Detection Probabilities of Karst Invertebrates

Abstract

Protection of federally listed endangered troglobites in central Texas focuses on caves that are occupied by the species. The determination of occupancy is based on presence/absence surveys for those taxa. Under current U.S. Fish and Wildlife Service recommendations, three surveys are used as a standard to determine presence or absence, and certain environmental and seasonal conditions must be met.

We used survey data from 23 caves on Camp Bullis Military Reservation, Bexar County, Texas to test the validity of the survey protocols. Presence/absence matrices were created for three cave species, Batrisodes uncicornis, Chinquipellobunus madlae, and Rhadine exilis. Eleven environmental and seasonal covariates that have been suggested to affect detection probability were tested for fit to the detection data. B. uncicornis and R. exilis were determined to have constant detection probabilities of 0.1226 and 0.1875. C. madlae was found to have a survey specific detection probability (average p = 0.2424), and in no case was detectability tied to any of the measured covariates. Parametric bootstrapping was used to simulate the number of surveys needed to have a 5% chance of not detecting the species if they were present at the site. The number of surveys needed ranged from 10 to 22.

These results indicate that more surveys should be performed before determining absence from a site. The results also indicate that most of the time cave species are not available to be surveyed, and we hypothesize that they retreat into humanly inaccessible cracks connected to the cave.