Premium
Using California Gnatcatcher to Test Underlying Models in Habitat Conservation Plans
Author(s) -
WINCHELL CLARK S.,
DOHERTY PAUL F.
Publication year - 2008
Publication title -
the journal of wildlife management
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.94
H-Index - 111
eISSN - 1937-2817
pISSN - 0022-541X
DOI - 10.2193/2006-356
Subject(s) - occupancy , habitat , sampling (signal processing) , vegetation (pathology) , abundance (ecology) , environmental science , probabilistic logic , scale (ratio) , ecology , species distribution , physical geography , geography , statistics , computer science , mathematics , cartography , biology , medicine , filter (signal processing) , pathology , computer vision
Habitat Conservation Plans are a widely used strategy to balance development and preservation of species of concern and have been used in southern California, USA, to protect the coastal California gnatcatcher (Polioptila californica). Few data exist on gnatcatcher abundance and distribution, and existing data have problems with issues of closure (i.e., sampling occurs in a short enough time period such that abundance or distribution are not changing), detectability, and proper attention to probability‐based sampling schemes. Thus, a habitat model has been relied upon in reserve design. California gnatcatchers are the flagship and umbrella species of many plans and we provide the first estimates that incorporate probabilistic sampling and test predictions from the habitat model. Probability of occurrence was 26% (SĚ = 0.06); however, occupancy varied by modeled habitat quality with slopes <40%, warm, and wet sagebrush habitat having higher occupancy probabilities. Interpreting abundance and occupancy probabilities by vegetation type was complicated by error detected in Geographic Information System vegetation metadata files. The slope (1.08, SĚ = 0.66), temperature (0.79, SĚ = 0.70), and precipitation (—2.62, SĚ = 1.21) variables associated with habitat models were stronger influences on occupancy than was patch size (0.48, SĚ = 0.66). Previous models weight patch size equal to slope and climate. Our work demonstrates that probabilistic sampling can be carried out on a large scale and the results provide better information for managers to make decisions about their reserves.