Forest Biomass Estimation Research Articles

Deep point cloud regression for above-ground forest biomass estimation from airborne LiDAR

Quantifying forest biomass stocks and their dynamics is important for implementing effective climate change mitigation measures by aiding local forest management, studying processes driving af-, re-, and deforestation, and improving the accuracy of carbon accounting. Owing to the 3-dimensional nature of forest structure, remote sensing using airborne LiDAR can be used to perform these measurements of vegetation structure at large scale. Harnessing the full dimensionality of the data, we present deep learning systems predicting wood volume and above ground biomass (AGB) directly from the full LiDAR point cloud and compare results to state-of-the-art approaches operating on basic statistics of the point clouds. For this purpose, we devise different neural network architectures for point cloud regression and evaluate them on remote sensing data of areas for which AGB estimates have been obtained from field measurements in the Danish national forest inventory. Our adaptation of Minkowski convolutional neural networks for regression give the best results. The deep neural networks produce significantly more accurate wood volume, AGB, and carbon stock estimates compared to state-of-the-art approaches. In contrast to other methods, the proposed deep learning approach does not require a digital terrain model and is robust to artifacts along the boundaries of the evaluated areas, which we demonstrate for the case where trees protrude into the area from the outside. We expect this finding to have a strong impact on LiDAR-based analyses of biomass dynamics.

Developing crown width model for mixed forests using soil, climate and stand factors

Abstract The tree crown is a useful measure of tree vigour and is highly relevant to a tree's environmental adaptability. Crown allometry depends on environmental and stand conditions. Several studies have focussed on the effects of climate change and competitive intensity on the crown, but the regulatory role of soil resources and diversity on crown allometry and carbon allocation has been neglected. Data from 20,994 trees in 232 mixed forests collected between 2011 and 2019 were located near four major mountain ranges in northeast China. The proposed crown width model includes the stand developmental stage, soil, climate, competition intensity, species mixture, species diversity, structural diversity and their interactions. We observed that the cross‐species allometric scaling exponent does not conform to the universal scaling law. Our results showed that crown width increased with increasing soil bulk density, quadratic mean diameter and coefficient of diameter variation but decreased with increasing de Martonne aridity index, basal area, Simpson index and species mixture. The interaction between quadratic mean diameter and soil bulk density had a significant negative effect on crown width. The influence of a particular factor within the interaction term on crown width was modulated by the gradients of other factors. Furthermore, soil bulk density contributed more to crown width modelling than the aridity index, and structural diversity had a greater effect on crown width than species diversity. Synthesis. Our results provide new insights into the environmental variability of crown allometry in mixed forests under global change, which is critical for improving regional and global estimates of forest biomass and carbon stocks.

Collaborative Estimation of Aboveground Forest Biomass Using P-Band and X-Band Interferometric Synthetic Aperture Radar Based on Feature Optimization

Collaborative Estimation of Aboveground Forest Biomass Using P-Band and X-Band Interferometric Synthetic Aperture Radar Based on Feature Optimization

A regional-level spatiotemporal perspective of land use and land cover change impact on forest aboveground biomass in three gorges reservoir region, China

This study used multi-platform remote sensing datasets to assess the impacts of land use and land cover (LULC) change on forest biomass density across the Three Gorges Reservoir (TGR) region along the Yangtze River. This study maps land use and land cover (LULC) change and forest aboveground biomass estimation (AGB) using Sentinel-2A and GEDI-L4B data based on observations between 2019 and 2022. The data analysis results of monitoring revealed a strong association frequency between land-use change and forest land cover since the significant relationship with forest aboveground biomass density has also been explored. The total deforestation decreased significantly in 2022. The accuracy assessment results show a classification accuracy of between 90.4% and 94.1%, with a kappa coefficient of 82.4%–92.1%. For the TGR region forest, the minimum to maximum mean above-ground biomass values are from 15.81 to 373.66 Mgha-1, respectively. Moreover, the fitting performance of the statistical analysis of aboveground biomass density of forest mean for the forest was quite decent with training (R 2: 0.92, RMSE: 12.30 Mg/ha) and testing (R 2: 0.79, RMSE: 18.07 Mg/ha) datasets. The results highlight LULC change impacts, particularly forest land class over forest AGB stands, and the effectiveness of the GEDI-L4B product in regional biomass estimation in the TGR region.

