scholarly journals Improving Accuracy Estimation of Forest Aboveground Biomass Based on Incorporation of ALOS-2 PALSAR-2 and Sentinel-2A Imagery and Machine Learning: A Case Study of the Hyrcanian Forest Area (Iran)

2018 ◽  
Vol 10 (2) ◽  
pp. 172 ◽  
Author(s):  
Sasan Vafaei ◽  
Javad Soosani ◽  
Kamran Adeli ◽  
Hadi Fadaei ◽  
Hamed Naghavi ◽  
...  
Keyword(s):  
Machine Learning ◽  
Aboveground Biomass ◽  
Forest Area ◽  
Hyrcanian Forest ◽  
Accuracy Estimation ◽  
Sentinel 2A ◽  
Improving Accuracy ◽  
Forest Aboveground Biomass

Assessment of dual polarization in Sentinel-1 data for estimating forest aboveground biomass: case study of Barru Regency, South Sulawesi

2019 ◽  
Author(s):  
Wahyu Hendardi Giri Ananto ◽  
Haeydar Anggara Hadi ◽  
Ade Febri Sandhini Putri ◽  
Difa Nisrina Hanum ◽  
Bayu Kurnia Puji Wiryawan ◽  
...  
Keyword(s):  
Aboveground Biomass ◽  
Dual Polarization ◽  
South Sulawesi ◽  
Forest Aboveground Biomass

scholarly journals An Evaluation of Eight Machine Learning Regression Algorithms for Forest Aboveground Biomass Estimation from Multiple Satellite Data Products

2020 ◽  
Vol 12 (24) ◽  
pp. 4015
Author(s):  
Yuzhen Zhang ◽  
Jun Ma ◽  
Shunlin Liang ◽  
Xisheng Li ◽  
Manyao Li
Keyword(s):  
Machine Learning ◽  
Satellite Data ◽  
Aboveground Biomass ◽  
Random Sampling ◽  
Stratified Sampling ◽  
Multivariate Adaptive Regression Splines ◽  
Support Vector ◽  
Regression Algorithms ◽  
Ensemble Algorithms ◽  
Forest Aboveground Biomass

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.

scholarly journals Improving non-representative-sample prediction of forest aboveground biomass maps: A combined machine learning and spatial statistical approach

2019 ◽  
Author(s):  
Shaoqing Dai ◽  
Xiaoman Zheng ◽  
Lei Gao ◽  
Shudi Zuo ◽  
Qi Chen ◽  
...  
Keyword(s):  
Machine Learning ◽  
Aboveground Biomass ◽  
Representative Sample ◽  
Large Scale ◽  
Statistical Approach ◽  
Low Cost ◽  
Ecological Modelling ◽  
Environmental Covariates ◽  
Remote Sensing Mapping ◽  
Forest Aboveground Biomass

Abstract. High-precision prediction of large-scale forest aboveground biomass (AGB) is important but challenging on account of the uncertainty involved in the prediction process from various sources, especially the uncertainty due to non-representative sample units. Usually caused by inadequate sampling, non-representative sample units are common and can lead to geographic clusters of localities. But they cannot fully capture complex and spatially heterogeneous patterns, in which multiple environmental covariates (such as longitude, latitude, and forest structures) affect the spatial distribution of AGB. To address this challenge, we propose herein a low-cost approach that combines machine learning with spatial statistics to construct a regional AGB map from non-representative sample units. The experimental results demonstrate that the combined methods can improve the accuracy of AGB mapping in regions where only non-representative sample units are available. This work provides a useful reference for AGB remote-sensing mapping and ecological modelling in various regions of the world.

Sign in / Sign up

Export Citation Format

Share Document