Table 3. Details of the model selection results of the N-mixture model for estimation of the relative abundance of sloth bears in Sanjay Tiger Reserve, Madhya Pradesh, India, during 2016–2017.
| Model description | nPars | AIC | ΔAIC | AICwt | Cumltv-wt |
|---|---|---|---|---|---|
| Poisson | |||||
| (p(fruit_density)λ(sal.forest+vildist+human_CR+mixed)) | 7 | 998.08 | 0.00 | 0.36 | 0.36 |
| (p(fruit_density)λ(sal.forest+vildist+human_CR+mixed+Agricultural.land)) | 8 | 999.52 | 1.44 | 0.17 | 0.53 |
| Null model (p(.)λ(.)) | 2 | 1,024.49 | 26.41 | 0.00 | 1.00 |
| NB | |||||
| (p(fruit_density)λ(sal.forest+vildist+human_CR+mixed)) | 8 | 983.00 | 0.00 | 0.11 | 0.11 |
| (p(fruit_density)λ(mixed +human_CR)) | 6 | 983.15 | 0.15 | 0.11 | 0.22 |
| (p(fruit_density)λ(mixed+human_CR+Agricultural. land)) | 7 | 983.43 | 0.43 | 0.09 | 0.31 |
| (p(fruit_density)λ(mixed+sal.forest+human_CR)) | 7 | 983.75 | 0.75 | 0.08 | 0.39 |
| (p(fruit_density)λ(mixed+vildist+human_CR)) | 7 | 983.81 | 0.81 | 0.08 | 0.47 |
| (p(fruit_density)λ(sal.forest+vildist+human_CR+mixed+Agricultural.land)) | 9 | 984.26 | 1.25 | 0.06 | 0.53 |
| (p(fruit_density)λ(mixed+Scrubland+human_CR)) | 7 | 984.60 | 1.59 | 0.05 | 0.58 |
| (p(fruit_density)λ(mixed+Agricultural.land)) | 6 | 984.71 | 1.70 | 0.05 | 0.63 |
| (p(fruit_density)λ(mixed+Water.body+human_CR)) | 7 | 984.96 | 1.96 | 0.04 | 0.67 |
| (p(fruit_density)λ(mixed+sal.forest)) | 6 | 984.99 | 1.99 | 0.04 | 0.71 |
| Null model (p(.)λ(.)) | 3 | 993.76 | 10.76 | 0.001 | 1.00 |
| ZIP | |||||
| (p(fruit_density)λ(sal.forest+vildist+human_CR+mixed)) | 8 | 995.09 | 0.00 | 0.36 | 0.36 |
| (p(fruit_density)λ(sal.forest+vildist+human_CR+mixed+Agricultural.land)) | 9 | 996.11 | 1.01 | 0.22 | 0.58 |
| Null model (p(.)λ(.)) | 3 | 1,022.10 | 27.00 | 0.00 | 1.00 |
Note:
Models include the most parsimonious models with the best-selected covariates and null models for the Poisson, negative binomial (NB) and zero-inflated Poisson (ZIP) distributions; Covariates considered: fruit_density- density of fruiting trees, mixed- area of mixed forest, sal.forest- area of sal forest, human_CR- photographic capture rate of humans, vildist–distance to the nearest village, Agricultural.land–area of agricultural land, Scrubland- area of scrubland and, Water.body–area of water body; Model selection was based on number of parameters (nPars), Akaike Information Criterion (AIC), the difference in AIC between best fit models (ΔAIC ≤ 2), AIC weight (AICwt) and cumulative AIC weight of models (Cumltv-wt); all models (including null models) are represented based on the lowest to the highest value of ΔAIC for each distribution.