The GROUNDED EO database: Extensive fiducial reference measurements of vegetation biophysical variables

This repository contains the fiducial reference database developed under the Ground Reference Observations Underlying Novel Decametric Vegetation Data Products from Earth Observation (GROUNDED EO) project, which was initiated within the context of the European Space Agency’s Living Planet Fellowship programme. The database contains over 16,000 fiducial reference measurements covering 81 National Ecological Observatory Network (NEON), Integrated Carbon Observation System (ICOS), and Terrestrial Ecosystem Research Network (TERN) sites between 2013 and 2022. The sites span cultivated crops, deciduous forest, evergreen forest, grasslands, mixed forest, pasture/hay, shrub/scrub and woody wetlands over the United States, Europe, and Australia. Fiducial reference measurements were generated from over 280,000 digital hemispherical photography (DHP) and digital cover photography (DCP) images. These were acquired in elementary sampling units (ESUs) of 20 m x 20 m (NEON), 30 m x 30 m (ICOS), and 100 m x 100 m (TERN), containing 9 (ICOS), 12 (NEON), and 36 (TERN) sampling points per ESU. DHP and DCP images were automatically processed using the HemiPy and CoverPy modules, respectively, which calculate estimates of gap fraction to derive plant area index (PAI), green area index (GAI), the fraction of vegetation cover (FCOVER), and for HemiPy, the fraction of intercepted photosynthetically active radiation (FIPAR). HemiPy and CoverPy adopt the recommendations of the Fiducial Reference Measurements for Vegetation (FRM4VEG) project, propagating uncertainties (due to variability in gap fraction) through the derivation of the considered variables, in line with the International Standards Organisation (ISO) Guide to the Expression of Uncertainty in Measurement (GUM). Each value, is, therefore, provided with an estimate of its standard uncertainty (delimited with +/-). Corrections for woody material, which is known to represent up to 35% of total plant area in forests, and may lead to errors of up to 61%, were applied, as were corrections for missing understory data at ICOS sites using previously published information. By applying these corrections and combining overstory and understory measurements, total leaf area index (LAI), fraction of absorbed photosynthetically active radiation (FAPAR), and FCOVER were derived from the values obtained from HemiPy and CoverPy. CSV files are provided containing i) all available fiducial reference measurements, and ii) only those fiducial reference measurements matched with contemporaneous Sentinel-2 L2A surface reflectance observations (i.e. those acquired within one day). More detailed description of the fiducial reference database is provided in the associated Remote Sensing of Environment paper. Please cite the paper in any work making use of these data: Brown, L.A., Fernandes, R., Verrelst, J., Morris, H., Djamai, N., Reyez-Muñoz, P., D.Kovács, D., Meier, C. GROUNDED EO: Data-driven Sentinel-2 LAI and FAPAR retrieval using Gaussian processes trained with extensive fiducial reference measurements, Remote Sens. Environ.本数据集仓库包含由地球观测新型十米级植被数据产品支撑的地面参考观测（Ground Reference Observations Underlying Novel Decametric Vegetation Data Products from Earth Observation，简称GROUNDED EO）项目开发的基准参考数据库（fiducial reference database），该项目依托欧洲空间局（European Space Agency）的生命星球奖学金计划（Living Planet Fellowship programme）发起。 该数据库包含2013至2022年间，覆盖81个国家生态观测站网络（National Ecological Observatory Network, NEON）、整合碳观测系统（Integrated Carbon Observation System, ICOS）及陆地生态系统研究网络（Terrestrial Ecosystem Research Network, TERN）观测站点的16000余组基准参考测量数据（fiducial reference measurements）。这些站点分布于美国、欧洲与澳大利亚，涵盖栽培作物、落叶林、常绿林、草原、混交林、牧场/干草田、灌丛/灌草丛以及木本湿地等多种生态类型。 基准参考测量数据源自超过28万张数字半球摄影（digital hemispherical photography, DHP）与数字覆盖摄影（digital cover photography, DCP）影像。这些影像采集于不同规格的基本采样单元（elementary sampling units, ESUs）：NEON站点为20m×20m，ICOS站点为30m×30m，TERN站点为100m×100m；每个基本采样单元内分别包含9个（ICOS）、12个（NEON）与36个（TERN）采样点。 DHP与DCP影像分别通过HemiPy与CoverPy模块完成自动化处理，通过计算间隙率推导得到植物面积指数（plant area index, PAI）、绿度面积指数（green area index, GAI）、植被覆盖度（fraction of vegetation cover, FCOVER），其中HemiPy还可计算光合有效辐射截获率（fraction of intercepted photosynthetically active radiation, FIPAR）。HemiPy与CoverPy遵循植被基准参考测量（Fiducial Reference Measurements for Vegetation, FRM4VEG）项目的规范，并依据国际标准化组织（International Standards Organisation, ISO）《测量不确定度表示指南（Guide to the Expression of Uncertainty in Measurement, GUM）》，在目标变量推导过程中传播由间隙率变异带来的不确定性。因此，每个测量值均附带其标准不确定度估计值（以±符号标注）。 针对木本物质的校正已被应用——森林中木本物质可占总植物面积的35%，若忽略该部分将带来最高达61%的计算误差；同时针对ICOS站点缺失的林下数据，研究团队采用已发表的公开资料完成了校正。通过上述校正并融合冠层与林下测量数据，研究人员基于HemiPy与CoverPy的输出结果，进一步推导得到总叶面积指数（leaf area index, LAI）、光合有效辐射吸收率（fraction of absorbed photosynthetically active radiation, FAPAR）以及植被覆盖度FCOVER。 本仓库提供两类逗号分隔值（Comma-Separated Values, CSV）格式文件：其一包含全部可用的基准参考测量数据，其二仅包含与同期哨兵-2号L2A地表反射率观测（Sentinel-2 L2A surface reflectance observations，即采集时间相差不超过1天的观测数据）匹配的基准参考测量数据。 该基准参考数据库的详细说明已发表于《环境遥感（Remote Sensing of Environment）》期刊的相关论文。若使用本数据集，请引用如下文献： Brown, L.A.、Fernandes, R.、Verrelst, J.、Morris, H.、Djamai, N.、Reyez-Muñoz, P.、D.Kovács, D.、Meier, C. 《GROUNDED EO: Data-driven Sentinel-2 LAI and FAPAR retrieval using Gaussian processes trained with extensive fiducial reference measurements》, Remote Sens. Environ.