Translate 
Abstract
Most previous works on aboveground biomass (AGB) estimation provide a single estimate of AGB rather than the probability distribution of the predicted values. However, the NGBoost algorithm provides a probabilistic regression and uncertainty estimation solution. In this study, we validate NGBoost for estimating AGB in mangrove forests in northeastern Vietnam. We use spectral bands and image indices extracted from WorldView-2 as independent variables and field data from eight plots as the basis for analysis. By applying a spatial scaling sampling strategy, we derived approximately 290 aggregated samples from the established plots, which served as the dependent variables in subsequent modeling. To augment the training dataset and capture a broader spatial context, window filters of varying sizes were applied, enabling the inclusion of adjacent pixels into the sampling matrix. NGBoost hyperparameters were optimized by the meta-heuristic Fox-inspired Optimization Algorithm using the Root Mean Square Error (RMSE) as the objective function. The trained model ended up at an RMSE of 1.8771, a Mean Absolute Error (MAE) of 1.2898, and an R2 of 0.924. We interpreted the trained model and found that the Green Leaf Index is the most influential factor in AGB estimation, far more than the following factors. Finally, we used the trained model to estimate AGB and its probability distribution for the entire study area.
Highlights
The estimation of carbon stock by land cover types is essential for effective land-use planning and implementing appropriate measures to control emissions, helping to minimize the greenhouse effect.Vol.: (0123456789) 43 Page 2 of 12Adding carbon to terrestrial ecosystems is considered the most effective method of reducing the level of CO2 in the atmosphere (Schimel et al 2001; Lettens et al 2005; Lal 2014)
FOX-inspired Optimization Algorithm (FOX-IA) searches for a set of hyper-parameters and injects these into Natural Gradient Boosting (NGBoost) for regression
aboveground biomass (AGB) is a crucial metric for mangrove forest management
Summary
Adding carbon to terrestrial ecosystems is considered the most effective method of reducing the level of CO2 in the atmosphere (Schimel et al 2001; Lettens et al 2005; Lal 2014). Combining optical and microwave datasets was found in various studies with Random Forest or ensemble algorithms (An et al 2019; Đạt et al 2020; Dũng & Kappas 2020; Li et al 2020; Nguyen & Nguyen 2021; Torre-Tojal et al 2022), models optimized by meta-heuristic algorithms (Bui et al 2024). In contrast to deterministic models that produce point estimates, probabilistic regression returns distributions over the outputs, which accounts for the variability in biomass estimates This is especially relevant concerning remote sensing-based AGB estimation, where uncertainties emerge from different spatial scales, environmental conditions, and sensor resolutions. In this regard, NGBoost can be a regression option to estimate the intervals around AGB predicted values.
Sign-in/Register to view highlights & summary 
Paper version not known (
Free)
Published Version
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
