作者: Jacopo Grilli , György Barabás , Stefano Allesina
DOI: 10.1371/JOURNAL.PCBI.1004251
关键词:
摘要: Habitat destruction and land use change are making the world in which natural populations live increasingly fragmented, often leading to local extinctions. Although might undergo extinction, a metapopulation may still be viable as long patches of suitable habitat connected by dispersal, so that empty can recolonized. Thus far, metapopulations models have either taken mean-field approach, or modeled empirically-based, realistic landscapes. Here we show an intermediate level complexity between these two extremes is consider random landscapes, randomly arranged area (or volume). Using methods borrowed from mathematics Random Geometric Graphs Euclidean Matrices, derive simple, analytic criterion for persistence fragmented Our results how density patches, variability their value, shape dispersal kernel, dimensionality landscape all contribute determining fate metapopulation. this framework, sufficient conditions population spatially localized, such confined clusters act source whole landscape. Finally, regular arrangement always detrimental persistence, compared patches. Given strong parallel contact processes, our also applicable disease spread on spatial networks.