Technical Abstract: Traditional and alternative energy development, logging and mining activities, together with off-highway vehicles (OHV) and exurban development, have increased the density of linear disturbances on public and private lands throughout the world. We argue that the dramatic increase in linear disturbances occurring globally has the potential to drastically alter landscape ecosystem processes, including soil and water conservation, and thus presents one of the greater challenges faced by natural resource scientists today -- a challenge we are poorly prepared to meet. Analytical tools and data are needed to systematically predict, assess, and minimize the impacts of these linear disturbances. The objective of this paper is to define the elements of analysis that can be used to systematically predict, assess, and minimize road impacts on ecosystem services across multiple spatial scales. The elements we outline are (1) identifying direct effects, (2) describing spatial interactions, (3) defining feedbacks among processes, (4) defining interacting effects of other stressors, (5) the extent to which these effects vary spatially (e.g. among units) and (6) thresholds and other non-linear responses. These elements can serve as the foundation for focusing monitoring efforts on those areas most likely to experience the greatest change in ecosystem processes and, similarly, target mitigation to areas with the greatest potential for recovery. The focus of this paper is on unimproved, unpaved roads in arid and semi-arid ecosystems; however, the general approach is relevant for most ecosystems and to all linear disturbances.