Predicting the seasonal dynamics of bird communities along an urban-rural gradient using NDVI

The normalized difference vegetation index (NDVI) is a strong and widely used proxy for bird species richness in urban environments. However, its potential to predict composition and seasonal dynamics of bird communities in urban areas remains unexplored. Our main objective was to analyze the power of NDVI to predict the seasonal dynamics of bird communities along urbanization gradients. Birds were surveyed in the breeding and non breeding seasons. Bird species richness was estimated using COMDYN, and bird community composition was summarized in two axes using Non-metric Multidimensional Scaling (NMDS). For each of the breeding and non breeding seasons, bird richness and the NMDS axis scores were related to both the mean NDVI of the season and the annual seasonal variation of NDVI (mean NDVI of the breeding season - mean NDVI of the non-breeding seasons). Results showed that NDVI decreased towards the most urbanized areas, with the highest seasonal variation occurring in agricultural areas. Single-family houses surrounded by yards held a high NDVI in both seasons and the lowest seasonal variation. Bird richness increased at intermediate levels of NDVI and was negatively related to the seasonal variations of the NDVI. Seasonal variation in bird community composition increased with annual mean and seasonal variation of the NDVI predicting higher seasonal stability of assemblage composition in urban than in non-urban areas. The results suggest that urbanization alters ecosystem functioning by reducing seasonal dynamics and the amount of primary productivity, which in turn promotes seasonal homogenization of urban avifauna relative to non-urban areas.