Root zone water storage capacity estimates from six datasets at 90 FLUXNET sites across CONUS

Dataset Overview This dataset accompanies the paper How data uncertainty influences estimates of root zone water storage capacity. It includes two primary datasets: InputDataCleaned_5yr and ForAnalysis_5yr. The “5yr” designation indicates that only sites with a minimum of five years of available FLUXNET data were retained. InputDataCleaned_5yr contains time series of evapotranspiration (ET) and precipitation (P) at each site and dastest combinations, at each dataset’s native temporal resolution. It also includes derived variables used throughout the analysis: ClimDiff, which corresponds to D(t) in the manuscript (the cumulative water balance, defined as cumulative P−ET), and CWD (cumulative water deficit), which represents deviations of D(t) from its running maximum. ForAnalysis_5yr contains processed data frames used in the analysis: SiteData: Aggregated site-level summaries of the input data Subset_PosBal: Site–dataset combinations identified as valid based on long-term positive water balance Sfinal: The final value of ClimDiff (D(t)) for each site–dataset combination, used to determine validity (i.e., to construct Subset_PosBal) Drawdown_Loc: Start date, end date, and magnitude of each CWD period Together, these datasets provide the inputs and intermediate products used to evaluate how uncertainty in precipitation and evapotranspiration influences estimates of root zone water storage capacity. Unless otherwise noted, all water balance and storage quantities, including P, ET, ClimDiff (D(t)), CWD, Sfinal, CWD magnitude, and SR, are reported in millimeters (mm). R scripts used to generate and process these datasets are included: 0_Functions.R, 1_Create_Data2.R, and 1_FNET_Input.R. Source Datasets The analysis integrates multiple observational and modeled datasets of precipitation and evapotranspiration: FLUXNET (FNET): Site-level observations of evapotranspiration and meteorological variables derived from eddy covariance flux tower measurements. Site records used in this analysis fall within 1991–2021, with start and end dates varying by site. PML: A remote sensing–based evapotranspiration product derived using the Penman–Monteith–Leuning framework, with data used in this analysis spanning 2000–2020. MODIS (MOD): Satellite-based evapotranspiration estimates derived from Moderate Resolution Imaging Spectroradiometer observations, with data used in this analysis spanning 2000–2022. PRISM: Gridded precipitation data based on station observations interpolated using climatologically aided methods, with data used in this analysis spanning 2000–2022. NLDAS: Land surface model outputs from the North American Land Data Assimilation System, providing gridded estimates of precipitation and evapotranspiration, with data used in this analysis spanning 2000–2022. Stage IV: Gridded precipitation estimates, with data used in this analysis spanning 2002–2022. These datasets and processing steps are described in detail in the associated manuscript and R scripts.