Uncertainty Analysis of the Variable Parameter McCarthy-Muskingum (VPMM) Method with Presence of Lateral Flow
Abstract:Uncertainty analysis of the estimate of a hydrological model is a required exercise for the risk management linked to the variable of interest. This study subjects the Variable Parameter McCarthy-Muskingum (VPMM) method recently proposed by Perumal and Price (2013) to uncertainty analysis. The VPMM method has been developed based on the assumption that there exists no lateral flow in the river stretch where it is employed for routing. But in this study this method is applied for the study of flood wave movement in a 24.2 km stretch between Rottweil and Oberndorf of Neckar River in Germany in the presence of lateral inflow. The study also proposes a general procedure for simulating flood events with the consideration of lateral flow in the reach. The cross sectioned information of the considered river stretch is estimated by the Robust Parameter Estimation (ROPE) algorithm. ROPE algorithm is used to get the best performing parameters set of bed width (Trapezoidal section) and side slope. As the evaluation of VPMM is done with the help of Nash-Sutcliffe efficiency criterion, this study uses it as an objective function to check the performance of the method with different data sets obtained using the ROPE algorithm. The uncertainty associated with parameter K and due to the presence of lateral flow is checked by the Jackknife method.
All the 26 flood events observed from the Neckar catchment from 1999 to 2004 have been used for the analysis of the VPMM method. When inflow and outflow hydrographs for lateral flow estimation are used, performance of the VPMM method as per N-S efficiency criterion can be up to 97.061 %. By the analysis of all 27 available flood events, a relationship between total rainfall and total loss is obtained, and the value of loss obtained from the developed relationship can be used to simulate outflow hydrograph with the maximum N-S efficiency of 93.812 %.