Estimating forest extent across Mexico

Information on forest extent and tree cover is required to evaluate the status of natural resources, conservation practices, and environmental policies. The challenge is that different forest definitions, remote sensing-based (RSB) products, and data availability can lead to discrepancies in reporting total forest area. Consequently, errors in forest extent can be propagated into forest biomass and carbon estimates. Here, we present a simple approach to compare forest extent estimates from seven regional and global land or tree cover RSB products at 30 m resolution across Mexico. We found substantial differences in forest extent estimates for Mexico, ranging from 387 607 km2 to 675 239 km2. These differences were dependent on the RSB product and forest definition used. Next, we compared these RSB products with two independent forest inventory datasets at national (n = 26 220 plots) and local scales (n = 754 plots). The greatest accuracy among RSB products and forest inventory data was within the tropical moist forest (range 82%–95%), and the smallest was within the subtropical desert (range <10%–80%) and subtropical steppe ecological zones (range <10%–60%). We developed a forest extent agreement map by combining seven RSB products and identifying a consensus in their estimates. We found a forest area of 288 749 km2 with high forest extent agreement, and 340 661 km2 with medium forest extent agreement. The high-to-medium forest extent agreement of 629 410 km2 is comparable to the official national estimate of 656 920 km2. We found a high forest extent agreement across the Yucatan Peninsula and mountain areas in the Sierra Madre Oriental and Sierra Madre Occidental. The tropical dry forest and subtropical mountain system represent the two ecological zones with the highest areas of disagreement among RSB products. These findings show discrepancies in forest extent estimates across ecological zones in Mexico, where additional ground data and research are needed. Dataset available at https://doi.org/10.3334/ORNLDAAC/2320.

A Compatible Estimation Method for Biomass Factors Based on Allometric Relationship: A Case Study on Pinus densata Natural Forest in Yunnan Province of Southwest China

Using various biomass factors, such as biomass expansion factor (BEF) and biomass conversion and expansion factor (BCEF), yields different results for estimating forest biomass. Therefore, ensuring compatibility between total biomass and its components when employing different biomass factors is crucial for developing a set of rapid and efficient models for large-scale biomass calculation. In this study, allometric equations were utilized to construct independent models and the proportional values (root-to-shoot ratio (Rra), crown-to-stem ratio (Rcs), bark-to-wood ratio (Rbw), foliage-to-bark ratio (Rfb), and wood biomass-to-wood volume (ρ)) by using the mean height (Hm) and the mean diameter at breast height (Dg) of 98 Pinus densata plots in Shangri-La, Yunnan province, China. The compatible methods were applied to reveal the compatibility between the total biomass and each component’s biomass. The results showed the following: (1) Both the independent model and compatible model had a higher accuracy. The values were greater than 0.7 overall, but the foliage biomass accuracy was only 0.2. The total biomass and the component biomass showed compatibility. (2) The accuracy of BEF and BCEF exceeded 0.87 and the total error was less than 0.1 for most components. (3) The mean BEF (1.6) was greater than that of the Intergovernmental Panel on Climate Change (IPCC) (M = 1.3), and the mean BCEF was smaller than that of the IPCC; the values were 0.6 and 0.7, respectively. The range of BEF (1.4–2.1) and BCEF (0.44–0.89) were all within the range of the IPCC (1.15–3.2, 0.4–1.0). This study provides a more convenient and accurate method for calculating conversion coefficients (BEF and BCEF), especially when only Rcs data is available.

Estimation of Forest Biomass and Absorbed CO2 by Remote Sensing in Can Gio, Vietnam

Estimation of Forest Biomass and Absorbed CO2 by Remote Sensing in Can Gio, Vietnam

Shifts in internal stem damage along a tropical precipitation gradient and implications for forest biomass estimation.

