Evaluation of groundwater contributions to long-term persistence of land-surface hydrological fluxes in the Upper Colorado River Basin

Abstract:

One of the most important characteristics of natural water storages (e.g. oceans, lakes, soil and groundwater) is their long-term persistence (LTP). LTP translates white-noise inputs, such as precipitation, to outputs as soil moisture content, evapotranspiration (ET), or streamflow which are autocorrelated over very long timescales. Understanding the LTP of key hydrological variables can benefit the forecast of natural extremes such as floods or droughts. While the LTP of individual hydrological variables has been extensively studied, the contribution of groundwater (GW) is less well understood. One possible reason may be the over-simplification of GW in traditional Earth system/hydrological models. Using a full-suite of outputs for nearly four decades from a land-surface and groundwater integrated model (ParFlow-CLM) for the Upper Colorado River Basin, the impacts of GW on the LTP of soil moisture, ET and streamflow are explored in this study.

The connection of soil moisture, ET, and streamflow with GW are firstly quantified both in space and with time (e.g., droughts or floods periods) through cross-correlation analysis. Then the LTP of different variables are determined by spectral analysis. Most importantly, the variations of LTP of variables with their connection with GW in space and with time are also quantified.