Uncovering the structures in ecological networks: Multiple techniques for multiple purposes

摘要: Landscape ecologists have increasingly turned to the use of landscape graphs in which a is represented as set nodes (habitat patches) connected by links representing inter-patch-dispersal. This study explores novel regionalization method, GraphRECAP (Graph-based REgionalization with Clustering And Partitioning), detect structural groups habitat patches (compartments) graph such that connections (i.e. movement individual organisms) within are greater than those across groups. Specifically, we mapped compartments using and dispersal data for ring-tailed lemurs (Lemur catta) an agricultural southern Madagascar both widely-used Girvan Newman method. Model performance was evaluated comparing compartment characteristics three measures network connectivity traversability: connection strength (modularity), potential ease organism movements (Harary Index), degree alternative route presence (Alpha Index). Compartments identified had stronger within1 The manuscript view from Ecology 20 connections, traversability, more routes, larger minimum number compartments, all desirable traits ecological networks. Our method could thus facilitate ecosystem resilience design nature reserves networks promote landscape-scale species fragmented habitats. 2.