The temperate savanna map in Northeastern China

The temperate savanna map in Northeastern China.This study map the temperate savanna on the sandy land in Northeastern Chinawithin a minimum identifiable area of 0.5 ha from Sentinel-1 and Sentinel-2 data by using Random Forest regression and Classification and Regression Tree algorithm (CART) algorithms.To improve the identification accuracy, training and validation datasets were collected from VeryHigh Resolution (VHR) UAV imagery. The proposed method achieved with the overall accuracy of 0.82 and Kappa coefficient of 0.63 for savanna identification. Temperate savanna in Northeastern China covering a total area of 156,000 km2.中国东北温带稀树草原分布图。本研究基于Sentinel-1与Sentinel-2遥感数据，采用随机森林回归（Random Forest regression）及分类与回归树（Classification and Regression Tree, CART）算法，针对中国东北沙质土地上的温带稀树草原开展制图，最小可识别面积为0.5公顷。为提升识别精度，研究团队从超高分辨率（Very High Resolution, VHR）无人机影像中采集训练与验证数据集。所提方法在稀树草原识别任务中取得了0.82的总体精度与0.63的Kappa系数。中国东北温带稀树草原总面积达156000平方千米。