Very high resolution Earth Observation features for testing the direct and indirect effects of landscape structure on local habitat quality

• Published in: International journal of applied earth observation and geoinformation Vol. 34; pp. 96 - 102
• Main Authors: Mairota, Paola, Cafarelli, Barbara, Labadessa, Rocco, Lovergine, Francesco P., Tarantino, Cristina, Nagendra, Harini, Didham, Raphael K.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.02.2015
• Subjects: Context-dependence; Habitat quality; Landscape structure; Mixed effects models; Landscape structure; Habitat quality; Mixed effects models; Context-dependence
• ISSN: 1569-8432; 1872-826X
• DOI: 10.1016/j.jag.2014.07.003

•We use VHR-EO features as habitat quality proxies in hierarchical models.•Mixed-effects models are apt to model LPIs/VHR-EO features causal relationships.•Landscape pattern influence on habitat quality can be both direct and indirect.•Landscape pattern influence on habitat quality varies across spatial scales.•We corroborate the notion of the spatial context-dependence of habitat quality. Modelling the empirical relationships between habitat quality and species distribution patterns is the first step to understanding human impacts on biodiversity. It is important to build on this understanding to develop a broader conceptual appreciation of the influence of surrounding landscape structure on local habitat quality, across multiple spatial scales. Traditional models which report that ‘habitat amount’ in the landscape is sufficient to explain patterns of biodiversity, irrespective of habitat configuration or spatial variation in habitat quality at edges, implicitly treat each unit of habitat as interchangeable and ignore the high degree of interdependence between spatial components of land-use change. Here, we test the contrasting hypothesis, that local habitat units are not interchangeable in their habitat attributes, but are instead dependent on variation in surrounding habitat structure at both patch- and landscape levels. As the statistical approaches needed to implement such hierarchical causal models are observation-intensive, we utilise very high resolution (VHR) Earth Observation (EO) images to rapidly generate fine-grained measures of habitat patch internal heterogeneities over large spatial extents. We use linear mixed-effects models to test whether these remotely-sensed proxies for habitat quality were influenced by surrounding patch or landscape structure. The results demonstrate the significant influence of surrounding patch and landscape context on local habitat quality. They further indicate that such an influence can be direct, when a landscape variable alone influences the habitat structure variable, and/or indirect when the landscape and patch attributes have a conjoined effect on the response variable. We conclude that a substantial degree of interaction among spatial configuration effects is likely to be the norm in determining the ecological consequences of habitat fragmentation, thus corroborating the notion of the spatial context dependence of habitat quality.