$not.retained # plot in geographical space par.old <- graphics::par(no.readonly=T) par(mfrow=c(1, 2)) plot(predictors[[1]], main="black = not retained", ext=extent(SpatialPoints(pres.thin3))) points(pres.env2, pch=20, col="black") points(pres.env1, pch=20, col="red") # plot in environmental space background.data <- data.frame(raster::extract(predictors, pres)) rda.result <- vegan::rda(X=background.data, scale=T) # select number of axes ax <- 2 while ( (sum(vegan::eigenvals(rda.result)[c(1:ax)])/ sum(vegan::eigenvals(rda.result))) < 0.95 ) {ax <- ax+1} rda.scores <- data.frame(vegan::scores(rda.result, display="sites", scaling=1, choices=c(1:ax))) rownames(rda.scores) <- rownames(pres) points.in <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.env1)), c(1:2)] points.out <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.env2)), c(1:2)] plot(points.out, main="black = not retained", pch=20, col="black", xlim=range(rda.scores[, 1]), ylim=range(rda.scores[, 2])) points(points.in, pch=20, col="red") graphics::par(par.old) ## removing outliers out.thin <- ensemble.outlierThin(pres, predictors.stack=predictors, k=10, return.outliers=T) pres.out1 <- out.thin$inliers pres.out2 <- out.thin$outliers # plot in geographical space par.old <- graphics::par(no.readonly=T) par(mfrow=c(1, 2)) plot(predictors[[1]], main="black = outliers", ext=extent(SpatialPoints(pres.thin3))) points(pres.out2, pch=20, col="black") points(pres.out1, pch=20, col="red") # plot in environmental space background.data <- data.frame(raster::extract(predictors, pres)) rda.result <- vegan::rda(X=background.data, scale=T) # select number of axes ax <- 2 while ( (sum(vegan::eigenvals(rda.result)[c(1:ax)])/ sum(vegan::eigenvals(rda.result))) < 0.95 ) {ax <- ax+1} rda.scores <- data.frame(vegan::scores(rda.result, display="sites", scaling=1, choices=c(1:ax))) rownames(rda.scores) <- rownames(pres) points.in <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.out1)), c(1:2)] points.out <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.out2)), c(1:2)] plot(points.out, main="black = outliers", pch=20, col="black", xlim=range(rda.scores[, 1]), ylim=range(rda.scores[, 2])) points(points.in, pch=20, col="red") graphics::par(par.old) ## End(Not run)## Not run: # get predictor variables, only needed for plotting library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17", "biome")) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[, -1] # number of locations nrow(pres) par.old <- graphics::par(no.readonly=T) par(mfrow=c(2,2)) pres.thin1 <- ensemble.spatialThin(pres, thin.km=100, runs=10, verbose=T) plot(predictors[[1]], main="5 runs", ext=extent(SpatialPoints(pres.thin1))) points(pres, pch=20, col="black") points(pres.thin1, pch=20, col="red") pres.thin2 <- ensemble.spatialThin(pres, thin.km=100, runs=10, verbose=T) plot(predictors[[1]], main="5 runs (after fresh start)", ext=extent(SpatialPoints(pres.thin2))) points(pres, pch=20, col="black") points(pres.thin2, pch=20, col="red") pres.thin3 <- ensemble.spatialThin(pres, thin.km=100, runs=100, verbose=T) plot(predictors[[1]], main="100 runs", ext=extent(SpatialPoints(pres.thin3))) points(pres, pch=20, col="black") points(pres.thin3, pch=20, col="red") pres.thin4 <- ensemble.spatialThin(pres, thin.km=100, runs=100, verbose=T) plot(predictors[[1]], main="100 runs (after fresh start)", ext=extent(SpatialPoints(pres.thin4))) points(pres, pch=20, col="black") points(pres.thin4, pch=20, col="red") graphics::par(par.old) ## thinning in environmental space env.thin <- ensemble.environmentalThin(pres, predictors.stack=predictors, thin.n=60, return.notRetained=T) pres.env1 <- env.thin$retained pres.env2 <- env.thin$not.retained # plot in geographical space par.old <- graphics::par(no.readonly=T) par(mfrow=c(1, 2)) plot(predictors[[1]], main="black = not retained", ext=extent(SpatialPoints(pres.thin3))) points(pres.env2, pch=20, col="black") points(pres.env1, pch=20, col="red") # plot in environmental space background.data <- data.frame(raster::extract(predictors, pres)) rda.result <- vegan::rda(X=background.data, scale=T) # select number of axes ax <- 2 while ( (sum(vegan::eigenvals(rda.result)[c(1:ax)])/ sum(vegan::eigenvals(rda.result))) < 0.95 ) {ax <- ax+1} rda.scores <- data.frame(vegan::scores(rda.result, display="sites", scaling=1, choices=c(1:ax))) rownames(rda.scores) <- rownames(pres) points.in <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.env1)), c(1:2)] points.out <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.env2)), c(1:2)] plot(points.out, main="black = not retained", pch=20, col="black", xlim=range(rda.scores[, 1]), ylim=range(rda.scores[, 2])) points(points.in, pch=20, col="red") graphics::par(par.old) ## removing outliers out.thin <- ensemble.outlierThin(pres, predictors.stack=predictors, k=10, return.outliers=T) pres.out1 <- out.thin$inliers pres.out2 <- out.thin$outliers # plot in geographical space par.old <- graphics::par(no.readonly=T) par(mfrow=c(1, 2)) plot(predictors[[1]], main="black = outliers", ext=extent(SpatialPoints(pres.thin3))) points(pres.out2, pch=20, col="black") points(pres.out1, pch=20, col="red") # plot in environmental space background.data <- data.frame(raster::extract(predictors, pres)) rda.result <- vegan::rda(X=background.data, scale=T) # select number of axes ax <- 2 while ( (sum(vegan::eigenvals(rda.result)[c(1:ax)])/ sum(vegan::eigenvals(rda.result))) < 0.95 ) {ax <- ax+1} rda.scores <- data.frame(vegan::scores(rda.result, display="sites", scaling=1, choices=c(1:ax))) rownames(rda.scores) <- rownames(pres) points.in <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.out1)), c(1:2)] points.out <- rda.scores[which(rownames(rda.scores) %in% rownames(pres.out2)), c(1:2)] plot(points.out, main="black = outliers", pch=20, col="black", xlim=range(rda.scores[, 1]), ylim=range(rda.scores[, 2])) points(points.in, pch=20, col="red") graphics::par(par.old) ## End(Not run)
The function ensemble.terra creates two consensus raster layers, one based on a (weighted) average of different suitability modelling algorithms, and a second one documenting the number of modelling algorithms that predict presence of the focal organisms. This function has the same behaviour as ensemble.raster.
ensemble.terra(xn = NULL, models.list = NULL, input.weights = models.list$output.weights, thresholds = models.list$thresholds, RASTER.species.name = models.list$species.name, RASTER.stack.name = "xnTitle", RASTER.filetype = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, RASTER.models.overwrite = TRUE, evaluate = FALSE, SINK = FALSE, p = models.list$p, a = models.list$a, pt = models.list$pt, at = models.list$at, CATCH.OFF = FALSE)ensemble.terra(xn = NULL, models.list = NULL, input.weights = models.list$output.weights, thresholds = models.list$thresholds, RASTER.species.name = models.list$species.name, RASTER.stack.name = "xnTitle", RASTER.filetype = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, RASTER.models.overwrite = TRUE, evaluate = FALSE, SINK = FALSE, p = models.list$p, a = models.list$a, pt = models.list$pt, at = models.list$at, CATCH.OFF = FALSE)
xn |
SpatRaster object ( |
models.list |
list with 'old' model objects such as |
input.weights |
array with numeric values for the different modelling algorithms; if |
thresholds |
array with the threshold values separating predicted presence for each of the algorithms. |
RASTER.species.name |
First part of the names of the raster files that will be generated, expected to identify the modelled species (or organism). |
RASTER.stack.name |
Last part of the names of the raster files that will be generated, expected to identify the predictor stack used. |
RASTER.filetype |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
RASTER.models.overwrite |
Overwrite the raster files that correspond to each suitability modelling algorithm (if |
evaluate |
if |
SINK |
Append the results to a text file in subfolder 'outputs' (if |
p |
presence points used for calibrating the suitability models, typically available in 2-column (x, y) or (lon, lat) dataframe; see also |
a |
background points used for calibrating the suitability models, typically available in 2-column (x, y) or (lon, lat) dataframe; see also |
pt |
presence points used for evaluating the suitability models, typically available in 2-column (lon, lat) dataframe; see also |
at |
background points used for calibrating the suitability models, typicall available in 2-column (lon, lat) dataframe; see also |
CATCH.OFF |
Disable calls to function |
The basic function ensemble.terra fits individual suitability models for all models with positive input weights. In subfolder "models" of the working directory, suitability maps for the individual suitability modelling algorithms are stored. In subfolder "ensembles", a consensus suitability map based on a weighted average of individual suitability models is stored. In subfolder "ensembles/presence", a presence-absence (1-0) map will be provided. In subfolder "ensembles/count", a consensus suitability map based on the number of individual suitability models that predict presence of the focal organism is stored.
Note that values in suitability maps are integer values that were calculated by multiplying probabilities by 1000 (see also trunc).
The basic function ensemble.terra mainly results in the creation of raster layers that correspond to fitted probabilities of presence of individual suitability models (in folder "models") and consensus models (in folder "ensembles"), and the number of suitability models that predict presence (in folder "ensembles"). Prediction of presence is based on a threshold usually defined by maximizing the sum of the true presence and true absence rates (see threshold.method and also ModelEvaluation).
Roeland Kindt (World Agroforestry Centre)
Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009
Buisson L, Thuiller W, Casajus N, Lek S and Grenouillet G. 2010. Uncertainty in ensemble forecasting of species distribution. Global Change Biology 16: 1145-1157
ensemble.raster,
evaluation.strip.plot, ensemble.calibrate.models, ensemble.calibrate.weights, ensemble.batch
## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "base" # make a SpatRaster object # Ideally this should not be created from files in the 'raster' grd format # (so a better method would be to create instead from 'tif' files). predictors.terra <- terra::rast(predictors) # predictors@title <- "base" crs(predictors.terra) <- c("+proj=longlat +datum=WGS84") predictors.terra # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # choose background points background <- dismo::randomPoints(predictors, n=1000, extf = 1.00) # if desired, change working directory where subfolders of "models" and # "ensembles" will be created # raster layers will be saved in subfolders of /models and /ensembles: getwd() # first calibrate the ensemble # calibration is done in two steps # in step 1, a k-fold procedure is used to determine the weights # in step 2, models are calibrated for all presence and background locations # Although a spatRaster (predictors.terra) object is used as input for # ensemble.calibrate.weights and ensemble.calibrate.models, # internally the spatRaster will be converted to a rasterStack for these # functions (among other things, to allow for dismo::prepareData) # step 1: determine weights through 4-fold cross-validation ensemble.calibrate.step1 <- ensemble.calibrate.weights( x=predictors.terra, p=pres, a=background, k=4, SINK=TRUE, species.name="Bradypus", MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, PROBIT=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=c(0.65, 0.7), Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 1 generated the weights for each algorithm model.weights <- ensemble.calibrate.step1$output.weights x.batch <- ensemble.calibrate.step1$x p.batch <- ensemble.calibrate.step1$p a.batch <- ensemble.calibrate.step1$a MAXENT.a.batch <- ensemble.calibrate.step1$MAXENT.a factors.batch <- ensemble.calibrate.step1$factors dummy.vars.batch <- ensemble.calibrate.step1$dummy.vars # step 2: calibrate models with all available presence locations # weights determined in step 1 calculate ensemble in step 2 ensemble.calibrate.step2 <- ensemble.calibrate.models( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, SINK=TRUE, species.name="Bradypus", models.keep=TRUE, input.weights=model.weights, ENSEMBLE.tune=FALSE, PROBIT=TRUE, Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 3: use previously calibrated models to create ensemble raster layers # re-evaluate the created maps at presence and background locations # (note that re-evaluation will be different due to truncation of raster layers # as they wered saved as integer values ranged 0 to 1000) ensemble.terra.results <- ensemble.terra(xn=predictors.terra, models.list=ensemble.calibrate.step2$models, input.weights=model.weights, SINK=TRUE, evaluate=TRUE, RASTER.species.name="Bradypus", RASTER.stack.name="base") ## End(Not run)## Not run: # based on examples in the dismo package # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors predictors@title <- "base" # make a SpatRaster object # Ideally this should not be created from files in the 'raster' grd format # (so a better method would be to create instead from 'tif' files). predictors.terra <- terra::rast(predictors) # predictors@title <- "base" crs(predictors.terra) <- c("+proj=longlat +datum=WGS84") predictors.terra # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # choose background points background <- dismo::randomPoints(predictors, n=1000, extf = 1.00) # if desired, change working directory where subfolders of "models" and # "ensembles" will be created # raster layers will be saved in subfolders of /models and /ensembles: getwd() # first calibrate the ensemble # calibration is done in two steps # in step 1, a k-fold procedure is used to determine the weights # in step 2, models are calibrated for all presence and background locations # Although a spatRaster (predictors.terra) object is used as input for # ensemble.calibrate.weights and ensemble.calibrate.models, # internally the spatRaster will be converted to a rasterStack for these # functions (among other things, to allow for dismo::prepareData) # step 1: determine weights through 4-fold cross-validation ensemble.calibrate.step1 <- ensemble.calibrate.weights( x=predictors.terra, p=pres, a=background, k=4, SINK=TRUE, species.name="Bradypus", MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, GLMNET=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, ENSEMBLE.tune=TRUE, PROBIT=TRUE, ENSEMBLE.best=0, ENSEMBLE.exponent=c(1, 2, 3), ENSEMBLE.min=c(0.65, 0.7), Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 1 generated the weights for each algorithm model.weights <- ensemble.calibrate.step1$output.weights x.batch <- ensemble.calibrate.step1$x p.batch <- ensemble.calibrate.step1$p a.batch <- ensemble.calibrate.step1$a MAXENT.a.batch <- ensemble.calibrate.step1$MAXENT.a factors.batch <- ensemble.calibrate.step1$factors dummy.vars.batch <- ensemble.calibrate.step1$dummy.vars # step 2: calibrate models with all available presence locations # weights determined in step 1 calculate ensemble in step 2 ensemble.calibrate.step2 <- ensemble.calibrate.models( x=x.batch, p=p.batch, a=a.batch, MAXENT.a=MAXENT.a.batch, factors=factors.batch, dummy.vars=dummy.vars.batch, SINK=TRUE, species.name="Bradypus", models.keep=TRUE, input.weights=model.weights, ENSEMBLE.tune=FALSE, PROBIT=TRUE, Yweights="BIOMOD", PLOTS=FALSE, formulae.defaults=TRUE) # step 3: use previously calibrated models to create ensemble raster layers # re-evaluate the created maps at presence and background locations # (note that re-evaluation will be different due to truncation of raster layers # as they wered saved as integer values ranged 0 to 1000) ensemble.terra.results <- ensemble.terra(xn=predictors.terra, models.list=ensemble.calibrate.step2$models, input.weights=model.weights, SINK=TRUE, evaluate=TRUE, RASTER.species.name="Bradypus", RASTER.stack.name="base") ## End(Not run)
Function ensemble.zones maps the zone of each raster cell within a presence map based on the minimum Mahalanobis distance (via mahalanobis) to different centroids. Function ensemble.centroids defines centroids within a presence map based on Principal Components Analysis (via rda) and K-means clustering (via kmeans).
