| Title: | Sediment Source Fingerprinting |
|---|---|
| Description: | Quantifies the provenance of the sediments in a catchment or study area. Based on a comprehensive characterization of the sediment sources and the end sediment mixtures a mixing model algorithm is applied to the sediment mixtures in order to estimate the relative contribution of each potential source. The package includes several statistical methods such as Kruskal-Wallis test, discriminant function analysis ('DFA'), principal component plot ('PCA') to select the optimal subset of tracer properties. The variability within each sediment source is also considered to estimate the statistical distribution of the sources contribution. |
| Authors: | Ivan Lizaga [aut, cre], Borja Latorre [aut], Leticia Gaspar [aut], Ana Navas [aut], Vince Q Vu [ctb] |
| Maintainer: | Ivan Lizaga <[email protected]> |
| License: | GPL (>= 2) |
| Version: | 1.1 |
| Built: | 2024-11-21 06:54:03 UTC |
| Source: | CRAN |
Soil erosion is one of the biggest challenges for food production and reservoirs siltation around the world. Information on sediment, nutrients and pollutant transport is required for effective control strategies. Source estimates are difficult to obtain using traditional monitoring techniques, but sediment source fingerprinting, has been proved to be a valuable tool. Sediment source fingerprinting offers the potential to assess sediment provenance as a basis to develop management plans and prevent erosion. The procedure focuses on developing methods that enable the apportionment of sediment sources to be identified from a composite sample of sediment mixture material. We developed an R-package as a tool to quantify the provenance of the sediments in a catchment. A mixing model algorithm is applied to the sediment mixture samples in order to estimate the relative contribution of each potential source. The package consists of a set of functions used to: i) characterise and pre-process the data, select the optimum subset of tracers; ii) unmix sediment samples and quantify the apportionment of each source; iii) assess the effect of the source variability; and iv) visualize and export the results.
Ivan Lizaga, Borja Latorre, Leticia Gaspar, Ana Navas
Maintainer: Ivan Lizaga <[email protected] // [email protected]>
https://github.com/eead-csic-eesa
#Created on 22/08/2018 #If you want to use your own data #setwd("the directory that contains your dataset") #data <- read.table('your dataset', header = T, sep = '\t') #install.packages("fingerPro") #library(fingerPro) #Example of the data included in the fingerPro package #Load the dataset called "catchment" # "Catchment": this dataset has been selected from a Mediterranean catchment for #this purpose and contains high-quality radionuclides and geochemistry data. #AG (cropland) #PI and PI1 (Pine forest, at first looks different but when you display de LDA plot #you will see that the wisher decision in join both pines as the same source) #SS (subsoil) data <- catchment #boxPlot(data, columns = 1:6, ncol = 3) #correlationPlot(data, columns = 1:5, mixtures = TRUE) LDAPlot(data, P3D=FALSE) #variables are collinear #select the optimum set of tracers by implementing the statistical tests data <- rangeTest(data) data <- KWTest(data) data <- DFATest(data) #Check how the selected tracers discriminate between sources LDAPlot(data, P3D=FALSE) #change P3D=FALSE to P3D=TRUE to visualize the 3D LDAPlot #2D and 3D LDAPlots suggest that two of the sources have to be combined #reload the original dataset "catchment" data <- catchment # Combine sources PI1 and PI based on the previous LDAPlot data$Land_Use[data$Land_Use == 'PI1'] <- 'PI' #select the optimum set of tracers by implementing the statistical tests data <- rangeTest(data) data <- KWTest(data) data <- DFATest(data) LDAPlot(data, P3D=FALSE) PCAPlot(data) #Now the optimum tracer properties selected discriminate well, so proceed with the unmix function result <- unmix(data, samples = 100L, iter =100L) #Display the results plotResults(result, y_high = 5, n = 1) writeResults(result)#Created on 22/08/2018 #If you want to use your own data #setwd("the directory that contains your dataset") #data <- read.table('your dataset', header = T, sep = '\t') #install.packages("fingerPro") #library(fingerPro) #Example of the data included in the fingerPro package #Load the dataset called "catchment" # "Catchment": this dataset has been selected from a Mediterranean catchment for #this purpose and contains high-quality radionuclides and geochemistry data. #AG (cropland) #PI and PI1 (Pine forest, at first looks different but when you display de LDA plot #you will see that the wisher decision in join both pines as the same source) #SS (subsoil) data <- catchment #boxPlot(data, columns = 1:6, ncol = 3) #correlationPlot(data, columns = 1:5, mixtures = TRUE) LDAPlot(data, P3D=FALSE) #variables are collinear #select the optimum set of tracers by implementing the statistical tests data <- rangeTest(data) data <- KWTest(data) data <- DFATest(data) #Check how the selected tracers discriminate between sources LDAPlot(data, P3D=FALSE) #change P3D=FALSE to P3D=TRUE to visualize the 3D LDAPlot #2D and 3D LDAPlots suggest that two of the sources have to be combined #reload the original dataset "catchment" data <- catchment # Combine sources PI1 and PI based on the previous LDAPlot data$Land_Use[data$Land_Use == 'PI1'] <- 'PI' #select the optimum set of tracers by implementing the statistical tests data <- rangeTest(data) data <- KWTest(data) data <- DFATest(data) LDAPlot(data, P3D=FALSE) PCAPlot(data) #Now the optimum tracer properties selected discriminate well, so proceed with the unmix function result <- unmix(data, samples = 100L, iter =100L) #Display the results plotResults(result, y_high = 5, n = 1) writeResults(result)
The boxplot compactly shows the distribution of a continuous variable. It displays five summary statistics (the median, two hinges and two whiskers), and all "outlying" points individually.
boxPlot(data, columns = 1:ncol(data) - 2, ncol = 3)boxPlot(data, columns = 1:ncol(data) - 2, ncol = 3)
data |
Data frame containing source and mixtures data |
columns |
Numeric vector containing the index of the columns in the chart (the first column refers to the first variable) |
ncol |
Number of charts per row |
A dataset containing the different tracer properties of the different land uses in a Mediterranean catchment and one mixture sample located at the output of the catchment. The variables are as follows:
catchmentcatchment
A data frame with 22 rows and 23 variables:
reference number id of each sample analysed
grouping variable, in this study refers to the different land uses in the catchment
value of the tracer property for each sample
The function displays a correlation matrix of each of the properties divided by the different sources to help the user in the decision.
correlationPlot(data, columns = c(1:ncol(data) - 1), mixtures = F)correlationPlot(data, columns = c(1:ncol(data) - 1), mixtures = F)
data |
Data frame containing source and mixtures data |
columns |
Numeric vector containing the index of the columns in the chart (the first column refers to the grouping variable) |
mixtures |
Boolean to include or exclude the mixture samples in the chart |
Performs a stepwise forward variable selection using the Wilk's Lambda criterion.
DFATest(data, niveau = 0.1)DFATest(data, niveau = 0.1)
data |
Data frame containing source and mixtures |
niveau |
level for the approximate F-test decision |
Data frame only containing the variables that pass the DFA test
Biplot for Principal Components using ggplot2
ggbiplot(pcobj, choices = 1:2, scale = 1, pc.biplot = TRUE, obs.scale = 1 - scale, var.scale = scale, groups = NULL, ellipse = FALSE, ellipse.prob = 0.68, labels = NULL, labels.size = 3, alpha = 1, var.axes = TRUE, circle = FALSE, circle.prob = 0.69, varname.size = 3, varname.adjust = 1.5, varname.abbrev = FALSE)ggbiplot(pcobj, choices = 1:2, scale = 1, pc.biplot = TRUE, obs.scale = 1 - scale, var.scale = scale, groups = NULL, ellipse = FALSE, ellipse.prob = 0.68, labels = NULL, labels.size = 3, alpha = 1, var.axes = TRUE, circle = FALSE, circle.prob = 0.69, varname.size = 3, varname.adjust = 1.5, varname.abbrev = FALSE)
pcobj |
an object returned by prcomp() or princomp() |
choices |
which PCs to plot |
scale |
covariance biplot (scale = 1), form biplot (scale = 0). When scale = 1, the inner product between the variables approximates the covariance and the distance between the points approximates the Mahalanobis distance. |
pc.biplot |
for compatibility with biplot.princomp() |
obs.scale |
scale factor to apply to observations |
var.scale |
scale factor to apply to variables |
groups |
optional factor variable indicating the groups that the observations belong to. If provided the points will be colored according to groups |
ellipse |
draw a normal data ellipse for each group? |
ellipse.prob |
size of the ellipse in Normal probability |
labels |
optional vector of labels for the observations |
labels.size |
size of the text used for the labels |
alpha |
alpha transparency value for the points (0 = transparent, 1 = opaque) |
var.axes |
draw arrows for the variables? |
circle |
draw a correlation circle? (only applies when prcomp was called with scale = TRUE and when var.scale = 1) |
varname.size |
size of the text for variable names |
varname.adjust |
adjustment factor the placement of the variable names, >= 1 means farther from the arrow |
varname.abbrev |
whether or not to abbreviate the variable names |
circle.prob |
size of the ellipse in Normal probability |
a ggplot2 plot
The function select and extract the sediment mixtures of the dataset.
