secrlinear (version 1.2.4)

Arvicola: Water Vole Capture Dataset

Description

Data from a trapping study of water voles (Arvicola amphibius) along a river in Oxfordshire, U.K.

Usage

arvicola

Format

secr capthist object

Details

Water voles were trapped monthly from May 1984 to May 1985 along 0.9 km of the River Glyme near Woodstock in Oxfordshire, U.K. (Efford 1985). Two sheet-aluminium traps were set at stations 20 m apart along one bank and checked morning and evening for 3 days. Traps were baited with slices of carrot and provisioned with bedding and additional carrot. Voles were marked with individual colour-coded ear tags. The dataset provided is from June 1984. This was early in the breeding season and most voles were overwintered adults; only 3 were young-of-the-year, and these were omitted.

Raw data files ``Jun84capt.txt'' and ``glymetrap.txt'' are provided in the `extdata' folder of the secrlinear installation. The vignette ../doc/secrlinear-vignette.pdf shows how to import the raw data.

The data comprise detections of 26 voles at 44 stations on 6 occasions. The two traps at each station were notionally labelled `A' and `B', but captures were recorded only by the station at which they occurred: captures were assigned label `A' or `B' effectively at random. Each trap could catch a single adult vole, but trap saturation was low (maximum 20.5% of traps caught a vole). No voles died in traps in June 1984.

Water voles in the U.K. restrict their activity to waterways and their immediate banks, except for some populations in more extensive (2-dimensional) wetlands. It is therefore natural to treat their habitat as linear in a spatially explicit capture--recapture model of these data. A suitable linear habitat mask is provided in the accompanying dataset glymemask.

See ../doc/secrlinear-vignette.pdf for more analysis of this dataset.

References

Efford, M. G. (1985) The structure and dynamics of water vole populations. D.Phil thesis, University of Oxford.

See Also

glymemask

Examples

head(traps(arvicola))

## for speed, pre-compute distance matrix
userd <- networkdistance (traps(arvicola), glymemask, glymemask)
## fit model
glymefit <- secr.fit(arvicola, mask = glymemask, trace = FALSE,
                     details = list(userdist = userd))
## estimates of 'real' parameters
predict(glymefit)

if (FALSE) {
summary(arvicola)

tmp <- secr.test(glymefit, nsim = 1000)
tmp
plot(tmp)

## More voles were caught only once than is predicted by the model.
## This is probably due to within-population variation in movement or
## capture probability.

}