Rodents and satellites: Predicting mice abundance and distribution with Sentinel-2 data

ANDREO, V; FRANK OSEI; DIANA BRITO HOYOS; ALFRED STEIN; MARIANA BELGIU; PROVENSAL, MARÍA CECILIA

ECOLOGICAL INFORMATICS

ELSEVIER

Lugar: Bologna; Año: 2019 vol. 51 p. 157 - 167

1574-9541

Remote sensing data is widely used in numerous ecological applications. The Sentinel-2 satellites (S2 A and B),recently launched by the European Spatial Agency´s (ESA), provide at present the best revisit time, spatial andspectral resolution among the freely available remote sensing optical data. In this study, we explored the potentialof S2 enhanced spectral and spatial resolution to explain and predict mice abundances and distribution inborder habitats of agroecosystems. We compared the predictive ability of different vegetation and water indicesderived from S2 and Landsat 8 (L8) imagery. Our analyses revealed that the best predictor of mice abundancewas L8-derived Enhanced Vegetation Index (EVI). S2-based indices, however, outperformed those computedfrom L8 bands for indices estimated simultaneously to mice trappings and for mice distribution models.Furthermore, indices including S2 red-edge bands were the best predictors of the distribution of the two mostcommon rodent species in the ensemble. The findings of this study can be used as guidelines when selecting thesensors and vegetation variables to be included in more complex models aimed at predicting the distribution andrisk of various vector-borne diseases, and especially rodents in other agricultural landscapes.