How Well Can We Model Seasonal Variations in Land Water Storage Globally?

Seasonal fluctuations dominate the climate signal and impact land water storages and fluxes underscoring the need to assess the reliability of modeled seasonal fluctuations globally. Here we compare modeled seasonal amplitudes in water storage from global hydrologic models (GHMs: PCR-GLOBWB and WGHM) and land surface models (LSMs: NOAH, VIC, CLM, and CLSM) with storage variations from GRACE satellite data in 180 river basins globally. While GHMs generally include surface water and groundwater storage compartments, many LSMs do not. Results show that seasonal signals account for about 50-75% of land total water storage signals globally. Most models underestimate seasonal amplitudes in tropical and (semi)arid basins and generally overestimate seasonal amplitudes in northern high latitude basins. Varying model storage capacity can improve simulated storage amplitudes in some regions relative to GRACE; however, it is difficult to determine which storage compartments are critical (e.g. surface water, soil moisture, groundwater). Analyzing model inputs (precipitation) and outputs (runoff and evapotranspiration fluxes) helps to better understand storage discrepancies between models and GRACE. The dominance of the seasonal signal highlights the importance of improving modeled seasonal storages and fluxes to enhance global models of the water cycle.