STOCHASTIC STREAMFLOW GENERATION USING DAILY RAINFALL DATA

Abstract

The adequate development of water resources requires the use of planning techniques which depend to a large extent on reliable estimates of the key hydrologic variables. One of the most important of these is the streamflow at the point of interest of the river. The information that is required very often are the quantity and availability and the frequency'' of occurance of floods and droughts. Although many streams have been gauged to provide continuous streamflow records, very often planners and designers face with little or no available streamflow information. Many investigators have developed techniques for synthetic generation of strearnflow sequences using the available data of streamflow. For areas with inadequate streamflow data, techniques have been develop-ed, which syntheize the sequences of rainfall data and use such generated sequences to obtain streamflow sequences using suitable rainfall-runoff relationships or conceptual models. Such a technique which combines ' synthesis' and 'simulation' enables synthetic generation of number of data samples of periods longer than that of historical data for better design of projects by providing possible patterns of extreme cases. Very often the available data consist of very short record of streamflow, say five to six years. The present study has been devoted to evolve a stochastic daily streamflow model using very limited data of rainfall and runoff and also to examine the performance of the approach in such a case of limited data.