Multiple habitat graphs: how connectivity brings forth landscape ecological processes

Abstract

Habitat connectivity is integral to current biodiversity science and conservation strategies. Originally, the connectivity concept stressed the role of individual movements for landscape-scale processes. Connectivity determines whether populations can survive in sub-optimal patches (i.e., source-sink effects), complete life cycles relying on different habitat types (i.e., landscape complementation), and benefit from supplementary resources distributed over the landscape (i.e., landscape supplementation). Although the past decades have witnessed major improvements in habitat connectivity modeling, most approaches have yet to consider the multiplicity of habitat types that a species can benefit from. Without doing so, connectivity analyses potentially fail to meet one of their fundamental purposes: revealing how complex individual movements lead to landscape-scale ecological processes. To bridge this conceptual and methodological gap, we propose to include multiple habitat types in spatial graph models of habitat connectivity, where nodes traditionally represent a single habitat type. Multiple habitat graphs will improve how we model connectivity and related landscape ecological processes, and how they are impacted by land cover changes. In three case studies, we use these graphs to model (i) source-sink effects, (ii) landscape supplementation, and (iii) complementation processes, in urban ecosystems, agricultural landscapes, and amphibian habitat networks, respectively. We show that multiple habitat graphs help addressing crucial conservation challenges (e.g., urban sprawl, biological control, climate change) by representing more accurately the dynamics of populations, communities, and their interactions. A new version of the Graphab open-source software implements the proposed approach, thereby extending the ecologist's toolbox and fostering the alignment between landscape ecology theory and practice.

How does dispersal shape the genetic patterns of animal populations in European cities? A simulation approach

“…Each forest or UGS habitat patch above 0.25 ha was a node of the graphs. Although a single type of habitat patch (node) is most often considered in graph-based connectivity analyzes, the special feature of our analyzes was the distinction between forests and UGS, considered as two distinct types of nodes in subsequent analyzes (Savary et al, 2024a).…”