EmpiricalExamples/PolarBearExample.R

In MarieAugerMethe/CCRWvsLW: Searching strategy movement models

######################
# Example to fit the models to a Polar Bear movement track
# This is to reproduce the analysis of Polar Bear P3 in 
# Auger-Methe et al. (2016) Evaluating random search strategies in three mammals
# from distinct feeding guils. Journal of Animal Ecology

# The results can be compared to those of the supporting information

library(CCRWvsLW)

##
# Load data
pb <- read.csv("SampleData/PolarBearSample.csv", stringsAsFactors = FALSE)

##
# Transform data into ltraj object
# We have the movement in the x and y direction
x <- c(0,cumsum(pb$dx)[-length(pb$dx)])
y <- c(0,cumsum(pb$dy)[-length(pb$dy)])
pbDate <- as.POSIXct(pb$Date,tz="UTC")

# Make ltraj object
cMov <- as.ltraj(xy=cbind(x,y), date=pbDate, id="pb", slsp="missing")

# Regularise the time series
gbDateRef <-  as.POSIXct(format(pbDate[1], "%Y-%m-%d 01:00"), tz="UTC")
cMov <- setNA(cMov, date.ref=gbDateRef, dt=4, units="hour")
cMov <- sett0(cMov, date.ref=gbDateRef, dt=4, units="hour")

#####
# Analyse the data
cRes <- movLikelihoods(cMov, PRdetails=TRUE, TAc=10, conts=FALSE, hspo=TRUE)
# CCRW_A, CCRW_L, TLW, BW, CRW
cAICc <- c(cRes$mleMov$CCRW["AICc"], cRes$mleMov$HSMMpo["AICc"], 
  cRes$mleMov$TLW["AICc"],
  cRes$mleMov$BW["AICc"], cRes$mleMov$CRW["AICc"])
# Delta AICc as in Table S2.3
cAICc - min(cAICc)

# p-value for test of absolute fit, based on Step length only (as in Table S2.3)
cRes$pseudoRes$Z["pval","SL_CCRW"]