First open access ensemble climate envelope predictions of Assamese macaque Macaca assamensis in Asia: a new role model and assessment of endangered species

Abstract

Species distribution models are a key component for understanding a species' potential occurrence, specifically in vastly undersampled landscapes. The current species distribution data for the Assamese macaque Macaca assamensis are outdated, but suggest a patchy distribution in moist broadleaved forests in South and Southeast Asia. Therefore, in this study, we used a species distribution model to explore the potential climatic niche of this species and assess its distribution and potential barriers in 12 South and Southeast Asian countries. We combined primary and secondary species occurrence records from different countries. We applied Classification and Regression Tree (CART), TreeNet (boosting), RandomForest (bagging) and Maximum Entropy (MaxEnt) machine-learning algorithms with elevation as well as 19 bioclimatic variables for the first ensemble predictions ever completed for this species. Our results suggested that the predicted distribution of the Assamese macaque is strongly associated with precipitation of warmest quarter (BIO18), temperature annual range (BIO7) and temperature seasonality (BIO4). Our prediction shows a continuous potential climatic niche of the species from east of the Kaligandaki River in Nepal to Lao People's Democratic Republic. There are also potential niche patches in Bhutan, Southeast China, Thailand and Cambodia, while Pakistan and Afghanistan have no potential niche for the species. We believe that our workflow presents a new conservationoriented open access research template to progress empirical primate conservation worldwide.

Parapatric subspecies of Macaca assamensis show a marginal overlap in their predicted potential distribution: Some elaborations for modern conservation management

“…The two parapatric subspecies are separated by a described zoogeographical barrier (defined here as a physical obstacle that prevents migration of M. assamensis ); the Brahmaputra river in northeastern India (Fooden, ; Roos et al., ). The distribution of the two subspecies is fairly well‐known, but quantitative mapping and the characteristics of their niche and distribution are lacking (Regmi et al., ).…”

“…Climatic variables, however, should compensate for its absence. The high‐resolution topographic data were not aggregated to match the coarse climate data because we did not use the raster file but instead used a point‐based method where raster values were extracted to points and analyzed (Kandel et al., ; Regmi et al., ). All the data used in this study are open access, and the variables we prepared (ABT, EQ, slope, aspect) as well as occurrence data have been made open access via a university repository http://hdl.handle.net/1956/16960.…”

“…We used open-access species occurrence data from Regmi et al (2018;collection from 1998to 2013 and Fooden (1982;museum records from 1849 to 1980; n = 186 for M. assamensis ssp. pelops and n = 184 for M. assamensis ssp.…”

“…The values of all environmental variables were extracted at occurrence points, background data points, and lattice points (points that are arranged in a grid at 3 arc minutes distance in the study area, total = 177,938, on which current and future distributions were predicted). All the analyses were performed as a point-based analysis (using environmental values extracted at points instead of raster files; e.g., Kandel et al, 2015;Regmi et al, 2018).…”

“…However, there are some model-wise discrepancies in their predictions (Figure 7). These problems are often tackled by making an ensemble of multiple models (Araújo & New, 2007;Regmi et al, 2018), but we did not do this here directly. Instead, we used one of the best algorithms in SDM (Aguirre-Gutiérrez et al, 2013;Craig & Huettmann, 2009;Mi et al, 2017), and, due to bagging, Random Forest being an ensemble model (Breiman, 2001a).…”