Annual hydroclimatology of the continental United States

The annual hydroclimatology of the continental United States is investigated. HydroClimatic time-series which account for the complex variations in hydrology and climate at 1337 watersheds in the U.S. are used to evaluate several approaches for estimating the long-term water balance and the interannual variability of streamflow. Budyko-type relations (Budyko, 1974), which are commonly used for understanding the long-term water balance, are shown to perform poorly for basins with low soil moisture storage capacity. The Koster and Suarez relationship, which predicts the interannual variability of streamflow, is also shown to be unable to predict the runoff variability ratio for basins with low soil moisture storage capacity. New relationships are derived for estimating the long-term water balance and the interannual variability of streamflow using a water balance model, which accounts for variations in soil moisture. The derived relationships depend only on an aridity index &phis; = PE/P and a soil moisture index. A physically based approach for estimating the soil moisture index is suggested using the monthly time series of precipitation, potential evaporation and average maximum soil moisture holding capacity. The resulting methodology can predict the long-term water balance and the interannual variability of streamflow without the need for either calibration and/or streamflow data.; Climate elasticity of streamflow (ϵ_p) provides a measure of the sensitivity of streamflow to changes in precipitation. Watershed model based estimates of ϵ_p are shown to be highly sensitive to model structure and calibration error.; This research introduces a general framework for the validation of deterministic watershed models. This framework tests the ability of the hypothesized model structure to reproduce the observed covariance structure of the input and output time-series without ever fitting the model to data or estimating the model parameters. The developed framework is tested on two annual watershed models (‘abc’ and ‘abcd’) by searching for the existence of unique parameter set(s) which can preserve the observed covariance structure. Both annual watershed models failed to reproduce the annual covariance structure, but aggregating from the monthly time scale using a monthly model better preserves the annual covariance structure. This study emphasises that model hypothesis testing (validation) should be performed prior to, and independent of, parameter estimation (calibration) yet deterministic models are usually validated after calibrating the model and observing the “goodness-of-fit” of the estimated model. (Abstract shortened by UMI.)...