Estimation of aboveground biomass in a tropical dry forest: An intercomparison of airborne, unmanned, and space laser scanning

TL;DR

This study addresses the need for precise forest carbon data under the Paris Agreement by comparing airborne, unmanned, and spaceborne LiDAR approaches for estimating aboveground biomass (AGB) in a tropical dry forest. Using field-derived AGB and regression models (OLS and SVM) across three LiDAR platforms, the authors perform variable selection, cross-validation, and platform-specific modeling to quantify biomass with high accuracy. SVM consistently outperforms OLS, with full-waveform spaceborne LiDAR (SLS_FW) achieving the lowest errors, especially for mean biomass ($AGBm$). The work demonstrates that even in data-limited tropical forests, laser scanning combined with robust machine learning can provide reliable biomass estimates, supporting carbon accounting and REDD+ monitoring, and highlights the need for atmospheric corrections and platform-aware modeling. The results offer a practical framework for integrating LiDAR-derived metrics with ground inventories to improve forest carbon assessments at local to regional scales.

Abstract

According to the Paris Climate Change Agreement, all nations are required to submit reports on their greenhouse gas emissions and absorption every two years by 2024. Consequently, forests play a crucial role in reducing carbon emissions, which is essential for meeting these obligations. Recognizing the significance of forest conservation in the global battle against climate change, Article 5 of the Paris Agreement emphasizes the need for high-quality forest data. This study focuses on enhancing methods for mapping aboveground biomass in tropical dry forests. Tropical dry forests are considered one of the least understood tropical forest environments; therefore, there is a need for accurate approaches to estimate carbon pools. We employ a comparative analysis of AGB estimates, utilizing different discrete and full-waveform laser scanning datasets in conjunction with Ordinary Least Squares and Bayesian approaches SVM. Airborne Laser Scanning, Unmanned Laser Scanning, and Space Laser Scanning were used as independent variables for extracting forest metrics. Variable selection, SVM regression tuning, and cross-validation via a machine-learning approach were applied to account for overfitting and underfitting. The results indicate that six key variables primarily related to tree height: Elevminimum, ElevL3, levMADmode, Elevmode, ElevMADmedian, and Elevskewness, are important for AGB estimation using ALSD and ULSD, while Leaf Area Index, canopy coverage and height, terrain elevation, and full-waveform signal energy emerged as the most vital variables. AGB values estimated from ten permanent tropical dry forest plots in Costa Rica Guanacaste province ranged from 26.02 Mg/ha to 175.43 Mg/ha. The SVM regressions demonstrated a 17.89 error across all laser scanning systems, with SLSF W exhibiting the lowest error 17.07 in estimating total biomass per plot.

Figures (17)

• Figure 1: Locations of permanent research plots of the Santa Rosa National Park---Environmental Monitoring SuperSite (SRNP-EMSS). Ecological succession, measured as a function of time since abandonment, is divided into Early (E), Intermediate (I), and Late (L) forests. Source: Worldview-2 Satellite Image, Maxar Inc., 01/09/2019.
• Figure 2: Aerial photographs capturing the SRNP-EMSS tropical dry forest canopy, illustrating (a) early stage, (b) intermediate--stage, and (c) late--stage tropical dry forest. All photographs were taken at the SRNP-EMSS in May 2021, using a Hasselblad H4D-50 aerial camera.
• Figure 3: Comparison of point densities between the ULS$_{D}$ (left) and ALS$_{D}$ (right) laser scanning systems in plot E1. The ALS$_{D}$ data exhibit a lower point density per square meter than the ULS$_{D}$ data. ULS$_{D}$ data were collected in March 2021, while ALS$_{D}$ data were collected in May 2021.
• Figure 4: Workflow diagram illustrating the acquisition, post-processing, and analysis of LiDAR data from three laser scanning systems (ALS, ULS, and SLS) in the SRNP-EMSS plots. The process begins with data collection and georeferencing (data acquisition), followed by preparation, pre-processing, normalization, and extraction of forest metrics (post-processing). Then, forest inventory data are integrated, and various statistical and validation methods are used to estimate above-ground biomass (data analysis). Finally, a comparative analysis was conducted to assess the differences and advantages among the three systems (comparative analysis).
• Figure 5: Three-dimensional point cloud representation of the ALS$_{D}$ data and simulated SLS$_{FW}$ signal derived from the ALS$_{D}$ data for all plots.