Aboveground Forest Biomass Estimation Based On Machine Learning Algorithms And Multi-source Data In A Typical Subtropical Region

Forest biomass is an important parameter for evaluating forest carbon sequestration capacity and ecological value.Although remote sensing has been widely used for quantitative estimation of forest biomass,how to effectively combine different sensors with different modeling algorithms to improve the biomass estimation accuracy is still poorly understood.This research conducted a comparative analysis of different datasets(e.g.,Landsat Thematic Mapper [TM],ALOS PALSAR L-band data,and their combinations)and modeling algorithms(e.g.,artificial neural network [ANN],support vector regression [SVR],Random Forest [RF],k-nearest neighbor [KNN],and linear regression [LR])for AGB estimation in a subtropical region under non-stratification and stratification of forest types.The results show that:(1)Landsat TM imagery provides more accurate AGB estimates(root mean squared error [RMSE] values in 27.7–29.3 Mg/ha)than ALOS PALSAR(RMSE values in 30.3–33.7 Mg/ha).The combination of TM and PALSAR data has similar performance for ANN and SVR(RMSE values from 27.6 Mg/ha and 28.2Mg/ha to 27.7 Mg/ha and 28.2 Mg/ha),worse performance for RF and KNN(RMSE values increase from 28.4 Mg/ha and 28.3 Mg/ha to 30.3 Mg/ha and 30.8 Mg/ha),and slightly improved performance for LR(RMSE values decrease from 29.3 Mg/ha to 27.7 Mg/ha).(2)Overestimation for small AGB values and underestimation for large AGB values are major problems when using the optical(e.g.,Landsat)or radar(e.g.,ALOS PALSAR)data.(3)LR is still an important tool for AGB modeling,especially for the AGB range of 40–120 Mg/ha.Machine learning algorithms have limited effects on improving AGB estimation overall,but ANN can improve AGB modeling when AGB values are greater than 120 Mg/ha(when biomass between 120-160Mg/ha,RMSE values decrease from 34.2Mg/ha to 29.1Mg/ha;and when biomass is greater than 160Mg/ha,RMSE values decrease from 61.0Mg/ha to 50.2Mg/ha).(4)Forest types and AGB ranges are important factors influencing AGB modeling performance.(5)Stratification based on forest types improved AGB estimation,especially when AGB was greater than 160 Mg/ha using LR approach(RMSE values decrease from 58.6Mg/ha to 41.5Mg/ha).This research provided new insights for remote sensing–based AGB modeling for the subtropical forest ecosystem through comprehensive analysis of different source data,modeling algorithms,and forest types.It is critical to develop an optimal AGB modeling procedure including collection of sufficient number of sample plots,extraction of suitable variables and modeling algorithms,and evaluation of the AGB estimates.