GNSS-PWV and Multi-Source Remote-Sensing Data and Code for Multiscale Forest Carbon–Water Response Analysis in Beijing–Tianjin–Hebei, China

This dataset and code archive supports a multiscale investigation of atmospheric water vapor, hydrothermal conditions, and forest carbon–water responses in the Beijing–Tianjin–Hebei region of China during 2023–2025. The archive contains derived hourly BeiDou precipitable water vapor (BDS-PWV) data, GNSS station information, forest-cover screening results, Sentinel-2 NDMI, MODIS land-surface temperature, MOD16 evapotranspiration and potential evapotranspiration, MOD17 gross primary productivity and net photosynthesis, ERA5-Land hydrothermal variables, and additional public-product validation data from PML-V2.2a, MCD15A3H, SMAP, GPM IMERG, and Dynamic World. It also provides analysis-ready hourly, daily, and 8-day panels, analysis workbooks, figure-source data, locked statistical results, Earth Engine scripts, Python code, data dictionaries, citation information, and SHA-256 manifests. The package supports numerical verification, buffer-sensitivity analysis, calendar-phase comparisons, moving-block robustness assessment, cross-product validation, and locked 2025 holdout evaluation. Raw GNSS receiver observations and RINEX files are not redistributed. PWV represents atmospheric column water vapor and is not equivalent to canopy water content. The supplied associations are observational and do not establish causal ecological effects. GPP, PsnNet, ET, LAI, FPAR, soil-moisture, precipitation, and tree-probability products have different definitions and uncertainties and should not individually be interpreted as direct measurements of forest carbon sinks.