bobcatSCR {multimark}

Bobcat spatial capture-recapture data

Description

Example spatial bobcat data for multimark package.

Format

These spatial capture-recapture data with multiple mark types are summarized in a list of length 3 containing the following objects:

Enc.Mat is a 42 x (noccas*ntraps) matrix containing observed encounter histories for 42 bobcats across noccas=187 sampling occasions and ntraps=30 traps. The first 187 columns correspond to trap 1, the second 187 columns corresopond to trap 2, etc.

trapCoords is a matrix of dimension ntraps x (2 + noccas) indicating the Cartesian coordinates and operating occasions for the traps, where rows correspond to trap, the first column the x-coordinate, and the second column the y-coordinate. The last noccas columns indicate whether or not the trap was operating on each of the occasions, where ‘1’ indciates the trap was operating and ‘0’ indicates the trap was not operating.

studyArea is a 3-column matrix containing the coordinates for the centroids of the contiguous grid of 1023 cells that define the study area and available habitat. Each row corresponds to a grid cell. The first 2 columns indicate the Cartesian x- and y-coordinate for the centroid of each grid cell, and the third column indicates whether the cell is available habitat (=1) or not (=0). The grid cells are 0.65x0.65km resolution.

Bobcats are bilaterially asymmetrical, and sampling was conducted using camera stations consisting of a single camera. Because the left-side cannot be reconciled with the right-side, the two types of “marks” in this case are the pelage patterns on the left- and right-side of each individual. Encounter type 0 corresponds to non-detection, encounter type 1 corresponds to left-sided detection, encounter type 2 corresponds to right-sided detection.

Both-sided encounters were never observed in this dataset, hence the most appropriate multimark data type is data.type="never".

The first 15 rows of bobcatSCR$Enc.Mat correspond to individuals for which both the left and right sides were known because they were physically captured for telemetry deployments prior to sampling surveys. The encounter histories for these 15 individuals are therefore known with certainty and should be specified as such using the known argument in processdataSCR and/or multimarkClosedSCR (see example below).

These data were obtained from the R package SPIM (Augustine et al. 2017) and modified by projecting onto a regular rectangular grid consisting of square grid cells (as is required by the spatial capture-recapture models in multimark).

Details

We thank B. Augustine and co-authors for making these data publicly available in the SPIM package (Augustine et al. 2017).

Source

Augustine, B., Royle, J.A., Kelly, M., Satter, C., Alonso, R., Boydston, E. and Crooks, K. 2017. Spatial capture-recapture with partial identity: an application to camera traps. bioRxiv doi: https://doi.org/10.1101/056804

See Also

multimarkClosedSCR, processdataSCR

Examples

data(bobcatSCR)
#plot the traps and available habitat within the study area
plotSpatialData(trapCoords=bobcatSCR$trapCoords,studyArea=bobcatSCR$studyArea)

# This example is excluded from testing to reduce package check time
# Example uses unrealistically low values for nchain, iter, and burnin

# Fit spatial model to tiger data
Enc.Mat <- bobcatSCR$Enc.Mat
trapCoords <- bobcatSCR$trapCoords
studyArea <- bobcatSCR$studyArea

# specify known encounter histories
known <- c(rep(1,15),rep(0,nrow(Enc.Mat)-15))

# specify prior bounds for sigma2_scr
sig_bounds <- c(0.1,max(diff(range(studyArea[,"x"])),diff(range(studyArea[,"y"]))))

mmsSCR <- processdataSCR(Enc.Mat,trapCoords,studyArea,known=known)
bobcatSCR.dot.type <- multimarkClosedSCR(mms=mmsSCR,iter=200,adapt=100,burnin=100,
                                         sigma_bounds=sig_bounds)
summary(bobcatSCR.dot.type$mcmc)