BETA-FOR_SP3_ForestStructure_Sentinel2_2015/2023

Sentinel-2 (multispectral) satellite time-series (2015-2023) at patch-level as proxy of canopy structure. The dataset is available for all 234 patches and stored in a single .xlsx file where each sheet name (BETA-FOR patch label) stores information for the respective patch as various metrics (combination of spectral index and spatial statistic to aggregate the pixels per patch, i.e. coefficient of variation). Based on the multispectral bands, spatial statistics of five different spectral indices (column name prefix: NMDI, MBWI, MSI, AVI, MLSWI26) were calculated (please find the full name and formula under https://github.com/awesome-spectral-indices/awesome-spectral-indices/blob/main/output/spectral-indices-table.csv and the publication of the spectral indices library under https://doi.org/10.1038/s41597-023-02096-0). The original time-series (column name ending *_original) was decomposed based on the BEAST model (https://doi.org/10.1016/j.rse.2019.04.034) into the seasonal (column name ending *_seasonality) and trend signal (column name ending *_trend). The aggregation of pixels at patch-level was calculated as coefficient of variation (column name *_cv_*). The seasonal and trend signal were modelled as weekly intervals from the original irregular time-series. Please find the data in the attachment since it is structured as nested .xlsx with multiple (234) sheets.哨兵二号（Sentinel-2）多光谱卫星2015-2023年的斑块级时间序列数据，作为冠层结构的替代表征指标。本数据集覆盖全部234个斑块，存储于单个.xlsx文件中，每个工作表的名称（BETA-FOR斑块标签）对应相应斑块的信息，内含多种聚合斑块内像素的指标——结合光谱指数与空间统计量生成，即变异系数。基于多光谱波段，研究人员计算了5种不同光谱指数的空间统计量，对应列名前缀分别为NMDI、MBWI、MSI、AVI、MLSWI26。完整的指数名称与计算公式可参见https://github.com/awesome-spectral-indices/awesome-spectral-indices/blob/main/output/spectral-indices-table.csv，该光谱指数库的相关研究文献可参考https://doi.org/10.1038/s41597-023-02096-0。原始时间序列（列名以*_original结尾）基于BEAST模型（https://doi.org/10.1016/j.rse.2019.04.034）分解为季节性分量（列名以*_seasonality结尾）与趋势分量（列名以*_trend结尾）。斑块内像素的聚合值以变异系数计算，对应列名格式为*_cv_*。研究人员将原始不规则时间序列中的季节性与趋势信号建模为周度间隔形式。本数据集为包含234个工作表的嵌套式Excel文件，完整数据请见附件。