Beyond Model Metrics: the Perils of Calibrating Hydrologic Models

The multi-metric assessment of model performance in a dominantly single-domain modeling approach (i.e., surface) may not be sufficient to gauge the validity of the model to represent the hydrologic system. Consequently, rating metrics can mathematically validate model results as satisfactory even when some of the simulated hydrologic processes are incorrectly represented during the calibration process. In this paper, we argue that to properly represent the hydrologic system, calibration tasks focused on modifying model parameters should account for equifinality, model inadequacy, and constraint inadequacy. To demonstrate our argument, we conducted a traditional calibration using the Soil and Water Assessment Tool (SWAT) coupled to the Modular Finite-difference Flow Model (MODFLOW) to simulate the hydrologic processes in the Fort Cobb Reservoir Experimental Watershed (FCREW) in central western Oklahoma. The results indicated that model calibration based on metrics from a single domain, in this case using SWAT, did not necessarily guarantee an appropriate representation of water circulation. Moreover, unsupervised calibration techniques based on the objective-function optimization (e.g., inverse calibration using SWAT-CUP SUFI2 algorithm) may not be sufficient if model constraints in multiple domains are not properly set, resulting in distorted parameters. This practice can result in the misrepresentation of water circulation and can have adverse impacts on environmental model results used to support policies and decisions.