int.est.dp: Double Platform Line Transect Method Abundance Estimation:...

In dill/wisp: Wildlife Simulation Package (wisp)

Description

This function estimates a confidence interval for group abundance for the double platform line transect method.

Usage

        int.est.dp(sampl, ci.type = "boot.nonpar", nboot = 999, vlevels = c(0.025, 0.975), model = "~distance", plot = FALSE, show.all = FALSE, seed = NULL, ...)

Arguments

sampl	object of class 'sample.dp´.
ci.type	="boot.nonpar" (no other methods implemented yet)
nboot	number bootstrap replicates
vlevels	percentage levels for confidence iterval
model	model specification allowing the selection of covariates
plot	=T if you want a plot of boostrap distribution of Nhat.grp
show.all	if TRUE, shows details for each bootstrap resample, as it is run.
seed	the number passed to set.seed() to initialise random number generator
...	other plot arguments

Details

'model' - The default setting is model='~distance' which takes the perpendicular distance of the recorded observation as the sole explanatory variable of the response variable 'seen'. That is, whether an animal was observed or not is solely a function of its perpendicular distance from the observer. The WiSP sample object 'sample.dp' contains the variables distance, exposure, groupsize and one factor level variable named types. A platform variable is also created to denote which of the two 'platforms' observed the animal. A model including all these explanatory variables would then be defined as: model="~platform + distance + exposure + size + types" Note that the explanatory variable size refers to groupsize in the sample object.

Value

An object of class 'int.est.dp´ containing the following items:

levels	percentage levels for confidence interval
ci	the confidence interval
boot.mean	mean of bootstrap estimates
boot.dbn	a list with components "Nhat.grp", "Nhat.ind", "Es", "prob.det", "mu", "nL", "average.g0", "log.Likelihood", and "AIC" (see point.est.dp) each containing the full set of nboot estimates of that parameter. (This is useful if you want to create a plot of the bootstrap distribution of the estimates and control bin width etc.)

Note

The nonparametric bootstrap method resamples transects with replacement (each resample has the same number of transects as were in the original sample). Details of the method are given in Borchers et al. (2002).

References

Borchers, D.L., Buckland, S.T. and Zucchini, W. 2002. Estimating animal abundance: closed populations. Springer. London. 314pp.

See Also

setpars.survey.dp, generate.sample.dp point.est.dp for point estimation

Examples

dp.reg <- generate.region(x.length = 100, y.width = 50)
dp.dens <- generate.density(dp.reg)

#heterogeneous population
dp.poppars<-setpars.population(density.pop = dp.dens, number.groups = 1000, size.method = "poisson", size.min = 1, size.max = 30, size.mean = 10, exposure.method = "beta", exposure.min = 0, exposure.max = 1, exposure.mean = 0.4, exposure.shape = 0.5, type.values=c("Male","Female"), type.prob=c(0.48,0.52))
dp.pop<-generate.population(dp.poppars)

dp.despars<-setpars.design.dp(dp.reg, n.transects=10, n.units=10, visual.range=2, percent.on.effort=1)
dp.des<-generate.design.dp(dp.despars, seed=3)
dp.survpars<-setpars.survey.dp(dp.pop, dp.des, disthalf.min=2, disthalf.max=4)
dp.samp<-generate.sample.dp(dp.survpars)

#This may take a minute or two to run.  
#For a quick viewing, press escape, and try again with a low nboot number
#At least 99 is recommended for estimates with real data.  999 is better.
dp.ci<-int.est.dp(dp.samp, nboot=99, model="~distance")
summary(dp.ci)
plot(dp.ci)

dill/wisp documentation built on May 15, 2019, 8:31 a.m.