Projected Changes in the Runoff Spectrum in Texas

We analyze projected changes in monthy temperature, precipitation, and other hydrologic variables across their entire range of variability using hydrologic model output driven by downscaled earth system models (LOCA-VIC) under a high-end emission scenario. We use quantile regression to identify differences in trends as a function of percentile rather than beginning with one or a handful of summary statistics for each variable.

Total monthly runoff trends in Texas are associated with changes in both surface runoff and baseflow. The mean surface runoff trends depend more strongly on trends in extreme precipitation than in mean precipitation, and extreme precipitation is expected to increase more rapidly (or decrease more slowly) than mean precipitation. Baseflow trends depend most strongly on trends in deep soil moisture, which are uniformly negative and decline much more sharply in wetter eastern basins than drier southern basins. The combination of these two effects leads to the greatest chance of overall runoff increases in southern basins and the greatest chance of overall runoff decreases in eastern basins.

The overall trend estimates mask a key aspect of the nature of projected runoff changes. Most models project that monthly runoff will become flashier, with declines likely across most total runoff percentiles and increases very likely in extremely wet months, with the primary question being at what percentile the total runoff trends switch from negative to positive. This sort of change will challenge surface water managers and reservoir operators, who will simultaneously have to deal with less reliable monthly inflows and occasional more intense flood flows. At the same time, the projections of declines in baseflow imply reductions in aquifer recharge, especially in eastern Texas, with declines of 1%-4% per decade possible.