Woody biomass is a large carbon store in terrestrial ecosystems. In calculating biomass, tree stems are assumed to be solid structures. However, decomposer agents such as microbes and insects target stem heartwood, causing internal wood decay which is poorly quantified. We investigated internal stem damage across five sites in tropical Australia along a precipitation gradient. We estimated the amount of internal aboveground biomass damaged in living trees and measured four potential stem damage predictors: wood density, stem diameter, annual precipitation, and termite pressure (measured as termite damage in downed deadwood). Stem damage increased with increasing diameter, wood density, and termite pressure and decreased with increasing precipitation. High wood density stems sustained less damage in wet sites and more damage in dry sites, likely a result of shifting decomposer communities and their differing responses to changes in tree species and wood traits across sites. Incorporating stem damage reduced aboveground biomass estimates by > 30% in Australian savannas, compared to only 3% in rainforests. Accurate estimates of carbon storage across woody plant communities are critical for understanding the global carbon budget. Future biomass estimates should consider stem damage in concert with the effects of changes in decomposer communities and abiotic conditions.

Aboveground Forest Biomass Estimation Using Tent Mapping Atom Search Optimized Backpropagation Neural Network with Landsat 8 and Sentinel-1A Data

Accurate forest biomass estimation serves as the foundation of forest management and holds critical significance for a comprehensive understanding of forest carbon storage and balance. This study aimed to integrate Landsat 8 OLI and Sentinel-1A SAR satellite image data and selected a portion of the Shanxia Experimental Forest in Jiangxi Province as the study area to establish a biomass estimation model by screening influencing factors. Firstly, we extracted spectral information, vegetation indices, principal component features, and texture features within 3 × 3-pixel neighborhoods from Landsat 8 OLI. Moreover, we incorporated Sentinel-1’s VV (vertical transmit–vertical receive) and VH (vertical transmit–horizontal receive) polarizations. We proposed an ensemble AGB (aboveground biomass) model based on a neural network. In addition to the neural network model, namely the tent mapping atom search optimized BP neural network (Tent_ASO_BP) model, partial least squares regression (PLSR), support vector machine (SVR), and random forest (RF) regression prediction techniques were also employed to establish the relationship between multisource remote sensing data and forest biomass. Optical variables (Landsat 8 OLI), SAR variables (Sentinel-1A), and their combinations were input into the four prediction models. The results indicate that Tent_ ASO_ BP model can better estimate forest biomass. Compared to pure optical or single microwave data, the Tent_ASO_BP model with the optimal combination of optical and microwave input features achieved the highest accuracy. Its R2 was 0.74, root mean square error (RMSE) was 11.54 Mg/ha, and mean absolute error (MAE) was 9.06 Mg/ha. Following this, the RF model (R2 = 0.54, RMSE = 21.33 Mg/ha, MAE = 17.35 Mg/ha), SVR (R2 = 0.52, RMSE = 17.66 Mg/ha, MAE = 15.11 Mg/ha), and PLSR (R2 = 0.50, RMSE = 16.52 Mg/ha, MAE = 12.15 Mg/ha) models were employed. In conclusion, the BP neural network model improved by tent mapping atom search optimization algorithm significantly enhanced the accuracy of AGB estimation in biomass studies. This will provide a new avenue for large-scale forest resource surveys.

Improving forest above-ground biomass estimation using genetic-based feature selection from Sentinel-1 and Sentinel-2 data (case study of the Noor forest area in Iran)

Biomass holds great importance in the environment, as it not only allows us to measure the carbon stored in forests but also facilitates the assessment of biodiversity and the evaluation of ecological integrity within these crucial ecosystems. In this study, we employed a Genetic Algorithm (GA) to estimate forest Above-Ground Biomass (AGB) by selecting the most applicable features from both Sentinel-2 optical and Sentinel-1 Synthetic Aperture Radar (SAR) images in the Noor forest. The study area was divided into four distinct regions (north, near north, middle, and south), and each region was documented with 100 sample plots through fieldwork to enable comprehensive analysis. In our workflow, Sentinel-2-derived features (i.e., spectral bands, vegetation indices (VIs), soil indices (SIs), and water indices (WIs), along with Sentinel-1 SAR features were initially extracted. Subsequently, GA was employed to select the most optimal features among them within both Random Forest (RF) and Multiple Linear Regression (MLR) models, leading to enhanced accuracy in the forest AGB estimation process. The experimental results demonstrated that the RF model outperformed the MLR model in estimating forest AGB. Furthermore, incorporating GA-based feature selection substantially improved the accuracy of both models, resulting in more dependable AGB estimations. The selected features from the combined Sentinel-1 and Sentinel-2 data also provided the best AGB estimation, surpassing the individual use of each dataset. The selected features from Sentinel-2 particularly played a more substantial role in achieving this overall enhanced performance in AGB estimation. The AGB estimates based on GA-RF were more accurate in all cases, with an average coefficient of determination (R2) of 0.5 and average RMSE of 13.17 Mg ha−1, while the MLR-based estimates were less accurate, with an average R2 value lower than 0.3 and average RMSE higher than 16 Mg ha−1. Furthermore, the GA-RF model selected a wider variety of features including spectral bands, indices, and SAR features compared to GA-MLR, resulting in accurate AGB estimation in the Noor forest.

