Prediction of Regional Ground Water Flow to Streams

Publisher Website
Google Scholar
Abstract
Database information on geology, hydrology, and hydrometeorology may form an excellent basis for studying ground water flow and seepage to surface water in a catchment. In a field case study of a 114 km2catchment, geological database information was used to determine layer thicknesses and boundary conditions as well as to parameterize a ground water flow model. Hydraulic head and stream flow data were used to estimate the model parameters by nonlinear regression. The uncertainty of the estimated parameters and of the predicted stream flow gains was quantified by individual likelihood method confidence intervals. During four stages of calibration, ranging from using only head data to also using an extensive set of measured stream flow gains, no parameter estimates changed significantly, but the number of parameters was increased from 12 to 14 in order to fit local stream flow gains. This indicates that the geology‐based parameterization is firm.Adding stream flows to the calibration data reduced the uncertainty of the estimated parameters significantly. However, the uncertainty of some of the parameters was significant even when an extensive set of measured stream flow gains and hydraulic heads was used to calibrate the model. Parameter uncertainty is reflected in the uncertainty of the predicted stream flow gains. When an extensive set of stream flow data was used during calibration, the prediction uncertainty is up to ±25% in large streams, and up to ±60% in smaller streams. The confidence intervals in general are skewed, and they are very skewed in the case where no stream flow measurements were used to calibrate the model.The case study shows that even when relatively extensive geological information and calibration data are available, there may be significant uncertainty connected with the prediction of local discharge of ground water to streams. Reducing uncertainty in such cases will require extensive field investigations in order to improve the definition of recharge areas and to describe the local fluxes and flow patterns in the aquifers.