ALTERNATIVE MEASURES OF ROAD NETWORK CHARACTERISTICS FOR MODELING ENVIRONMENTAL CHANGE

Investigations of ecosystem change in human-affected landscapes require the incorporation of parameters that estimate the extent and scale of human influence. Road networks provide clear evidence of human presence and potential use in the landscape. Road networks are significantly related to numerous changes in ecosystems, including land cover change, the fragmentation of animal populations, and the spread of invasive species. Not surprisingly, parameters that estimate road networks are commonly used as proxies for human influence in ecosystem change models. Typically, a few measures such as road density, distance-to-road, and occasionally road surface type and traffic are incorporated into models that include “road effects.” Few studies have considered alternative ways to measure and characterize road networks that relate to ways that people actually use the road network, or consider the spaces adjacent to roads in their estimation of the potential effects of the system of roads. This paper presents two alternative approaches for including parameters of road networks in models of landscape change. The first approach uses network analysis to abstract the road network as connected sets of weighted elements that quantitatively describe changes in structure and function. The second approach considers the spatial isolation of the road in relation to all other roads within a relevant neighborhood. Examples of road networks in north-central Florida and the Colorado Front Range were modeled using these techniques. In both areas, the road networks were measured in approximately twenty year intervals to record changes as the population in these regions grew and the transportation systems expanded. Changes in road network characteristics included changes in extent, connectivity and accessibility. An important change in the structure of the road networks included the proliferation of relatively short, local, dead-end roads, and an overall increase in their proportion within the networks. This characteristic was also associated with land use intensification at those points, as the creation of dead-end roads preceded land cover transformation in many places. This suggests that such topological changes to the network may provide leading indicators of change, including increases in road density or impervious surface area. Over time, spatial isolation decreased as spaces between the roads were invaded by small penetrating roads suggesting that the invulnerability of the landscape to human-induced ecological change decreased. Regardless of the methods used to characterize them, road networks are a fundamental and critical piece of information in analyzing and predicting the human dimensions of ecological change. One of the greatest challenges in using these data is the acquisition of accurate, timely geographic data of the locations and attributes of all roads, including small, local roads. While such maps may exist for one county or state, they may be non-existent or outdated for others. Given the relative importance of these data for predicting human-induced environmental changes, we suggest the adoption of comprehensive, national road mapping standards, which clarify the mapping and attribution of road networks, including local roads.