ensemble.zones(presence.raster = NULL, centroid.object = NULL, x = NULL, ext = NULL, RASTER.species.name = centroid.object$name, RASTER.stack.name = x@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, CATCH.OFF = FALSE) ensemble.centroids(presence.raster = NULL, x = NULL, categories.raster = NULL, an = 10000, ext = NULL, name = "Species001", pca.var = 0.95, centers = 0, use.silhouette = TRUE, plotit = FALSE, dev.new.width = 7, dev.new.height = 7)ensemble.zones(presence.raster = NULL, centroid.object = NULL, x = NULL, ext = NULL, RASTER.species.name = centroid.object$name, RASTER.stack.name = x@title, RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767, CATCH.OFF = FALSE) ensemble.centroids(presence.raster = NULL, x = NULL, categories.raster = NULL, an = 10000, ext = NULL, name = "Species001", pca.var = 0.95, centers = 0, use.silhouette = TRUE, plotit = FALSE, dev.new.width = 7, dev.new.height = 7)
presence.raster |
RasterLayer object ( |
centroid.object |
Object listing values for centroids and covariance to be used with the |
x |
RasterStack object ( |
ext |
an Extent object to limit the predictions and selection of background points to a sub-region of |
RASTER.species.name |
First part of the names of the raster file that will be generated, expected to identify the modelled species (or organism) |
RASTER.stack.name |
Last part of the names of the raster file that will be generated, expected to identify the predictor stack used |
RASTER.format |
Format of the raster files that will be generated. See |
RASTER.datatype |
Format of the raster files that will be generated. See |
RASTER.NAflag |
Value that is used to store missing data. See |
CATCH.OFF |
Disable calls to function |
categories.raster |
RasterLayer object ( |
an |
Number of presence points to be used for Principal Components Analysis (via |
name |
Name for the centroid object, for example identifying the species and area for which centroids are calculated |
pca.var |
Minimum number of axes based on the fraction of variance explained (default value of 0.95 indicates that at least 95 percent of variance will be explained on the selected number of axes). Axes and coordinates are obtained from Principal Components Analysis ( |
centers |
Number of centers (clusters) to be used for K-means clustering ( |
use.silhouette |
If |
plotit |
If |
dev.new.width |
Width for new graphics device ( |
dev.new.height |
Heigth for new graphics device ( |
Function ensemble.zones maps the zone of each raster cell of a predefined presence map, whereby the zone is defined as the centroid with the smallest Mahalanobis distance. The function returns a RasterLayer object (raster) and possibly a KML layer.
Function ensemble.centroid provides the centroid locations in environmental space and a covariance matrix (cov) to be used with mahalanobis. Also provided is information on the analogue presence location that is closest to the centroid in environmental space.
Function ensemble.centroid returns a list with following objects:
centroids |
Location of centroids in environmental space |
centroid.analogs |
Location of best analogs to centroids in environmental space |
cov.mahal |
Covariance matrix |
Roeland Kindt (World Agroforestry Centre)
## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio7", "bio8", "bio12", "bio16", "bio17")) predictors predictors@title <- "base" # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) # predicted presence from GLM ensemble.calibrate.step1 <- ensemble.calibrate.models( x=predictors, p=pres, a=background, species.name="Bradypus", MAXENT=0, MAXLIKE=0, MAXNET=0, CF=0, GBM=0, GBMSTEP=0, RF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=0, DOMAIN=0, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", models.keep=TRUE) ensemble.raster.results <- ensemble.raster(xn=predictors, models.list=ensemble.calibrate.step1$models, RASTER.species.name="Bradypus", RASTER.stack.name="base") # get presence map as for example created with ensemble.raster in subfolder 'ensemble/presence' # presence values are values equal to 1 presence.file <- paste("ensembles//presence//Bradypus_base.tif", sep="") presence.raster <- raster(presence.file) # let cascadeKM decide on the number of clusters dev.new() centroids <- ensemble.centroids(presence.raster=presence.raster, x=predictors, an=1000, plotit=T) ensemble.zones(presence.raster=presence.raster, centroid.object=centroids, x=predictors, RASTER.species.name="Bradypus") dev.new() zones.file <- paste("ensembles//zones//Bradypus_base.tif", sep="") zones.raster <- raster(zones.file) max.zones <- maxValue(zones.raster) plot(zones.raster, breaks=c(0, c(1:max.zones)), col = grDevices::rainbow(n=max.zones), main="zones") ensemble.zones(presence.raster=presence.raster, centroid.object=centroids, x=predictors, RASTER.species.name="Bradypus") # manually choose 6 zones dev.new() centroids6 <- ensemble.centroids(presence.raster=presence.raster, x=predictors, an=1000, plotit=T, centers=6) ensemble.zones(presence.raster=presence.raster, centroid.object=centroids6, x=predictors, RASTER.species.name="Bradypus6") dev.new() zones.file <- paste("ensembles//zones//Bradypus6_base.tif", sep="") zones.raster <- raster(zones.file) max.zones <- maxValue(zones.raster) plot(zones.raster, breaks=c(0, c(1:max.zones)), col = grDevices::rainbow(n=max.zones), main="six zones") ## End(Not run)## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio7", "bio8", "bio12", "bio16", "bio17")) predictors predictors@title <- "base" # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) # predicted presence from GLM ensemble.calibrate.step1 <- ensemble.calibrate.models( x=predictors, p=pres, a=background, species.name="Bradypus", MAXENT=0, MAXLIKE=0, MAXNET=0, CF=0, GBM=0, GBMSTEP=0, RF=0, GLM=1, GLMSTEP=0, GAM=0, GAMSTEP=0, MGCV=0, MGCVFIX=0, EARTH=0, RPART=0, NNET=0, FDA=0, SVM=0, SVME=0, GLMNET=0, BIOCLIM.O=0, BIOCLIM=0, DOMAIN=0, MAHAL=0, MAHAL01=0, Yweights="BIOMOD", models.keep=TRUE) ensemble.raster.results <- ensemble.raster(xn=predictors, models.list=ensemble.calibrate.step1$models, RASTER.species.name="Bradypus", RASTER.stack.name="base") # get presence map as for example created with ensemble.raster in subfolder 'ensemble/presence' # presence values are values equal to 1 presence.file <- paste("ensembles//presence//Bradypus_base.tif", sep="") presence.raster <- raster(presence.file) # let cascadeKM decide on the number of clusters dev.new() centroids <- ensemble.centroids(presence.raster=presence.raster, x=predictors, an=1000, plotit=T) ensemble.zones(presence.raster=presence.raster, centroid.object=centroids, x=predictors, RASTER.species.name="Bradypus") dev.new() zones.file <- paste("ensembles//zones//Bradypus_base.tif", sep="") zones.raster <- raster(zones.file) max.zones <- maxValue(zones.raster) plot(zones.raster, breaks=c(0, c(1:max.zones)), col = grDevices::rainbow(n=max.zones), main="zones") ensemble.zones(presence.raster=presence.raster, centroid.object=centroids, x=predictors, RASTER.species.name="Bradypus") # manually choose 6 zones dev.new() centroids6 <- ensemble.centroids(presence.raster=presence.raster, x=predictors, an=1000, plotit=T, centers=6) ensemble.zones(presence.raster=presence.raster, centroid.object=centroids6, x=predictors, RASTER.species.name="Bradypus6") dev.new() zones.file <- paste("ensembles//zones//Bradypus6_base.tif", sep="") zones.raster <- raster(zones.file) max.zones <- maxValue(zones.raster) plot(zones.raster, breaks=c(0, c(1:max.zones)), col = grDevices::rainbow(n=max.zones), main="six zones") ## End(Not run)
These functions provide a dataframe which can subsequently be used to evaluate the relationship between environmental variables and the fitted probability of occurrence of individual or ensemble suitability modelling algorithms. The biomod2 package provides an alternative implementation of this approach (response.plot2).
evaluation.strip.data(xn = NULL, ext = NULL, models.list = NULL, input.weights = models.list$output.weights, steps=200, CATCH.OFF = FALSE ) evaluation.strip.plot(data, TrainData=NULL, variable.focal = NULL, model.focal = NULL, ylim=c(0, 1.25), dev.new.width = 7, dev.new.height = 7, ... )evaluation.strip.data(xn = NULL, ext = NULL, models.list = NULL, input.weights = models.list$output.weights, steps=200, CATCH.OFF = FALSE ) evaluation.strip.plot(data, TrainData=NULL, variable.focal = NULL, model.focal = NULL, ylim=c(0, 1.25), dev.new.width = 7, dev.new.height = 7, ... )
xn |
RasterStack object ( |
ext |
an Extent object to limit the prediction to a sub-region of |
models.list |
list with 'old' model objects such as |
input.weights |
array with numeric values for the different modelling algorithms; if |
steps |
number of steps within the range of a continuous explanatory variable |
CATCH.OFF |
Disable calls to function |
data |
data set with ranges of environmental variables and fitted suitability models, typically returned by |
TrainData |
Data set representing the calibration data set. If provided, then a boxplot will be added for presence locations via |
variable.focal |
focal explanatory variable for plots with evaluation strips |
model.focal |
focal model for plots with evaluation strips |
ylim |
range of Y-axis |
dev.new.width |
Width for new graphics device ( |
dev.new.height |
Heigth for new graphics device ( |
... |
Other arguments passed to |
These functions are mainly intended to be used internally by the ensemble.raster function.
evaluation.strip.data creates a data frame with variables (columns) corresponding to the environmental variables encountered in the RasterStack object (x) and the suitability modelling approaches that were defined. The variable of focal.var is an index of the variable for which values are ranged. The variable of categorical is an index for categorical (factor) variables.
A continuous (numeric) variable is ranged between its minimum and maximum values in the number of steps defined by argument steps. When a continuous variable is not the focal variable, then the average (mean) is used.
A categorical (factor) variable is ranged for all the encountered levels (levels) for this variable. When a categorical variable is not the focal variable, then the most frequent level is used.
function evaluation.strip.data creates a data frame, function evaluation.strip.data allows for plotting.
Roeland Kindt (World Agroforestry Centre)
Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009
Elith J, Ferrier S, Huettmann F & Leathwick J. 2005. The evaluation strip: A new and robust method for plotting predicted responses from species distribution models. Ecological Modelling 186: 280-289
ensemble.calibrate.models and ensemble.raster
## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors <- stack(predictors) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/5) and training (4/5) data groupp <- kfold(pres, 5) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 5) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # calibrate the models # MAXLIKE not included as does not allow predictions for data.frames # ENSEMBLE.min and ENSEMBLE.weight.min set very low to explore all # algorithms. # If focus is on actual ensemble, then set ENSEMBLE.min and # ENSEMBLE.weight.min to more usual values ensemble.calibrate <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, ENSEMBLE.min=0.5, ENSEMBLE.weight.min = 0.001, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, Yweights="BIOMOD", PLOTS=FALSE, models.keep=TRUE) # obtain data for plotting the evaluation strip strip.data <- evaluation.strip.data(xn=predictors, steps=500, models.list=ensemble.calibrate$models) # in case predictions for DOMAIN failed # however, ENSEMBLE should also be recalculated DOMAIN.model <- ensemble.calibrate$models$DOMAIN strip.data$plot.data[, "DOMAIN"] <- dismo::predict(object=DOMAIN.model, x=strip.data$plot.data) # in case predictions for MAHAL01 failed predict.MAHAL01 <- function(model, newdata, MAHAL.shape) { p <- dismo::predict(object=model, x=newdata) p <- p - 1 - MAHAL.shape p <- abs(p) p <- MAHAL.shape / p return(as.numeric(p)) } MAHAL01.model <- ensemble.calibrate$models$MAHAL01 MAHAL.shape1 <- ensemble.calibrate$models$formulae$MAHAL.shape strip.data$plot.data[, "MAHAL01"] <- predict.MAHAL01(model=MAHAL01.model, newdata=strip.data$plot.data, MAHAL.shape=MAHAL.shape1) # create graphs evaluation.strip.plot(data=strip.data$plot.data, variable.focal="bio6", TrainData=strip.data$TrainData, type="o", col="red") evaluation.strip.plot(data=strip.data$plot.data, model.focal="ENSEMBLE", TrainData=strip.data$TrainData, type="o", col="red") ## End(Not run)## Not run: # get predictor variables library(dismo) predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''), pattern='grd', full.names=TRUE) predictors <- stack(predictor.files) # subset based on Variance Inflation Factors predictors <- subset(predictors, subset=c("bio5", "bio6", "bio16", "bio17")) predictors <- stack(predictors) predictors predictors@title <- "base" # presence points presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='') pres <- read.table(presence_file, header=TRUE, sep=',')[,-1] # the kfold function randomly assigns data to groups; # groups are used as calibration (1/5) and training (4/5) data groupp <- kfold(pres, 5) pres_train <- pres[groupp != 1, ] pres_test <- pres[groupp == 1, ] # choose background points background <- randomPoints(predictors, n=1000, extf=1.00) colnames(background)=c('lon', 'lat') groupa <- kfold(background, 5) backg_train <- background[groupa != 1, ] backg_test <- background[groupa == 1, ] # calibrate the models # MAXLIKE not included as does not allow predictions for data.frames # ENSEMBLE.min and ENSEMBLE.weight.min set very low to explore all # algorithms. # If focus is on actual ensemble, then set ENSEMBLE.min and # ENSEMBLE.weight.min to more usual values ensemble.calibrate <- ensemble.calibrate.models(x=predictors, p=pres_train, a=backg_train, pt=pres_test, at=backg_test, ENSEMBLE.min=0.5, ENSEMBLE.weight.min = 0.001, MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1, GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1, Yweights="BIOMOD", PLOTS=FALSE, models.keep=TRUE) # obtain data for plotting the evaluation strip strip.data <- evaluation.strip.data(xn=predictors, steps=500, models.list=ensemble.calibrate$models) # in case predictions for DOMAIN failed # however, ENSEMBLE should also be recalculated DOMAIN.model <- ensemble.calibrate$models$DOMAIN strip.data$plot.data[, "DOMAIN"] <- dismo::predict(object=DOMAIN.model, x=strip.data$plot.data) # in case predictions for MAHAL01 failed predict.MAHAL01 <- function(model, newdata, MAHAL.shape) { p <- dismo::predict(object=model, x=newdata) p <- p - 1 - MAHAL.shape p <- abs(p) p <- MAHAL.shape / p return(as.numeric(p)) } MAHAL01.model <- ensemble.calibrate$models$MAHAL01 MAHAL.shape1 <- ensemble.calibrate$models$formulae$MAHAL.shape strip.data$plot.data[, "MAHAL01"] <- predict.MAHAL01(model=MAHAL01.model, newdata=strip.data$plot.data, MAHAL.shape=MAHAL.shape1) # create graphs evaluation.strip.plot(data=strip.data$plot.data, variable.focal="bio6", TrainData=strip.data$TrainData, type="o", col="red") evaluation.strip.plot(data=strip.data$plot.data, model.focal="ENSEMBLE", TrainData=strip.data$TrainData, type="o", col="red") ## End(Not run)
This dataset describes the abundance (number of trees with diameter at breast height equal or larger than 10 cm) of the tree species Faramea occidentalis as observed in a 1-ha quadrat survey from the Barro Colorada Island of Panama. For each quadrat, some environmental characteristics are also provided.
data(faramea)data(faramea)
A data frame with 45 observations on the following 8 variables.
