A COMPARISON OF REGRESSION-BASED ESTIMATES OF DUNG DECAY IN TWO AFRICAN FOREST DUIKER SPECIES (PHILANTOMBA MONTICOLA, CEPHALOPHUS OGILBYI)

摘要

Reliable dung decay rate estimates are essential for indirect surveys of forest ungulates, but there are still few reports of this parameter in the literature and there is a lack of consistency in their analytical assessment. In contrast to the standardized and statistically robust protocol for analyzing sign surveys of forest elephants, most duiker dung decay studies provide simple means only. We here report decay rates based on logistic regression on prospective data, collected at Korup National Park, Cameroon, based on 45 dung piles from Ogilbys duiker {Cephalophus ogilby) and on 23 dung piles from blue duiker (Philantomba monticola). Estimates were computed using maximum likelihood-based (R functions jglm' as fixed and 'lmer' as mixed model specification), and quasi-likelihood-based ('Irm' and 'glmmpql') logistic regressions. Ogilbys and blue duiker dung decay rates were estimated at 17 and 11 days on average, respectively. Decay rate estimates did not differ between mixed and standard models in blue duikers and only slightly in Ogilbys duiker. Our model-based decay rate estimates differed considerably from previously published dung decay rates and from arithmetic mean decay rates. Therefore we strongly suggest performing site- and time- specific dung decay experiments and utilizing logistic regression models to estimate dung decay to ensure robust duiker density estimations based on dung density.