Calibration and validation of a general in®ltration model Surendra Kumar Mishra, 1 Shashi Ranjan Kumar 1 and Vijay P. Singh 2, * 1 National Institute of Hydrology, Roorkee-247667, U.P., India 2 Department of Civil and Environmental Engineering, Louisiana State University, Baton Rouge, LA 70803-6405, USA Abstract: A general in®ltration model proposed by Singh and Yu (1990) was calibrated and validated using a split sampling approach for 191 sets of in®ltration data observed in the states of Minnesota and Georgia in the USA. Of the ®ve model parameters, f c (the ®nal in®ltration rate), S o (the available storage space) and exponent `n' were found to be more predictable than the other two parameters: m (exponent) and a (proportionality factor). A critical examination of the general model revealed that it is related to the Soil Conservation Service (1956) curve number (SCS-CN) method and its parameter S o is equivalent to the potential maximum retention of the SCS-CN method and is, in turn, found to be a function of soil sorptivity and hydraulic conductivity. The general model was found to describe in®ltration rate with time varying curve number. Copyright # 1999 John Wiley & Sons, Ltd. KEY WORDS in®ltration model; calibration; validation; split sampling; curve number; SCS-CN method; sorptivity; hydraulic conductivity INTRODUCTION In®ltration is an integral part of the rainfall±runo process whose modelling is required for planning and design of water resource systems. A number of in®ltration models have been developed over the years. These models can be classi®ed as physically based models which are based on soil physics, conceptual models based on systems theory and empirical models based on ®eld or laboratory experiments. Examples of physically based models are those of Philip (1957, 1969), Mein and Larson (1971), Morel-Seytoux and Khanji (1974), Parlange (1974), and Smith and Parlange (1978). The models of Green and Ampt (1911), Kostiakov (1932), Horton (1938), Holtan (1961) and Overton (1964), Dooge (1973), Smith (1972), Collis-Groege (1977) and Singh and Yu (1990) are conceptual. The Soil Conservation Service (1956) curve number (SCS-CN)-based in®ltration model (Mishra and Singh, 1998), the Hydrologic Engineering Center (HEC, 1981) loss model, the hydrograph model (Dunin, 1969) and the Snyder (1971) model are examples of empirical models. Conceptual models have been derived using simple hypotheses and veri®ed with the results of ®eld or laboratory experimentation. These models have gained much popularity in recent years (Singh and Yu, 1990). The usefulness of a model in ®eld applications depends on how precisely the parameters of the model can be predicted from the information on physically measurable soil characteristics and ®eld conditions. It is, therefore, useful to calibrate and validate a model which is more general in form and broadly represents a wide variety of available in®ltration models. The Singh and Yu (1990) conceptual model, derived using a systems approach, is one such general in®ltration model, for it specializes in the models of Horton, Kostiakov, Overton, Green and Ampt and Philip, among others. The objective of this paper is to calibrate and validate this model for various soil types. CCC 0885±6087/99/111691±28$17 . 50 Received 12 June 1998 Copyright # 1999 John Wiley & Sons, Ltd. Revised 29 October 1998 Accepted 25 November 1998 HYDROLOGICAL PROCESSES Hydrol. Process. 13, 1691±1718 (1999) *Correspondence to: Professor Singh, Department of Civil and Environmental Engineering, Louisiana State University, Baton Rouge, LA 70803-6405, USA.