Objective Forest biomass is the key factor to measure forest carbon reserves. Therefore, accurate estimation of forest biomass is helpful for forest management and resource utilization. The data from Sentinel-2A provide new opportunities for biomass estimation and monitoring due to rich spectral information and high spatial resolution. In this paper, we evaluated the ability of various parameters to estimate aboveground biomass of coniferous forest , and completed regional-scale forest biomass estimation based on Sentinel-2A.
Method Selecting the coniferous forest of Wangyedian Forest Farm in Chifeng City, Inner Mongolia of northern China as the research object, this research extracted spectral reflectance, vegetation index and biophysical parameters in Sentinel-2A. Then we set up multiple stepwise regression equations by data types to estimate biomass. In addition, the elevation factor was added to improve the accuracy of model.
Result The results showed that: (1) the model built with multiple types of parameters had the highest accuracy, with R2 reaching 0.765 and RMSE was 39.49 t/ha; (2) among the models established by spectral reflectance, vegetation index and biophysical parameters, the accuracy based on the vegetation index was higher, indicating that the vegetation index had a greater impact on coniferous forest aboveground biomass estimation than the spectral reflectance and biophysical parameters; (3) for all the models of our research, elevation always improved accuracy.
Conclusion The retrieved aboveground biomass from Sentinel-2A spatial distribution is basically consistent with the actual situation, which shows that the coniferous forest aboveground biomass inversion is meaningful for regional forest resource monitoring.