A systematic approach for a multi-scale evaluation of dominant hydrological processes (Invited)

The concept of “dominant processes” suggests that simplification is the key for an effective/efficient way to model hydrological systems. However, the use of calibration-based hydrological models can lead to incorrect interpretation of the significance of some processes. We argue that dominant processes can be identified through the use of 1) calibration-free hydrological models that simulate the physics of hydrological systems across multiple scales, and 2) a more holistic view of basin response. In this work we systematically build calibration-free hydrological models by adding complexity and evaluate the importance of different hydrological processes on peak flow spatial-scaling. All the parameterizations in the model are data-based. We illustrate this concept using a case study for the 250 km2 Clear Creek basin during the Iowa 2008 extreme flood event. The simplest model applied is the Geomorphological Unit Hydrograph with time variable rainfall. We add complexities (incrementally) to flow propagation dynamics in the river network, spatio-temporal variability of rainfall and runoff generation, hillslope morphology and functioning, and soil dynamics. The results are evaluated across a large range of spatial scales. Finally we take the processes and parameters obtained at Clear Creek and apply them to the 32,000 km2 Iowa River Basin and compare the un-calibrated results to observed data. We find a close agreement, indicating that the approach successfully captured dominant processes.