Abstract
According to the Paris Climate Change Agreement, all nations are required tosubmit reports on their greenhouse gas emissions and absorption every two yearsby 2024. Consequently, forests play a crucial role in reducing carbonemissions, which is essential for meeting these obligations. Recognizing thesignificance of forest conservation in the global battle against climatechange, Article 5 of the Paris Agreement emphasizes the need for high-qualityforest data. This study focuses on enhancing methods for mapping abovegroundbiomass in tropical dry forests. Tropical dry forests are considered one of theleast understood tropical forest environments; therefore, there is a need foraccurate approaches to estimate carbon pools. We employ a comparative analysisof AGB estimates, utilizing different discrete and full-waveform laser scanningdatasets in conjunction with Ordinary Least Squares and Bayesian approachesSVM. Airborne Laser Scanning, Unmanned Laser Scanning, and Space Laser Scanningwere used as independent variables for extracting forest metrics. Variableselection, SVM regression tuning, and cross-validation via a machine-learningapproach were applied to account for overfitting and underfitting. The resultsindicate that six key variables primarily related to tree height:Elev\.minimum, Elev\.L3, lev\.MAD\.mode, Elev\.mode, Elev\.MAD\.median, andElev\.skewness, are important for AGB estimation using ALSD and ULSD, whileLeaf Area Index, canopy coverage and height, terrain elevation, andfull-waveform signal energy emerged as the most vital variables. AGB valuesestimated from ten permanent tropical dry forest plots in Costa Rica Guanacasteprovince ranged from 26.02 Mg/ha to 175.43 Mg/ha. The SVM regressionsdemonstrated a 17.89 error across all laser scanning systems, with SLSF Wexhibiting the lowest error 17.07 in estimating total biomass per plot.