Simulating complex flow and transport dynamics in an integrated surface-subsurface modeling framework

A fully-integrated surface-subsurface flow and transport model is applied to a 17 km 2 subcatchment of the Laurel Creek Watershed within the Grand River basin in Southern Ontario, Canada. Through past and ongoing field studies, the subcatchment is reasonably well characterized and is being monitored on an ongoing basis. In addition to diverse land-usage and surface cover and more than 65 m of topographic relief, the watershed is underlain by a complex interconnected sequence of sand and gravel aquifers that are separated by discontinuous clayey aquitards. A steady-state condition was achieved in the model by calibrating the subsurface flow field to 16 observation wells where long-term hydraulic head data were available, while simultaneously establishing a level of baseflow discharge on the surface regime approximating the level observed at the beginning of the transient simulation period. The model is then subjected to several hundred hours of rainfall data and the resulting discharge hydrographs are compared with the measured hydrographs. The calculated subsurface hydraulic head distribution and surficial rainfall-runoff responses, respectively, were shown to agree moderately well with those observed in the system during this period. The impact of an upland surficial contaminant source discharging along a reach of a small stream within the subcatchment was also examined. Results showed that short-duration, high-intensity concentration peaks were not captured if annual or monthly average rainfall was used as input. The hydraulic head and concentration variations due to short-duration rainfall variations showed a muted response with increasing depth below the streambed due to the natural smoothing in the hydraulic response and to dispersion and diffusion of the solute, respectively. Discrete daily precipitation events were also found to cause rapid changes in the calculated water and solute exchange fluxes. The variability and sensitivity of these near-stream processes to the temporal resolution of rainfall input, specifically the concentration and solute exchange flux responses, may be significant in the prediction of health risks to aquatic habitats. Overall, it is concluded that the model is capable of reproducing surface and subsurface hydrodynamic processes at the subcatchment scale although the results could be better through improved parameterization of the subcatchment and the manner in which the model simulates evapotranspiration processes.