Date of Award:


Document Type:


Degree Name:

Doctor of Philosophy (PhD)


Biological and Irrigation Engineering




The main objective of this research was to develop and utilize a coupled surface water groundwater model to simulate hydrological responses of watersheds. This was achieved by coupling the U.S. Geological Survey (USGS) groundwater flow model, MODFLOW, and the rainfall runoff model, TOPMODEL, in one case study and coupling MODFLOW with a networked version of TOPMODEL called TOPNET in another case study. The model coupling was achieved using the InCouple approach, which utilizes Potential Coupling Interfaces (PCIs) that are abstractions from model flow diagrams that expose only those aspects of a model relevant to coupling. Coupling the rainfall-runoff models to MODFLOW involved development of a routine relating the spatial discretization of MODFLOW to TOPMODEL and similarly MODFLOW to TOPNET and development of a feedback scheme where groundwater and surface water interact in the soil zone. The key coupling concept was replacing the wetness index-based depth-to-water table concept of TOPMODEL with the groundwater heads simulated by MODFLOW. In the MODFLOW-TOPMODEL coupling, using data for the Tenmile Creek watershed, for the period, 1968 to 1972, it was concluded that the coupled model was able to continuously simulate the stream flow. However, the coupled model under predicted stream flow and did not agree well with observations in a point wise comparison. A mean coefficient of efficiency of 0.54 was obtained between simulated and measured stream flow. Only 24% of received precipitation was observed as baseflow and this shows that there is limited interaction between surface water and groundwater in the watershed. It was demonstrated using the coupled model that the lateral flow processes and the interactions between groundwater and surface water have a major importance for the water balance. For the Big Darby watershed, for the period 1992 to 2000, the coupled model adequately predicts the stream and groundwater flow distribution in the watershed. After model calibration, simulated groundwater showed the greatest residual variance, attributed to model error and uncertainty in model parameters. Model fit efficiencies of 0.61 and 0.69 were obtained for simulating stream flow measured at two gaging stations. The overall watershed hydrologic budget also showed small mass balance errors using the coupled model. However, the study also shows the need for further research in regard to constraining the groundwater recharge parameter which links the models.


This work was revised and made publicly available electronically on July 21, 2011