CS(AR)-206 : Relationship between frequency of rainfall and frequency of flood for a catchment of upper Narmada and Tapi subzone-3(C)

Abstract

Floods of various return periods have been estimated for the catchment defined by Bridge No. 253 of the Upper Narmada and Tapi Subzone 3 (c) using various methods involving frequency analysis of rainfall and frequency analysis of annual maximum peak floods computed from the annual maximum excess rainfall of the design storm duration. The floods of various return periods have also been computed using the regional flood frequency analysis approach based on the observed annual maximum peak flood record for 13 gauging sites of the Subzone 3(c). Sensitivity analysis has also been conducted by increasing and decreasing the peak of the unit hydrograph, which has been used to convert the excess rainfall hyetographs into direct surface runoff hydrographs for identifying the peak values of floods.

The analysis carried out based on rainfall data shows that the floods are under estimated by 4.3% to 5.7% for the return periods of 2 to 200 years, by frequency analysis of floods as compared to the frequency analysis of rainfall. The rainfall data used in the study is of the limited record length of 19 years of one raingauge station only; hence the results of the study may be considered as indicative only, and detailed studies with long term record for a large number of catchments should be carried out for drawing more realistic conclusions.

The regional flood frequency methods used in the study, viz. SREV1 SRGEV and SRWAKE are based on at site and regional data; whereas, RGEV method is based on 'regional data alone. Flood estimates obtained by these methods show a deviation of -5.4% for SREV1, 0.4% for SRGEV, -1.4% for SRWAKE and 7.5% for RGEV methods for the return period of 50 years. The deviation varies from -3.1% to 13.5% for the return period of 100 years. For the return period of 200 years the deviation is -1.4% for SREV 1, 11.3% for SRGEV, 2.8% for SRWAKE and 19.1% for RGEV method. Percentage deviations for flood estimates by the SRWAKE method, for the return periods of 50, 100 and 200 years are -1.4%, 1.2% and 2.8% respectively; which show that the flood frequency estimates obtained by SRWAKE method are very close to the flood frequency estimates obtained by the method based on frequency of rainfall [RAIN(CS) .

While conducting sensitivity analysis, when peak of the regional unit hydrograph is increased by 20%, it is observed that the flood estimates for the various return periods increase with respect to flood estimates computed by the respective methods, considered with the actual peak of the regional unit hydrograph by about 11.5% in case of the RAIN(CS) method, by about 6.5% for FLOD(CS) method, by about 14.5% for RAIN method and by about 10.5% for FLOD method. When peak of the regional unit hydrograph is decreased by 20%, it is observed that the flood estimates for the various return periods decrease with respect to flood estimates computed by the respective methods, considered with the actual peak of the regional unit hydrograph by about 14% in case of the RAIN(CS) method, by about 8.5% for FLOD(CS), by about 13% for RAIN method and by about 13% for FLOD method.