Towards quantifying tropical tree species richness in tropical forests

This letter reports tests of whether the normalized difference vegetation index (NDVI) from Landsat Enhanced Thematic Mapper Plus (ETM+) and radar backscatter (C‐, L‐ and P‐bands) from Airborne Synthetic Aperture Radar (AIRSAR) imagery can be used to estimate tree species richness from 25 1‐ha plots within continuous lowland forest types (dry, moist, wet) in the Panama Canal Zone Watershed. Species richness for trees ranged from 49 to 98 species per ha. There was no evidence that density, basal area, above‐ground biomass or P‐backscatter was correlated with tree species richness. Mean and standard deviations (SDs) of the NDVI could explain 39% and 37% of the variance in tree species richness, respectively (41% when combined), while mean radar backscatter in the shorter wavelength C‐ and L‐bands could explain 44% and 33% of the variance in tree species richness, respectively (50% when combined).