Bridging the gap: optimising connectivity solutions for an arboreal gliding mammal.
Abstract
Connectivity modelling tools are important for developing mitigation strategies to alleviate negative impacts on animal movement caused by road networks. Arboreal gliding mammals are especially vulnerable to road widening as they are unable to cross gaps beyond their gliding capacity. However, there are limitations of conventional raster-based modelling techniques when applied on this group. We developed and applied a new model that quantifies the changes in connectivity for an arboreal gliding species, the Sunda colugo, in Singapore using information on species-specific glide performance through a vector-based approach. We also incorporated a genetic algorithm to determine optimal locations for sets of glide poles that could be installed to improve connectivity. Expectedly, connectivity was heavily impacted by the road widening works, with a 93.3% decrease in the total number of feasible glide paths connecting roadside trees. Nine glide poles installed during the construction phase of the project were initially uninformed by the connectivity model and they provided 26 additional connections only. Comparatively, the top nine pole locations identified through the genetic algorithm provided 247 additional connections, almost 10 times more than the amount supplied by the initial nine glide pole locations determined through qualitative methods. The multi-objective capacity of the genetic algorithm also reduced large connectivity gaps post-development, with simulated glide poles increasing the percentage of pixels covered with connectivity from 16.1% to 27.7%. Synthesis and applications. Our model fills a knowledge gap in connectivity modelling for arboreal gliding mammals whose movements are affected by varying habitat alterations. We demonstrated that locations of mitigation structures greatly influence the success of mitigation efforts and the locations can be optimised by incorporating the glide performance of the species into the genetic algorithm. Given limited conservation resources, this approach would benefit managers in formulating cost-effective efforts for arboreal gliding mammals.