Hierarchical models improve the use of alligator abundance as an indicator

Abstract

Indicator species are species which can be monitored as an index to measure the overall health of an ecosystem. Crocodylians have been shown to be good indicators of wetland condition as they respond to changes in hydrology, can be efficiently monitored, and are a key part of ecosystem trophic relationships. Eye shine surveys at night are a standard method used to sample alligators, but because some individuals that are present in a study area may go undetected and the proportion of individuals counted is not constant over time, appropriate modeling is required to convert counts to estimates of abundance. We analyzed 13 years of American alligator (Alligator mississippiensis) survey count data from South Florida using an N-mixture model. Alligator abundance estimates were assigned to quartiles that were then represented as color coded categories of red, yellow, or green to provide a straightforward rating of Everglades restoration based on familiar stoplight coloring. These results were then compared to a previously used method in which unadjusted counts of these same data were assigned to color coded quartile categories. Water depth played a major role in the detection probability of alligators and the stoplight colors between the two methods matched 76% of the time. This suggests that the original stoplight score method provided a good overall snapshot of the trends in alligator abundance in the Everglades; however, the hierarchical models estimate abundance and trends of alligator abundance by incorporating detection probability thus providing unbiased estimates of abundance.

Keywords

• Comprehensive Everglades restoration
• Hydrology
• N-mixture models
• Stoplight report