Forest Height Inversion by Combining Single-Baseline TanDEM-X InSAR Data with External DTM Data

Forest canopy height estimation is essential for forest management and biomass estimation. In this study, we aimed to evaluate the capacity of TanDEM-X interferometric synthetic aperture radar (InSAR) data to estimate canopy height with the assistance of an external digital terrain model (DTM). A ground-to-volume ratio estimation model was proposed so that the canopy height could be precisely estimated from the random-volume-over-ground (RVoG) model. We also refined the RVoG inversion process with the relationship between the estimated penetration depth (PD) and the phase center height (PCH). The proposed method was tested by TanDEM-X InSAR data acquired over relatively homogenous coniferous forests (Teruel test site) and coniferous as well as broadleaved forests (La Rioja test site) in Spain. Comparing the TanDEM-X-derived height with the LiDAR-derived height at plots of size 50 m × 50 m, the root-mean-square error (RMSE) was 1.71 m (R2 = 0.88) in coniferous forests of Teruel and 1.97 m (R2 = 0.90) in La Rioja. To demonstrate the advantage of the proposed method, existing methods based on ignoring ground scattering contribution, fixing extinction, and assisting with simulated spaceborne LiDAR data were compared. The impacts of penetration and terrain slope on the RVoG inversion were also evaluated. The results show that when a DTM is available, the proposed method has the optimal performance on forest height estimation.

Inversion of Forest Biomass Based on Multi-Source Remote Sensing Images

Ecological forests are an important part of terrestrial ecosystems, are an important carbon sink and play a pivotal role in the global carbon cycle. At present, the comprehensive utilization of optical and radar data has broad application prospects in forest parameter extraction and biomass estimation. In this study, tree and topographic data of 354 plots in key nature reserves of Liaoning Province were used for biomass analysis. Remote sensing parameters were extracted from Landsat 8 OLI and Sentinel-1A radar data. Based on the strong correlation factors obtained via Pearson correlation analysis, a linear model, BP neural network model and PSO neural network model were used to simulate the biomass of the study area. The advantages of the three models were compared and analyzed, and the optimal model was selected to invert the biomass of Liaoning province. The results showed that 44 factors were correlated with forest biomass (p &lt; 0.05), and 21 factors were significantly correlated with forest biomass (p &lt; 0.01). The comparison between the prediction results of the three models and the real results shows that the PSO-improved neural network simulation results are the best, and the coefficient of determination is 0.7657. Through analysis, it is found that there is a nonlinear relationship between actual biomass and remote sensing data. Particle swarm optimization (PSO) can effectively solve the problem of low accuracy in traditional BP neural network models while maintaining a good training speed. The improved particle swarm model has good accuracy and speed and has broad application prospects in forest biomass inversion.

LIDAR-Based Forest Biomass Remote Sensing: A Review of Metrics, Methods, and Assessment Criteria for the Selection of Allometric Equations

The increasing level of atmospheric carbon dioxide and its effects on our climate system has become a global environment issue. The forest ecosystem is essential for the stability of carbon in the atmosphere as it operates as a carbon sink and provides a habitat for numerous species. Therefore, our understanding of the structural elements of the forest ecosystem is vital for the estimation of forest biomass or terrestrial carbon stocks. Over the last two decades, light detection and ranging (LIDAR) technology has significantly revolutionized our understanding of forest structures and enhanced our ability to monitor forest biomass. This paper presents a review of metrics for forest biomass estimation, outlines metrics selection methods for biomass modeling, and addresses various assessment criteria for the selection of allometric equations for the aboveground forest biomass estimations, using LIDAR data. After examining one hundred publications written by different authors between 1999 and 2023, it was observed that LIDAR technology has become a dominant data collection tool for aboveground biomass estimation with most studies focusing on the use of airborne LIDAR data for the plot-level analysis on a local scale. Parametric-based models dominated in most studies with coefficient of determination (R2) and root mean square error (RMSE) as assessment criteria. In addition, mean top canopy height (MCH) and quadratic mean height (QMH) were reported as strong predictors for aboveground biomass (AGB) estimation. Pixel-based uncertainty analysis was found to be a reliable method for assessing spatial variations in uncertainties.

