Increased irrigation in Kansas and other regions during the last several decades has caused serious water depletion, making the development of comprehensive strategies and tools to resolve such problems increasingly important. This paper makes the case for an intermediate complexity, quasi-distributed, comprehensive, large-watershed model, which falls between the fully distributed, physically based hydrological modeling system of the type of the SHE model and the lumped, conceptual rainfall-runoff modeling system of the type of the Stanford watershed model. This is achieved by integrating the quasi-distributed watershed model SWAT with the fully-distributed ground-water model MODFLOW. The advantage of this approach is the appreciably smaller input data requirements and the use of readily available data (compared to the fully distributed, physically based models), the statistical handling of watershed heterogeneities by employing the hydrologic-response-unit concept, and the significantly increased flexibility in handling stream–aquifer interactions, distributed well withdrawals, and multiple land uses. The mechanics of integrating the component watershed and ground-water models are outlined, and three real-world management applications of the integrated model from Kansas are briefly presented. Three different aspects of the integrated model are emphasized: (1) management applications of a Decision Support System for the integrated model (Rattlesnake Creek subbasin); (2) alternative conceptual models of spatial heterogeneity related to the presence or absence of an underlying aquifer with shallow or deep water table (Lower Republican River basin); and (3) the general nature of the integrated model linkage by employing a watershed simulator other than SWAT (Wet Walnut Creek basin). These applications demonstrate the practicality and versatility of this relatively simple and conceptually clear approach, making public acceptance of the integrated watershed modeling system much easier. This approach also enhances model calibration and thus the reliability of model results.
View Complete Article
You will need the Acrobat PDF Reader, available free from Adobe, to read the article.
Back to main page...