Modeling Hydrological Extremes in the Colorado River Basin at Various Watershed Scales.

Over the past century, the Colorado River Basin (CRB) experienced substantial warming and interannual climate variations, with extreme flood, heatwave, and drought events. Models project more frequent and intense extreme hydroclimate events, with major consequences for water resources in the southwestern U.S. Climate change has already reduced snowpack and streamflow, and induced earlier snowmelt and peak streamflow. Observational studies suggest that these effects have in turn amplified plant water stress and shortened growing seasons, but more work is needed to investigate the modulating role of watershed scale-dependencies, and various seasonal climate regimes and site conditions. To evaluate future projections appropriately, it is thus important to first quantify the regional hydrologic response to historic climate variability across a wider range of represented environments of the CRB. In the current effort, we force the Variable Infiltration Capacity (VIC) land surface hydrology model and a river routing model with a gridded dataset of historical meteorological observations (1976-2005) to estimate water balance components and naturalized streamflow response at 1/16^o spatial resolution and at an hourly time step. We employ VIC version 5 (VIC-5), the latest model version, which has yet to be employed in CRB studies. We also incorporate recent model modifications to account for (1) spatiotemporal variations in vegetation parameters using satellite remote sensing and (2) evapotranspiration from croplands and urban areas using a novel irrigation scheme. We conduct three analyses: (1) ground-truthing of the gridded meteorological data products against observations from meteorological stations across the CRB, (2) evaluating the effects of employing the modified VIC-5 by comparing our model results to those of earlier studies that used VIC version 4 without the current enhancements, and (3) quantifying spatiotemporal variability and watershed scale-dependencies of major water balance terms simulated from our model runs and their linkages to historic extreme meteorological events. For this final analysis, we fit statistical distributions to monthly, seasonal, and annual maximums, minimums, and extreme percentiles of temperature and major water balance terms at the grid-cell, basin-wide, and various sub-watershed scales. We use the fitted distribution parameters as indices to characterize and compare extreme events and their variable effects under the various climate regimes and site conditions. Through these novel model-based and statistical approaches, our study highlights the importance of (1) using quantifiable metrics to explicitly examine the effects of extreme events on water resources across the full range of CRB environments and (2) accounting for the role of seasonality and watershed scale in modifying these impacts.