Novel Features of Canopy Height Distribution for Aboveground Biomass Estimation Using Machine Learning: A Case Study in Natural Secondary Forests

Identifying important factors (e.g., features and prediction models) for forest aboveground biomass (AGB) estimation can provide a vital reference for accurate AGB estimation. This study proposed a novel feature of the canopy height distribution (CHD), a function of canopy height, that is useful for describing canopy structure for AGB estimation of natural secondary forests (NSFs) by fitting a bimodal Gaussian function. Three machine learning models (Support Vector Regression (SVR), Random Forest (RF), and eXtreme Gradient Boosting (Xgboost)) and three deep learning models (One-dimensional Convolutional Neural Network (1D-CNN4), 1D Visual Geometry Group Network (1D-VGG16), and 1D Residual Network (1D-Resnet34)) were applied. A completely randomized design was utilized to investigate the effects of four feature sets (original CHD features, original LiDAR features, the proposed CHD features fitted by the bimodal Gaussian function, and the LiDAR features selected by the recursive feature elimination algorithm) and models on estimating the AGB of NSFs. Results revealed that the models were the most important factor for AGB estimation, followed by the features. The fitted CHD features significantly outperformed the other three feature sets in most cases. When employing the fitted CHD features, the 1D-Renset34 model demonstrates optimal performance (R2 = 0.80, RMSE = 9.58 Mg/ha, rRMSE = 0.09), surpassing not only other deep learning models (e.g.,1D-VGG16: R2 = 0.65, RMSE = 18.55 Mg/ha, rRMSE = 0.17) but also the best machine learning model (RF: R2 = 0.50, RMSE = 19.42 Mg/ha, rRMSE = 0.16). This study highlights the significant role of the new CHD features fitting a bimodal Gaussian function and the effects between the models and the CHD features, which provide the sound foundations for effective estimation of AGB in NSFs.

Advances in remote sensing for sustainable forest management: monitoring and protecting natural resources

Remote sensing has emerged as a powerful tool for the monitoring and management of forests, contributing to the sustainable utilization and protection of natural resources. This paper presents a review of recent advances in remote sensing techniques and technologies for forest management, highlighting their role in monitoring and protecting forests. The integration of remote sensing with other geospatial methods enhances the accuracy and efficiency of data acquisition, aiding in the assessment of forest cover, biomass estimation, disturbance detection, and biodiversity monitoring. Furthermore, the potential of remote sensing for supporting decision-making processes in sustainable forest management is explored, emphasizing its versatility, cost-effectiveness, and ability to provide invaluable insights at both local and global scales. The paper concludes by discussing current challenges and future opportunities in remote sensing applications for sustainable forest management.

Sampling Estimation and Optimization of Typical Forest Biomass Based on Sequential Gaussian Conditional Simulation

The traditional classical sampling statistics method ignores the spatial location relationship of survey samples, which leads to many problems. This study aimed to propose a spatial sampling method for sampling estimation and optimization of forest biomass, achieving a more efficient and effective monitoring system. In this paper, we used Sequential Gaussian Conditional Simulation (SGCS) to obtain the biomass of four typical forest types in Shangri-La, Yunnan Province, China. In addition, we adopted a geostatistical sampling method for sample point layout and optimization to achieve the purpose of improving sampling efficiency and accuracy, and compared with the traditional sampling method. The main results showed that (1) the Gaussian model, exponential model, and spherical model were used to analyze the variogram of the four typical forests biomass, among which the exponential model had the best fitting effect (R2 = 0.571, RSS = 0.019). The range of the exponential model was 8700 m, and the nugget coefficient (C0/(C0 + C)) was 11.67%, which showed that the exponential model could be used to analyze the variogram of forest biomass. (2) The coefficient of variation (CV) based on 323 biomass field plots was 0.706, and the CV based on SGCS was 0.366. In addition, the Overall Estimate Consistency (OEC) of the simulation result was 0.871, which can be used for comparative analysis of traditional and spatial sampling. (3) Based on the result of SGCS, with 95% reliability, the sample size of traditional equidistant sampling (ES) was 191, and the sampling accuracy was 95.16%. But, the spatial sampling method based on the variation scale needed 92 samples, and the sampling accuracy was 93.12%. On the premise of satisfying sampling accuracy, spatial sampling efficiency was better than traditional ES. (4) The accuracy of stratified sampling (SS) of four typical forest areas based on 191 samples was 97.46%. However, the sampling accuracy of the biomass variance stratified space based on the SGCS was 93.89%, and the sample size was 52. Under the premise of satisfying the sampling accuracy, the sampling efficiency was obviously better than the traditional SS. Therefore, we can obtain the conclusion that the spatial sampling method is superior to the traditional sampling method, as it can reduce sampling costs and solve the problem of sample redundancy in traditional sampling, improving the sampling efficiency and accuracy, which can be used for sampling estimation of forest biomass.

