Skip to main content
Dryad logo

Data from: Satellite image texture for the assessment of tropical anuran communities

Citation

Sugai, Larissa Sayuri Moreira; Sugai, José Luiz Massao Moreira; Ferreira, Vanda Lucia; Silva, Thiago Sanna Freire (2019), Data from: Satellite image texture for the assessment of tropical anuran communities, Dryad, Dataset, https://doi.org/10.5061/dryad.3tq2v6m

Abstract

The relationship between environmental heterogeneity and biodiversity represents a cornerstone of ecological research. While environmental descriptors over large extents usually have medium to low spatial resolution, in-situ measures provide accurate information for limited areas, and a gap remains in providing remote descriptors that represent local environmental structure. Texture from satellite images can represent fine-scale heterogeneity over wide spatial coverage, but to date, it has mostly been used to predict general aspects of species diversity, such as richness. Here, we assess the utility of image textures from high resolution satellite images (RapidEye 3A) and in-situ variables to predict differences in the composition of anuran communities in a tropical savanna (Cerrado) of Brazil. While in-situ measures accounted for compositional differences of the whole community, two measures of image textures were associated only with the variation of species within the Hylidae family (adj. R² = 0.16 and 0.14). Comparatively, image textures predicted ~2/3 of the variation explained by in-situ­ measures (adj. R² = 0.23). When both approaches were combined, a greater compositional variation was achieved (adj. R² = 0.28), with 1/5 of it shared by both in-situ and textures, and 1/5 attributed solely to texture. Our findings suggest that image texture can complement the assessment of environmental heterogeneity acting on the assembly of local anuran communities. This approach can be valuable for explicitly including spatial heterogeneity in biological assessments over broad spatial extents, especially for biological groups strongly filtered by environmental conditions.

Usage Notes