FLOW PATH DIMENSIONALITY AND HYDROLOGICAL MODELING

Increasingly, research is indicating that subsurface flow paths govern ion transport within river catchments. Distributed prediction of these solute flow paths in typically heterogeneous catchments must inevitably be highly uncertain without some identification of a spatial structure relating small-scale measurements of soil properties to flow predictions distributed over large catchments. To date, the evidence for profile and catenal structure within soil hydrological properties and resultant solute flow paths is not fully embraced by the hydrological community. As a consequence soil parameters are often poorly distributed within catchment-scale distributed models. This paper seeks, first, to generalize the disparate sources of evidence of parameter and flow path structure within the profile-ward and catena-ward dimensions. Second, to outline how much of this structure has been incorporated into previous hydrological simulations using distributed models, and third, to examine the physical basis of attempts to simplify parameter and flow path dimensions using pedological classifications. The available evidence suggests that a considerable number of world soils show profile-ward structure within their hydrological properties and resultant flow paths. Changes in profile-ward patterns along catenal sequences remain uncertain. The Plynlimon region of mid-Wales has been the focus for many detailed studies of solute flow paths, catchment-scale model simulations, soil property characterizations and soil classification. Comparison of these studies suggests that most model simulations and hydromorphic classifications of soil taxa fail to distinguish adequately between soil horizons and soil types with markedly different property distributions. Preliminary analysis, however, suggests that by using a catena based criterion to classify the hydromorphic characteristics of soils, soil elements with distinct patterns of properties and flow paths can be identified. This might suggest that the accuracy of distributed predictions of ion movements within river catchments could be greatly improved by the derivation of profile-specific patterns in soil properties. These profile-specific effective parameters need to be derived from measurements over a range of scales, including individual layers, profiles and complete catenas.