Understanding the patterns of abundance and richness of species is key when predicting how biodiversity will respond to environmental global change. The combination of terrestrial measurements with indices based on remote sensing is a targeted approach for such biodiversity assessments. We applied the Dynamic Habitat Indices (DHIs) based on MODIS time series of the years 2003 to 2014 to predict the richness of breeding birds across the conterminous United States. We calculated DHIs annually and summarized three components that we expect structure biodiversity: (1) cumulative productivity (DHI Cum), because sites with more available energy are generally more biodiverse, (2) minimum productivity (DHI Min), because sites with high minima are more biodiverse and (3) seasonality expressed as the coefficient of variation in productivity (DHI Var), because sites with less intra-annual variability are generally more biodiverse. We tested a range of MODIS vegetation phenology input data (EVI, NDVI, fPAR, LAI, and GPP) to investigate the relationship between bird species distribution and the DHIs. We used bird occurrence data from the North American Breeding Bird Survey for the same time period for correlation analysis and the development of statistical models.
Results/Conclusions
We found that the relationship between the DHIs and bird species richness was dependent on the breeding habitat affiliation (woodland, early successional/scrub and grassland or woodland), nest placement within a habitat (ground/low-nesting and mid-story/canopy-nesting species) and migratory behavior. Highest correlations were found with habitat affiliation guilds, such as grassland breeding species (R2~0.6, P<0.001 for the composite DHI; R2~0.5, P<0.001 for DHI Min) and woodland breeding species (R2~0.5, P<0.001 for the composite DHI; R2~0.5, P<0.001 for DHI Cum). Models were calculated annually, but showed the same trends over the whole analyzed period. The type of MODIS vegetation input data used did matter, with DHIs based on GPP showing the strongest relationship with bird species richness. The good relationship between the different components of the DHIs and the bird diversity data holds potential for similar biodiversity assessments for other taxa such as mammals or amphibians. Our approach and the resulting maps of bird biodiversity provide an effective tool for conservation planning at regional and broader scales.