UTM.EWa numeric vector
UTM.NSa numeric vector
Precipitationa numeric vector
Elevationa numeric vector
Agea numeric vector
Age.cata factor with levels c1 c2 c3
Geologya factor with levels pT Tb Tbo Tc Tcm Tgo Tl
Faramea.occidentalisa numeric vector
Although the original survey documented tree species composition of all 1-ha subplots of larger (over 1 ha) sample plot, only the first (and sometimes the last) quadrats of the larger plots were included. This selection was made to avoid that larger sample plots dominated the analysis. This selection of sites is therefore different from the selection of the 50 1-ha quadrats of the largest sample plot of the same survey (BCI and BCI.env)
This dataset is the main dataset used for the examples provided in chapters 6 and 7 of the Tree Diversity Analysis manual (Kindt & Coe, 2005).
Pyke CR, Condit R, Aguilar S and Lao S. (2001). Floristic composition across a climatic gradient in a neotropical lowland forest. Journal of Vegetation Science 12: 553-566.
Condit, R, Pitman, N, Leigh, E.G., Chave, J., Terborgh, J., Foster, R.B., Nunez, P., Aguilar, S., Valencia, R., Villa, G., Muller-Landau, H.C., Losos, E. & Hubbell, S.P. (2002). Beta-diversity in tropical forest trees. Science 295: 666-669.
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
data(faramea)data(faramea)
This data set contains information on the number of stems (individuals) and basal areas for 34 vegetation plots inventoried in February 1997 in Lothlorien forest, 37 vegetation plots inventoried in February 1996 in May Creek Forest and 36 vegetation plots inventoried in May 1995 in Yellowwood State Forest. All three sites are in Indiana, USA. Data were gathered through IFRI inventory protocols to record any tree, palm and woody climber with diameter at breast height greater than or equal to 10 cm in 10-m radius circular plots; only tree species data were kept in the example data sets (IFRI research instruments and IFRI manual section P: Forest Plot Form, section D1: Tree, Palm and Woody Climber Information).
data(ifri)data(ifri)
A data frame with 486 observations on the following 5 variables.
foresta factor with 3 levels: "LOT" (Lothlorien forest), "MCF" (May Creek Forest) and "YSF" (Yellowwood State Forest)
plotIDa factor with 107 levels providing an identification code for a 314.16 square metres (10 m radius) vegetation plot
speciesa factor with 50 levels providing an 8 character code for a tree species
counta numeric vector providing the number of stems (individuals) for each species in each vegetation plot
basala numeric vector providing the basal area (calculated from the diameter at breast height) in square cm for each species in each vegetation plot
IFRI (2014) Data from the International Forestry Resources and Institutions (IFRI) research network. http://ifri.forgov.org/
data(ifri)data(ifri)
Calculates the importance values of tree species based on frequency (calculated from number of plots), density (calculated from number of individuals) and dominance (calculated from basal area). See details.
importancevalue(x, site="plotID", species="species", count="count", basal="basal", factor="forest", level="") importancevalue.comp(x, site="plotID", species="species", count="count", basal="basal", factor="forest")importancevalue(x, site="plotID", species="species", count="count", basal="basal", factor="forest", level="") importancevalue.comp(x, site="plotID", species="species", count="count", basal="basal", factor="forest")
x |
data frame with information on plot identities, species identities, number of individuals and basal areas |
site |
factor variable providing the identities of survey plots |
species |
factor variable providing the identities of tree species |
count |
number of individuals for each tree species in each survey plot |
basal |
basal area for each tree species in each survey plot |
factor |
factor variable used to define subsets (typically different forest reserves) |
level |
level of the factor variable used to create a subset from the original data |
The importance value is calculated as the sum from (i) the relative frequency; (ii) the relative density; and (iii) the relative dominance. The importance value ranges between 0 and 300.
Frequency is calculated as the number of plots where a species is observed divided by the total number of survey plots. Relative frequency is calculated by dividing the frequency by the sum of the frequencies of all species, multiplied by 100 (to obtain a percentage).
Density is calculated as the total number of individuals of a species. Relative density is calculated by dividing the density by the sum of the densities of all species, multiplied by 100 (to obtain a percentage).
Dominance is calculated as the total basal area of a species. Relative dominance is calculated by dividing the dominance by the sum of the dominance of all species, multiplied by 100 (to obtain a percentage).
Functions importancevalue.comp applies function importancevalue to all available levels of a factor variable.
Provides information on the importance value for all tree species
Roeland Kindt (World Agroforestry Centre), Peter Newton (University of Michigan)
Curtis, J.T. & McIntosh, R. P. (1951) An Upland Forest Continuum in the Prairie-Forest Border Region of Wisconsin. Ecology 32: 476-496.
Kent, M. (2011) Vegetation Description and Data Analysis: A Practical Approach. Second edition. 428 pages.
data(ifri) importancevalue(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest', level='YSF') importancevalue.comp(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest') # When all survey plots are the same size, importance value # is not affected. Counts and basal areas now calculated per square metre ifri$count <- ifri$count/314.16 ifri$basal <- ifri$basal/314.16 importancevalue(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest', level='YSF') importancevalue.comp(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest') # Calculate diversity profiles from importance values imp <- importancevalue.comp(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest') vals <- imp[["values"]] for (i in 1:length(vals)) { imp.i <- data.frame(imp[[vals[i]]]) name.i <- paste(vals[[i]], ".Renyi", sep="") imp[[name.i]] <- renyi(imp.i$importance.value) } # LOT more diverse imp$LOT.Renyi - imp$MCF.Renyi imp$LOT.Renyi - imp$YSF.Renyi # YSF and MCF different richness and evenness imp$YSF.Renyi - imp$MCF.Renyidata(ifri) importancevalue(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest', level='YSF') importancevalue.comp(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest') # When all survey plots are the same size, importance value # is not affected. Counts and basal areas now calculated per square metre ifri$count <- ifri$count/314.16 ifri$basal <- ifri$basal/314.16 importancevalue(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest', level='YSF') importancevalue.comp(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest') # Calculate diversity profiles from importance values imp <- importancevalue.comp(ifri, site='plotID', species='species', count='count', basal='basal', factor='forest') vals <- imp[["values"]] for (i in 1:length(vals)) { imp.i <- data.frame(imp[[vals[i]]]) name.i <- paste(vals[[i]], ".Renyi", sep="") imp[[name.i]] <- renyi(imp.i$importance.value) } # LOT more diverse imp$LOT.Renyi - imp$MCF.Renyi imp$LOT.Renyi - imp$YSF.Renyi # YSF and MCF different richness and evenness imp$YSF.Renyi - imp$MCF.Renyi
This function provides citation information for all loaded packages.
loaded.citations()loaded.citations()
The function checks for the loaded packages via .packages. Citation information is provided for the base package and for all the non-standard packages via citation.
The function provides a list of all loaded packages and the relevant citation information.
Roeland Kindt (World Agroforestry Centre)
Makes a community data set from a stacked dataset (with separate variables for the site identities, the species identities and the abundance).
makecommunitydataset(x, row, column, value, factor="", level="", drop=F) stackcommunitydataset(comm, remove.zeroes=FALSE, order.sites=FALSE, order.species=FALSE)makecommunitydataset(x, row, column, value, factor="", level="", drop=F) stackcommunitydataset(comm, remove.zeroes=FALSE, order.sites=FALSE, order.species=FALSE)
x |
Data frame. |
row |
Name of the categorical variable for the rows of the crosstabulation (typically indicating sites) |
column |
Name of the categorical variable for the columns of the crosstabulation (typically indicating species) |
value |
Name of numerical variable for the cells of the crosstabulation (typically indicating abundance). The cells provide the sum of all values in the data frame. |
factor |
Name of the variable to calculate a subset of the data frame. |
level |
Value of the subset of the factor variable to calculate a subset of the data frame. |
drop |
Drop rows without species (species with total abundance of zero are always dropped) |
comm |
Community data set |
remove.zeroes |
Should rows with zero abundance be removed? |
order.sites |
Should sites be ordered alphabetically? |
order.species |
Should species be ordered alphabetically? |
makecommunitydataset calculates a cross-tabulation from a data frame, summing up all the values of the numerical variable identified as variable for the cell values. If factor="", then no subset is calculated from the data frame in the first step.
stackcommunitydataset reverses the actions of makecommunitydataset and recreates the data in stacked format.
The function provides a community dataset from another data frame.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
## Not run: dune.file <- normalizePath(paste(system.file(package="BiodiversityR"), '/etc/dunestacked.csv', sep='')) dune.stacked <- read.csv(dune.file) # dune.stacked has different variables for sites, species and abundance head(dune.stacked) dune.comm2 <- makecommunitydataset(dune.stacked, row='sites', column='species', value='abundance') # recreate the original stack dune.stacked2 <- stackcommunitydataset(dune.comm2, remove.zeroes=T) ## End(Not run)## Not run: dune.file <- normalizePath(paste(system.file(package="BiodiversityR"), '/etc/dunestacked.csv', sep='')) dune.stacked <- read.csv(dune.file) # dune.stacked has different variables for sites, species and abundance head(dune.stacked) dune.comm2 <- makecommunitydataset(dune.stacked, row='sites', column='species', value='abundance') # recreate the original stack dune.stacked2 <- stackcommunitydataset(dune.comm2, remove.zeroes=T) ## End(Not run)
This function implements pairwise comparisons for categorical variable through capscale, cca, dbrda or rda followed by anova.cca. The function simply repeats constrained ordination analysis by selecting subsets of data that correspond to two factor levels.
multiconstrained(method="capscale", formula, data, distance = "bray" , comm = NULL, add = FALSE, multicomp="", contrast=0, ...)multiconstrained(method="capscale", formula, data, distance = "bray" , comm = NULL, add = FALSE, multicomp="", contrast=0, ...)
method |
Method for constrained ordination analysis; one of "rda", "cca", "dbrda" or "capscale". |
formula |
Model formula as in |
data |
Data frame containing the variables on the right hand side of the model formula as in |
distance |
Dissimilarity (or distance) index in vegdist used if the LHS of the formula is a data frame instead of dissimilarity matrix; used only with function |
comm |
Community data frame which will be used for finding species scores when the LHS of the formula was a dissimilarity matrix as only allowed for |
add |
Logical indicating if an additive constant should be computed, and added to the non-diagonal dissimilarities such that all eigenvalues are non-negative in underlying Principal Co-ordinates Analysis; only applicable in |
multicomp |
Categorical variable used to construct the contrasts from. In case that this variable is missing, then the first explanatory variable of the formula will be used. |
contrast |
Return the ordination results for the particular contrast indicated by this number (e.g. with 5 levels, one can choose in between contrast 1-10). In case=0, then the first row of the |
... |
Other parameters passed to |
This function provides a simple expansion of capscale, cca and rda by conducting the analysis for subsets of the community and environmental datasets that only contain two levels of a categoricl variable.
When the choice is made to return results from all contrasts (contrast=0), then the first row of the anova.cca tables for each contrast are provided. It is therefore possible to compare differences in results by modifying the "by" argument of this function (i.e. obtain the total of explained variance, the variance explained on the first axis or the variance explained by the variable alone).
When the choice is made to return results from a particular contrast (contrast>0), then the ordination result is returned and two new datasets ("newcommunity" and "newenvdata") are created that only contain data for the two selected contrasts.
The function returns an ANOVA table that contains the first rows of the ANOVA tables obtained for all possible combinations of levels of the first variable. Alternatively, it returns an ordination result for the selected contrast and creates two new datasets ("newcommunity" and "newenvdata")
Roeland Kindt (World Agroforestry Centre)
Legendre, P. & Anderson, M.J. (1999). Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecological Monographs 69: 1-24.
Anderson, M.J. & Willis, T.J. (2003). Canonical analysis of principal coordinates: a useful method of constrained ordination for ecology. Ecology 84: 511-525.
## Not run: library(vegan) library(MASS) data(dune) data(dune.env) multiconstrained(method="capscale", dune~Management, data=dune.env, distance="bray",add=TRUE) multiconstrained(method="capscale", dune~Management, data=dune.env, distance="bray", add=TRUE, contrast=3) ## End(Not run)## Not run: library(vegan) library(MASS) data(dune) data(dune.env) multiconstrained(method="capscale", dune~Management, data=dune.env, distance="bray",add=TRUE) multiconstrained(method="capscale", dune~Management, data=dune.env, distance="bray", add=TRUE, contrast=3) ## End(Not run)
The functions provide nested analysis of variance for a two-level hierarchical model. The functions are implemented by estimating the correct F-ratio for the main and nested factors (assuming the nested factor is random) and using the recommended permutation procedures to test the significance of these F-ratios. F-ratios are estimated from variance estimates that are provided by distance-based redundancy analysis (capscale) or non-parametric multivariate analysis of variance (adonis2).
nested.anova.dbrda(formula, data, method="euc", add=FALSE, permutations=100, warnings=FALSE) nested.npmanova(formula, data, method="euc", permutations=100, warnings=FALSE)nested.anova.dbrda(formula, data, method="euc", add=FALSE, permutations=100, warnings=FALSE) nested.npmanova(formula, data, method="euc", permutations=100, warnings=FALSE)
formula |
Formula with a community data frame (with sites as rows, species as columns and species abundance as cell values) or (for |
data |
Environmental data set. |
method |
Method for calculating ecological distance with function |
add |
Should a constant be added to the off-diagonal elements of the distance-matrix (TRUE) or not. |
permutations |
The number of permutations for significance testing. |
warnings |
Should warnings be suppressed (TRUE) or not. |
The functions provide two alternative procedures for multivariate analysis of variance on the basis of any distance measure. Function nested.anova.dbrda proceeds via capscale, whereas nested.npmanova proceeds via adonis2. Both methods are complementary to each other as nested.npmanova always provides correct F-ratios and estimations of significance, whereas nested.anova.dbrda does not provide correct F-ratios and estimations of significance when negative eigenvalues are encountered or constants are added to the distance matrix, but always provides an ordination diagram.
The F-ratio for the main factor is estimated as the mean square of the main factor divided by the mean square of the nested factor. The significance of the F-ratio of the main factor is tested by permuting entire blocks belonging to levels of the nested factor. The significance of the F-ratio of the nested factor is tested by permuting sample units within strata defined by levels of the main factor.
The functions provide an ANOVA table.
Roeland Kindt (World Agroforestry Centre)
Legendre, P. & Anderson, M. J. (1999). Distance-based redundancy analysis: testing multispecies responses in multifactorial ecological experiments. Ecological Monographs 69, 1-24.
Anderson, M.J. (2001). A new method for non-parametric multivariate analysis of variance. Austral Ecology, 26: 32-46.
McArdle, B.H. and M.J. Anderson. (2001). Fitting multivariate models to community data: A comment on distance-based redundancy analysis. Ecology, 82: 290-297.
## Not run: library(vegan) data(warcom) data(warenv) # use larger number of permutations for real studies nested.npmanova(warcom~rift.valley+popshort, data=warenv, method="jac", permutations=5) nested.anova.dbrda(warcom~rift.valley+popshort, data=warenv, method="jac", permutations=5) ## End(Not run)## Not run: library(vegan) data(warcom) data(warenv) # use larger number of permutations for real studies nested.npmanova(warcom~rift.valley+popshort, data=warenv, method="jac", permutations=5) nested.anova.dbrda(warcom~rift.valley+popshort, data=warenv, method="jac", permutations=5) ## End(Not run)
This function provides a simplified version of the method of calculating NMS results implemented by the function metaMDS (vegan).
