pdissf: Fits the probability of detection integrated step selection...

In jlombard314159/Probability-of-Detection-integrated-Step-Selection-Function: Probability of Detection integrated Step Selection Function (PDiSSF)

Fits the probability of detection integrated step selection function (PDiSSF)

Description

This package fits the probability of detection integrated detection function presented in Vales et al. (2022) and presented as the "RSF for GPS fix success" in Nielson et al. (2009). As of now, this package relies on totally pre-processed data. See help(habitat) and help(locations) for more details and examples.

Usage

pdissf(
habitatDF,
habitatCellID,
CellID,
maximumGap = 3,
iSSFCovars = NULL,
probDetCovars = NULL,
distColumns = NULL
)

Arguments

habitatDF	Dataframe of habitat information
CellID	Column name in the locations file that contains the ID for cells in the study area as they were used by the animal (chronological order is important). Should have IDs found in habitatCellID. Missing fixes should be 'NA'.
habitatCellID	Column in the habitat DF that contains unique IDs for the habitat data. Some will have been used and thus also in the vector of CellID. Missing values are not allowed.
maximumGap	Maximum allowable number of consecutive missing fixes. Default is 3 (3 consecutive missing locations). Large gaps will cause an exponential increase in memory requirements and computing time. This may result in convergence issues.
iSSFCovars	covariates for step selection
probDetCovars	Character string of covariate names for estimating the probability of detection. Set to NULL for fitting a standard iSSF.
distColumns	If distance from previous location (step length) is to be included in the model, identify the columns in the habitat DF that contain x-y coordinates in 2-D planar coordinates such as UTM or State Plane.

Details

pdissf

The system time is printed to the console to indicate that model estimation is ongoing and can be used to identify when the process has stalled. The system time is not printed when running in parallel.

Value

Results of model estimation. Converged = 0 if the algorithim converged, and 1 if there was not covergence.

Examples

library(PDiSSF)
data(habitat)
data(locations)

# Fix success rate
mean(!is.na(locations$unitID))

# Computing time for larger data sets may vary

# integrated step selection function (iSSF) using step length,
pdissf(habitatDF = habitat,
      CellID = locations$unitID, 
      habitatCellID = 'unitID',
      iSSFCovars = c('distance', 'prctSage', 'elevation'), 
      probDetCovars = NULL, 
      distColumns = c('utmX','utmY'))

# PDiSSF including step length
pdissf(habitatDF = habitat, 
      CellID = locations$unitID, 
      habitatCellID = 'unitID',
      iSSFCovars = c('distance', 'prctSage', 'elevation'), 
      probDetCovars = 'prctSage', distColumns = c('utmX','utmY'),
      maximumGap = 3)

jlombard314159/Probability-of-Detection-integrated-Step-Selection-Function documentation built on July 7, 2022, 11:55 p.m.