Scale dependency in effectiveness, isolation, and social-ecological spillover of protected areas.

Protected areas are considered vital for the conservation of biodiversity. Given their central role in many conservation strategies, it is important to know whether they adequately protect biodiversity within their boundaries; whether they are becoming more isolated from other natural areas over time; and whether they play a role in facilitating or reducing land-cover change in their surroundings. We used matching methods and national and local analyses of land-cover change to evaluate the combined effectiveness (i.e., avoided natural-cover loss), isolation (i.e., changes in adjacent areas), and spillover effects (i.e., impacts on adjacent areas) of 19 national parks in South Africa from 2000 to 2009. All parks had either similar or lower rates of natural-cover loss than matched control samples. On a national level, mean net loss of natural cover and mean net gain of cultivation cover decreased with distance from park boundary, but there was considerable variation in trends around individual parks, providing evidence for both increased isolation and buffering of protected areas. Fourteen parks had significant positive spillover and reduced natural-cover loss in their surroundings, whereas five parks experienced elevated levels of natural-cover loss. Conclusions about social-ecological spillover effects from protected areas depended heavily on the measures of land-cover change used and the scale at which the results were aggregated. Our findings emphasize the need for high-resolution data when assessing spatially explicit phenomena such as land-cover change and challenge the usefulness of large-scale (coarse grain, broad extent) studies for understanding social-ecological dynamics around protected areas.