Improving landscape connectivity in forest districts: A two-stage process for prioritizing agricultural patches for reforestation

Connectivity is a key concern in natural resource planning. Many studies have focused on the development of methods, tools and indices for the assessment of both components of connectivity: structural and functional. In particular, approaches based on graph theory principles have been recently proposed and are being increasingly applied to guide landscape connectivity conservation. However, forest planners and managers still need effective and operational methodologies to detect those landscapes where connectivity should be treated as a particularly critical conservation concern. In addition, in the Mediterranean, as in other parts of the world, socioeconomic changes in the last decades have driven the abandonment of many formerly cultivated lands. This poses both a challenge and an opportunity for managers intending to restore ecological connectivity in forested areas. In this context, setting adequate priorities for the reforestation of agricultural lands is of outmost importance. Here we show how a two-stage hierarchical methodology based on network analysis can be used to meet these needs. In particular, we apply a graph metric based on the measurement of habitat availability at the landscape scale (the Integral Index of Connectivity) to two Mediterranean forest districts in Spain with different management objectives and environmental heterogeneity. First, we identify those landscapes where efforts to improve forest connectivity should be concentrated. In a second stage, we prioritize within those landscapes the individual patches of agricultural lands that, being available for a potential reforestation program, would contribute most to uphold connectivity and ecological flows at wide spatial scales. We show how the extent of the agricultural patches is not strictly related to the contribution to connectivity they would provide if reforested, and how the results of such analysis vary with species traits (dispersal capabilities). We discuss the suitability of the proposed approach for forest landscape planning purposes and conclude that it can provide a useful diagnosis and helpful guidelines for the development of efficient reforestation programs that might be applied in a variety of situations for improving the ecological coherence of forest landscapes.