sdmProfiling: Run SDM profiling for unsampled sites

In charliem2003/sdmProfiling: A Spatially-explicit Species Distribution Model Evaluation Tool

Run SDM profiling for unsampled sites

Description

Run SDM profiling for unsampled sites

Usage

sdmProfiling(
  unsampledCoords,
  sampledCoords,
  origSDM,
  envStack,
  sdmFun = "sdmModelling",
  sdmFunArgs = list(samples = NULL, envStack = NULL, modFormula = NULL, ntrees = 500),
  parallel = FALSE,
  noCores = 2
)

Arguments

unsampledCoords	data frame of unsampled cells for which we wish to run SDM profiling on, with column names "x", "y" and "ID"
sampledCoords	data frame of sampled cells used to build original SDM, with column names "presence" (0 = absence, 1 = presence), "x" and "y"
origSDM	raster file of the probability of occurrences of the original SDM built using the sampled cells (i.e. those in 'sampledCoords')
envStack	raster stack of environmental variables for building SDM
sdmFun	name of modelling function to use. The default, 'sdmModelling', is the built-in random forest model. The user can use their own defined function as long as the output is a single raster object of the predicted probability of occurrences
sdmFunArgs	a list with the arguments for the sdmFun (which is called through do.call). The first argument should be the coordinates with to build the model, and will be automatically generated and replaced the first list item. See the examples for how it works with the default random forest SDM function.
parallel	TRUE or FALSE to whether to run profiling on multiple cores
noCores	if parallel = TRUE then the number of cores to share jobs to. If noCores is larger than values retrieved through parallel::detectCores(), then the values from parallel::detectCores() (the total number of cores on the machine) will be used. Be wary of memory requirements when thinking about how many cores to use/
modFormula	model formula if using built-in random forest model
ntrees	number of trees for generating forest if using built-in random forest model

Value

A data frame containing results of the SDM profile where each row is an unsampled cell and eight columns: the 'x' and 'y' coordinates of the unsampled cells we profiled; the sum standardised channge in probability of occurrencces when the unsampled point was changed to an absence ('diff0') or presence ('diff1'); the leverage values ('redundancy', 'presence_leverage', 'absence_leverage' and 'dual_leverage') measured as the distance to the four corners of the profile plot.

Author(s)

Charlie Marsh ([e-mail]) & Yoni Gavish

Examples

set.seed(9999)
envSet <- create_env_nsets(cellDims = c(100, 100),
                           sets     = c(4, 4, 3, 1),
                           model    = "Sph",
                           psill    = 1.5,
                           dep1     = 1,
                           rangeFun = function() exp(runif(1, 1, 6)),
                           propSamp = 0.25)

### generate a virtual species from the variables
sp <- create_sp(envStack = envSet,
                spFun    = "x[1] * x[5] * x[9]",
                spModel  = "Sph",
                spPsill  = 1,
                spRange  = 500,
                propSamp = 0.5,
                prev     = 0.1)

### an initial 'sample' of the species (assuming perfect detection)
sampPts <- data.frame(sampleRandom(sp$presence, 50, na.rm = TRUE, xy = TRUE))

### a formula to fit to random forest (additive for all vars + quadratics)
form <- paste0("presence ~ ", paste(names(envSet), collapse = " + "), "+ I(",
                paste(names(envSet), collapse = " ^ 2) + I("), " ^ 2)")

### run the initial model
spMod <- sdmModelling(samples = sampPts,
                      envStack = envSet,
                      modFormula = form,
                      ntrees = 500,
                      plot = FALSE)

### a random set of 500 points to profile
unsampPts <- data.frame(x = runif(500, 1, 100), y = runif(500, 1, 100))
unsampPts <- unsampPts[!paste(unsampPts$x, unsampPts$y) %in%
                         paste(sampPts$x, sampPts$y), ]

profile <- sdmProfiling(unsampledCoords = unsampPts,
                        sampledCoords   = sampPts,
                        origSDM         = spMod,
                        envStack        = envSet,
                        sdmFun          = "sdmModelling",
                        sdmFunArgs      = list(samples    = NULL,
                                               envStack   = envSet,
                                               modFormula = form,
                                               ntrees     = 50),
                        parallel = FALSE)

charliem2003/sdmProfiling documentation built on June 13, 2022, 4:43 a.m.