boreal_wetland_recovery_sentinel

These data were used to train and validate predictive models for the temporal change of four vegetation metrics. Metrics are expressed as a proportion of the associated Sentinel-2 cells. Proportions of jack pine, tamarack, and exposed herbaceous vegetation were derived from aerial field estimates. The height metric derives from a LiDAR CHM expressed as the maximum at 1 m resolution, then expressed as a proportion of the Sentinel-2 cell over 2 m in height. Spatial attributes have been removed, but training values extracted from Sentinel-2 bands and spectral indices are included ('Training Features'). Individual Sentinel-2 scenes used in the study are listed. The data dictionary provides additional details about each attribute.本数据集用于训练并验证针对四项植被指标时间变化的预测模型。各项植被指标以对应哨兵二号（Sentinel-2）像元的占比形式进行表示。其中，短叶松、美洲落叶松及裸露草本植被的占比均通过航空野外估算得到。高度参数源自激光雷达（LiDAR）冠层高度模型（CHM），该模型以1米分辨率下的最大值形式呈现，随后转换为哨兵二号像元中高度超过2米的区域占比。数据集已移除空间属性，但包含从哨兵二号波段与光谱指数中提取的训练特征值（'Training Features'）。本研究中使用的所有哨兵二号影像场景均已逐一列出，数据字典提供了各属性的详细补充信息。