computeMESS: Compute MESS rasters

In peterbat1/fitMaxnet: Fit MaxEnt Niche Models Using maxnet

Compute MESS rasters

Description

Compute set of MESS rasters for a maxnet model

Usage

computeMESS(
  thisModel = NULL,
  envPath = "",
  occSWD,
  bkgSWD,
  varImpThreshold = 0,
  outPath = "",
  MESSonly = TRUE
)

Arguments

thisModel	A maxnet model object
envPath	String (character). Full specified path to the environmental predictor raster layers used to fit the maxnet model
occSWD	Data.frame. Reference environmental data at the occurrence locations used to fit the model in MaxEnt SWD format
bkgSWD	Data.frame. Reference environmental data for locations used as background points used to fit the model in MaxEnt SWD format
varImpThreshold	Numeric. Value used to filter variables by their importance scores; percentage importance therefore ranges from 0 to 100.
outPath	Character. Path to folder into which output raster layers will be written
MESSonly	Logical. Compute only the MESS raster layer? Default is TRUE. If FALSE, then MESS component rasters for each variable plus the MESS raster layer are computed

Details

Multivariate Environmental Similarity Surfaces (MESS) were introduced by Elith, Kearney & Phillips (2010; The art of modelling range-shifting species. Methods in Ecology and Evolution 1: 330-342) to assist in identifying the degree of similarity between the ranges of environmental predictors presented to model training (the 'reference' set) and the values of those predictors across other areas. It's purpose is to identify areas of extrapolation. In the present implementation, the reference set is extracted from the tables of variable maximum and minimumthose in areas or times to which the model is projected.

A MESS computation is performed using the supplied maxnet model and set of environmental rasters.

To facilitate the processing of very large, high-resolution rasters, only those variables which pass the variable importance threshold are used in the final model are included in the computation of MESS output.

Please note that this function is tailored to work with data structures and workflows associated with the primary model fitting function in this package, namely fit_maxnet. It can be used for MESS computations in other settings, but users will have to transform data into the required structure for each parameter. An alternate and more generic MEWS function in found in the R-package modEVA.

If output rasters are to be written, note that the following options are set as defaults for terra::writeRaster(): 'overwrite = TRUE', and 'gdal = "COMPRESS=DEFLATE"'

Value

A named list:

MESS

Matrix of OVERALL MESS scores

varMESS

Matrix of MESS scores for each included variable

peterbat1/fitMaxnet documentation built on Sept. 17, 2024, 10:50 p.m.