NMSrandom(x,perm=100,k=2,stressresult=F,method="isoMDS")NMSrandom(x,perm=100,k=2,stressresult=F,method="isoMDS")
x |
Distance matrix. |
perm |
Number of permutations to select the configuration with the lowest stress. |
k |
Number of dimensions for the non metric scaling result; passed to |
stressresult |
Provide the calculated stress for each permutation. |
method |
This function is an easier method of calculating the best NMS configuration after various random starts than implemented in the metaMDS function (vegan). The function uses a distance matrix (as calculated for example by function vegdist from a community data set) and calculates random starting positions by function initMDS (vegan) analogous to metaMDS.
The function returns the NMS ordination result with the lowest stress (calculated by isoMDS or sammon.), or the stress of each NMS ordination.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) library(MASS) data(dune) distmatrix <- vegdist(dune) Ordination.model1 <- NMSrandom(distmatrix, perm=100, k=2) Ordination.model1 <- add.spec.scores(Ordination.model1, dune, method='wa.scores') Ordination.model1library(vegan) library(MASS) data(dune) distmatrix <- vegdist(dune) Ordination.model1 <- NMSrandom(distmatrix, perm=100, k=2) Ordination.model1 <- add.spec.scores(Ordination.model1, dune, method='wa.scores') Ordination.model1
This function provides the best solution from various calls to the nnet feed-forward artificial neural networks function (nnet).
nnetrandom(formula,data,tries=10,leave.one.out=F,...)nnetrandom(formula,data,tries=10,leave.one.out=F,...)
formula |
Formula as passed to |
data |
Data as passed to |
tries |
Number of calls to |
leave.one.out |
Calculate leave-one-out predictions. |
... |
Other arguments passed to |
This function makes various calls to nnet. If desired by the user, leave-one-out statistics are provided that report the prediction if one particular sample unit was not used for iterating the networks.
The function returns the same components as nnet, but adds the following components:
range |
Summary of the observed "values". |
tries |
Number of different attempts to iterate an ANN. |
CV |
Predicted class when not using the respective sample unit for iterating ANN. |
succesful |
Test whether leave-one-out statistics provided the same class as the original class. |
Roeland Kindt (World Agroforestry Centre)
## Not run: data(faramea) faramea <- na.omit(faramea) faramea$presence <- as.numeric(faramea$Faramea.occidentalis > 0) attach(faramea) library(nnet) result <- nnetrandom(presence ~ Elevation, data=faramea, size=2, skip=FALSE, entropy=TRUE, trace=FALSE, maxit=1000, tries=100, leave.one.out=FALSE) summary(result) result$fitted.values result$value result2 <- nnetrandom(presence ~ Elevation, data=faramea, size=2, skip=FALSE, entropy=TRUE, trace=FALSE, maxit=1000, tries=50, leave.one.out=TRUE) result2$range result2$CV result2$successful ## End(Not run)## Not run: data(faramea) faramea <- na.omit(faramea) faramea$presence <- as.numeric(faramea$Faramea.occidentalis > 0) attach(faramea) library(nnet) result <- nnetrandom(presence ~ Elevation, data=faramea, size=2, skip=FALSE, entropy=TRUE, trace=FALSE, maxit=1000, tries=100, leave.one.out=FALSE) summary(result) result$fitted.values result$value result2 <- nnetrandom(presence ~ Elevation, data=faramea, size=2, skip=FALSE, entropy=TRUE, trace=FALSE, maxit=1000, tries=50, leave.one.out=TRUE) result2$range result2$CV result2$successful ## End(Not run)
A graph is produced that summarizes (through GAM as implemented by gam) how the abundance of all species of the community data set change along an ordination axis (based on the position of sites along the axis and the information from the community data set).
ordicoeno(x, ordiplot, axis = 1, legend = FALSE, cex = 0.8, ncol = 4, ...)ordicoeno(x, ordiplot, axis = 1, legend = FALSE, cex = 0.8, ncol = 4, ...)
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
ordiplot |
Ordination plot created by |
axis |
Axis of the ordination graph (1: horizontal, 2: vertical). |
legend |
if |
cex |
the amount by which plotting text and symbols should be magnified relative to the default; see also |
ncol |
the number of columns in which to set the legend items; see also |
... |
This functions investigates the relationship between the species vectors and the position of sites on an ordination axis. A GAM (gam) investigates the relationship by using the species abundances of each species as response variable, and the site position as the explanatory variable. The graph shows how the abundance of each species changes over the gradient of the ordination axis.
The function plots coenoclines and provides the expected degrees of freedom (complexity of the relationship) estimated for each species by GAM.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) library(mgcv) data(dune) Ordination.model1 <- rda(dune) plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=1) ordicoeno(dune, ordiplot=plot1, legend=TRUE)library(vegan) library(mgcv) data(dune) Ordination.model1 <- rda(dune) plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=1) ordicoeno(dune, ordiplot=plot1, legend=TRUE)
Functions to add some other graphical itmes to ordination diagrams than provided within vegan by ordihull, ordispider, ordiarrows, ordisegments, ordigrid, ordiellipse, ordicluster and lines.spantree.
ordisymbol(ordiplot, y, factor, col = 1, colors = TRUE, pchs = TRUE, rainbow_hcl = TRUE, rainbow_hcl.c = 90, rainbow_hcl.l = 50, rainbow = TRUE, heat.colors = FALSE, terrain.colors = FALSE, topo.colors = FALSE, cm.colors = FALSE, legend = TRUE, legend.x = "topleft", legend.ncol = 1, ...) ordibubble(ordiplot,var,...) ordicluster2(ordiplot, cluster, mingroups = 1, maxgroups = nrow(ordiplot$sites), ...) ordinearest(ordiplot, dist,...) ordivector(ordiplot, spec, lty=2,...)ordisymbol(ordiplot, y, factor, col = 1, colors = TRUE, pchs = TRUE, rainbow_hcl = TRUE, rainbow_hcl.c = 90, rainbow_hcl.l = 50, rainbow = TRUE, heat.colors = FALSE, terrain.colors = FALSE, topo.colors = FALSE, cm.colors = FALSE, legend = TRUE, legend.x = "topleft", legend.ncol = 1, ...) ordibubble(ordiplot,var,...) ordicluster2(ordiplot, cluster, mingroups = 1, maxgroups = nrow(ordiplot$sites), ...) ordinearest(ordiplot, dist,...) ordivector(ordiplot, spec, lty=2,...)
ordiplot |
An ordination graph created by |
y |
Environmental data frame. |
factor |
Variable of the environmental data frame that defines subsets to be given different symbols. |
var |
Continous variable of the environmental dataset or species from the community dataset. |
col |
Colour (as |
colors |
Apply different colours to different factor levels |
pchs |
Apply different symbols (plotting characters) to different factor levels (as in |
rainbow_hcl |
Use rainbow_hcl colours ( |
rainbow_hcl.c |
Set the chroma value |
rainbow_hcl.l |
Set the luminance value |
rainbow |
Use rainbow colours |
heat.colors |
Use heat colours |
terrain.colors |
Use terrain colours |
topo.colors |
Use topo colours |
cm.colors |
Use cm colours |
legend |
Add the legend. |
legend.x |
Location of the legend; see also |
legend.ncol |
the number of columns in which to set the legend items; see also |
cluster |
Cluster object. |
mingroups |
Minimum of clusters to be plotted. |
maxgroups |
Maximum of clusters to be plotted.. |
dist |
Distance matrix. |
spec |
Species name from the community dataset. |
lty |
Line type as specified for |
... |
Other arguments passed to functions |
Function ordisymbol plots different levels of the specified variable in different symbols and different colours. In case more than one colour palettes are selected, the last palette selected will be used.
Function ordibubble draws bubble diagrams indicating the value of the specified continuous variable. Circles indicate positive values, squares indicate negative values.
Function ordicluster2 provides an alternative method of overlaying information from hierarchical clustering on an ordination diagram than provided by function ordicluster. The method draws convex hulls around sites that are grouped into the same cluster. You can select the minimum and maximum number of clusters that are plotted (i.e. the range of clustering steps to be shown).
Function ordinearest draws a vector from each site to the site that is nearest to it as determined from a distance matrix. When you combine the method with lines.spantree using the same distance measure, then you can evaluate in part how the minimum spanning tree was constructed.
Function ordivector draws a vector for the specified species on the ordination diagramme and draws perpendicular lines from each site to a line that connects the origin and the head of species vector. This method helps in the biplot interpretation of a species vector as described by Jongman, ter Braak and van Tongeren (1995).
These functions add graphical items to an existing ordination diagram.
Roeland Kindt (World Agroforestry Centre) and Jari Oksanen (ordinearest)
Jongman, R.H.G, ter Braak, C.J.F & van Tongeren, O.F.R. (1987). Data Analysis in Community and Landscape Ecology. Pudog, Wageningen.
Kindt, R. & Coe, R. (2005). Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune) data(dune.env) Ordination.model1 <- rda(dune) plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=2) ordisymbol(plot1, dune.env, "Management", legend=TRUE, legend.x="topleft", legend.ncol=1) plot2 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=1) distmatrix <- vegdist(dune, method='bray') cluster <- hclust(distmatrix, method='single') ordicluster2(plot2, cluster) ordinearest(plot2, distmatrix, col=2) ordivector(plot2, "Agrostol", lty=2)library(vegan) data(dune) data(dune.env) Ordination.model1 <- rda(dune) plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=2) ordisymbol(plot1, dune.env, "Management", legend=TRUE, legend.x="topleft", legend.ncol=1) plot2 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=1) distmatrix <- vegdist(dune, method='bray') cluster <- hclust(distmatrix, method='single') ordicluster2(plot2, cluster) ordinearest(plot2, distmatrix, col=2) ordivector(plot2, "Agrostol", lty=2)
Calculates the number of significant axes from a Principal Components Analysis based on the broken-stick criterion, or adds an equilibrium circle to an ordination diagram.
PCAsignificance(pca,axes=8) ordiequilibriumcircle(pca,ordiplot,...)PCAsignificance(pca,axes=8) ordiequilibriumcircle(pca,ordiplot,...)
pca |
Principal Components Analysis result as calculated by |
axes |
Number of axes to calculate results for. |
ordiplot |
Ordination plot created by |
... |
Other arguments passed to function |
These functions provide two methods of providing some information on significance for a Principal Components Analysis (PCA).
Function PCAsignificance uses the broken-stick distribution to evaluate how many PCA axes are significant. This criterion is one of the most reliable to check how many axes are significant. PCA axes with larger percentages of (accumulated) variance than the broken-stick variances are significant (Legendre and Legendre, 1998).
Function ordiequilibriumcircle draws an equilibirum circle to a PCA ordination diagram. Only species vectors with heads outside of the equilibrium circle significantly contribute to the ordination diagram (Legendre and Legendre, 1998). Vectors are drawn for these species. The function considers the scaling methods used by rda for scaling=1. The method should only be used for scaling=1 and PCA calculated by function rda.
Function PCAsignificance returns a matrix with the variances that are explained by the PCA axes and by the broken-stick criterion.
Function ordiequilibriumcircle plots an equilibirum circle and returns a list with the radius and the scaling constant used by rda.
Roeland Kindt (World Agroforestry Centre)
Legendre, P. & Legendre, L. (1998). Numerical Ecology. 2nd English Edition. Elsevier.
Kindt, R. & Coe, R. (2005). Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune) Ordination.model1 <- rda(dune) PCAsignificance(Ordination.model1) plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=1) ordiequilibriumcircle(Ordination.model1,plot1)library(vegan) data(dune) Ordination.model1 <- rda(dune) PCAsignificance(Ordination.model1) plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling=1) ordiequilibriumcircle(Ordination.model1,plot1)
Provides alternative methods of obtaining rank abundance curves than provided by functions radfit, fisherfit and prestonfit (vegan), although these same functions are called.
radfitresult(x,y="",factor="",level,plotit=T)radfitresult(x,y="",factor="",level,plotit=T)
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
y |
Environmental data frame. |
factor |
Variable of the environmental data frame that defines subsets to calculate fitted rank-abundance curves for. |
level |
Level of the variable to create the subset to calculate fitted rank-abundance curves. |
plotit |
Plot the results obtained by |
These functions provide some alternative methods of obtaining fitted rank-abundance curves, although functions radfit, fisherfit and prestonfit (vegan) are called to calculate the actual results.
The function returns the results from three methods of fitting rank-abundance curves:
radfit |
results of |
fisherfit |
results of |
prestonfit |
results of |
Optionally, a plot is provided of the radfit results by plot.radfit.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(BCI) BCIall <- t(as.matrix(colSums(BCI))) radfitresult(BCIall)library(vegan) data(BCI) BCIall <- t(as.matrix(colSums(BCI))) radfitresult(BCIall)
Provides methods of calculating rank-abundance curves.
rankabundance(x, y="", factor="", level, digits=1, t=qt(0.975, df=n-1)) rankabunplot(xr, addit=F, labels="", scale="abundance", scaledx=F, type="o", xlim=c(min(xpos), max(xpos)), ylim=c(0, max(x[,scale])), specnames=c(1:5), srt=0, ...) rankabuncomp(x, y="", factor, return.data=T, specnames=c(1:3), scale="abundance", scaledx=F, type="o", rainbow=T, legend=T, xlim=c(1, max1), ylim=c(0, max2), ...)rankabundance(x, y="", factor="", level, digits=1, t=qt(0.975, df=n-1)) rankabunplot(xr, addit=F, labels="", scale="abundance", scaledx=F, type="o", xlim=c(min(xpos), max(xpos)), ylim=c(0, max(x[,scale])), specnames=c(1:5), srt=0, ...) rankabuncomp(x, y="", factor, return.data=T, specnames=c(1:3), scale="abundance", scaledx=F, type="o", rainbow=T, legend=T, xlim=c(1, max1), ylim=c(0, max2), ...)
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
y |
Environmental data frame. |
factor |
Variable of the environmental data frame that defines subsets to calculate rank abundance curves for. |
level |
Level of the variable to create the subset to calculate rank abundance curves. |
digits |
Number of digits in the results. |
t |
t-value to calculate confidence interval limits for the species proportion for cluster sampling (following Hayek and Buzas 1997). |
xr |
Result from |
addit |
Add rank abundance curve to an existing graph. |
labels |
Labels to plot at left of the rank abundance curves. |
scale |
Method of scaling the vertical axis. Method "abundance" uses abundance, "proportion" uses proportional abundance (species abundance / total abundance), "logabun" calculates the logarithm of abundance using base 10 and "accumfreq" accumulates the proportional abundance. |
scaledx |
Scale the horizontal axis to 100 percent of total number of species. |
type |
Type of plot (as in function |
xlim |
Limits for the horizontal axis. |
ylim |
Limits for the vertical axis. |
specnames |
Vector positions of species names to add to the rank-abundance curve. |
srt |
The string rotation in degrees of the species names (as in |
return.data |
Return the data used for plotting. |
rainbow |
Use rainbow colouring for the different curves. |
legend |
Add the legend (you need to click in the graph where the legend needs to be plotted). |
... |
These functions provide methods of calculating and plotting rank-abundance curves.
