Using Dung to Estimate Gorilla Density: Modeling Dung Production Rate

There is an urgent need for information on western gorilla population sizes and distribution to improve present and plan future conservation actions. Researchers traditionally have estimated gorilla densities on the basis of nest counts despite demonstrated variation in nest production and decay rates. The variation may lead to large biases in estimates of gorilla abundance. We investigated the use of an alternate index of gorilla abundance, via defecation data collected from habituated gorillas at Bai Hokou, Central African Republic. Our sample of 274/370 defecation events/dung piles produced a production rate of ca. 5 dung piles/d: comparable to previous estimates based on much smaller sample sizes. Heuristic models that failed to account for imperfect dung pile detection produced a lower defecation rate estimate than that of a maximum likelihood model that explicitly modeled detection probability. Generalized linear modeling (GLM) showed that dung pile production rate was strongly linked to rainfall, suggesting that failure to correct for seasonal variation in dung pile production rates could lead to substantial biases in gorilla abundance estimates. In our study, failing to distinguish between the number of defecation events and the number of dung piles produced would lead to a ca. 31% overestimate of true gorilla numbers. The use of dung as an index of gorilla abundance shows potential, but more fieldwork and modeling on seasonal variation in dung production rates is required.