HYDROLOGICAL MODELING OF BHOPAL LAKE CATCHMENT

Abstract

Like many other cities/villages in India, Bhopal, the capital of State of Madhya Pradesh also relies heavily for its water supply from a lake named as Bhopal Lake. Two rivers, Kolans and Niljhi, besides recessing flows from its urban catchments primarily feed this lake. Prior to this work, like other lakes in the country, the lake was seldorr investigated for its hydrological behavior, even the water balance of lake was much subjective in the absence of such hydrological investigations. With the financial support from Ministry of Water Resources of Government of India, a detailed hydrological inv igation of The Bhopal lake catchment was undertaken for four years beginning 2002. Information regarding catchment characteristics including land use pattern, geomorphology , rainfall, runoff, sediment yield of the catchment was collected. The present work primarily deals with modeling of rainfall-runoff and runoffsediment yield from Bhopal Lake catchment. As the appropriate data was not available for the catchment, a considerable time and effort was spent on generating the data. Use of Geomorphological Instantaneous Unit Hydrograph (GIUH) approach was given a priority as it accounts for heterogeneous characteristics of the catchment. As GIUH can be derived using different approaches available in literature, only some of selected approaches were considered in the work. Prediction using such GIUH were made to assess resulting runoff from different rainfall events. Since the existing approaches considering basin geomorphology did not satisfy in predicting runoff from the Bhopal Lake catchment, a new distributed approach considering geomorphology of the catchment was proposed for GIUH derivation for the Bhopal Lake catchment. In this work, catchment of Bhopal Lake has been first sub-divided into smaller subcatchments by identifying the drainage area for each sub-catchment on the basis of geo morphology of the sub-catchment. In total, 23 sub catchments are considered. First 12 subcatchments contribute water to river Kolans, sub-catchment 13 contributes water to river Niljhi, sub-catchments 14 to 16 are agricultural catchments and sub-catchments 17 to 23 are urban ii catchments. Contributions from sub-catchments 14-23 are non-point in nature and contribute along the periphery of the lake. For any particular sub-catchment, rainfall has been assumed to be uniform over space and time. Stream area of each stream has been considered to generate runoff due to rainfall-excess. Runoff generated by area of each stream first comes to the stream from whose area the runoff is getting generated and then reaches the outlet of drainage area or sub-catchment of which this stream belongs to. From the outlet of this drainage area, the flow generated by each area is routed through streams of higher order to reach the outlet of the final stream to give a GIUH at outlet of the higher order stream due to flow generating from an area of the particular stream considered. In this way GIUH from all streams have been generated first at the outlet of the sub-catchment to whom the stream belongs and then at final outlet of the drainage area of which the considered sub-catchment is a part of. Catchment characteristics, such as length, width, and mean slope of each stream and overland slope of the areal plane were considered in the analysis. This method has advantage that spatial variability of rainfall can be accounted for, and runoff can be generated from any particular pocket of the catchment over which the rainfall is occurring. Dispersion of flow at the outlet of the drainage area due to flow from generating area of each stream has been termed as geo-morphological dispersion. By the method adopted in this work, geo-morphological dispersion at any point in the catchment can be found out, which is useful in watershed conservation, development and management. To assess the accuracy of runoff predictions, lake water balance was carried out. In addition to runoff predictions using GIUH approach, sediment yield from the catchment of Bhopal Lake was assessed using the existing Singh and Krstanovic (1987) and proposed empirical models. As the results based on empirical models proved quiet encouraging, no attempt was made to test other approaches of sediment yield estimation. Considering that use of ANN models is getting popular in modeling of hydrological processes, its potential has also been tested in this work. For better performance ofANN models, need for having larger data base is projected