Ecological Modelling 251 (2013) 279–287 Contents lists available at SciVerse ScienceDirect Ecological Modelling jo ur n al homep ag e: www.elsevier.com/locate/ecolmodel Forest patch connectivity diagnostics and prioritization using graph theory B.S. Shanthala Devi a,∗,1 , M.S.R. Murthy a,2 , Bijan Debnath a , C.S. Jha a a Forestry and Ecology Group, National Remote Sensing Centre, ISRO, Department of Space, Hyderabad 500 625, India a r t i c l e i n f o Article history: Received 30 June 2012 Received in revised form 5 December 2012 Accepted 18 December 2012 Available online 29 January 2013 Keywords: Fragmentation Connectivity indices Graph theory Patch prioritization a b s t r a c t 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 km 2 which is distributed over 10,807 km 2 . 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. © 2013 Elsevier B.V. All rights reserved. 1. Introduction Species dispersal and movement within the landscape is sig- nificant for biodiversity conservation. Heterogeneity of landscape makes biodiversity conservation planning an alarming challenge (Fahrig, 2007). Coherence of habitat patch network within the land- scape depends on patch characteristics and spatial configuration. Along with the location centered analysis, indices based on the spa- tial patterns are also required (Mas et al., 2012). Intermediate patch that exists between them creates prospective functional links for the biota. Changes in forest composition resulting from seasonal variability, habitat fragmentation due to anthropogenic impacts and climate change also affects the connectivity. Landscape man- agement must put together spatial and behavioral aspects so as to enhance the understanding of patch connectivity parameters. It is based on choosing different combinations of (a) patch size, (b) shape, (c) inter-intra patch distribution and (d) distance between ∗ Corresponding author at: Forestry and Ecology Group, National Remote Sensing Centre, ISRO, Department of Space, Hyderabad 500625, India. E-mail address: shanthaladevi100@gmail.com (B.S. Shanthala Devi). 1 Pursuing PhD in Jawaharlal Nehru Technological University, Department of Spa- tial Information Technology, Kukatpally, Hyderabad: 500085, India. 2 Current Address: Mountain Environment and Natural Resources Information System (MENRIS), International Center for Integrated Mountain Development (ICI- MOD), GPO Box 3226, Kathmandu, Nepal. habitat patches. Arrays of carefully preferred indices provide the most useful information for connectivity by characterizing different aspects of the positional importance of patches (Peng et al., 2010). Network-based models and metrics have been suggested to possess a convenient benefit to effort ratio for conservation prob- lems that require characterization of connectivity at relatively large scales (Calabrese and Fagan, 2004). Diverse modelling approaches have been used to determine the potentiality of the elements for habitat connectivity or network. Quantitative methods in land- scape ecology and increasing need of objective methods for mea- suring connectivity have stimulated the development of a wide set of connectivity-related indices (Saura and Pascual-Hortal, 2007b). Simple spatial metrics and software like Fragstats, connectance index, patch cohesion and buffer metrics are used for descriptive analysis, but has limitations with reference to decision making. Highest importance is assigned to key stepping-stone patches, since their loss divides the remnant habitat into two or more isolated components (Pascual-Hortal and Saura, 2006). However, since each connectivity metrics is based on different criteria, patch prioritiza- tion (as well as its scale robustness) is different for each analyzed metrics. According to Baranyi et al. (2011), most of the variability in the patch ranking system provided by the wide set of indices is captured by three different aspects: (a) amount of flux a patch is estimated to receive, (b) degree to which a patch is valuable to uphold the connectivity between other habitat areas different from itself, and (c) intrinsic patch attributes (e.g. habitat area or quality) 0304-3800/$ – see front matter © 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ecolmodel.2012.12.022