ensemble.zones {BiodiversityR}

Mapping of environmental zones based on the Mahalanobis distance from centroids in environmental space.

Description

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).

Usage

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)

Arguments

presence.raster

RasterLayer object (raster) documenting presence (coded 1) of an organism

centroid.object

Object listing values for centroids and covariance to be used with the mahalanobis distance (used internally by the prediction function called from predict).

x

RasterStack object (stack) containing all environmental layers that correspond to explanatory variables

ext

an Extent object to limit the predictions and selection of background points to a sub-region of presence.raster and x, typically provided as c(lonmin, lonmax, latmin, latmax). See also randomPoints and extent.

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 writeFormats and writeRaster.

RASTER.datatype

Format of the raster files that will be generated. See dataType and writeRaster.

RASTER.NAflag

Value that is used to store missing data. See writeRaster.

CATCH.OFF

Disable calls to function tryCatch.

categories.raster

RasterLayer object (raster) documenting predefined zones such as vegetation types. In case this object is provided, then centroids will be calculated for each zone.

an

Number of presence points to be used for Principal Components Analysis (via rda); see also prepareData and extract

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 (scores).

centers

Number of centers (clusters) to be used for K-means clustering (kmeans). In case a value smaller than 1 is provided, function cascadeKM is called to determine the optimal number of centers via the Calinski-Harabasz criterion.

use.silhouette

If TRUE, then centroid values are only based on presence points that have silhouette values (silhouette) larger than 0.

plotit

If TRUE, then a plot is provided that shows the locations of centroids in geographical and environmental space. Plotting in geographical space is based on determination of the presence location (analogue) with smallest Mahalanobis distance to the centroid in environmental space.

dev.new.width

Width for new graphics device (dev.new). If < 0, then no new graphics device is opened.

dev.new.height

Heigth for new graphics device (dev.new). If < 0, then no new graphics device is opened.

Details

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.

Value

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

Author(s)

Roeland Kindt (World Agroforestry Centre)

See Also

ensemble.raster

Examples

## 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)