map.sdm: Predict species distributions.

View source: R/map.R

map.sdmR Documentation

Predict species distributions.

Description

Prediction of potential species distributions using maximum entropy (maxent).

Usage

map.sdm(
  longlat,
  layers,
  error = NULL,
  year = NULL,
  idconf = NULL,
  categorical = NULL,
  thres = 0,
  testpercentage = 0,
  mcp = TRUE,
  points = FALSE,
  eval = TRUE,
  runs = 0,
  subset = 0
)

Arguments

longlat

Matrix of longitude and latitude or eastness and northness (two columns in this order) of each occurrence record.

layers

Predictor variables, a Raster* object as defined by package raster.

error

Vector of spatial error in longlat (one element per row of longlat) in the same unit as longlat. Used to move any point randomly within the error radius.

year

Vector of sampling years in longlat (one element per row of longlat). Used to exclude old records with a given probability proportional to time passed since sampling (never excluded only for current year).

idconf

Vector of identification confidence in longlat (one element per row of longlat). Used to exclude uncertain records with a given probability. Can be on any scale where max values are certain (e.g. from 1 - very uncertain to 10 - holotype).

categorical

Vector of layer indices of categorical (as opposed to quantitative) data. If NULL the package will try to find them automatically based on the data.

thres

Threshold of logistic output used for conversion of probabilistic to binary (presence/absence) maps. If 0 this will be the value that maximizes the sum of sensitivity and specificity.

testpercentage

Percentage of records used for testing only. If 0 all records will be used for both training and testing.

mcp

Used for a precautionary approach. If TRUE, all areas predicted as present but outside the minimum convex hull polygon encompassing all occurrence records are converted to absence. Exceptions are cells connected to other areas inside the polygon.

points

If TRUE, force map to include cells with presence records even if suitable habitat was not identified.

eval

If TRUE, build a matrix with AUC, Kappa, TSS, EOO (from raw data), EOO (from model), AOO (from raw data) and AOO (from model).

runs

If <= 0 no ensemble modelling is performed. If > 0, ensemble modelling with n runs is made. For each run, a new random sample of occurrence records (if testpercentage > 0), background points and predictive variables (if subset > 0) are chosen. In the ensemble model, each run is weighted as max(0, (runAUC - 0.5)) ^ 2.

subset

Number of predictive variables to be randomly selected from layers for each run if runs > 0. If <= 0 all layers are used on all runs. Using a small number of layers is usually better than using many variables for rare species, with few occurrence records (Lomba et al. 2010, Breiner et al. 2015).

Details

Builds maxent (maximum entropy) species distribution models (Phillips et al. 2004, 2006; Elith et al. 2011) using function maxent from R package dismo (Hijmans et al. 2017). Dismo requires the MaxEnt species distribution model software, a java program that can be downloaded from http://biodiversityinformatics.amnh.org/open_source/maxent. Copy the file 'maxent.jar' into the 'java' folder of the dismo package. That is the folder returned by system.file("java", package="dismo"). You need MaxEnt version 3.3.3b or higher. Please note that this program (maxent.jar) cannot be redistributed or used for commercial or for-profit purposes.

Value

List with either one or two raster objects (depending if ensemble modelling is performed, in which case the second is a probabilistic map from all the runs) and, if eval = TRUE, a matrix with AUC, Kappa, TSS, EOO (from raw data), EOO (from model), AOO (from raw data) and AOO (from model). Aggregate values are taken from maps after transformation of probabilities to incidence, with presence predicted for cells with ensemble values > 0.5.

References

Breiner, F.T., Guisan, A., Bergamini, A., Nobis, M.P. (2015) Overcoming limitations of modelling rare species by using ensembles of small models. Methods in Ecology and Evolution, 6: 1210-1218.

Hijmans, R.J., Phillips, S., Leathwick, J., Elith, J. (2017) dismo: Species Distribution Modeling. R package version 1.1-4. https://CRAN.R-project.org/package=dismo

Lomba, A., Pellissier, L., Randin, C.F., Vicente, J., Moreira, F., Honrado, J., Guisan, A. (2010) Overcoming the rare species modelling paradox: a novel hierarchical framework applied to an Iberian endemic plant. Biological Conservation, 143: 2647-2657.

Phillips, S.J., Dudik, M., Schapire, R.E. (2004) A maximum entropy approach to species distribution modeling. Proceedings of the Twenty-First International Conference on Machine Learning. p. 655-662.

Phillips, S.J., Anderson, R.P., Schapire, R.E. (2006) Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190: 231-259.

Elith, J., Phillips, S.J., Hastie, T., Dudik, M., Chee, Y.E., Yates, C.J. (2011) A statistical explanation of MaxEnt for ecologists. Diversity and Distributions, 17: 43-57.


red documentation built on Sept. 9, 2023, 1:07 a.m.

Related to map.sdm in red...