inputSample(data)inputSample(data)
data |
Data frame containing source and mixtures data |
The function select and extract the source samples of the dataset.
inputSource(data)inputSource(data)
data |
Data frame containing source and mixtures data |
This function excludes from the original data frame the properties which do not show significant differences between sources.
KWTest(data, pvalue = 0.05)KWTest(data, pvalue = 0.05)
data |
Data frame containing source and mixtures |
pvalue |
p-value threshold |
Data frame only containing the variables that pass the Kruskal-Wallis test
The function performs a linear discriminant analysis and displays the data in the relevant dimensions.
LDAPlot(data, P3D = FALSE)LDAPlot(data, P3D = FALSE)
data |
Data frame containing source and mixtures data |
P3D |
Boolean to switch between 2 to 3 dimensional chart |
The function performs a principal components analysis on the given data matrix and displays a biplot using vqv.ggbiplot package of the results for each different source to help the user in the decision.
PCAPlot(data, components = c(1:2))PCAPlot(data, components = c(1:2))
data |
Data frame containing source and mixtures data |
components |
Numeric vector containing the index of the two principal components in the chart |
The function performs a density chart of the relative contribution of the potential sediment sources for each sediment mixture in the dataset.
plotResults(data, y_high = 6.5, n = 1)plotResults(data, y_high = 6.5, n = 1)
data |
Data frame containing the relative contribution of the potential sediment sources for each sediment mixture in the dataset |
y_high |
Number of the vertical height of the y-axis |
n |
Number of charts per row |
Function that excludes the properties of the sediment mixture/s outside the minimum and maximum values in the sediment sources.
rangeTest(data)rangeTest(data)
data |
Data frame containing source and mixtures |
Data frame containing sediment sources and mixtures
Asses the relative contribution of the potential sediment sources for each sediment mixture in the dataset.
unmix(data, samples = 100L, iter = 100L, seed = 123456L)unmix(data, samples = 100L, iter = 100L, seed = 123456L)
data |
Data frame containing sediment source and mixtures |
samples |
Number of samples in each hypercube dimension |
iter |
Iterations in the source variability analysis |
seed |
Seed for the random number generator |
Data frame containing the relative contribution of the sediment sources for each sediment mixture and iterations
Mixing model
unmix_c(sources, samples, trials = 100L, iter = 100L, seed = 69512L)unmix_c(sources, samples, trials = 100L, iter = 100L, seed = 69512L)
sources |
Data frame containing sediment sources data |
samples |
Data frame containing sediment mixtures data |
trials |
Number of samples in each hypercube dimension |
iter |
Iterations in the source variability analysis |
seed |
Seed for the random number generator |
Data frame containing the relative contribution of the sediment sources for each sediment mixture and iterations
The function saves the results in the workspace file for all the sediment mixture samples and for each sediment mixture sample separately
writeResults(data)writeResults(data)
data |
Data frame containing the relative contribution of the potential sediment sources for each sediment mixture in the dataset |