EVALUATION OF THE CONCEPTUAL BASIS OF THE CLARK AND NASH MODELS

Abstract

A study to evaluate the conceptual basis of the Clark and Nash models for modelling rainfall-runoff process is presented herein. These popular models link only the excess rainfall and the corresponding surface runoff. Therefore, to adhere to these requirements, and to avoid any observation errors in the rainfall-runoff data, it was considered appropriate to simulate a number of rainfall events, using the Kinematic wave model on hypothetical impervious V-catchments of sizes varying from 0.1 king to 1.0 km2 for the evaluation study proposed herein. Hypothetical first order V-catchments were developed following the established law governing the geomorphological processes of such catchments. Study of these hypothetical rainfall-runoff events using the Nash and Clark models reveals that while the Clark model could successfully simulate many of these events very closely, the Nash model failed to do so based on the measure of Nash-Sutcliffe criterion. More over the Nash model under estimated the Kinematic wave model hydrograph. The plausible reasonings for such performances of these models may be attributed to the appropriate simulation of the translation and attenuation characteristics of the runoff process by the Clark model by, respectively, the time of concentration, T. and the storage coefficient, K parameters of the model; whereas, the Nash model is mainly attempting to simulate the attenuation process, and the translation process only as a consequence of attenuation process and not independently as modelled by the Clark model. Consequently the Nash model always underestimated the El peak flow. As a corollary of this study, it is revealed that for small catchments of the order of less than 0.3 km2 area, the Clark's storage coefficient is less than 0.05 Tc and, thus, leaving the translation as the dominant runoff process. Under such condition, the runoff simulated by the V-catchment resembles closely to the triangular hydrograph, a solution given by the popular Rational method used in urban hydrology for design peak flood estimation.