Predicting the impacts of edge effects in fragmented habitats: Laurance and Yensen’s core area model revisited

An accurate estimate of the impact of edge effects is essential for the conservation management of threatened species in isolated habitat fragments. In a seminal paper, Laurance and Yensen (1991) developed the first generalized predictive model estimating the amount of edge-affected area and remaining unmodified core area in isolated fragments. Despite their conclusion that the core-area model should be applicable to all fragments, regardless of size or shape, the model has never been validated beyond the limited range of parameters under which the model was constructed. Here, we conduct a sensitivity test of the core-area model accounting for ecologically-realistic variation in fragment Shape Index (SI), and collinearity among shape, area and depth of edge influence (DEI). Using well-resolved GIS data for the South Island of New Zealand, we find substantial core-area model underestimates for large fragments (>10,000 ha) with high SI values (>10.0), particularly at high DEI (>100 m). To overcome this we reformulate the core-area model with a more accurate shape-specific exponent function, and test our new model against Laurance and Yensen’s core-area model and a null geometric model using an independently-derived GIS dataset for the North Island, New Zealand. Our new model provides a significantly better fit to the data. We conclude that past predictive applications of the core-area model have underestimated true core area, and we recommend our simple re-formulated model for estimating the degree of population and community responses to habitat fragmentation under ecologically-realistic scenarios of high patch shape complexity and large spatial extent of edge influence.

► Accurate prediction of edge impact is essential for conservation management.
► We test limitations of the core-area model in highly-fragmented landscapes.
► We reformulate a more accurate generalized core-area model for complex landscapes.
► It accurately models ecologically-realistic scenarios of patch shape and edge extent.
► Our new model will simplify and improve predictions of core area and edge impact.