The vertical axis can be scaled by various methods. Method "abundance" uses abundance, "proportion" uses proportional abundance (species abundance / total abundance), "logabun" calculates the logarithm of abundance using base 10 and "accumfreq" accumulates the proportional abundance.
The horizontal axis can be scaled by the total number of species, or by 100 percent of all species by option "scaledx".
The method of calculating the confidence interval for species proportion is described in Hayek and Buzas (1997).
Functions rankabundance and rankabuncomp allow to calculate rank abundance curves for subsets of the community and environmental data sets. Function rankabundance calculates the rank abundance curve for the specified level of a selected environmental variable. Method rankabuncomp calculates the rank abundance curve for all levels of a selected environmental variable separatedly.
The functions provide information on rankabundance curves. Function rankabundance provides information on abundance, proportional abundance, logarithmic abundance and accumulated proportional abundance. The function also provides confidence interval limits for the proportion of each species (plower, pupper) and the proportion of species ranks (in percentage).
Roeland Kindt (World Agroforestry Centre)
Hayek, L.-A. C. & Buzas, M.A. (1997). Surveying Natural Populations. Columbia University Press.
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune.env) data(dune) RankAbun.1 <- rankabundance(dune) RankAbun.1 rankabunplot(RankAbun.1, scale='abundance', addit=FALSE, specnames=c(1,2,3)) rankabunplot(RankAbun.1, scale='logabun', addit=FALSE, specnames=c(1:30), srt=45, ylim=c(1,100)) rankabuncomp(dune, y=dune.env, factor='Management', scale='proportion', legend=FALSE) ## CLICK IN THE GRAPH TO INDICATE WHERE THE LEGEND NEEDS TO BE PLACED ## IF YOU OPT FOR LEGEND=TRUE. ## Not run: # ggplot2 plotting method # Only label the two most abundant species RA.data <- rankabuncomp(dune, y=dune.env, factor='Management', return.data=TRUE, specnames=c(1:2), legend=FALSE) library(ggplot2) library(ggrepel) # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) plotgg1 <- ggplot(data=RA.data, aes(x = rank, y = abundance)) + scale_x_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(aes(colour=Grouping), size=1) + geom_point(aes(colour=Grouping, shape=Grouping), size=5, alpha=0.7) + geom_text_repel(data=subset(RA.data, labelit == TRUE), aes(colour=Grouping, label=species), angle=45, nudge_x=1, nudge_y=1, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x = "rank", y = "abundance", colour = "Management", shape = "Management") plotgg1 # use different facets # now label first 10 species RA.data <- rankabuncomp(dune, y=dune.env, factor='Management', return.data=TRUE, specnames=c(1:10), legend=FALSE) plotgg2 <- ggplot(data=RA.data, aes(x = rank, y = abundance)) + scale_x_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(aes(colour=Grouping), size=1) + geom_point(aes(colour=Grouping), size=5, alpha=0.7) + geom_text_repel(data=subset(RA.data, labelit == TRUE), aes(label=species), angle=45, nudge_x=1, nudge_y=1, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + facet_wrap(~ Grouping) + labs(x = "rank", y = "abundance", colour = "Management") plotgg2 ## End(Not run) # dontrunlibrary(vegan) data(dune.env) data(dune) RankAbun.1 <- rankabundance(dune) RankAbun.1 rankabunplot(RankAbun.1, scale='abundance', addit=FALSE, specnames=c(1,2,3)) rankabunplot(RankAbun.1, scale='logabun', addit=FALSE, specnames=c(1:30), srt=45, ylim=c(1,100)) rankabuncomp(dune, y=dune.env, factor='Management', scale='proportion', legend=FALSE) ## CLICK IN THE GRAPH TO INDICATE WHERE THE LEGEND NEEDS TO BE PLACED ## IF YOU OPT FOR LEGEND=TRUE. ## Not run: # ggplot2 plotting method # Only label the two most abundant species RA.data <- rankabuncomp(dune, y=dune.env, factor='Management', return.data=TRUE, specnames=c(1:2), legend=FALSE) library(ggplot2) library(ggrepel) # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) plotgg1 <- ggplot(data=RA.data, aes(x = rank, y = abundance)) + scale_x_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(aes(colour=Grouping), size=1) + geom_point(aes(colour=Grouping, shape=Grouping), size=5, alpha=0.7) + geom_text_repel(data=subset(RA.data, labelit == TRUE), aes(colour=Grouping, label=species), angle=45, nudge_x=1, nudge_y=1, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x = "rank", y = "abundance", colour = "Management", shape = "Management") plotgg1 # use different facets # now label first 10 species RA.data <- rankabuncomp(dune, y=dune.env, factor='Management', return.data=TRUE, specnames=c(1:10), legend=FALSE) plotgg2 <- ggplot(data=RA.data, aes(x = rank, y = abundance)) + scale_x_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(expand=c(0, 1), sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(aes(colour=Grouping), size=1) + geom_point(aes(colour=Grouping), size=5, alpha=0.7) + geom_text_repel(data=subset(RA.data, labelit == TRUE), aes(label=species), angle=45, nudge_x=1, nudge_y=1, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + facet_wrap(~ Grouping) + labs(x = "rank", y = "abundance", colour = "Management") plotgg2 ## End(Not run) # dontrun
These functions may assist to ensure that the sites of the community dataset are the same sites as those from the environmental dataset, something that is assumed to be the case for the BiodiversityR and vegan packages.
same.sites(x, y) check.datasets(x, y) check.ordiscores(x, ord, check.species = TRUE) removeNAcomm(x, y, variable) removeNAenv(x, variable) removezerospecies(x) subsetcomm(x, y, factor, level, returncomm = TRUE) import.with.readxl(file = file.choose(), data.type = "community", sheet = NULL, sitenames = "sites", column = "species", value = "abundance", factor = "", level = "", cepnames = FALSE, write.csv = FALSE, csv.file = paste(data.type, ".csv", sep=""))same.sites(x, y) check.datasets(x, y) check.ordiscores(x, ord, check.species = TRUE) removeNAcomm(x, y, variable) removeNAenv(x, variable) removezerospecies(x) subsetcomm(x, y, factor, level, returncomm = TRUE) import.with.readxl(file = file.choose(), data.type = "community", sheet = NULL, sitenames = "sites", column = "species", value = "abundance", factor = "", level = "", cepnames = FALSE, write.csv = FALSE, csv.file = paste(data.type, ".csv", sep=""))
x |
Data frame assumed to be the community dataset with variables corresponding to species. |
y |
Data frame assumed to be the environmental dataset with variables corresponding to descriptors of sites. |
ord |
Ordination result. |
check.species |
Should the species scores be checked (TRUE) or not. |
variable |
Name of the variable from the environmental dataset with NA values that indicate those sites that should be removed. |
factor |
Variable of the environmental data frame that defines subsets for the data frame. |
level |
Level of the variable to create the subsets for the data frame. |
returncomm |
For the selected sites, return the community dataset (TRUE) or the environmental dataset. |
file |
Location of the Excel (or Access) file. |
data.type |
Type of the data set to be imported: one of "community", "environmental" or "stacked". |
sheet |
Name of the sheet of the Excel file to import from (if missing, then |
sitenames |
Name of categorical variable that provides the names for the sites. |
column |
Name of the categorical variable for the columns of the crosstabulation (typically indicating species); passed to |
value |
Name of numerical variable for the cells of the crosstabulation (typically indicating abundance). The cells provide the sum of all values in the data frame; passed to |
cepnames |
Should the names of columns be abbreviated via |
write.csv |
Create a comma-delimited text file in the working directory (if |
csv.file |
Name of the comma-delimited text file to be created. |
Function same.sites provides a new data frame that has the same row names as the row names of the environmental data set and the same (species) variables as the original community data set. Sites from the original community data set that have no corresponding sites in the environmental data set are not included in the new community data set. (Hint: this function can be especially useful when some sites do not contain any species and where a community dataset was generated by the makecommunitydataset function.)
Function check.datasets checks whether the community and environmental data sets have the same number of rows, and (if this was the case) whether the rownames of both data sets are the same. The function also returns the dimensions of both data sets.
Function check.ordiscores checks whether the community data set and the ordination result have the same number of rows (sites) and columns (species, optional for check.species==TRUE), and (if this was the case) whether the row and column names of both data sets are the same. Site and species scores for the ordination result are obtained via function scores (vegan).
Functions removeNAcomm and removeNAenv provide a new data frame that does not contain NA for the specified variable. The specifed variable is part of the environmental data set. These functions are particularly useful when using community and environmental datasets, as new community and environmental datasets can be calculated that contain information from the same sample plots (sites). An additional result of removeNAenv is that factor levels of any categorical variable that do not occur any longer in the new data set are removed from the levels of the categorical variable.
Function replaceNAcomm substitutes cells containing NA with 0 in the community data set.
Function removezerospecies removes species from a community dataset that have total abundance that is smaller or equal to zero.
Function subsetcomm makes a subset of sites that contain a specified level of a categorical variable from the environmental data set. The same functionality of selecting subsets of the community or environmental data sets are implemented in various functions of BiodiversityR (for example diversityresult, renyiresult and accumresult) and have the advantage that it is not necessary to create a new data set. If a community dataset is returned, species that did not contain any individuals were removed from the data set. If an environmental dataset is returned, factor levels that did not occur were removed from the data set.
Function import.with.readxl provides methods of importing community or environmental datasets through read_excel.
For stacked datasets, a community data set is created with function makecommunitydataset. For community data with more species than the limited number of columns in Excel, this may be the only option of importing a community dataset.
An additional advantage of the function is that the community and environmental data can be stored in the same file.
You may want to check compatibility of the community and environmental datasets with functions check.datasets and modify the community dataset through same.sites.
The functions return a data frame or results of tests on the correspondence between community and environmental data sets.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(vegan) data(dune.env) data(dune) dune.env2 <- dune.env dune.env2[1:4,"Moisture"] <- NA dune2 <- removeNAcomm(dune,dune.env2,"Moisture") dune.env2 <- removeNAenv(dune.env2,"Moisture") dune3 <- same.sites(dune,dune.env2) check.datasets(dune,dune.env2) check.datasets(dune2,dune.env2) check.datasets(dune3,dune.env2) dune4 <- subsetcomm(dune,dune.env,"Management","NM",returncomm=TRUE) dune.env4 <- subsetcomm(dune,dune.env,"Management","NM",returncomm=FALSE) dune5 <- same.sites(dune,dune.env4) check.datasets(dune4,dune5)library(vegan) data(dune.env) data(dune) dune.env2 <- dune.env dune.env2[1:4,"Moisture"] <- NA dune2 <- removeNAcomm(dune,dune.env2,"Moisture") dune.env2 <- removeNAenv(dune.env2,"Moisture") dune3 <- same.sites(dune,dune.env2) check.datasets(dune,dune.env2) check.datasets(dune2,dune.env2) check.datasets(dune3,dune.env2) dune4 <- subsetcomm(dune,dune.env,"Management","NM",returncomm=TRUE) dune.env4 <- subsetcomm(dune,dune.env,"Management","NM",returncomm=FALSE) dune5 <- same.sites(dune,dune.env4) check.datasets(dune4,dune5)
Provides some alternative methods of obtaining results on Renyi diversity profile values than provided by renyi (vegan).
renyiresult(x, y=NULL, factor, level, method = "all", scales = c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), evenness = FALSE ,...) renyiplot(xr, addit=FALSE, pch = 1, xlab = "alpha", ylab = "H-alpha", ylim = NULL, labelit = TRUE, legend = TRUE, legend.x="topleft", legend.ncol = 8, col = 1, cex = 1, rainbow = TRUE, evenness = FALSE, ...) renyiaccumresult(x, y=NULL, factor, level, scales=c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), permutations = 100,...) renyicomp(x, y, factor, sites=Inf, scales = c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), permutations = 100, plotit = FALSE, ...)renyiresult(x, y=NULL, factor, level, method = "all", scales = c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), evenness = FALSE ,...) renyiplot(xr, addit=FALSE, pch = 1, xlab = "alpha", ylab = "H-alpha", ylim = NULL, labelit = TRUE, legend = TRUE, legend.x="topleft", legend.ncol = 8, col = 1, cex = 1, rainbow = TRUE, evenness = FALSE, ...) renyiaccumresult(x, y=NULL, factor, level, scales=c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), permutations = 100,...) renyicomp(x, y, factor, sites=Inf, scales = c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), permutations = 100, plotit = FALSE, ...)
x |
Community data frame with sites as rows, species as columns and species abundance as cell values. |
y |
Environmental data frame. |
factor |
Variable of the environmental data frame that defines subsets to calculate diversity profiles for. |
level |
Level of the variable to create the subset to calculate diversity profiles. |
method |
Method of calculating the diversity profiles: "all" calculates the diversity of the entire community (all sites pooled together), "s" calculates the diversity of each site separatedly. |
scales |
Scale parameter values as in function |
evenness |
Calculate or plot the evenness profile. |
xr |
Result from |
addit |
Add diversity profile to an existing graph. |
pch |
Symbol used for drawing the diversity profiles (as in function |
xlab |
Label for the horizontal axis. |
ylab |
Label for the vertical axis. |
ylim |
Limits of the vertical axis. |
labelit |
Provide site labels (site names) at beginning and end of the diversity profiles. |
legend |
Add the legend (you need to click in the graph where the legend needs to be plotted). |
legend.x |
Location of the legend; see also |
legend.ncol |
number of columns for the legend (as in function |
col |
Colour for the diversity profile (as in function |
cex |
Character expansion factor (as in function |
rainbow |
Use rainbow colours for the diversity profiles. |
sites |
Number of accumulated sites to provide profile values. |
permutations |
Number of permutations for the Monte-Carlo simulations for accumulated renyi diversity profiles (estimated by |
plotit |
Plot the results (you need to click in the graph where the legend should be plotted). |
... |
These functions provide some alternative methods of obtaining results with diversity profiles, although function renyi is always used to calculate the diversity profiles.
The method of calculating the diversity profiles: "all" calculates the diversity profile of the entire community (all sites pooled together), whereas "s" calculates the diversity profile of each site separatedly. The evenness profile is calculated by subtracting the profile value at scale 0 from all the profile values.
Functions renyiresult, renyiaccumresult and renyicomp allow to calculate diversity profiles for subsets of the community and environmental data sets. functions renyiresult and renyiaccumresult calculate the diversity profiles for the specified level of a selected environmental variable. Method renyicomp calculates the diversity profile for all levels of a selected environmental variable separatedly.
Functions renyicomp and renyiaccumresult calculate accumulation curves for the Renyi diversity profile by randomised pooling of sites and calculating diversity profiles for the pooled sites as implemented in renyiaccum. The method is similar to the random method of species accumulation (specaccum). If the number of "sites" is not changed from the default, it is replaced by the sample size of the level with the fewest number of sites.
The functions provide alternative methods of obtaining Renyi diversity profiles.
Roeland Kindt (World Agroforestry Centre)
Kindt R., Degrande A., Turyomurugyendo L., Mbosso C., Van Damme P., Simons A.J. (2001). Comparing species richness and evenness contributions to on-farm tree diversity for data sets with varying sample sizes from Kenya, Uganda, Cameroon and Nigeria with randomised diversity profiles. Paper presented at IUFRO conference on forest biometry, modeling and information science, 26-29 June, University of Greenwich, UK
Kindt R. (2002). Methodology for tree species diversification planning for African agroecosystems. Thesis submitted in fulfilment of the requirement of the degree of doctor (PhD) in applied biological sciences. Faculty of agricultural and applied biological sciences, Ghent University, Ghent (Belgium), 332+xi pp.
