Optimal road design through ecologically sensitive areas considering animal migration dynamics

• Published in: Transportation research. Part C, Emerging technologies Emerging technologies, 2017-04, Vol.77, p.478-494
• Main Authors: Davey, Nicholas, Dunstall, Simon, Halgamuge, Saman
• Format: Article
• Language: English
• Subjects: Dimensionality reduction; Ecological constraints; Genetic algorithm; Road design; Spatially-explicit model; Surrogate model
• ISSN: 0968-090X; 1879-2359
• DOI: 10.1016/j.trc.2017.02.016

•Incorporation of an animal migration model into optimal road design is proposed.•Computational tractability is achieved via a surrogate model.•Problem dimensionality is reduced via a novel reduced-order parameter.•The model allows finding low cost roads that maintain a minimum animal population. With increasing land transportation requirements in both urban and rural areas, roads are encroaching ever more on animal habitats, where collisions with vehicles are a leading contributor to wildlife mortality. While road designers recognise the importance of accounting for such impacts at the design level, existing approaches simply either ignore viable habitat or avoid such regions entirely. Respectively, this can result in road alignments that are overly damaging to vulnerable species or prohibitively expensive to build and operate. The research presented in this paper investigates the effects of explicitly accounting for animal mortality on the design of a road through an ecologically sensitive area. The model presented achieves this by incorporating a spatially-explicit animal migration and road mortality model into an accepted optimal road alignment algorithm to propose low-cost roads that maintain the animal population above a minimum threshold by the end of a specified design horizon. The new method was applied to an example scenario to demonstrate the effect of setting a minimum required animal population on the road design. This model was able to consistently produce a road that met a minimum required species conservation benefit. This reflected a major improvement over the model that ignored animal habitats while only requiring a minor increase in construction and operating costs compared to the model that avoids habitat.