On The Relationship Between Runoff Coefficient And Catchment Initial Conditions A. Longobardi a , P. Villani a , R.B. Grayson b and A.W. Western b a Department of Civil Engineering, University of Salerno, Fisciano (SA), Italy, alongobardi@unisa.it b Centre for Environmental Applied Hydrology and Cooperative Research Centre for Catchment Hydrology, Department of Civil and Environmental Engineering, The University of Melbourne, Vic 3010, Australia. Abstract: The runoff coefficient, estimated as the ratio between quick flow and rainfall volume, on an event basis, has been analyzed in an empirical framework, as a function of the initial catchment state conditions prior to an event, such as pre-event soil moisture and pre-event base flow. The resulting relationships have been tested for their utility in runoff prediction, focusing on the variability of the catchment response, dependent on the initial condition itself. The variability of the catchment response is also addressed in a conceptually based stochastic model for stream flow simulation and forecasting. The watershed response to effective rainfall is considered as deriving from the response of linear reservoirs in parallel, representing contribution to stream flow from the fast (surface flow and shallow subsurface flow) and slow (deep subsurface flow and groundwater flow) components. The parameter that indicates how the hydrograph is split between these components, responsible for the non linearity in the rainfall-runoff transformation and for difficulties in the runoff ratio estimation, has been estimated. This shows a seasonal dependence and, generally, a dependence on the initial state, prior to an event. Keywords: antecedent conditions; soil moisture; base flow; runoff coefficient 1. INTRODUCTION The proportion of total rainfall that becomes runoff during a storm event, represents the runoff coefficient. In the classical ‘rational method’ it is considered to be a constant, depending on characteristics of the drainage basin, such as surface cover (eg. Dooge, 1957). Some authors proposed a dependence of runoff ratio on the percentage of impermeable catchment area (eg. Schaake et al., 1967; Boughton, 1987). Hebson and Wood (1982), in their study assumed a constant runoff coefficient, interpreted as the percentage of contributing area to runoff generation. The runoff ratio variability is also well documented in the literature, even thought there is no clear conclusion about what factors govern this variability (eg. Gottschalk et al., 1998; Wainwright et al., 2002). The rainfall– runoff transformation is a non linear process. The most important cause of non linearity is represented by the effect of antecedent conditions, consequently the runoff coefficient depends on the initial conditions. It is well known that soil moisture is a major control on catchment response: antecedent soil moisture conditions have been used to represent variability of the CN in the SCS method for predicting direct runoff (eg. Rallison and Miller, 1981). But soil moisture measurements are usually not widely available and a lot of point measurements are needed to get either an average or an antecedent catchment state. Antecedent soil moisture conditions can be simulated using models (either lumped or distributed) provided standard climate data are available. Alternatively, some other easily measured hydrological variables that are even if only approximately related to soil moisture conditions, may be considered. In this study, we assess the role of antecedent conditions, using directly observed soil moisture and base flow (derived from daily flow records), as predictors of catchment state. The dependence of runoff ratio on antecedent conditions can also be assessed by means of a conceptual framework. We expect catchment state to be variable from event to event and we might expect seasonal variability to be particularly important in many climates. In this paper we show that the parameter of a conceptually based stochastic model for stream flow simulation, which considers the watershed response as deriving from the response of two linear reservoirs in parallel, is found to be seasonal dependent. The result in each section will be discussed distinguishing between different climatic regions.