RESUMO
To offset the declining timber supply by shifting towards more sustainable forestry practices, industrial tree plantations are expanding in tropical production forests. The conversion of natural forests to tree plantation is generally associated with loss of biodiversity and shifts towards more generalist and disturbance tolerant communities, but effects of mixed-landuse landscapes integrating natural and plantation forests remain little understood. Using camera traps, we surveyed the medium-to-large bodied terrestrial wildlife community across two mixed-landuse forest management areas in Sarawak, Malaysia Borneo which include areas dedicated to logging of natural forests and adjacent planted Acacia forests. We analyzed data from a 25-wildlife species community using a Bayesian community occupancy model to assess species richness and species-specific occurrence responses to Acacia plantations at a broad scale, and to remote-sensed local habitat conditions within the different forest landuse types. All species were estimated to occur in both landuse types, but species-level percent area occupied and predicted average local species richness were slightly higher in the natural forest management areas compared to licensed planted forest management areas. Similarly, occupancy-based species diversity profiles and defaunation indices for both a full community and only threatened and endemic species suggested the diversity and occurrence were slightly higher in the natural forest management areas. At the local scale, forest quality was the most prominent predictor of species occurrence. These associations with forest quality varied among species but were predominantly positive. Our results highlight the ability of a mixed-landuse landscape with small-scale Acacia plantations embedded in natural forests to retain terrestrial wildlife communities while providing an alternate source of timber. Nonetheless, there was a tendency towards reduced biodiversity in planted forests, which would likely be more pronounced in plantations that are larger or embedded in a less natural matrix.
RESUMO
The integration of Bayesian analysis into existing great ape survey methods could be used to generate precise and reliable population estimates of Bornean orang-utans. We used the Marked Nest Count (MNC) method to count new orang-utan nests at seven previously undocumented study sites in Sarawak, Malaysia. Our survey teams marked new nests on the first survey and revisited the plots on two more occasions; after about 21 and 42 days respectively. We used the N-mixture models to integrate suitability, abundance and detection models which account for zero inflation and imperfect detection for the analysis. The result was a combined estimate of 355 orang-utans with the 95% highest density interval (HDI) of 135 to 602 individuals. We visually inspected the posterior distributions of our parameters and compared precisions between study sites. We subsequently assess the strength or reliability of the generated estimates using identifiability tests. Only three out of the seven estimates had <35% overlap to indicate strong reliability. We discussed the limitations and advantages of our study design, and made recommendations to improve the sampling scheme. Over the course of this research, two of the study sites were gazetted as extensions to the Lanjak-Entimau Wildlife Sanctuary for orang-utan conservation.