Individual Tree Aboveground Biomass Estimation Based on UAV Stereo Images in a Eucalyptus Plantation

As one of the three fastest-growing tree species in the world, eucalyptus grows rapidly, with a monthly growth rate of up to 1 m and a maximum annual growth rate of up to 10 m. Therefore, ways to accurately and quickly obtain the aboveground biomass (AGB) of eucalyptus in different growth stages at a low cost are the foundation of achieving eucalyptus growth-change monitoring and precise management. Although Light Detection and Ranging (LiDAR) can achieve high-accuracy estimations of individual eucalyptus tree biomasses, the cost of data acquisition is relatively high. While the AGB estimation accuracy of high-resolution images may be affected by a lack of forest vertical structural information, stereo images obtained using unmanned aerial vehicles (UAVs) can not only provide horizontal structural information but also vertical structural information through derived point data, demonstrating strong application potential in estimating the biomass of eucalyptus plantations. To explore the potential of UAV stereo images for estimating the AGB of individual eucalyptus trees and further investigate the impact of stereo-image-derived features on the construction of biomass models, in this study, UAVs equipped with consumer-grade cameras were used to obtain multitemporal stereo images. Different features, such as spectral features, texture, tree height, and crown area, were extracted to estimate the AGB of individual eucalyptus trees of five different ages with three algorithms. The different features extracted based on the UAV images had different effects on estimating AGB in individual eucalyptus trees. By estimating eucalyptus AGB using only spectrum features, we found that tree height had the greatest impact, with its R2 value increasing by 0.28, followed by forest age. Other features, such as spectrum, texture, and crown area, had relatively small effects. For the three algorithms, the estimation accuracy of the CatBoost algorithm was the highest, with an R2 ranging from 0.65 to 0.90, and the normalized root-mean-square error (NRMSE) ranged from 0.08 to 0.15. This was followed by the random forest algorithm. The ridge regression algorithm had the lowest accuracy, with an R2 ranging from 0.34 to 0.82 and an NRMSE value ranging from 0.11 to 0.21. The AGB model that we established with forest age, TH, crown area, and HOM-B feature variables using the CatBoost algorithm had the best estimation accuracy, with an R2 of 0.90 and an NRMSE of 0.08. The results indicated that accurately estimating the AGB of individual eucalyptus trees can be achieved based on stereo images obtained using UAVs equipped with affordable, consumer-grade cameras. This paper can provide methodological references and technical support for estimating forest biomass, carbon storage, and other structural parameters based on UAV images.

