H2CM - inputs and targets

Dataset description This Zarr archive contains the global‑scale time‐series and static data used to force and constrain H2CM. It is organized into one folder (Zarr group) per variable: Meteorological forcing (daily time series used to force H2CM) Static inputs (time‑invariant geospatial encodings of static data) Model constraints (data that is used to constrain H2CM) Each group stores a regularly gridded array on the same global grid, with coordinate variables for latitude, longitude (and time for the temporally varying fields). All names below correspond to top‑level Zarr groups in the archive. Summary of Zarr groups: Group name Variable Frequency Source product Meteorological forcing prec Precipitation Daily GPCP 1dd v1.2 rn Net surface radiation Daily CERES SYN1deg Ed4A tair 2 m air temperature Daily CRUNCEP v8 vpd Vapor pressure deficit Daily ERA5 co2 Atmospheric CO₂ concentration Daily CAMS Static inputs static (all features) Condensed representation of soil properties, land‑cover, DEM, wetlands – SoilGrids v2, GlobLand30 v1, GTOPO30, Tootchi map Model constraints tws Terrestrial water storage anomaly Monthly GRACE Tellus JPL RL06M v1 fapar Fraction of absorbed PAR (fAPAR) 8‑daily MOD15A2H swe_bc Snow Water Equivalent (bias‑corrected) Monthly GlobSnow v3 et_xbase Evapotranspiration Monthly FLUXCOM‑X‑BASE q_v2 Runoff Monthly GRUN v2 gpp Gross Primary Productivity Monthly FLUXCOM‑X‑BASE nee Net Ecosystem Exchange (monthly) Monthly OCO‑2 v10 MIP nee_iav Net Ecosystem Exchange (interannual variability) Monthly Jena CarboScope cue Mean annual carbon use efficiency (NPP / GPP) Monthly TRENDY v11 Utility variables mask Land grids used for training – grid_area Grid‑cell surface area (m²) – Derived from spherical geometry Additional information For full details on these datasets—including their development, validation, and primary references—please see: Baghirov, Z., Reichstein, M., Kraft, B., Ahrens, B., Körner, M., & Jung, M. (2025). H2CM (v1.0): Hybrid modeling of global water–carbon cycles constrained by atmospheric and land observations. EGUsphere, 2025, 1–33. https://doi.org/10.5194/egusphere-2025-3123 (preprint at https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3123/) Contact If you have any questions about these data, need help getting started, or encounter any bugs or other issues, please don’t hesitate to reach out: GitHub issuesOpen an issue or feature request at the H2CM repository:https://github.com/zavud/h2cm/issues EmailSend us a message at [e-mail]数据集说明 本Zarr归档文件（Zarr archive）包含用于驱动并约束H2CM模型的全球尺度时序数据与静态数据。该归档按变量划分为多个文件夹（即Zarr组）： 气象强迫数据（用于驱动H2CM的逐日时序数据） 静态输入数据（时不变的静态地理空间编码数据） 模型约束数据（用于约束H2CM模型的数据） 每个组均存储同一全球网格下的规则网格化阵列，包含纬度、经度坐标变量（时变场额外包含时间坐标变量）。下文所有名称均对应该归档中的顶级Zarr组。 Zarr组汇总： 组名称 | 变量 | 数据频率 | 源产品 气象强迫数据 prec | 降水 | 逐日 | GPCP 1dd v1.2 rn | 地表净辐射 | 逐日 | CERES SYN1deg Ed4A tair | 2米气温 | 逐日 | CRUNCEP v8 vpd | 水汽压亏缺 | 逐日 | ERA5 co2 | 大气CO₂浓度 | 逐日 | CAMS 静态输入数据 static（全特征集） | 土壤属性、土地覆盖、数字高程模型（DEM）、湿地的整合表征 | 无固定频率 | SoilGrids v2、GlobLand30 v1、GTOPO30、Tootchi map 模型约束数据 tws | 陆地水储量异常 | 逐月 | GRACE Tellus JPL RL06M v1 fapar | 吸收光合有效辐射占比（fAPAR） | 每8天 | MOD15A2H swe_bc | 偏差校正雪水当量 | 逐月 | GlobSnow v3 et_xbase | 蒸散发 | 逐月 | FLUXCOM-X-BASE q_v2 | 径流 | 逐月 | GRUN v2 gpp | 总初级生产力 | 逐月 | FLUXCOM-X-BASE nee | 月尺度生态系统净交换 | 逐月 | OCO-2 v10 MIP nee_iav | 生态系统净交换（年际变率） | 逐月 | Jena CarboScope cue | 年均碳利用效率（净初级生产力(NPP)/总初级生产力(GPP)） | 逐月 | TRENDY v11 工具变量 mask | 训练所用陆地区域网格 | 无固定频率 | 无 grid_area | 网格单元表面积（单位：平方米） | 无固定频率 | 由球面几何推导得到 补充说明 如需了解这些数据集的完整细节——包括其开发、验证与主要参考文献，请参阅： Baghirov, Z., Reichstein, M., Kraft, B., Ahrens, B., Körner, M., & Jung, M. (2025). H2CM (v1.0): Hybrid modeling of global water–carbon cycles constrained by atmospheric and land observations. EGUsphere, 2025, 1–33. https://doi.org/10.5194/egusphere-2025-3123 (预印本：https://egusphere.copernicus.org/preprints/2025/egusphere-2025-3123/) 联系方式 若您对这些数据有任何疑问、需要入门协助，或遇到任何程序错误或其他问题，请随时联系我们： GitHub 问题反馈：前往H2CM仓库提交问题或功能请求：https://github.com/zavud/h2cm/issues 电子邮件：请发送邮件至 [e-mail]