Fig. 3
From: Sumatran tiger survival threatened by deforestation despite increasing densities in parks
Variation in tiger densities through time and between habitats. Results from mixed model meta-regressions explaining variation in log(\(\hat D\)*) showing 90% CI (thick lines) and 95% CI (thin lines). Top model (white fill) based on AICc values included habitat and degradation (secondary forest). Full model (black fill) including time and deforestation rates had marginally higher AICc values (+3.3). Malaysia estimates were not significantly different than Sumatra’s accounting for forest type and disturbance (LRT: χ1 = 0.67, P = 0.41) and the top Sumatra-only model also included forest type and disturbance (LRTS: χ2 = 7.93, P = 0.02, R2 = 0.50) and supported lower densities in peat and montane forests (−48.6%; MRMMS: z = −2.90, P = 0.01), and logged areas (−27.2%; MMRMS: z = 1.29, P = 0.21). Including deforestation and time significantly improved the Sumatra-only model fit (LRTS: χ2 = 8.65, P = 0.01, R2 = 0.77) and indicated densities increased from 2000 to 2012 by 47.6% (MMRMS: z = 2.04, P = 0.07). All models included a random effect to account for multiple estimates from the same landscape