Comparing Algorithms for Estimation of Aboveground Biomass in Pinus yunnanensis

Comparing algorithms are crucial for enhancing the accuracy of remote sensing estimations of forest biomass in regions with high heterogeneity. Herein, Sentinel 2A, Sentinel 1A, Landsat 8 OLI, and Digital Elevation Model (DEM) were selected as data sources. A total of 12 algorithms, including 7 types of learners, were utilized for estimating the aboveground biomass (AGB) of Pinus yunnanensis forest. The results showed that: (1) The optimal algorithm (Extreme Gradient Boosting, XGBoost) was selected as the meta-model (referred to as XGBoost-stacking) of the stacking ensemble algorithm, which integrated 11 other algorithms. The R2 value was improved by 0.12 up to 0.61, and RMSE was decreased by 4.53 Mg/ha down to 39.34 Mg/ha compared to the XGBoost. All algorithms consistently showed severe underestimation of AGB in the Pinus yunnanensis forest of Yunnan Province when AGB exceeded 100 Mg/ha. (2) XGBoost-Stacking, XGBoost, BRNN (Bayesian Regularized Neural Network), RF (Random Forest), and QRF (Quantile Random Forest) have good sensitivity to forest AGB. QRNN (Quantile Regression Neural Network), GP (Gaussian Process), and EN (Elastic Network) have more outlier data and their robustness was poor. SVM-RBF (Radial Basis Function Kernel Support Vector Machine), k-NN (K Nearest Neighbors), and SGB (Stochastic Gradient Boosting) algorithms have good robustness, but their sensitivity was poor, and QRF algorithms and BRNN algorithm can estimate low values with higher accuracy. In conclusion, the XGBoost-stacking, XGBoost, and BRNN algorithms have shown promising application prospects in remote sensing estimation of forest biomass. This study could provide a reference for selecting the suitable algorithm for forest AGB estimation.

Improving the Potential of Coniferous Forest Aboveground Biomass Estimation by Integrating C- and L-Band SAR Data with Feature Selection and Non-Parametric Model

Forests play a significant role in terrestrial ecosystems by sequestering carbon, and forest biomass is a crucial indicator of carbon storage potential. However, the single-frequency SAR estimation of forest biomass often leads to saturation issues. This research aims to improve the potential for estimating forest aboveground biomass (AGB) by feature selection based on a scattering mechanism and sensitivity analysis and utilizing a non-parametric model that combines the advantage of dual-frequency SAR data. By employing GF-3 and ALOS-2 data, this study explores the scattering mechanism within a coniferous forest by using results of target decomposition and the pixel statistics method. By selecting an appropriate feature (backscatter coefficients and polarization parameters) and using stepwise regression models and a non-parametric model (the random forest adaptive genetic algorithm (RF-AGA)), the results revealed that the RF-AGA model with feature selection exhibited excellent AGB estimation performance without obvious saturation (RMSE = 10.42 t/ha, R2 = 0.93, leave-one-out cross validation). The σHV, σVH, Pauli three-component decomposition, Yamaguchi three-component decomposition, and VanZyl3 component decomposition of thee C-band and σHV, σVH,σHH, Yamaguchi three-component decomposition, and VanZyl3 component decomposition of the L-band are suited for estimating the AGB of coniferous forests. Volume scattering was the dominant mechanism, followed by surface scattering, while double-bounce scattering had the smallest proportion. This study highlights the potential of investigating scattering mechanisms, sensitivity factors, and parameter selection in the C- and L-band SAR data for improved forest AGB estimation.

Modelling aboveground biomass of a multistage managed forest through synergistic use of Landsat-OLI, ALOS-2 L-band SAR and GEDI metrics

To better understand forest carbon budgets and design forest-based climate change mitigation solutions, reliable biomass estimation is critical. Integrating multi-sensor remote sensing data can improve forest monitoring and evaluation. This study adopted a one-hectare plot size sampling design to improve the integration of GEDI footprints with in-situ, optical, and SAR data for the estimation of forest above-ground biomass (AGB). The study was carried out for a managed tropical forest in the Himalayan foothills of India. The space-borne GEDI retrieved the canopy height of the study area with an RMSE of 3.36 m and Adj. R2 of 0.70. Extrapolating GEDI footprint heights (RH95) with Landsat 8 indices using Random Forest (RF) yielded a spatial canopy height of the study area with an R2 of 0.97 and RMSE of 2.32 m. Using the GEDI canopy height, foliage density, and plant area index, the AGB at GEDI footprints level was estimated using RF, with an R2 of 0.91 and an RMSE of 20.10 Mg ha−1. The spatial AGB model that only considered ALOS-2 SAR variables had an R2 of 0.61 and an RMSE of 18.27 Mg ha−1, whereas the RF model that considered both SAR variables and canopy height had a superior R2 of 0.77 and a lower RMSE of 13.86 Mg ha−1. The addition of canopy height data reduced the RMSE of the AGB model by 4.41 Mg ha−1 and also predicted a higher range of AGB. The study demonstrates the effectiveness of combining GEDI data with other sensors (optical and SAR data) to provide precise AGB of multistage managed forests using a one-hectare plot size sampling design.