Machine learning algorithms are increasingly used in various remote sensing applications due to their ability to identify nonlinear correlations. Ensemble algorithms have been included in many practical applications to improve prediction accuracy. We provide an overview of three widely used ensemble techniques: bagging, boosting, and stacking. We first identify the underlying principles of the algorithms and present an analysis of current literature. We summarize some typical applications of ensemble algorithms, which include predicting crop yield, estimating forest structure parameters, mapping natural hazards, and spatial downscaling of climate parameters and land surface temperature. Finally, we suggest future directions for using ensemble algorithms in practical applications.

This study provided a comprehensive evaluation of eight machine learning regression algorithms for forest aboveground biomass (AGB) estimation from satellite data based on leaf area index, canopy height, net primary production, and tree cover data, as well as climatic and topographical data. Some of these algorithms have not been commonly used for forest AGB estimation such as the extremely randomized trees, stochastic gradient boosting, and categorical boosting (CatBoost) regression. For each algorithm, its hyperparameters were optimized using grid search with cross-validation, and the optimal AGB model was developed using the training dataset (80%) and AGB was predicted on the test dataset (20%). Performance metrics, feature importance as well as overestimation and underestimation were considered as indicators for evaluating the performance of an algorithm. To reduce the impacts of the random training-test data split and sampling method on the performance, the above procedures were repeated 50 times for each algorithm under the random sampling, the stratified sampling, and separate modeling scenarios. The results showed that five tree-based ensemble algorithms performed better than the three nonensemble algorithms (multivariate adaptive regression splines, support vector regression, and multilayer perceptron), and the CatBoost algorithm outperformed the other algorithms for AGB estimation. Compared with the random sampling scenario, the stratified sampling scenario and separate modeling did not significantly improve the AGB estimates, but modeling AGB for each forest type separately provided stable results in terms of the contributions of the predictor variables to the AGB estimates. All the algorithms showed forest AGB were underestimated when the AGB values were larger than 210 Mg/ha and overestimated when the AGB values were less than 120 Mg/ha. This study highlighted the capability of ensemble algorithms to improve AGB estimates and the necessity of improving AGB estimates for high and low AGB levels in future studies.

Accurately quantifying the aboveground biomass (AGB) of forests is crucial for understanding global change-related issues such as the carbon cycle and climate change. Many studies have estimated AGB from multiple remotely sensed datasets using various algorithms, but substantial uncertainties remain in AGB predictions. In this study, we aim to explore whether diverse algorithms stacked together are able to improve the accuracy of AGB estimates. To build the stacking framework, five base learners were first selected from a series of algorithms, including multivariate adaptive regression splines (MARS), support vector regression (SVR), multilayer perceptron (MLP) model, random forests (RF), extremely randomized trees (ERT), stochastic gradient boosting (SGB), gradient-boosted regression tree (GBRT) algorithm, and categorical boosting (CatBoost), based on diversity and accuracy metrics. Ridge and RF were utilized as the meta learner to combine the outputs of base learners. In addition, six important features were selected according to the feature importance values provided by the CatBoost, ERT, GBRT, SGB, MARS and RF algorithms as inputs of the meta learner in the stacking process. We then used stacking models with 3–5 selected base learners and ridge or RF to estimate AGB. The AGB data compiled from plot-level forest AGB, high-resolution AGB data derived from field and lidar data and the corresponding predictor variables extracted from the satellite-derived leaf area index, net primary production, forest canopy height, tree cover data, and Global Multiresolution Terrain Elevation Data 2010, as well as climate data, were randomly split into groups of 80% for training the model and 20% for model evaluation. The evaluation results showed that stacking generally outweighed the optimal base learner and provided improved AGB estimations, mainly by decreasing the bias. All stacking models had relative improvement (RI) values in bias of at least 22.12%, even reaching more than 90% under some scenarios, except for deciduous broadleaf forests, where an optimal algorithm could provide low biased estimations. In contrast, the improvements of stacking in R² and RMSE were not significant. The stacking of MARS, MLP, and SVR provided improved results compared with the optimal base learner, and the average RI in R² was 3.54% when we used all data without separating forest types. Finally, the optimal stacking model was used to generate global forest AGB maps.

Forest biomass quantification is essential to the global carbon cycle and climate studies. Many studies have estimated forest biomass from a variety of data sources, and consequently generated some regional and global maps. However, these forest biomass maps are not well known and evaluated. In this paper, we reviewed an extensive list of currently available forest biomass maps. For each map, we briefly introduced the data sources, the algorithms used, and the associated uncertainties. Large-scale biomass datasets were compared across Europe, the conterminous United States, Southeast Asia, tropical Africa and South America. Results showed that these forest biomass datasets were almost entirely inconsistent, particularly in woody savannas and savannas across these regions. The uncertainties in biomass maps could be from a variety of sources including the chosen allometric equations used to calculate field data, the choice and quality of remotely sensed data, as well as the algorithms to map forest biomass or extrapolation techniques, but these uncertainties have not been fully quantified. We suggested the future directions for generating more accurate large-scale forest biomass maps should concentrate on the compilation of field biomass data, novel approaches of forest biomass mapping, and comprehensively addressing the accuracy of generated biomass maps.

China has experienced intense land use and land cover changes during the past several decades, which have exerted significant influences on climate change. Previous studies exploring related climatic effects have focused mainly on one or two specific land use changes, or have considered all land use and land cover change types together without distinguishing their individual impacts, and few have examined the physical processes of

Many advanced satellite estimation methods have been developed, but global forest aboveground biomass (AGB) products remain largely uncertain. In this study, we explored data fusion techniques to generate a global forest AGB map for the 2000s at 0.01-degree resolution with improved accuracy by integrating ten existing local or global maps. The error removal and simple averaging algorithm, which is efficient and makes no assumption about the data and associated errors, was proposed to integrate these ten forest AGB maps. We first compiled the global reference AGB from in situ measurements and high-resolution AGB data that were originally derived from field data and airborne lidar data and determined the errors of each forest AGB map at the pixels with corresponding reference AGB values. Based on the errors determined from reference AGB data, the pixel-by-pixel errors associated with each of the ten AGB datasets were estimated from multiple predictors (e.g., leaf area index, forest canopy height, forest cover, land surface elevation, slope, temperature, and precipitation) using the random forest algorithm. The estimated pixel-by-pixel errors were then removed from the corresponding forest AGB datasets, and finally, global forest AGB maps were generated by combining the calibrated existing forest AGB datasets using the simple averaging algorithm. Cross-validation using reference AGB data showed that the accuracy of the fused global forest AGB map had an R-squared of 0.61 and a root mean square error (RMSE) of 53.68 Mg/ha, which is better than the reported accuracies (R-squared of 0.56 and RMSE larger than 80 Mg/ha) in the literature. Intercomparison with previous studies also suggested that the fused AGB estimates were much closer to the reference AGB values. This study attempted to integrate existing forest AGB datasets for generating a global forest AGB map with better accuracy and moved one step forward for our understanding of the global terrestrial carbon cycle by providing improved benchmarks of global forest carbon stocks.