Bias files
The ‘presence only’ SDMs are affected by bias sampling, to overcome this one of the common methods employed is the use of a target group bias file to restrict background sampling area (Phillips et al. 2009; Syfert et al. 2013; Vollering et al. 2019). We pooled all occurrence coordinates of the fish species and adult odonate species with 2 or more occurrence coordinates to produce target group bias files for both fish and odonate species respectively. However, we excluded Epiophlebia laidlawi from the odonate target group since its distribution in Bhutan is known only from its larval distribution while all other odonate species are mainly known from their adult occurrence (Appendix 1). Therefore, we developed a separate bias file for the E. laidlawiusing all the sampling sites from Dorji (2015) along with occurrence coordinates from Brockhaus and Hartmann (2008) and Dupchu, S (Personal communication). We used Hydrosheds (hybas_as_lev12_v1c available inwww.hydrosheds.org, Lehner and Grill, 2013) clipped to the Bhutan boundary to bound the occurrence instead of regular sized grid cells of the environmental variables (Phillips et al. 2009; Syfert et al. 2013) to develop sampling bias grids.