Use of datasets derived from time-series AVHRR imagery as surrogates for land cover maps in predicting species' distributions

Abstract

We hypothesized that NDVI time-series composite imagery or clustered data derived from the NDVI time series could serve as effective surrogates for land cover data in predictive modeling of species’ ecological niches and potential geographic distributions. Using two Mexican bird species, we examined our hypothesis with GARP, the Genetic Algorithm for Rule-set Prediction. Inputs included topographic and climate data, as well as the NDVI and clustered NDVI datasets. We used a land cover map previously derived from the NDVI dataset for comparison testing. Considering only topographic factors, we found that the NDVI or clustered NDVI data performed as well as or better than the land cover data. When climate data were added, the land cover data performed better than the NDVI data, but improvements were slight.