Validation approaches for field-, basin-, and regional-scale water quality models

Environmental issues such as global warming, hypoxia, and non-point source pollution of rivers and aquifers occur at scales which include the entire earth, the Gulf of Mexico, the Baltic Sea, the Mississippi and Amazon River basins, as well as smaller regions which cover states, provinces, and counties. The increasing availability of data layers at these scales through remote sensing and Geographic Information Systems (GIS) makes it possible to model transport processes at scales far removed from the traditional plot and field scales at which most transport models in soil physics were developed. This paper reviews and synthesizes the general approaches and concepts governing the use of water and solute transport models over a wide range of scales. Topics discussed include model selection, criteria for model calibration and validation, sources of error in modeling, non-linearity, spatial variability, non-uniqueness, and scale-transition techniques. The paper concludes that rigorous validation of models at the scale of large regions, basins, or continents is difficult for a variety of reasons. This does not preclude the value of modeling transport processes over large regions. If proper procedures are followed for model selection and calibration, then there can still be great value in using the model to investigate various scenarios, which would be impossible to study experimentally.