9406436 McDonnell The knowledge of the hydrologic flowpaths is critical to understanding runoff of generation and the transport of pollutants on to modeling the linkages between hydrologic flux and geochemical flux. However, models and field studies are not well linked in hydrological investigations. Furthermore, attempts to introduce such field observations into models have faced the inevitable compromise between the value of additional data and the identifiability of parameters required to take account of the data. New research on runoff flowpaths has shown that spatially- distributed models are required to capture the key hydrological and geochemical processes. This work will develop a new synergistically linked hydrologic/geochemical model of how: (1) waters evolve chemically and isotopically along the different flowpaths and (2) specific flowpaths produce distinctive "signatures" related to the topographic position and hydrologic history of the water. Uncertainty analysis will be used to assess uncertainty in the model parameters. Our objectives are: (i) to develop a hydrogeochemical model within which the dynamics of stream chemistry and flowpaths can be related to fieldwork, (ii) to assess model structure/parameter uncertainties and identify fieldwork strategies to reduce such uncertainties, (iii) to make spatially-distributed measurements in the field to validate spatially-distributed estimates of soil moisture, water table elevations, isotope and solute concentrations, (iv) to examine the controls on geochemical/isotopic evolution along flowpaths by using physical characteristics and responses of the catchment, and (v) to use the new model as a tool for testing specific hypotheses on flowpaths and mixing overall this study represents an innovative attempt (i) to link a evolutionary reaction-path model constrained by both solute isotope and geochemical data within a topographically-driven hydrologic model at the hillslope scale and (ii) to de velop a new type of hydrogeochemical model in a form constrained by the interaction of the field campaign and the assessment of model uncertainty. The work would make it easier to apply fully distributed models in other study areas.
