Forest patch connectivity diagnostics and prioritization using graph theory

Landscape level forest connectivity regulates species level biodiversity, wildlife movement, seed dispersal and ecological factors. Geospatial assessment of forest connectivity at the landscape level is realized as one of the important frameworks to prioritize the biodiversity conservation strategies. The paper presents an approach to identify the optimal threshold distance and set of forest patches (component) using graph theory to propose potential connectivity alternatives over fragmented tropical deciduous forest tracts in parts of Eastern Ghats of India. The study analyzed 598 forest patches (derived from remote sensing data) constituting an area of 3502.87 km2 which is distributed over 10,807 km2. Optimal threshold distance and components were derived using graph theory based connectivity indices namely, Integral Index of Connectivity importance value (dIIC), Landscape Coincidence Probability importance value (dLCP) and also addressing patch size, number, inter distance and relative importance of each patch in the totality of the landscape. The study identified a total of 191 components at an optimal threshold distance of 250 m for potential connectivity. A component consisting of 145 patches of which 5 patches having high relative importance value within the component were identified for prioritizing the connectivity efforts. Further analysis on the three fractions of dIIC (dIICintra, dIICflux and dIICconnector) justified the importance of individual patches for connectivity. Proposed connectivity could enhance the forest habitat network through these potential patches.

► We evaluated patch size, number, inter distance, and connectivity indices.
► Prioritized action for forest connectivity of the deciduous forest landscape is analyzed.
► 191 components at optimal distance of 250 m using graph theory is obtained.
► Five patches within the component were identified for prioritization.
► Proposed connectivity could enhance the forest habitat network through these potential patches.