Water Resources Management (2005) 19: 505–520 DOI: 10.1007/s11269-005-6811-2 C  Springer 2005 Discharge Rating Curve Extension – A New Approach CHANDRASEKARAN SIVAPRAGASAM 1,∗ and NITIN MUTTIL 2 1 Department of Civil Engineering, Mepco Schlenk Engineering College, Tamilnadu, India 626005; 2 Department of Civil Engineering, University of Hong Kong, Pokfulam Road, Hong Kong ( ∗ author for correspondence, e-mail: sivapragasam@yahoo.com) (Received: 20 April 2004; in final form: 17 November 2004) Abstract. It is often necessary to have stage discharge curve extended (extrapolated) beyond the highest (and sometimes lowest) measured discharges, for river forecasting, flood control and water supply for agricultural/industrial uses. During the floods or high stages, the river may become inac- cessible for discharge measurement. Rating curves are usually extended using log–log axes, which are reported to have a number of problems. This paper suggests the use of Support Vector Machine (SVM) in the extrapolation of rating curves, which works on the principle of linear regression on a higher dimensional feature space. SVM is applied to extend the rating curves developed at three gauging stations in Washington, namely Chehalis River at Dryad and Morse Creek at Four Seasons Ranch (for extension of high stages) and Bear Branch near Naselle (for extension of low stages). The results obtained are significantly better as compared with widely used logarithmic method and higher order polynomial fitting method. A comparison of SVM results with ANN (Artificial Neural Network) indicates that SVM is better suited for extrapolation. Key words: artificial neural network, forecasting, rating curve, regression, support vector machine 1. Introduction Stage-discharge relations have been used for well over a century and many authors have reported the results of investigations of different issues related to the devel- opment and application of these relations. The rating curves or stage-discharge relationships are usually developed from periodic measurements of stage and dis- charge taken around the mean stage in the river. Discharges are computed from velocity measurements and are plotted against the stages to obtain the rating curve. During floods or high stages, measurement conditions offer serious practical dif- ficulties, and therefore, it is often necessary to extend the rating curve beyond measured highest discharges. Low flow extrapolation is required for management of domestic, industrial and agricultural water needs (Fenton, 2001). Of the different methods of rating curve extension, methods accounting for hydraulic conditions are preferred because of theoretical support in explaining the phenomena. Consequently, many investigators have recommended use of such methods. Herschy (1999) recommended the ‘Velocity–area method’ in which a curve fitted to relate the stage to the mean velocity is extrapolated and combined with a curve fitted to relate the stage to the cross-sectional area to extrapolate the