simcapt.bvn: Simulate Detections from Elliptical Home Ranges
In MurrayEfford/secrBVN: Bivariate Normal Home Range SECR

Description Usage Arguments Details Value Note See Also Examples

This extends the functionality of sim.capthist to elliptical home ranges, with some limitations.

1 2	simcapt.bvn(traps, popn, type = c('uniform','BVN'), g0 = 0.2, lambda0 = 0.2, p = 0.95, noccasions = 5, renumber = TRUE)

`traps`	object of class 'traps'
`popn`	object of class 'popn' with attributes 's2xy' and 'theta'
`type`	uniform ellipse vs bivariate normal
`g0`	numeric intercept of detection function (type = 'uniform')
`lambda0`	numeric intercept of hazard function (type = 'BVN')
`p`	numeric probability corresponding to edge of ellipse (type = 'uniform')
`noccasions`	integer number of sampling occasions
`renumber`	logical; if FALSE row names are carried over from `popn`; otherwise the n detected animals are renumbered 1:n

traps may have detector type ‘multi’ or ‘proximity’; other types will be coerced to detector type 'proximity'.

popn will have the necessary attributes if it was generated with simpopn.bvn.

Two detection functions are offered. In one (type = 'BVN') the hazard of detection h by a detector at point x is proportional to the probability density of the bivariate normal distribution at x specified by s2xy and theta (these may vary among individuals in popn). The proportionality is determined by lambda0, which is the maximum detection hazard (i.e. when detector is at HR centre). The probability of detection is 1 - exp(-h).

With the other detection function (type = 'uniform') a uniform probability of detection (g0) applies throughout a home range ellipse, with zero probability of detection elsewhere. The boundary is determined by both the shape, orientation and location (using s2xy and theta as before) and the argument p that determines the notional bivariate-normal probability contour to use. (The simulated distribution is not bivariate-normal - this is just a device to specify an ellipse in a familiar way).

For detector type ‘multi’, the hazard of detection is modelled as bivariate normal or elliptical uniform.

capthist object suitable for analysis in secr

Arguments g0 and p have no effect when type = 'BVN'. Argument lambda0 has no effect when type = 'uniform'.

simpopn.bvn, secr.fit

tempgrid <- make.grid(detector='proximity')
pop <- simpopn.bvn(s2xy=c(225/2,225*2), core = tempgrid,
    buffer = 100, D = 10)
temp <- simcapt.bvn(tempgrid, pop, 'uniform')
plot (tempgrid)
plotpopn.bvn(pop)
plot(temp, add = TRUE)