Application of RS Data for Reservoir Sediment Profiling using Latin HypercubeOne at Time (LH-OAT) Technique

Abstract

Assessment of reservoir sediment and its distribution are essential for irrigation releases, reservoir operation and assessment of soil conservation measures in the catchment. The empirical area reduction method suggested by Borland and Miller, 1960 is commonly used to determine sediment prediction based on limited data from reservoirs in USA. In the present study, an attempt has been made to optimize the parameters (C, m and n) of empirical area reduction method using Latin Hypercube, one parameter at a time (LH-OAT) technique for determination of sediment distribution in Ravishankar Sagar reservoir situated on river Mahanadi in India having gross storage of 909.54 Mm3 and used mainly for irrigation purposes. The results obtained from optimized sediment distribution were compared with observed distribution through hydrographic survey (HS-2003) and further used to predict future sediment in reservoir. The proposed optimization technique conferred root mean square error of 10.79 Mm3 in comparison with conventional technique which provided the same as 11.85 Mm3 when both were compared with hydrographic survey. The best-fit revised water spread line/curve obtained from remote sensing results was by extrapolating in dead storage zone to determine new zero elevation and further computation of revised capacities at different levels. The results of remote sensing approach (RS-2003) were used in optimization technique to compare with hydrographic survey and found a close agreement of proposed remote sensing approach for determination capacity loss in dead storage zone also