Evaluation of NDVI to assess avian abundance and richness along the upper San Pedro River

Remote-sensing models have become increasingly popular for identifying, characterizing, monitoring, and predicting avian habitat but have largely focused on single bird species. The Normalized Difference Vegetation Index (NDVI) has been shown to positively correlate with avian abundance and richness and has been successfully applied to southwestern riparian systems which are uniquely composed of narrow bands of vegetation in an otherwise dry landscape. Desert riparian ecosystems are important breeding and stopover sites for many bird species but have been degraded due to altered hydrology and land management practices. Here we investigated the use of NDVI, coupled with vegetation, to model the avian community structure along the San Pedro River, Arizona. We also investigated how vegetation and physical features measured locally compared to those data that can be gathered through remote-sensing. We found that NDVI has statistically significant relationships with both avian abundance and species richness, although is better applied at the individual species level. However, the amount of variation explained by even our best models was quite low, suggesting that NDVI habitat models may not presently be an accurate tool for extensive modeling of avian communities. We suggest additional studies in other watersheds to increase our understanding of these bird/NDVI relationships.