06-Stage Discharge Analysis of Dhond River Using Adaptive Neuro-Fuzzy Inference System (ANFIS).

Abstract

Stream flow information is important for effective and reliable planning and management of various water resources activities and the assessment, management and control of water resources can be effective if accurate and continuous information on river-flow is available. Generally a network of river gauging stations provides continuous information on river stage and sparse information of corresponding discharges. Thus, the continuous discharge data corresponding to observed gauge can be obtained by developing a stage discharge relationship and using this relationship to convert the recorded stages into corresponding discharges. This relationship is determined by correlating measurements of discharge with the corresponding observations of stage. However, under certain conditions (flatter gradients and constricted channels) the discharge for a flood on a rising stage differs from that on the falling stage. This phenomenon is called hysteresis and results in a looped Stage—discharge curve (Tawfik et al., 1997) for floods with different stage— discharge relations for rising and falling water stages. Rating curve development approaches can be categorized into three main groups: the single curve approach, the rising and falling approach, and the Jone's approach (Tawfik et al., 1997). DeGagne et al. (1996) documented the process of developing a decision support system for the analysis and use of stage—discharge rating curve while other possible models have been proposed by Gawne and Simonovic (1994) and Yu (2000). The functional relationship between stage and discharge is complex and cannot always be captured by these traditional modeling techniques (Bhattacharya and Solomatine, 2005). In the real world, stage and discharge relationship do not exhibit simple structure and are difficult to analyze and model accurately. Therefore, it seems necessary that soft computing methods e.g. artificial neural network (ANN) and fuzzy logic, which are suited to complex non-linear models, be used for the analysis. There are several applications of ANNs in stage—discharge modeling. Jain and Chalisgaonkar (2000) used three layer feed forward ANNs to establish stage-discharge relationship. Bhattacharya and Solomatine (2005) have found that the predictive accuracy of ANN model is superior than the traditional rating curves. The effectiveness of an ANN with a radial basis function was explored by Sudheer and Jain (2003). The ANN based approaches have also provided promising results in modeling loop rating curves (Jain and Chalisgaonkar, 2000; Sudheer and Jain, 2003).

The purpose of this study is to investigate and explore the potential of an alternate soft computing technique for stage discharge modeling based on fuzzy logic. The ability of fuzzy logic to model nonlinear events makes it even more important to investigate its ability to model stage discharge relationship. Uncertainty in conventional gauge—discharge rating curves involves a variety of components such as measurement noise, inadequacy of the model, insufficiency of river flow conditions, etc. Fuzzy logic based modeling approach has a significant potential to tackle the uncertainty problem in this field and to model nonlinear functions of arbitrary complexity. Other advantage of fuzzy logic is its flexibility and tolerance to imprecise data (Zadeh, 1999). Fuzzy rule based modeling is a qualitative modeling scheme where the system behavior is described using a natural language (Sugeno and Yasukawa, 1993). The transparency of the fuzzy rules provides explicit qualitative and quantitative insights into the physical behavior of the system (Coppolaet al., 2002). The application of fuzzy logic as a modeling tool in the field of water resources is a relatively new concept although some studies have been carried out to a limited extent and these studies have generated considerable enthusiasm. Fuzzy rule based modeling has been attempted in water resources management, reservoir operation, flood forecasting and other areas of water resources analysis (Bardossy and Duckstein, 2002; Fontane et al., 1997; Kindler,1992; Mujumdar. and Sasikumar, 2002; Panigrahi and Mujumdar, 2000; Sasikumar and Mujumdar, 1998; Deka and Chandramouli, 2003; Lohani et al., 2005). This paper is concerned with the application of an emerging, powerful soft computing technique fuzzy logic to stage discharge rating curves.