1. Synergy of Landsat, climate and LiDAR data for aboveground biomass mapping in medium-stature tropical forests of the Yucatan Peninsula, Mexico
- Author
-
Gregorio Ángeles-Pérez, Alicia Peduzzi, Héctor M. de los Santos-Posadas, Laura Schneider, José René Valdez-Lazalde, Carlos Arturo Aguirre-Salado, and Alma Delia Ortiz-Reyes
- Subjects
Yucatan peninsula ,Biomass (ecology) ,Lidar ,Ecology ,Mean squared error ,Environmental science ,Forestry ,Spatial variability ,Ecosystem ,Explained variation ,Random forest - Abstract
Introduction: Tropical forests represent complex and dynamic ecosystems that cover extensive areas, hence the importance of determining biomass content and representing spatial variability. Objective: Estimating and mapping aboveground biomass and its associated uncertainty for medium-stature semi-evergreen (SMSP) and semi-deciduous (SMSC) tropical forests of the Yucatan Peninsula. Materials and methods: Aboveground biomass was estimated as a function of explanatory variables taken from Landsat images and climatic variables, using the random Forest algorithm. Aboveground biomass was mapped from previous biomass estimates for stripes of the territory with the presence of LiDAR (Light Detection And Ranging) and field data. Uncertainty at the pixel level was estimated as the coefficient of variation. Results and discussion: A combination of climatic and spectral variables showed acceptable capacity to estimate biomass in the medium-stature semi-evergreen and semi-deciduous tropical forest with an explained variance of 50 % and RMSE (root mean squared error) of 34.2 Mg·ha -1 and 26.2 Mg·ha -1 , respectively, prevailing climate variables. SMSP biomass ranged from 4.0 to 185.7 Mg·ha -1 and SMSC ranged from 11.7 to 117 Mg·ha -1 . The lowest values of uncertainty were recorded for the medium-stature semi-evergreen tropical forest, being higher in areas with lower amounts of aboveground biomass. Conclusion: Aboveground biomass was estimated and mapped by the combined use of auxiliary variables with an acceptable accuracy, against uncertainty of predictions, which represents an opportunity for future improvement.
- Published
- 2021