Abstract

In northern countries, moose vehicle collisions (MVCs) are often associated with the presence of salt pools near highways. Mitigation measures such as designing compensation salt pools further away from the highway and building fences along highways can be used. Fences are very efficient in reducing MVCs, but wildlife passages are required in order to increase highway permeability. Agent-based modelling (ABM) was used to study the effect on moose movement behaviour near highways of roadside salt pool removal and displacement and to estimate the required density of underpasses in fenced areas. ABM was applied to Highway 175 (Québec) where an extensive telemetry dataset on moose movement was available. The movement rules were based on cover and food quality in GIS forest polygons. Model moose had salt pool spatial memory (SPSM) and, in most cases, road avoidance (RA) behaviour, the opposing effect of which on the number of road crossings was investigated. Completely removing roadside salt pools with no compensation salt pools resulted in the highest highway crossing reductions (by 79%). A conceptual framework was also designed for investigating the movement of moose along fences using ABM. The current spacing of wildlife passages along Highway 175 is markedly larger than the recommended allometric spacing (i.e. based on home range size). The objective was to assess the impact of wildlife crossing distances on highway permeability. Because of the lack of telemetry data near fences, probabilistic and fuzzy-logic approaches were proposed to determine movement rules of the model moose. Once the model is implemented, it is anticipated that permeability will increase with decreasing distances between passages, and that a plateau may be observed once the allometric distance is reached.