1365-2664/asset/olbannerleft.gif?v=1&s=8b608cc23970983efcf0bf9354181123ee4feba9)
1365-2664/asset/olbannerright.gif?v=1&s=01405a21098d64198820bdbe2e30807b513e69f2)
A cautionary note on substituting spatial subunits for repeated temporal sampling in studies of site occupancy
Article first published online: 12 NOV 2009
DOI: 10.1111/j.1365-2664.2009.01732.x
© 2009 The Authors. Journal compilation © 2009 British Ecological Society
Additional Information
How to Cite
Kendall, W. L. and White, G. C. (2009), A cautionary note on substituting spatial subunits for repeated temporal sampling in studies of site occupancy. Journal of Applied Ecology, 46: 1182–1188. doi: 10.1111/j.1365-2664.2009.01732.x
Publication History
- Issue published online: 26 NOV 2009
- Article first published online: 12 NOV 2009
- Received 10 February 2009; accepted 18 October 2009 Handling Editor: Fangliang He
Keywords:
- closure assumption;
- landscape;
- metapopulation;
- patch occupancy;
- sampling with replacement;
- species occurrence
Summary
1. Assessing the probability that a given site is occupied by a species of interest is important to resource managers, as well as metapopulation or landscape ecologists. Managers require accurate estimates of the state of the system, in order to make informed decisions. Models that yield estimates of occupancy, while accounting for imperfect detection, have proven useful by removing a potentially important source of bias. To account for detection probability, multiple independent searches per site for the species are required, under the assumption that the species is available for detection during each search of an occupied site.
2. We demonstrate that when multiple samples per site are defined by searching different locations within a site, absence of the species from a subset of these spatial subunits induces estimation bias when locations are exhaustively assessed or sampled without replacement.
3. We further demonstrate that this bias can be removed by choosing sampling locations with replacement, or if the species is highly mobile over a short period of time.
4. Resampling an existing data set does not mitigate bias due to exhaustive assessment of locations or sampling without replacement.
5. Synthesis and applications. Selecting sampling locations for presence/absence surveys with replacement is practical in most cases. Such an adjustment to field methods will prevent one source of bias, and therefore produce more robust statistical inferences about species occupancy. This will in turn permit managers to make resource decisions based on better knowledge of the state of the system.