Influence of Covariates:
The null model that did not include any covariates for detection or
occupancy performed poorly and ranked the lowest (AIC=175.48). Model
performance improved after we included the covariates alone or in
combination according to our priori hypothesis. Summed AIC weight of
covariates from the most competitive models was highest for termite
followed by fruit, disturbance, tree cover and equal for EVI
Wi =0.12 and TRI (Figure 3). Average model specific
β-coefficient value from the top competitive models for termites, fruit,
disturbance, terrain ruggedness and vegetation productivity indicated
their positive influence on sloth bear occupancy whereas negative beta
coefficient for tree cover indicated its negative association with sloth
bear habitat occupancy (Figure 4) .