Effects of spatial resolution imagery for aboveground biomass and species richness estimation in tropical dry forest.

Abstract:

Tropical Dry Forest (TDF) has high diversity levels and great potential for sequestering atmospheric carbon on its aboveground biomass (ABG). Accurate estimations of species richness and AGB are needed for effective conservation and forest management strategies. However, the accuracy attained depends on spatial resolution of imagery. We evaluated the effects of Landsat 8 (30m), RapidEye (5m) and Very High-Resolution imagery (VHR, 18cm) on the accuracy of estimation of aboveground biomass and species richness, in twenty plots (0.1 ha) of a TDF at the Yucatan Peninsula. We obtained spectral information, vegetation indices, and textures measurements that were used as a surrogate of plant productivity and habitat structure to explain species diversity and AGB of TDF. VHR imagery had a different level of processing because we wanted to obtain information about structure vegetation. We fitted regression models for each type of imagery. We found that the VHR images have better performance for estimating species richness (R²=0.83) and biomass (R²=0.92) compared with Landsat 8 OLI and the Rapideye data. Besides, Landsat showed a higher accuracy for ABG estimation (R²=0.62) than RapidEye (R²=0.51), but RapidEye (R²=0.73) has a better performance to estimate species richness compared to Landsat (R²=0.62). On the other hand, we found that texture measures of red 4 and IR 5, increases the accuracy of estimations for ABG and species richness. VHR imagery characterize very well the horizontal structure of canopies in forest of complex vegetation structure allowing an improvement of the accuracy of ABG and species richness estimations that can be used for forest management decisions. However, Landsat and RapidEye imagery allows precise and cost-effective estimations of species richness and AGB.