On the testing of fully integrated surface–subsurface hydrological models

Studies employing integrated surface–subsurface hydrological models (ISSHMs) have utilized a variety of test cases to demonstrate model accuracy and consistency between codes. Here, we review the current state of ISSHM testing and evaluate the most popular ISSHM test cases by comparing the hydrodynamic processes simulated in each case to the processes found in well‐characterized, real‐world catchments and by comparing their general attributes to those of successful benchmark problems from other fields of hydrogeology. The review reveals that (1) ISSHM testing and intercode comparison have not adopted specific test cases consistently; (2) despite the wide range of ISSHM metrics available for model testing, only two model performance diagnostics are typically adopted: the catchment outflow hydrograph and the catchment water balance; (3) in intercode comparisons, model performance is usually judged by evaluating only one performance diagnostic: the catchment outflow hydrograph; and (4) ISSHM test cases evaluate a small number of hydrodynamic processes that are largely uniform across the model domain, representing a limited selection of the processes of interest in well‐characterized, real‐world catchments. ISSHM testing would benefit from more intercode comparisons using a consistent set of test cases, aimed at evaluating more catchment processes (e.g. flooding) and using a wider range of simulation diagnostics (e.g. pressure head distributions). To achieve this, a suite of test case variations is required to capture the relevant catchment processes. Finally, there is a need for additional ISSHM test problems that compare model predictions with hydrological observations from intensively monitored field sites and controlled laboratory experiments. Copyright © 2012 John Wiley & Sons, Ltd.