ovenbird: Ovenbird Mist-netting Dataset

Description

Data from a multi-year mist-netting study of ovenbirds (Seiurus aurocapilla) at a site in Maryland, USA.

Usage

ovenbird

Arguments

Details

From 2005 to 2009 D. K. Dawson and M. G. Efford conducted a capture--recapture survey of breeding birds in deciduous forest at the Patuxent Research Refuge near Laurel, Maryland, USA. The forest was described by Stamm, Davis & Robbins (1960), and has changed little since. Analyses of data from previous mist-netting at the site by Chan Robbins were described in Efford, Dawson & Robbins (2004) and Borchers & Efford (2008).

Forty-four mist nets (12 m long, 30-mm mesh) spaced 30 m apart on the perimeter of a 600-m x 100-m rectangle were operated for approximately 9 hours on each of 9 or 10 non-consecutive days during late May and June in each year. Netting was passive (i.e. song playback was not used to lure birds into the nets). Birds received individually numbered bands, and both newly banded and previously banded birds were released at the net where captured. Sex was determined in the hand from the presence of a brood patch (females) or cloacal protuberance (males). A small amount of extra netting was done by other researchers after the main session in some years.

This dataset comprises all records of adult (after-hatch-year) ovenbirds caught during the main session in each of the five years 2005--2009. One ovenbird was killed by a predator in the net in 2009, as indicated by a negative net number in the dataset. Sex was determined in the hand from the presence of a brood patch (females) or cloacal protuberance (males). Birds are listed by their band number (4-digit prefix, `.', and 5-digit number).

The data are provided as a multi-session capthist object `ovenCHp'. Sex is coded as a categorical individual covariate ("M" or "F").

Recaptures at the same site within a day are not included in this dataset, so ovenCHp has detector type `proximity'. Previous versions of secr provided only a trimmed version of these data, retaining only one capture per bird per day (ovenCH with detector type `multi'). That may be obtained from ovenCHp as shown in the examples.

Although several individuals were captured in more than one year, no use is made of this information in the analyses presently offered in secr.

An analysis of the data for males in the first four years showed that they tended to avoid nets after their first capture within a season (Dawson & Efford 2009). While the species was present consistently, the number of detections in any one year was too small to give reliable estimates of density; pooling of detection parameters across years helped to improve precision.

Included with the data are a mask and four models fitted to ovenCH as in Examples.

Object	Description
ovenCH	multi-session capthist object
ovenbird.model.1	fitted secr model -- null
ovenbird.model.1b	fitted secr model -- g0 net shyness
ovenbird.model.1T	fitted secr model -- g0 time trend within years
ovenbird.model.h2	fitted secr model -- g0 finite mixture
ovenbird.model.D	fitted secr model -- trend in density across years
ovenmask	mask object

References

Borchers, D. L. and Efford, M. G. (2008) Spatially explicit maximum likelihood methods for capture-recapture studies. Biometrics 64, 377--385.

Dawson, D. K. and Efford, M. G. (2009) Bird population density estimated from acoustic signals. Journal of Applied Ecology 46, 1201--1209.

Efford, M. G., Dawson, D. K. and Robbins C. S. (2004) DENSITY: software for analysing capture-recapture data from passive detector arrays. Animal Biodiversity and Conservation 27, 217--228.

Stamm, D. D., Davis, D. E. and Robbins, C. S. (1960) A method of studying wild bird populations by mist-netting and banding. Bird-Banding 31, 115--130.

Examples

Run this code

# NOT RUN {
# }
# NOT RUN {
## commands used to create ovenCH from the input files
## "netsites0509.txt" and "ovencapt.txt"
## for information only - these files not distributed
netsites0509 <- read.traps(file = "netsites0509.txt",
    skip = 1, detector = "proximity")
temp <- read.table("ovencapt.txt", colClasses=c("character",
    "character", "numeric", "numeric", "character"))
ovenCHp <- make.capthist(temp, netsites0509, covnames=c("Sex"))
ovenCHp <- updateCH(ovenCHp)   # drop repeat detections
# }
# NOT RUN {
par(mfrow = c(1,5), mar = c(1,1,4,1))
plot(ovenCHp, tracks = TRUE, varycol = TRUE)
par(mfrow = c(1,1), mar = c(5,4,4,2) + 0.1)  ## defaults

counts(ovenCHp, "n")

# }
# NOT RUN {
## trimmed version of data - for consistency with earlier versions

ovenCH <- reduce(ovenCHp, outputdetector = "multi", dropunused = FALSE)

## array constant over years, so build mask only once
ovenmask <- make.mask(traps(ovenCH)[["2005"]], type="pdot", buffer=400,
    spacing=15, detectpar=list(g0=0.03, sigma=90), nocc=10)

## fit constant-density model
ovenbird.model.1 <- secr.fit(ovenCH, mask = ovenmask)
ovenbird.model.1

## fit net avoidance model
ovenbird.model.1b <- secr.fit(ovenCH, mask = ovenmask, model =
    list(g0~b))
ovenbird.model.1b

## fit model with time trend in detection
ovenbird.model.1T <- secr.fit(ovenCH, mask = ovenmask, model =
    list(g0 ~ T))
ovenbird.model.1T

## fit model with 2-class mixture for g0
ovenbird.model.h2 <- secr.fit(ovenCH, mask = ovenmask, model =
    list(g0~h2))
ovenbird.model.h2

# }
# NOT RUN {
## compare & average pre-fitted models
AIC (ovenbird.model.1, ovenbird.model.1b, ovenbird.model.1T,
    ovenbird.model.h2)
model.average (ovenbird.model.1,ovenbird.model.1b, ovenbird.model.1T,
    ovenbird.model.h2, realnames = "D")

## select one year to plot
plot(ovenbird.model.1b, newdata = data.frame(session = "2005",
    b = 0))

# }