point.sim.lt: Line Transect Method : Simulation

In dill/wisp: Wildlife Simulation Package (wisp)

Description

This function simulates sample objects of class 'sample.lt' and estimates abundance and related parameters for each simulated sample object.

Usage

        point.sim.lt(pop.spec, survey.spec, design.spec, B = 999, model.sel = FALSE, plot = FALSE, 
				title = FALSE, conditional = TRUE, model = "half.normal", seed = NULL, show = FALSE, ...) 

Arguments

pop.spec	population specification; either an object of class 'population' or 'pars.population'
survey.spec	survey specification; an object of class 'pars.survey.lt'
design.spec	design specification; either an object of class 'design.lt' or 'pars.design.lt'
B	number of simulations required
model.sel	if TRUE, the half-normal and hazard rate models will be fitted to each simulation and the model chosen based on the lowest AIC value will be selected.
plot	if TRUE, a density histogram of observed radial distances with the estimated probability density function superimposed, is plotted
title	Flag telling if you want the title "Perpendicular distance distribution and fitted detection function"
conditional	argument in point.est.lt - if FALSE, the full likelihood (Equation (7.10) of Borchers et al. 2002) is maximized; if TRUE the conditional likelihood (Equation (7.8) of Borchers et al. 2002) is maximized and abundance is estimated using Equation (7.12) of Borchers et al. 2002.
model	argument in point.est.lt - model: ="half.normal" or ="hazard.rate" for detection function form
plot	argument in point.est.lt - if TRUE, a density histogram of observed radial distances with the estimated probability density function superimposed, is plotted
seed	Number passed to set.seed() to initialise random number generator
show	if TRUE displays the histograms of observations, and the fitted distribution function for each simulated survey as it is run
...	extra plot arguments

Details

This function simulates sample objects of class 'sample.lt' by simulating from the observation model (using survey.spec) and if pop.spec is of class 'pars.population' from the state model (using pop.spec to generate new populations on each simulation ) and if design.spec is of class 'pars.design.lt' from the design (using design.spec to generate new design realisations on each simulation.)

Value

An object of class point.sim.lt with the following elements: A results matrix, each row of which contains the following values:

See Also

setpars.population, setpars.design.lt setpars.survey.lt, point.est.lt set.seed

Examples

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

#heterogeneous population
lt.poppars<-setpars.population(density.pop = lt.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))
lt.pop<-generate.population(lt.poppars)

lt.despars<-setpars.design.lt(lt.reg, n.transects=10, n.units=10, visual.range=4, percent.on.effort=1)
lt.des<-generate.design.lt(lt.despars, seed=3)

lt.survpars<-setpars.survey.lt(lt.pop, lt.des, disthalf.min=2, disthalf.max=4)
lt.samp<-generate.sample.lt(lt.survpars)

# simulate (design and population randomisation):
lt.sim<-point.sim.lt(pop.spec=lt.poppars, design.spec=lt.despars, survey.spec=lt.survpars, B=9,
                     seed=NULL, plot=FALSE, model.sel=TRUE)
summary(lt.sim)
plot(lt.sim)
plot(lt.sim, type="hist")
plot(lt.sim, type="box")

# simulate (design randomisation only) half normal:
lt.sim<-point.sim.lt(pop.spec=lt.pop, design.spec=lt.despars, survey.spec=lt.survpars, B=99,
                     seed=NULL, plot=FALSE)
summary(lt.sim)
plot(lt.sim)

# simulate (design randomization only) hazard rate:
lt.sim<-point.sim.lt(pop.spec=lt.pop, design.spec=lt.despars, survey.spec=lt.survpars, B=99,
                     seed=NULL, plot=FALSE, model="hazard.rate")
summary(lt.sim)
plot(lt.sim)

# simulate (design randomization only, automated model selection):
lt.sim<-point.sim.lt(pop.spec=lt.pop, design.spec=lt.despars, survey.spec=lt.survpars, B=99,
                     seed=NULL, plot=FALSE, model.sel=TRUE)
summary(lt.sim)
plot(lt.sim)

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