Paradise_shelduck: Mulstistate Live-Dead Paradise Shelduck Data

In marked: Mark-Recapture Analysis for Survival and Abundance Estimation

Mulstistate Live-Dead Paradise Shelduck Data

Description

Paradise shelduck recapture and recovery data in multistrata provided by Richard Barker and Gary White.

Format

A data frame with 1704 observations of 3 variables

ch

a character vector containing the capture history (each is 2 character positions LD) for 6 occasions

freq

capture history frequency

sex

Male or Female

Author(s)

Jeff Laake

References

Barker, R.J, White,G.C, and M. McDougall. 2005. MOVEMENT OF PARADISE SHELDUCK BETWEEN MOLT SITES: A JOINT MULTISTATE-DEAD RECOVERY MARK–RECAPTURE MODEL. JOURNAL OF WILDLIFE MANAGEMENT 69(3):1194–1201.

Examples

# In the referenced article, there are 3 observable strata (A,B,C) and 3 unobservable 
# strata (D,E,F). This example is setup by default to use only the 3 observable strata
# to avoid problems with multiple modes in the likelihood.
# Code that uses all 6 strata are provided but commented out. 
# With unobservable strata, simulated annealing should
# be used (options="SIMANNEAL")
data("Paradise_shelduck")
# change sex reference level to Male to match design matrix used in MARK
ps$sex=relevel(ps$sex,"Male")
# Process data with MSLiveDead model using sex groups and specify only observable strata
ps_dp=process.data(ps,model="MSLD",groups="sex",strata.labels=c("A","B","C"))
# Process data with MSLiveDead model using sex groups and specify observable and 
# unboservable strata
# ps_dp=process.data(ps,model="MSLD",groups="sex",strata.labels=c("A","B","C","D","E","F"))
# Make design data and specify constant PIM for Psi to reduce parameter space. No time 
#variation was allowed in Psi in the article.
ddl=make.design.data(ps_dp)
# Fix p to 0 for unobservable strata (only needed if they are included)
ddl$p$fix=NA
ddl$p$fix[ddl$p$stratum%in%c("D","E","F")]=0
# Fix p to 0 for last occasion
ddl$p$fix[ddl$p$time%in%6:7]=0.0
# Fix survival to 0.5 for last interval to match MARK file (to avoid confounding)
ddl$S$fix=NA
ddl$S$fix[ddl$S$time==6]=0.5
# create site variable for survival which matches A with D, B with E and C with F 
ddl$S$site="A"
ddl$S$site[ddl$S$stratum%in%c("B","C")]=as.character(ddl$S$stratum[ddl$S$stratum%in%c("B","C")])
ddl$S$site[ddl$S$stratum%in%c("E")]="B"
ddl$S$site[ddl$S$stratum%in%c("F")]="C"
ddl$S$site=as.factor(ddl$S$site)
# create same site variable for recovery probability (r)
ddl$r$site="A"
ddl$r$site[ddl$r$stratum%in%c("B","C")]=as.character(ddl$r$stratum[ddl$r$stratum%in%c("B","C")])
ddl$r$site[ddl$r$stratum%in%c("E")]="B"
ddl$r$site[ddl$r$stratum%in%c("F")]="C"
ddl$r$site=as.factor(ddl$r$site)
# Specify formula used in MARK model
S.1=list(formula=~-1+sex+time+site)
p.1=list(formula=~-1+stratum:time)
r.1=list(formula=~-1+time+sex+site)
Psi.1=list(formula=~-1+stratum:tostratum)
# Run top model from paper but only for observable strata
# commented out to prevent dll being built - problem with CRAN check
#crmmod=crm(ps_dp,ddl,model.parameters=list(S=S.1,p=p.1,r=r.1,Psi=Psi.1),
#              method="nlminb",hessian=TRUE)
# Run top model from paper for all strata using simulated annealing (commented out)
# crmmod=crm(ps_dp,ddl,model.parameters=list(S=S.1,p=p.1,r=r.1,Psi=Psi.1),
#                     method="SANN",itnmax=6e6,hessian=TRUE)
unlink("*.cpp")
unlink("*.o")
unlink("symbols.rds")

marked documentation built on Oct. 19, 2023, 5:06 p.m.