Kindt R., Van Damme P. & Simons A.J. (2006). Tree diversity in western Kenya: using diversity profiles to characterise richness and evenness. Biodiversity and Conservation 15: 1253-1270.
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
https://rpubs.com/Roeland-KINDT
library(vegan) data(dune.env) data(dune) Renyi.1 <- renyiresult(dune, y=dune.env, factor='Management', level='NM', method='s') Renyi.1 renyiplot(Renyi.1, evenness=FALSE, addit=FALSE, pch=1,col='1', cex=1, legend=FALSE) ## CLICK IN THE GRAPH TO INDICATE WHERE THE LEGEND NEEDS TO BE PLACED ## IN CASE THAT YOU OPT FOR LEGEND=TRUE ## Not run: # ggplot2 plotting method Renyi.2 <- renyicomp(dune, y=dune.env, factor='Management', scales=c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), permutations=100, plotit=F) Renyi.2 library(ggplot2) # change the theme # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) renyi.long2 <- renyicomp.long(Renyi.2, label.freq=1) plotgg1 <- ggplot(data=renyi.long2, aes(x=Scales, y=Diversity, ymax=UPR, ymin=LWR)) + scale_x_discrete() + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(data=renyi.long2, aes(x=Obs, colour=Grouping), size=2) + geom_point(data=subset(renyi.long2, labelit==TRUE), aes(colour=Grouping, shape=Grouping), size=5) + geom_ribbon(data=renyi.long2, aes(x=Obs, colour=Grouping), alpha=0.2, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x=expression(alpha), y = "Diversity", colour = "Management", shape = "Management") plotgg1 # calculate a separate diversity profile for each site Renyi.3 <- renyiresult(dune, evenness=FALSE, method="s", scales=c(0, 0.25, 0.5, 1, 2, 4, 8, Inf)) Renyi.3 renyi.long3 <- renyi.long(Renyi.3, env.data=dune.env, label.freq=2) plotgg2 <- ggplot(data=renyi.long3, aes(x=Scales, y=Diversity, group=Grouping)) + scale_x_discrete() + scale_y_continuous(sec.axis = dup_axis(name=NULL)) + geom_line(aes(colour=Management), size=2) + geom_point(data=subset(renyi.long3, labelit==TRUE), aes(colour=Management, shape=Management), size=5) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x=expression(alpha), y="Diversity", colour="Management") plotgg2 plotgg3 <- ggplot(data=renyi.long3, aes(x=Scales, y=Diversity, group=Grouping)) + scale_x_discrete() + scale_y_continuous(sec.axis = dup_axis(name=NULL)) + geom_line(aes(colour=Management), size=1) + geom_point(data=subset(renyi.long3, labelit==TRUE), aes(colour=Management, shape=Management), size=2) + BioR.theme + scale_color_brewer(palette = "Set1") + facet_wrap(~ Management) + labs(x=expression(alpha), y="Diversity", colour="Management") plotgg3 ## End(Not run) # dontrunlibrary(vegan) data(dune.env) data(dune) Renyi.1 <- renyiresult(dune, y=dune.env, factor='Management', level='NM', method='s') Renyi.1 renyiplot(Renyi.1, evenness=FALSE, addit=FALSE, pch=1,col='1', cex=1, legend=FALSE) ## CLICK IN THE GRAPH TO INDICATE WHERE THE LEGEND NEEDS TO BE PLACED ## IN CASE THAT YOU OPT FOR LEGEND=TRUE ## Not run: # ggplot2 plotting method Renyi.2 <- renyicomp(dune, y=dune.env, factor='Management', scales=c(0, 0.25, 0.5, 1, 2, 4, 8, Inf), permutations=100, plotit=F) Renyi.2 library(ggplot2) # change the theme # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) renyi.long2 <- renyicomp.long(Renyi.2, label.freq=1) plotgg1 <- ggplot(data=renyi.long2, aes(x=Scales, y=Diversity, ymax=UPR, ymin=LWR)) + scale_x_discrete() + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_line(data=renyi.long2, aes(x=Obs, colour=Grouping), size=2) + geom_point(data=subset(renyi.long2, labelit==TRUE), aes(colour=Grouping, shape=Grouping), size=5) + geom_ribbon(data=renyi.long2, aes(x=Obs, colour=Grouping), alpha=0.2, show.legend=FALSE) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x=expression(alpha), y = "Diversity", colour = "Management", shape = "Management") plotgg1 # calculate a separate diversity profile for each site Renyi.3 <- renyiresult(dune, evenness=FALSE, method="s", scales=c(0, 0.25, 0.5, 1, 2, 4, 8, Inf)) Renyi.3 renyi.long3 <- renyi.long(Renyi.3, env.data=dune.env, label.freq=2) plotgg2 <- ggplot(data=renyi.long3, aes(x=Scales, y=Diversity, group=Grouping)) + scale_x_discrete() + scale_y_continuous(sec.axis = dup_axis(name=NULL)) + geom_line(aes(colour=Management), size=2) + geom_point(data=subset(renyi.long3, labelit==TRUE), aes(colour=Management, shape=Management), size=5) + BioR.theme + scale_color_brewer(palette = "Set1") + labs(x=expression(alpha), y="Diversity", colour="Management") plotgg2 plotgg3 <- ggplot(data=renyi.long3, aes(x=Scales, y=Diversity, group=Grouping)) + scale_x_discrete() + scale_y_continuous(sec.axis = dup_axis(name=NULL)) + geom_line(aes(colour=Management), size=1) + geom_point(data=subset(renyi.long3, labelit==TRUE), aes(colour=Management, shape=Management), size=2) + BioR.theme + scale_color_brewer(palette = "Set1") + facet_wrap(~ Management) + labs(x=expression(alpha), y="Diversity", colour="Management") plotgg3 ## End(Not run) # dontrun
These functions organize outputs from ordiplot, accumcomp and renyicomp so these can be plotted with ggplot.
sites.long(x, env.data = NULL) species.long(x, spec.data = NULL) centroids.long(y, grouping, FUN = mean, centroids.only = FALSE) vectorfit.long(z) ordisurfgrid.long(z) ordiellipse.long(z, grouping.name = "Grouping") pvclust.long(cl, cl.data) axis.long(w, choices = c(1, 2), cmdscale.model=FALSE, CAPdiscrim.model=FALSE) accumcomp.long(x, ci = 2, label.freq = 1) renyicomp.long(x, label.freq = 1) renyi.long(x, env.data=NULL, label.freq = 1)sites.long(x, env.data = NULL) species.long(x, spec.data = NULL) centroids.long(y, grouping, FUN = mean, centroids.only = FALSE) vectorfit.long(z) ordisurfgrid.long(z) ordiellipse.long(z, grouping.name = "Grouping") pvclust.long(cl, cl.data) axis.long(w, choices = c(1, 2), cmdscale.model=FALSE, CAPdiscrim.model=FALSE) accumcomp.long(x, ci = 2, label.freq = 1) renyicomp.long(x, label.freq = 1) renyi.long(x, env.data=NULL, label.freq = 1)
x |
|
env.data |
Environmental descriptors for each site. |
spec.data |
Descriptors for each species. |
y |
Result of function sites.long. |
grouping |
Variable defining the centroids |
FUN |
A function to compute the summary statistics which can be applied to all data subsets, as in |
centroids.only |
Return the coordinates for the centroids |
z |
Result of |
grouping.name |
Name for the categorical variable, expected as the factor used in the earlier ordiellipse call. |
cl |
Result of |
cl.data |
Result of |
w |
Ordination object from which the |
choices |
Ordination axes selected, as in |
cmdscale.model |
Use |
CAPdiscrim.model |
Use |
ci |
Multiplier for confidence intervals as in |
label.freq |
Frequency of labels along the x-axis (count between labels). |
Examples for ordiplot results are shown below.
Function pvclust.long combines data from pvclust with coordinates of nodes and branches from ordicluster. The variable of prune allows to remove higher levels of nodes and branches in the clustering hierarchy in a similar way as argument prune for ordicluster - see examples.
See also section: see examples for species accumulation curves and Renyi diversity profiles
These functions produce data.frames that can subsequentially be plotted via ggplot methods.
Roeland Kindt
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
https://rpubs.com/Roeland-KINDT
## Not run: # ggplot2 plotting method library(ggplot2) library(ggforce) library(concaveman) library(ggrepel) library(ggsci) library(dplyr) library(vegan) data(dune) data(dune.env) attach(dune) attach(dune.env) Ordination.model1 <- capscale(dune ~ Management, data=dune.env, distance='kulczynski', sqrt.dist=F, add='cailliez') plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling='species') # obtain 'long' data from the ordiplot object sites1 <- sites.long(plot1, env.data=dune.env) species1 <- species.long(plot1) centroids1 <- centroids.long(sites1, Management, FUN=median) centroids2 <- centroids.long(sites1, Management, FUN=median, centroids.only=TRUE) # information on percentage variation from the fitted ordination axislabs <- axis.long(Ordination.model1, choices=c(1 , 2)) # change the theme # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) # no species scores # centroids calculated directly via the centroids.long function plotgg1 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_mark_hull(data=sites1, aes(x=axis1, y=axis2, colour = Management), concavity = 0.1, alpha=0.8, size=0.2, show.legend=FALSE) + geom_point(data=sites1, aes(x=axis1, y=axis2, colour=Management, shape=Management), size=5) + # geom_segment(data=species1, aes(x=0, y=0, xend=axis1*2, yend=axis2*2), # size=1.2, arrow=arrow()) + # geom_label_repel(data=species1, aes(x=axis1*2, y=axis2*2, label=labels)) + geom_point(data=centroids.long(sites1, grouping=Management, centroids.only=TRUE), aes(x=axis1c, y=axis2c, colour=Centroid, shape=Centroid), size=10, show.legend=FALSE) + geom_segment(data=centroids.long(sites1, grouping=Management), aes(x=axis1c, y=axis2c, xend=axis1, yend=axis2, colour=Management), size=1, show.legend=FALSE) + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg1 # select species to plot based on goodness of fit spec.envfit <- envfit(plot1, env=dune) spec.data1 <- data.frame(r=spec.envfit$vectors$r, p=spec.envfit$vectors$pvals) species2 <- species.long(plot1, spec.data=spec.data1) species2 <- species2[species2$r > 0.6, ] # after_scale introduced in ggplot2 3.3.0 plotgg2 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_mark_ellipse(data=sites1, aes(x=axis1, y=axis2, colour=Management, fill=after_scale(alpha(colour, 0.2))), expand=0, size=0.2, show.legend=TRUE) + geom_point(data=sites1, aes(x=axis1, y=axis2, colour=Management, shape=Management), size=5) + geom_segment(data=centroids.long(sites1, grouping=Management), aes(x=axis1c, y=axis2c, xend=axis1, yend=axis2, colour=Management), size=1, show.legend=FALSE) + geom_segment(data=species2, aes(x=0, y=0, xend=axis1*2, yend=axis2*2), size=1.2, arrow=arrow()) + geom_label_repel(data=species2, aes(x=axis1*2, y=axis2*2, label=labels)) + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg2 # Add contour and vector for a continuous explanatory variable Ordination.model2 <- capscale(dune ~ Management, data=dune.env, distance='kulczynski', sqrt.dist=F, add='cailliez') plot2 <- ordiplot(Ordination.model2, choices=c(1,2), scaling='species') sites2 <- sites.long(plot2, env.data=dune.env) axislabs <- axis.long(Ordination.model2, choices=c(1 , 2)) dune.envfit <- envfit(plot2, dune.env) vectors2 <- vectorfit.long(dune.envfit) A1.surface <- ordisurf(plot2, y=A1) A1.surface A1.grid <- ordisurfgrid.long(A1.surface) plotgg3 <- ggplot() + geom_contour_filled(data=A1.grid, aes(x=x, y=y, z=z)) + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_point(data=sites2, aes(x=axis1, y=axis2, size=A1), shape=21, colour="black", fill="red") + geom_segment(data=subset(vectors2, vector=A1), aes(x=0, y=0, xend=axis1*2, yend=axis2*2), size=1.2, arrow=arrow()) + geom_label_repel(data=subset(vectors2, vector=A1), aes(x=axis1*2, y=axis2*2, label=vector), size=5) + BioR.theme + scale_fill_viridis_d() + scale_size(range=c(1, 20)) + labs(fill="A1") + coord_fixed(ratio=1) plotgg3 # after_stat introduced in ggplot2 3.3.0 plotgg4 <- ggplot() + geom_contour(data=A1.grid, aes(x=x, y=y, z=z, colour=factor(after_stat(level))), size=2) + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_point(data=sites2, aes(x=axis1, y=axis2, size=A1), shape=21, colour="black", fill="red") + geom_label_repel(data=sites2, aes(x=axis1, y=axis2, label=labels), colour='black', size=4) + BioR.theme + scale_colour_viridis_d() + scale_size(range=c(1, 20)) + labs(colour="A1") + coord_fixed(ratio=1) plotgg4 # example of Voronoi segmentation plotgg5 <- ggplot(data=sites2, aes(x=axis1, y=axis2)) + geom_voronoi_tile(aes(fill = Management, group=-1L), max.radius=0.2) + geom_voronoi_segment(colour="grey50") + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_point() + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg5 # adding ellipse via ordiellipse plot3 <- ordiplot(Ordination.model1, choices=c(1,2), scaling='species') axislabs <- axis.long(Ordination.model1, choices=c(1 , 2)) Management.ellipses <- ordiellipse(plot3, groups=Management, display="sites", kind="sd") Management.ellipses.long2 <- ordiellipse.long(Management.ellipses, grouping.name="Management") plotgg6 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_polygon(data=Management.ellipses.long2, aes(x=axis1, y=axis2, colour=Management, fill=after_scale(alpha(colour, 0.2))), size=0.2, show.legend=FALSE) + geom_point(data=sites1, aes(x=axis1, y=axis2, colour=Management, shape=Management), size=5) + geom_segment(data=centroids.long(sites1, grouping=Management), aes(x=axis1c, y=axis2c, xend=axis1, yend=axis2, colour=Management), size=1, show.legend=FALSE) + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg6 # adding cluster results via pvclust.long library(pvclust) # transformation as pvclust works with Euclidean distance dune.Hellinger <- disttransform(dune, method='hellinger') dune.pv <- pvclust(t(dune.Hellinger), method.hclust="mcquitty", method.dist="euclidean", nboot=1000) plot(dune.pv) pvrect(dune.pv, alpha=0.89, pv="au") # Model fitted earlier plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling='species') cl.data1 <- ordicluster(plot1, cluster=as.hclust(dune.pv$hclust)) sites1 <- sites.long(plot1, env.data=dune.env) axislabs <- axis.long(Ordination.model1, choices=c(1 , 2)) cl.data1 <- ordicluster(plot2, cluster=as.hclust(dune.pv$hclust)) pvlong <- pvclust.long(dune.pv, cl.data1) # as in example for ordicluster, prune higher level hierarchies plotgg7 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_segment(data=subset(pvlong$segments, pvlong$segments$prune > 3), aes(x=x1, y=y1, xend=x2, yend=y2, colour=au>=0.89, size=au), show.legend=TRUE) + geom_point(data=subset(pvlong$nodes, pvlong$nodes$prune > 3), aes(x=x, y=y, fill=au>=0.89), shape=21, size=2, colour="black") + geom_point(data=sites1, aes(x=axis1, y=axis2, shape=Management), colour="darkolivegreen4", alpha=0.9, size=5) + geom_text(data=sites1, aes(x=axis1, y=axis2, label=labels)) + BioR.theme + ggsci::scale_colour_npg() + scale_size(range=c(0.3, 2)) + scale_shape_manual(values=c(15, 16, 17, 18)) + guides(shape = guide_legend(override.aes = list(linetype = 0))) + coord_fixed(ratio=1) plotgg7 ## End(Not run) # dontrun## Not run: # ggplot2 plotting method library(ggplot2) library(ggforce) library(concaveman) library(ggrepel) library(ggsci) library(dplyr) library(vegan) data(dune) data(dune.env) attach(dune) attach(dune.env) Ordination.model1 <- capscale(dune ~ Management, data=dune.env, distance='kulczynski', sqrt.dist=F, add='cailliez') plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling='species') # obtain 'long' data from the ordiplot object sites1 <- sites.long(plot1, env.data=dune.env) species1 <- species.long(plot1) centroids1 <- centroids.long(sites1, Management, FUN=median) centroids2 <- centroids.long(sites1, Management, FUN=median, centroids.only=TRUE) # information on percentage variation from the fitted ordination axislabs <- axis.long(Ordination.model1, choices=c(1 , 2)) # change the theme # possibly need for extrafont::loadfonts(device="win") to have Arial # as alternative, use library(ggThemeAssist) BioR.theme <- theme( panel.background = element_blank(), panel.border = element_blank(), panel.grid = element_blank(), axis.line = element_line("gray25"), text = element_text(size = 12, family="Arial"), axis.text = element_text(size = 10, colour = "gray25"), axis.title = element_text(size = 14, colour = "gray25"), legend.title = element_text(size = 14), legend.text = element_text(size = 14), legend.key = element_blank()) # no species scores # centroids calculated directly via the centroids.long function plotgg1 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_mark_hull(data=sites1, aes(x=axis1, y=axis2, colour = Management), concavity = 0.1, alpha=0.8, size=0.2, show.legend=FALSE) + geom_point(data=sites1, aes(x=axis1, y=axis2, colour=Management, shape=Management), size=5) + # geom_segment(data=species1, aes(x=0, y=0, xend=axis1*2, yend=axis2*2), # size=1.2, arrow=arrow()) + # geom_label_repel(data=species1, aes(x=axis1*2, y=axis2*2, label=labels)) + geom_point(data=centroids.long(sites1, grouping=Management, centroids.only=TRUE), aes(x=axis1c, y=axis2c, colour=Centroid, shape=Centroid), size=10, show.legend=FALSE) + geom_segment(data=centroids.long(sites1, grouping=Management), aes(x=axis1c, y=axis2c, xend=axis1, yend=axis2, colour=Management), size=1, show.legend=FALSE) + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg1 # select species to plot based on goodness of fit spec.envfit <- envfit(plot1, env=dune) spec.data1 <- data.frame(r=spec.envfit$vectors$r, p=spec.envfit$vectors$pvals) species2 <- species.long(plot1, spec.data=spec.data1) species2 <- species2[species2$r > 0.6, ] # after_scale introduced in ggplot2 3.3.0 plotgg2 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_mark_ellipse(data=sites1, aes(x=axis1, y=axis2, colour=Management, fill=after_scale(alpha(colour, 0.2))), expand=0, size=0.2, show.legend=TRUE) + geom_point(data=sites1, aes(x=axis1, y=axis2, colour=Management, shape=Management), size=5) + geom_segment(data=centroids.long(sites1, grouping=Management), aes(x=axis1c, y=axis2c, xend=axis1, yend=axis2, colour=Management), size=1, show.legend=FALSE) + geom_segment(data=species2, aes(x=0, y=0, xend=axis1*2, yend=axis2*2), size=1.2, arrow=arrow()) + geom_label_repel(data=species2, aes(x=axis1*2, y=axis2*2, label=labels)) + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg2 # Add contour and vector for a continuous explanatory variable Ordination.model2 <- capscale(dune ~ Management, data=dune.env, distance='kulczynski', sqrt.dist=F, add='cailliez') plot2 <- ordiplot(Ordination.model2, choices=c(1,2), scaling='species') sites2 <- sites.long(plot2, env.data=dune.env) axislabs <- axis.long(Ordination.model2, choices=c(1 , 2)) dune.envfit <- envfit(plot2, dune.env) vectors2 <- vectorfit.long(dune.envfit) A1.surface <- ordisurf(plot2, y=A1) A1.surface A1.grid <- ordisurfgrid.long(A1.surface) plotgg3 <- ggplot() + geom_contour_filled(data=A1.grid, aes(x=x, y=y, z=z)) + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_point(data=sites2, aes(x=axis1, y=axis2, size=A1), shape=21, colour="black", fill="red") + geom_segment(data=subset(vectors2, vector=A1), aes(x=0, y=0, xend=axis1*2, yend=axis2*2), size=1.2, arrow=arrow()) + geom_label_repel(data=subset(vectors2, vector=A1), aes(x=axis1*2, y=axis2*2, label=vector), size=5) + BioR.theme + scale_fill_viridis_d() + scale_size(range=c(1, 20)) + labs(fill="A1") + coord_fixed(ratio=1) plotgg3 # after_stat introduced in ggplot2 3.3.0 plotgg4 <- ggplot() + geom_contour(data=A1.grid, aes(x=x, y=y, z=z, colour=factor(after_stat(level))), size=2) + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_point(data=sites2, aes(x=axis1, y=axis2, size=A1), shape=21, colour="black", fill="red") + geom_label_repel(data=sites2, aes(x=axis1, y=axis2, label=labels), colour='black', size=4) + BioR.theme + scale_colour_viridis_d() + scale_size(range=c(1, 20)) + labs(colour="A1") + coord_fixed(ratio=1) plotgg4 # example of Voronoi segmentation plotgg5 <- ggplot(data=sites2, aes(x=axis1, y=axis2)) + geom_voronoi_tile(aes(fill = Management, group=-1L), max.radius=0.2) + geom_voronoi_segment(colour="grey50") + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_point() + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg5 # adding ellipse via ordiellipse plot3 <- ordiplot(Ordination.model1, choices=c(1,2), scaling='species') axislabs <- axis.long(Ordination.model1, choices=c(1 , 2)) Management.ellipses <- ordiellipse(plot3, groups=Management, display="sites", kind="sd") Management.ellipses.long2 <- ordiellipse.long(Management.ellipses, grouping.name="Management") plotgg6 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_polygon(data=Management.ellipses.long2, aes(x=axis1, y=axis2, colour=Management, fill=after_scale(alpha(colour, 0.2))), size=0.2, show.legend=FALSE) + geom_point(data=sites1, aes(x=axis1, y=axis2, colour=Management, shape=Management), size=5) + geom_segment(data=centroids.long(sites1, grouping=Management), aes(x=axis1c, y=axis2c, xend=axis1, yend=axis2, colour=Management), size=1, show.legend=FALSE) + BioR.theme + ggsci::scale_colour_npg() + coord_fixed(ratio=1) plotgg6 # adding cluster results via pvclust.long library(pvclust) # transformation as pvclust works with Euclidean distance dune.Hellinger <- disttransform(dune, method='hellinger') dune.pv <- pvclust(t(dune.Hellinger), method.hclust="mcquitty", method.dist="euclidean", nboot=1000) plot(dune.pv) pvrect(dune.pv, alpha=0.89, pv="au") # Model fitted earlier plot1 <- ordiplot(Ordination.model1, choices=c(1,2), scaling='species') cl.data1 <- ordicluster(plot1, cluster=as.hclust(dune.pv$hclust)) sites1 <- sites.long(plot1, env.data=dune.env) axislabs <- axis.long(Ordination.model1, choices=c(1 , 2)) cl.data1 <- ordicluster(plot2, cluster=as.hclust(dune.pv$hclust)) pvlong <- pvclust.long(dune.pv, cl.data1) # as in example for ordicluster, prune higher level hierarchies plotgg7 <- ggplot() + geom_vline(xintercept = c(0), color = "grey70", linetype = 2) + geom_hline(yintercept = c(0), color = "grey70", linetype = 2) + xlab(axislabs[1, "label"]) + ylab(axislabs[2, "label"]) + scale_x_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + scale_y_continuous(sec.axis = dup_axis(labels=NULL, name=NULL)) + geom_segment(data=subset(pvlong$segments, pvlong$segments$prune > 3), aes(x=x1, y=y1, xend=x2, yend=y2, colour=au>=0.89, size=au), show.legend=TRUE) + geom_point(data=subset(pvlong$nodes, pvlong$nodes$prune > 3), aes(x=x, y=y, fill=au>=0.89), shape=21, size=2, colour="black") + geom_point(data=sites1, aes(x=axis1, y=axis2, shape=Management), colour="darkolivegreen4", alpha=0.9, size=5) + geom_text(data=sites1, aes(x=axis1, y=axis2, label=labels)) + BioR.theme + ggsci::scale_colour_npg() + scale_size(range=c(0.3, 2)) + scale_shape_manual(values=c(15, 16, 17, 18)) + guides(shape = guide_legend(override.aes = list(linetype = 0))) + coord_fixed(ratio=1) plotgg7 ## End(Not run) # dontrun
Spatial sampling within a polygon provides several methods of selecting rectangular sample plots within a polygon. Using a GIS package may be preferred for actual survey design.
spatialsample(x,method="random",n=5,xwidth=0.5,ywidth=0.5,xleft=0, ylower=0,xdist=0,ydist=0,plotit=T,plothull=F)spatialsample(x,method="random",n=5,xwidth=0.5,ywidth=0.5,xleft=0, ylower=0,xdist=0,ydist=0,plotit=T,plothull=F)
x |
2-column matrix with the coordinates of the vertices of the polygon. The first column contains the horizontal (x) position, the second column contains the vertical (y) position. |
method |
Method of sampling, any of "random", "grid" or "random grid". |
n |
Number of sample plots to be selected, or number of horizontal and vertical grid positions. |
xwidth |
Horizontal width of the sample plots. |
ywidth |
Vertical width of the sample plots. |
xleft |
Horizontal starting position of the grid. |
ylower |
Vertical starting position of the grid. |
xdist |
Horizontal distance between grid locations. |
ydist |
Vertical distance between grid locations. |
plotit |
Plot the sample plots on the current graph. |
plothull |
Plot a convex hull around the sample plots. |
Spatial sampling within a polygon provides several methods of selecting the position of sample plots.
Method "random" selects random positions of the sample plots using simple random sampling.
Method "grid" selects sample plots from a grid defined by "xleft", "ylower", "xdist" and "ydist". In case xdist=0 or ydist=0, then the number of grid positions are defined by "n". In case "xleft" or "ylower" are below the minimum position of any vertix of the polygon, then a random starting position is selected for the grid.
Method "random grid" selects sample plots at random from the sampling grid using the same methods of defining the grid as for method "grid".
The function returns a list of centres of rectangular sample plots.
Roeland Kindt (World Agroforestry Centre)
Kindt, R. & Coe, R. (2005) Tree diversity analysis: A manual and software for common statistical methods for ecological and biodiversity studies.
https://www.worldagroforestry.org/output/tree-diversity-analysis
library(splancs) area <- array(c(10,10,15,35,40,35,5,35,35,30,30,10), dim=c(6,2)) landuse1 <- array(c(10,10,15,15,30,35,35,30), dim=c(4,2)) landuse2 <- array(c(10,10,15,15,35,30,10,30,30,35,30,15), dim=c(6,2)) landuse3 <- array(c(10,10,30,35,40,35,5,10,15,30,30,10), dim=c(6,2)) plot(area[,1], area[,2], type="n", xlab="horizontal position", ylab="vertical position", lwd=2, bty="l") polygon(landuse1) polygon(landuse2) polygon(landuse3) spatialsample(area, method="random", n=20, xwidth=1, ywidth=1, plotit=TRUE, plothull=FALSE) spatialsample(area, method="grid", xwidth=1, ywidth=1, plotit=TRUE, xleft=12, ylower=7, xdist=4, ydist=4) spatialsample(area, method="random grid", n=20, xwidth=1, ywidth=1, plotit=TRUE, xleft=12, ylower=7, xdist=4, ydist=4)library(splancs) area <- array(c(10,10,15,35,40,35,5,35,35,30,30,10), dim=c(6,2)) landuse1 <- array(c(10,10,15,15,30,35,35,30), dim=c(4,2)) landuse2 <- array(c(10,10,15,15,35,30,10,30,30,35,30,15), dim=c(6,2)) landuse3 <- array(c(10,10,30,35,40,35,5,10,15,30,30,10), dim=c(6,2)) plot(area[,1], area[,2], type="n", xlab="horizontal position", ylab="vertical position", lwd=2, bty="l") polygon(landuse1) polygon(landuse2) polygon(landuse3) spatialsample(area, method="random", n=20, xwidth=1, ywidth=1, plotit=TRUE, plothull=FALSE) spatialsample(area, method="grid", xwidth=1, ywidth=1, plotit=TRUE, xleft=12, ylower=7, xdist=4, ydist=4) spatialsample(area, method="random grid", n=20, xwidth=1, ywidth=1, plotit=TRUE, xleft=12, ylower=7, xdist=4, ydist=4)
This dataset documents the site sequence of 19 sites on a gradient determined from unimodal species distributions. The dataset is accompanied by transfspecies that documents the species composition of the sites. This is a hypothetical example that allows to investigate how well ecological distance measures or ordination methods recover the expected best sequence of sites.
data(transfgradient)data(transfgradient)
A data frame with 19 observations on the following variable.
gradienta numeric vector
Legendre, P. & Gallagher, E.D. (2001) Ecologically meaningful transformations for ordination of species data. Oecologia 129: 271-280.
Figure 3a.
data(transfspecies) data(transfgradient) plot(transfspecies[,1]~transfgradient[,1],xlab="gradient", ylab="species abundance",type="n",ylim=c(0.5,8.5)) for (i in 1:9) {points(transfgradient[,1],transfspecies[,i],type="o",pch=i)}data(transfspecies) data(transfgradient) plot(transfspecies[,1]~transfgradient[,1],xlab="gradient", ylab="species abundance",type="n",ylim=c(0.5,8.5)) for (i in 1:9) {points(transfgradient[,1],transfspecies[,i],type="o",pch=i)}
This dataset documents the species composition of 19 sites that follow a specific sequence of sites as determined from unimodal species distributions. The dataset is accompanied by transfgradient that documents the gradient in species turnover. This is a hypothetical example that allows to investigate how well ecological distance measures or ordination methods recover the expected best sequence of sites.
data(transfspecies)data(transfspecies)
A data frame with 19 observations on the following 9 variables.
species1a numeric vector
species2a numeric vector
species3a numeric vector
species4a numeric vector
species5a numeric vector
species6a numeric vector
species7a numeric vector
species8a numeric vector
species9a numeric vector
The example in the Tree Diversity Analysis manual only looks at the ecological distance from the first site. Hence, only the first 10 sites that share some species with this site should be selected.
This dataset enables investigations of how well ecological distance measures and ordination diagrams reconstruct the gradient (sequence of sites). The gradient expresses how the sites would be arranged based on their species composition.
