LONG TERM HYDROLOGICAL SIMULATION USING SCS-CN METHOD

Abstract

A model capable of simulating runoff response of watershed resulting from long term rainfall could be an essential tool for the study of quantitative and qualitative assessment of stream flow and water resource management. In the present study, a lumped long term hydrologic simulation model based on SCS-CN method is developed. The developed model operates on daily time step and requires daily rainfall as input to generate stream flow. The model incorporates different sub-components for computation of surface runoff, evapo-transpiration, interflow and base flow. MATLAB is used to implement the developed model. The developed model is linked with the Shuffled Complex Evolution algorithm (SCE-UA) for optimizing parameters of the model. To test the performance of the developed model, five watersheds located in North-East state of Tripura and one watershed in Western Ghats in India having different topography and climatic conditions are used. It is worth mentioning here that, no studies are reported on the mountainous river basins of Tripura on a long term basis to assess the hydrological behaviour of watersheds in this region. Collected daily rainfall-runoff data from study watersheds are first analysed to check consistency, trend and runoff coefficient. Based on average runoff coefficient of study watersheds, these are categorised into dry, average and wet to express their runoff producing characteristics to assess performance of the model vis-a-vis runoff coefficient. A split sampling technique is used to partition available data into two sets to calibrate the model using one set of data and validate the model on other set of observed data. From the study, it is observed that the model performed better on wet watersheds and satisfactory on average watersheds. Whereas for dry watersheds, its performance is less than satisfactory with significant differences between annual simulated and observed runoff. Using the results from the proposed model, dominant/dormant processes involved in watersheds runoff VIII generating mechanism have also been identified. Study of outputs from different sub-components of the model reveals that study watersheds being located in the humid region, evapotranspiration is a dominant process. Study also suggests that the ground water flow to be quite significant as a major contribution to the total flow is contributed from the ground water store. The sensitivity analysis is also performed for the parameters used in the developed model to obtain information about the model's responses to perturbations in parameter values. The five watersheds taken up for study cover all range of varied watershed characteristics of the region and it is hoped that the developed model would be useful in understanding the runoff behaviour of similar watersheds in North-Eastern region. ix