global, remote sensing‐based characterization of terrestrial habitat heterogeneity for biodiversity and ecosystem modelling

• Published in: Global ecology and biogeography Vol. 24; no. 11; pp. 1329 - 1339
• Main Authors: Tuanmu, Mao‐Ning, Jetz, Walter
• Format: Journal Article
• Language: English
• Published: Oxford Blackwell Science 01-11-2015; Blackwell Publishing Ltd; John Wiley & Sons Ltd; Wiley Subscription Services, Inc
• Subjects: Biodiversity; Birds; ecosystems; Environmental heterogeneity; Habitats; heterogeneity-diversity relationship; image texture; land cover; landscapes; moderate resolution imaging spectroradiometer; MODIS; monitoring; prediction; Remote sensing; species diversity; species richness; surveys; topography; vegetation index; United States; USA
• ISSN: 1466-822X; 1466-8238
• DOI: 10.1111/geb.12365

AIM: Habitat heterogeneity has long been recognized as a key landscape characteristic determining biodiversity patterns. However, a lack of standardized, large‐scale, high‐resolution and temporally updatable heterogeneity information based on direct observations has limited our understanding of this connection and its effective use for biodiversity conservation. To address this, we develop here remote sensing‐based metrics to characterize global habitat heterogeneity at 1‐km resolution and assess their value for biodiversity modelling. LOCATION: Global. METHODS: We develop 14 heterogeneity metrics (available at http://www.earthenv.org) based on the textural features of the enhanced vegetation index (EVI) imagery from the Moderate Resolution Imaging Spectroradiometer (MODIS), and closely examine a complementary core set of six of these metrics. We evaluate their ability to provide fine‐grain habitat heterogeneity by comparing the heterogeneity information captured by them with that measured by 30‐m Landsat‐based land‐cover data. Using spatial autoregressive models, we then compare their utility with that of more conventional metrics (derived from topography or categorical land‐cover data) for modelling the species richness of bird communities across the conterminous United States based on Breeding Bird Survey data. RESULTS: The newly derived metrics capture different aspects of habitat heterogeneity and provide fine‐grain information for locations deemed homogeneous by traditional land‐cover classifications at both continental and global extents. Most of them strongly exceed conventional heterogeneity variables in capturing the spatial variation in bird species richness, with Homogeneity emerging as the strongest predictor. MAIN CONCLUSIONS: This study develops and validates the performance of readily usable metrics of textural measures capturing fine‐grain habitat heterogeneity. The presented metrics outperform conventional measures in capturing detailed spatial variation in habitats and in predicting key biodiversity patterns. They provide a rigorous and comparable basis for understanding heterogeneity–diversity relationships, and offer a powerful tool for monitoring and understanding the responses of biodiversity and ecosystems to the changing environment.