IENE 2016

Landscape fragmentation due to road infrastructure is very high in Flanders (= northern region of Belgium). Within the framework of environmental, climate adaptation, spatial and mobility policy plans, the Flemish government has the objective to mitigate the ecological impact of this infrastructure.  A first priority-atlas for defragmentation in Flanders was made in 2001, using policy maps and species vulnerability maps. Because of the complex and static characteristic of this atlas, a quick and easy update with new information was impossible. Therefore, we decided to make a model based and dynamic defragmentation tool, which can automatically evaluate the current defragmentation and suggest the best locations for new defragmentation measures (with priority) including various spatial scenarios. The defragmentation tool is based on a constrained cellular automata land use model of Flanders region. This model enables calculating and assessing the effectiveness of various spatial scenarios in Flanders, and has been used in many government studies, including spatial policy plans, nature reporting, and as a base for the implementation of Natura 2000 objectives. The model operates on a 100m resolution. It incorporates numerous GIS data layers and computes land use dynamics based on spatial interaction rules among some 50 land uses. The ecological input data layers in the defragmentation tool were (1) suitability maps for about 20 species and (2) the current and already planned defragmentation measures (e.g. fauna passages) with species specific quality values for ‘permeability’ of the road. The species suitability maps were made with dynamic scripting, using quality requirements for habitat type, landscape properties, slope, soil type, action range and minimum area for viable populations. We also incorporated an extra valuation in these suitability maps, specifically with land use properties and data of actual (known) species distribution including road-kills. The combination of all the ecological data layers and the road transport infrastructure map, resulted in clusters of possible ‘migration habitat’ and/or ‘optimal habitat’ for the species, separated by roads. A value of the number of optimal habitat raster cells within each cluster was automatically calculated for each species or group of species. Assuming that one or several defragmentation measures can (re-)connect two clusters, the result was a new value of optimal habitat raster cells in the combined larger cluster. With this method, the model performed an iterative calculation to find the qualitatively best locations for future defragmentation measures, for individual species and groups of species. The results were then presented in a priority map for Flanders. The tool can automatically recalculate the result with new input layers and for different scenarios. Optionally, policy maps (e.g. protected areas) can be included afterwards. The application of this dynamic tool can also be used in other regions around the world.