IENE 2016

Transportation networks cover entire landscapes, and their impacts extend from the scale of individual plants and animals and their habitat to ecosystems. The issue of scale is particularly important to consider when planning new projects meant to expand or improve existing road networks. Yet, assessment of ecological impacts across the range of scales at which road effects are manifested remains a major challenge to road planners and ecologists alike. Environmental impact assessments, although an integral part of road projects, typically remain limited to localized effect, i.e. effects at the site of construction and immediate surrounding area. There is a particular need for better tools and methods for assessing ecological impacts at large scales, e.g. in connection with strategic planning and concept evaluations.

At the network-level, roads fragment the landscape, essentially creating a mosaic of more or less isolated habitat tiles. These tiles (“road islands”) and their interactions are intuitive units for analysis and prediction of future impacts, as they capture local effects and at the same time are themselves constituents of the larger network. We develop a spatially-explicit predictive model that allows the user to evaluate and compare the environmental impacts of alternative road projects prior to implementation, with a primary focus on biodiversity effects. The underlying rational for our approach is that biodiversity within a tile in the road network is related to known or remotely measurable attributes of that tile, such as habitat composition, tile size and shape, surrounding road type and traffic volume, and other human disturbance factors. Road construction and other changes to the network alter the configuration of one or more tiles in the mosaic, potentially leading to predictable effects on biodiversity and ecosystem services. We demonstrate the functionality of the approach by applying it to Norwegian road network and biodiversity data and estimating the effect of real and simulated network manipulations. Finally, we discuss the potential of the model as a national planning tool that enables assessment and comparison of the environmental cost of of infrastructure development at all relevant spatial scales.