There are a growing number of hydrologic models with fully 2D and 3D formulations that seek to combine surface and subsurface flow. These complex models fall into two broad categories: integrated and coupled, with about ten of these models available. Additionally, many of these models are coupled to land-surface energy balances, biogeochemical and ecological process models and to atmospheric models. While there is an increasing need and application for large scale, complex models of environmental function for hydrologic prediction and environmental understanding, no formal verification and/or benchmarking of these models has been performed.
This workshop will gather many of the authors of these coupled, integrated environmental models for a first-of-a-kind model intercomparison. The workshop will perform a suite of benchmarking problems with each model and will run a series of idealized problems that emphasize and identify various model components and interactions. The benchmarking problems can then be used in subsequent future benchmarking intercomparison exercises and will be provided freely to anyone in the hydrologic community. This intercomparison will be published in a peer reviewed journal.
The field of hydrology is increasingly integrated and interdisciplinary. Hydrologists tasked with addressing complicated problems involving many interconnected systems often turn to computer programs, called models, to simulate the flow of water in our natural and built environments. Today, the number of hydrologic computer models and their applications is constantly growing. Current models can be applied to areas ranging from small watersheds to large river basins and can simulate the movement of water on the ground (such as streams and rivers) and below it (such as groundwater). In addition to simulating the flow of water on and below the earth’s surface, complex hydrologic models are even capable of predicting movement of water in plants, as snow and through the atmosphere. Although they are increasingly applied to answer a range of environmental questions, very little model benchmarking (i.e. comparison of model results for a standardized set of inputs) has been performed. For this project, a community of hydrologic modelers gathered at the Colorado School of Mines for a two-day workshop to compare seven hydrologic models based on a series of benchmark problems. The selected models were chosen because they all solve the same basic equations governing the flow of water on and below ground. Although these models are all fundamentally the same, they use slightly different mathematical approaches. The community agreed upon the benchmark problems and chose to start very simple and increase in complexity. The inputs needed to run every benchmark problem are published on the web and in a peer-reviewed, scientific journal so that others scientists who were not at the workshop may also use them. In general, the models generated answers to the benchmark problem that agree. For the simpler benchmark cases, all of the model answers are basically the same. For the more complicated benchmarks, where groundwater and streamflow interact more rapidly, the models agree qualitatively, though there were some significant quantitative differences. While there were differences between models, the fact that all the models demonstrate the same qualitative behavior builds confidence in their use for water and environmental applications. The test cases generated through this effort provide a framework for model evaluation as the field of hydrologic modeling advances and new tools are introduced. However, there is still more work to be done; this intercomparison exercise is the first step in a continuing effort that will develop additional benchmarks of increasing complexity. Future efforts will continue to build the community of hydrologic modelers, helping to guide scientific understanding as hydrologic models, and the problems they are used to address, increase in complexity.
