Abstract:
The estimation of the extreme flood at a site is required for the design of a variety of urban planning and river engineering works. An accurate estimatio of extreme flows for the associated recurrence interval is difficult to obtain if the length of the available stream flow records at the site of concern is shorter than the recurrence interval of interest. An even greater difficulty occurs if there is no flow record available at the site of interest. To cater the problem arising due to an insufficient length of data record, the trade off between the spatial and the temporal characterization of extreme flows can be effected through the use of regional flood frequency analysis. Regional flood frequency analysis facilitate the estimation of an extreme flow value at a location for which limited flow data exist, based on an extreme flow relationship derived using the information from basins with similar hydrologic responses.
This report describes the study of regional flood frequency analysis using peak flood series data of 34 gauging stations of Krishna basin using the following methods,
i) Index-Flood method
ii) PWM based EVI distribution
iii)PWM based GEV distribution
iv) PWM based Wakeby distribution based on at site and regional data combined.
Out of 34 sites, 6 sites were omitted after USGS homogeneity test. From the remaining 28 sites, 26 sites were considered for analysis under 3 different sub-group i) medium catchments, ii) large catchments and iii) considering the basin as whole. In order to evaluate the fitting performance of different methods used, some of the error functions respectively their descriptive ability are computed. The results indicate that, the Index-Flood approach and PWM based EVI distribution are best suited for medium catchments for the basin as whole. However, it is recommended to include the other physiographic characteristics also for developing more rational regional flood formulae using good data base.
