Patuxent Science Meeting 2006 Poster Abstract

Estimating site occupancy and detection probability parameters for

meso- and large mammals in a coastal ecosystem

O'Connell AF, Sauer JR, Gilbert AT, Bailey LL (NCSU - USGS CF&WRU)

Long-term monitoring programs for wildlife populations are being implemented world-wide in an

effort to assess changes that occur across time and space. Unfortunately, many programs fail

to consider two important sources of variability, spatial variation and detectability, often resulting

in index-type information where the relationship between the count and the target population is

unknown, making it difficult to make meaningful inferences. As part of the U.S. National Park

Service's (NPS) nationwide effort to develop science-based monitoring programs, we used a new

technique (MacKenzie et al. 2002, 2003, 2004, 2005) to estimate site occupancy and detection

probabilities for medium-sized and large mammals on Cape Cod, Massachusetts. Our results

demonstrated that detectability is not equal to 1 nor is it likely to be constant due to a variety of

factors including sampling methods, behavior and ecology of the target species, and/or local

environmental characteristics. In this study, species-specific detectability varied considerably

across detection methods and resulted in questionable site occupancy estimates when not

accounting for this variation. As a result, we question site occupancy estimates (and resulting

inferences) when detectability falls below 0.15 and we consider probabilities < 0.05

unacceptable for occupancy estimation and inference. Future studies will benefit from

incorporating detection probabilities into their findings because of the potential variation in

detectability. Site occupancy surveys that incorporate detection probabilities provide the

necessary foundation for conducting efficient biological inventories and subsequent long-term

Friday, September 22, 2006

Return to Posters page