Legendre, P. & Gallagher, E.D. (2001) Ecologically meaningful transformations for ordination of species data. Oecologia 129: 271-280.
Figure 3a.
data(transfspecies) data(transfgradient) plot(transfspecies[,1]~transfgradient[,1],xlab="gradient", ylab="species abundance",type="n",ylim=c(0.5,8.5)) for (i in 1:9) {points(transfgradient[,1],transfspecies[,i],type="o",pch=i)}data(transfspecies) data(transfgradient) plot(transfspecies[,1]~transfgradient[,1],xlab="gradient", ylab="species abundance",type="n",ylim=c(0.5,8.5)) for (i in 1:9) {points(transfgradient[,1],transfspecies[,i],type="o",pch=i)}
Function treegoer.score calculates a climate score via a similar algorithm that is used internally in the GlobalUsefulNativeTrees (GlobUNT) database (Kindt et al. 2023, doi:10.1038/s41598-023-39552-1). The function depends on treegoer.filter and requires a data set (argument treegoer.wide) as created from the Tree Globally Observed Environmental Ranges (TreeGOER) database (Kindt 2023, doi:10.1111/gcb.16914) via treegoer.widen.
treegoer.score(site.data, site.species=treegoer.wide$species, treegoer.wide, filter.vars=c("bio05", "bio14", "climaticMoistureIndex")) treegoer.filter(site.data, treegoer.wide, filter.vars=c("bio05", "bio14", "climaticMoistureIndex"), limit.vars=c("Q05", "Q95")) treegoer.widen(treegoer, species=unique(treegoer$species)[1:100], filter.vars=c("bio05", "bio14", "climaticMoistureIndex"))treegoer.score(site.data, site.species=treegoer.wide$species, treegoer.wide, filter.vars=c("bio05", "bio14", "climaticMoistureIndex")) treegoer.filter(site.data, treegoer.wide, filter.vars=c("bio05", "bio14", "climaticMoistureIndex"), limit.vars=c("Q05", "Q95")) treegoer.widen(treegoer, species=unique(treegoer$species)[1:100], filter.vars=c("bio05", "bio14", "climaticMoistureIndex"))
site.data |
Data set with 1 row, containing the environmental conditions at the planting site for the selected environmental variables of the TreeGOER database. This data set can be set by selecting a city from the CitiesGOER datase (https://zenodo.org/records/10004594) or a weather station from the ClimateForecasts database (https://zenodo.org/records/10726088). |
site.species |
Species for which the climate score will be calculated. |
treegoer.wide |
Data set created by |
filter.vars |
Environmental variables for which ranges (minimum, maximum and 0.05, 0.25, 0.75 and 0.95 quantile) are documented in the treegoer.wide data set. |
limit.vars |
Selection of the lower and upper limits for the environmental ranges, typically set as |
treegoer |
Data set with environmental limits that was locally downloaded file ( |
species |
Selection of species to document in the wide format. |
The calculation of the climate score uses an expanded version of the algorithms used by BIOCLIM (Booth 2018, doi:10.1111/aec.12628).
- A score of 3 indicates that for all selected variables, the planting site has environmental conditions that are within the middle (0.25 to 0.75 quantiles) of the species range.
- A score of 2 indicates that for all selected variables, the planting site has environmental conditions that are within the core (0.05 to 0.95 quantiles) of the species range. For some variables, the planting conditions are outside the middle of the species range.
- A score of 1 indicates that for all selected variables, the planting site has environmental conditions that are within the documented limits (minimum to maximum) of the species range. For some variables, the planting conditions are outside the core of the species range; the BIOCLIM algorithm defines this domain as the 'marginal domain'.
- A score of 0 indicates that for some of the selected variables, the planting site has environmental conditions that are outside the documented limits (< minimum or > maximum) of the species range.
- A score of -1 indicates that there was no information on the environmental ranges of the species.
The same algorithm and similar scripts are used internally in the GlobalUsefulNativeTrees database (see Kindt et al. 2023). The internal scripts also resemble scripts provided here: https://rpubs.com/Roeland-KINDT/1114902.
Function treegoer.score returns a data set that includes a climate score representing the position of the planting site within the environmental range of species documented by the Tree Globally Observed Environmental Ranges database.
Roeland Kindt (World Agroforestry, CIFOR-ICRAF)
Booth TH. 2018. Why understanding the pioneering and continuing contributions of BIOCLIM to species distribution modelling is important. Austral Ecology 43: 852-860. doi:10.1111/aec.12628
Kindt R. 2023. TreeGOER: A database with globally observed environmental ranges for 48,129 tree species. Global Change Biology. doi:10.1111/gcb.16914
Kindt R., Graudal L, Lilleso J.P.-B. et al. 2023. GlobalUsefulNativeTrees, a database documenting 14,014 tree species, supports synergies between biodiversity recovery and local livelihoods in landscape restoration. Scientific Reports. doi:10.1038/s41598-023-39552-1
Kindt R. 2023. Using the Tree Globally Observed Environmental Ranges and CitiesGOER databases to Filter GlobalUsefulNativeTrees Species lists. https://rpubs.com/Roeland-KINDT/1114902
Kindt R. 2023. CitiesGOER: Globally Observed Environmental Data for 52,602 Cities with a Population >= 5000 (version 2023.10). doi:10.5281/zenodo.10004594
Kindt R. 2024. ClimateForecasts: Globally Observed Environmental Data for 15,504 Weather Station Locations (version 2024.03). doi:10.5281/zenodo.10776414
## Not run: # Example adapted from https://rpubs.com/Roeland-KINDT/1114902 # treegoer.file <- choose.files() # Provide the location where the TreeGOER file was downloaded locally # (https://zenodo.org/records/10008994: TreeGOER_2023.txt) treegoer <- fread(treegoer.file, sep="|", encoding="UTF-8") nrow(treegoer) length(unique(treegoer$species)) # 48129 # A data set of tree species # Example has useful tree species filtered # for Kenya and human food from the GlobalUsefulNativeTrees database # (https://worldagroforestry.org/output/globalusefulnativetrees) # globunt.file <- choose.files() globunt <- fread(globunt.file, sep="|", encoding="UTF-8") nrow(globunt) # 461 # Environmental variables used for filtering or scoring species focal.vars <- c("bio01", "bio12", "climaticMoistureIndex", "monthCountByTemp10", "growingDegDays5", "bio05", "bio06", "bio16", "bio17", "MCWD") # Use treegoer.widen() treegoer.wide <- treegoer.widen(treegoer=treegoer, species=globunt$Switchboard, filter.vars=focal.vars) names(treegoer.wide) # Environmental conditions at the planting site # Provide the locations where the CitiesGOER files were downloaded locally # (https://zenodo.org/records/10004594: CitiesGOER_baseline.xlsx). # Alternatively, the ClimateForecasts database can be used # (https://zenodo.org/records/10726088: ClimateForecasts_baseline.xlsx) # baseline.file <- choose.files() site.baseline <- data.frame(read_excel(baseline.file, sheet="Cities data", skip=6)) # Set the planting location in Nairobi site.planting <- site.baseline[site.baseline$Name == "Nairobi", ] site.planting # Calculate the climate scores treegoer.scores <- treegoer.score(site.species=globunt$Switchboard, treegoer.wide=treegoer.wide, filter.vars=focal.vars, site.data=site.planting) # Calculate the climate score for a single environmental variable treegoer.score <- treegoer.score(site.species=globunt$Switchboard, treegoer.wide=treegoer.wide, filter.vars="bio01", site.data=site.planting) ## End(Not run)## Not run: # Example adapted from https://rpubs.com/Roeland-KINDT/1114902 # treegoer.file <- choose.files() # Provide the location where the TreeGOER file was downloaded locally # (https://zenodo.org/records/10008994: TreeGOER_2023.txt) treegoer <- fread(treegoer.file, sep="|", encoding="UTF-8") nrow(treegoer) length(unique(treegoer$species)) # 48129 # A data set of tree species # Example has useful tree species filtered # for Kenya and human food from the GlobalUsefulNativeTrees database # (https://worldagroforestry.org/output/globalusefulnativetrees) # globunt.file <- choose.files() globunt <- fread(globunt.file, sep="|", encoding="UTF-8") nrow(globunt) # 461 # Environmental variables used for filtering or scoring species focal.vars <- c("bio01", "bio12", "climaticMoistureIndex", "monthCountByTemp10", "growingDegDays5", "bio05", "bio06", "bio16", "bio17", "MCWD") # Use treegoer.widen() treegoer.wide <- treegoer.widen(treegoer=treegoer, species=globunt$Switchboard, filter.vars=focal.vars) names(treegoer.wide) # Environmental conditions at the planting site # Provide the locations where the CitiesGOER files were downloaded locally # (https://zenodo.org/records/10004594: CitiesGOER_baseline.xlsx). # Alternatively, the ClimateForecasts database can be used # (https://zenodo.org/records/10726088: ClimateForecasts_baseline.xlsx) # baseline.file <- choose.files() site.baseline <- data.frame(read_excel(baseline.file, sheet="Cities data", skip=6)) # Set the planting location in Nairobi site.planting <- site.baseline[site.baseline$Name == "Nairobi", ] site.planting # Calculate the climate scores treegoer.scores <- treegoer.score(site.species=globunt$Switchboard, treegoer.wide=treegoer.wide, filter.vars=focal.vars, site.data=site.planting) # Calculate the climate score for a single environmental variable treegoer.score <- treegoer.score(site.species=globunt$Switchboard, treegoer.wide=treegoer.wide, filter.vars="bio01", site.data=site.planting) ## End(Not run)
This data set contains scores for 185 loci for 100 individuals of the Warburgia ugandensis tree species (a medicinal tree species native to Eastern Africa). Since the data set is a subset of a larger data set that originated from a study of several Warburgia species, some of the loci did not produce bands for W. ugandensis (i.e. some loci only contain zeroes). This data set is accompanied by warenv that describes population and regional structure of the 100 individuals.
data(warcom)data(warcom)
A data frame with 100 observations on the following 185 variables.
locus001a numeric vector
locus002a numeric vector
locus003a numeric vector
locus004a numeric vector
locus005a numeric vector
locus006a numeric vector
locus007a numeric vector
locus008a numeric vector
locus009a numeric vector
locus010a numeric vector
locus011a numeric vector
locus012a numeric vector
locus013a numeric vector
locus014a numeric vector
locus015a numeric vector
locus016a numeric vector
locus017a numeric vector
locus018a numeric vector
locus019a numeric vector
locus020a numeric vector
locus021a numeric vector
locus022a numeric vector
locus023a numeric vector
locus024a numeric vector
locus025a numeric vector
locus026a numeric vector
locus027a numeric vector
locus028a numeric vector
locus029a numeric vector
locus030a numeric vector
locus031a numeric vector
locus032a numeric vector
locus033a numeric vector
locus034a numeric vector
locus035a numeric vector
locus036a numeric vector
locus037a numeric vector
locus038a numeric vector
locus039a numeric vector
locus040a numeric vector
locus041a numeric vector
locus042a numeric vector
locus043a numeric vector
locus044a numeric vector
locus045a numeric vector
locus046a numeric vector
locus047a numeric vector
locus048a numeric vector
locus049a numeric vector
locus050a numeric vector
locus051a numeric vector
locus052a numeric vector
locus053a numeric vector
locus054a numeric vector
locus055a numeric vector
locus056a numeric vector
locus057a numeric vector
locus058a numeric vector
locus059a numeric vector
locus060a numeric vector
locus061a numeric vector
locus062a numeric vector
locus063a numeric vector
locus064a numeric vector
locus065a numeric vector
locus066a numeric vector
locus067a numeric vector
locus068a numeric vector
locus069a numeric vector
locus070a numeric vector
locus071a numeric vector
locus072a numeric vector
locus073a numeric vector
locus074a numeric vector
locus075a numeric vector
locus076a numeric vector
locus077a numeric vector
locus078a numeric vector
locus079a numeric vector
locus080a numeric vector
locus081a numeric vector
locus082a numeric vector
locus083a numeric vector
locus084a numeric vector
locus085a numeric vector
locus086a numeric vector
locus087a numeric vector
locus088a numeric vector
locus089a numeric vector
locus090a numeric vector
locus091a numeric vector
locus092a numeric vector
locus093a numeric vector
locus094a numeric vector
locus095a numeric vector
locus096a numeric vector
locus097a numeric vector
locus098a numeric vector
locus099a numeric vector
locus100a numeric vector
locus101a numeric vector
locus102a numeric vector
locus103a numeric vector
locus104a numeric vector
locus105a numeric vector
locus106a numeric vector
locus107a numeric vector
locus108a numeric vector
locus109a numeric vector
locus110a numeric vector
locus111a numeric vector
locus112a numeric vector
locus113a numeric vector
locus114a numeric vector
locus115a numeric vector
locus116a numeric vector
locus117a numeric vector
locus118a numeric vector
locus119a numeric vector
locus120a numeric vector
locus121a numeric vector
locus122a numeric vector
locus123a numeric vector
locus124a numeric vector
locus125a numeric vector
locus126a numeric vector
locus127a numeric vector
locus128a numeric vector
locus129a numeric vector
locus130a numeric vector
locus131a numeric vector
locus132a numeric vector
locus133a numeric vector
locus134a numeric vector
locus135a numeric vector
locus136a numeric vector
locus137a numeric vector
locus138a numeric vector
locus139a numeric vector
locus140a numeric vector
locus141a numeric vector
locus142a numeric vector
locus143a numeric vector
locus144a numeric vector
locus145a numeric vector
locus146a numeric vector
locus147a numeric vector
locus148a numeric vector
locus149a numeric vector
locus150a numeric vector
locus151a numeric vector
locus152a numeric vector
locus153a numeric vector
locus154a numeric vector
locus155a numeric vector
locus156a numeric vector
locus157a numeric vector
locus158a numeric vector
locus159a numeric vector
locus160a numeric vector
locus161a numeric vector
locus162a numeric vector
locus163a numeric vector
locus164a numeric vector
locus165a numeric vector
locus166a numeric vector
locus167a numeric vector
locus168a numeric vector
locus169a numeric vector
locus170a numeric vector
locus171a numeric vector
locus172a numeric vector
locus173a numeric vector
locus174a numeric vector
locus175a numeric vector
locus176a numeric vector
locus177a numeric vector
locus178a numeric vector
locus179a numeric vector
locus180a numeric vector
locus181a numeric vector
locus182a numeric vector
locus183a numeric vector
locus184a numeric vector
locus185a numeric vector
Muchugi, A.N. (2007) Population genetics and taxonomy of important medicinal tree species of the genus Warburgia. PhD Thesis. Kenyatta University, Kenya.
data(warcom)data(warcom)
This data set contains population and regional locations for 100 individuals of the Warburgia ugandensis tree species (a medicinal tree species native to Eastern Africa). This data set is associated with warcom that contains scores for 185 AFLP loci.
data(warenv)data(warenv)
A data frame with 100 observations on the following 4 variables.
populationa factor with levels Kibale Kitale Laikipia Lushoto Mara
popshorta factor with levels KKIT KLAI KMAR TLUS UKIB
countrya factor with levels Kenya Tanzania Uganda
rift.valleya factor with levels east west
Muchugi, A.N. (2007) Population genetics and taxonomy of important medicinal tree species of the genus Warburgia. PhD Thesis. Kenyatta University, Kenya.
data(warenv)data(warenv)