TR(BR)-4/97-98 : Sensitivity analysis for hydrological parameters of flood hydrograph

Abstract

In design flood estimation approach using unit hydrograph techniques, the design unit hydrograph parameters are used together with design storm for estimation of design flood. The standard project storm (SPS)/probable maximum precipitation (PMP) depth obtained are in the form of accumulated amounts of the given durations. These are to be distributed in time according to the realistic time distribution which has been observed for the severe storms in the region. In order to obtain the most critical sequence of the rainfall depths, these are arranged in some chronological pattern of rainfall increments characterizing severe storms occurred in a catchment in producing maximum probable peak flood for the involved rainfall amount. In general, it is observed that maximum duration of severe spell of rainfall possible within severe most long duration storms is of the order of 10-15 hours. The arrangement of rainfall in one bell is aimed at arriving the maximum possible peak for the given volume of rainfall excess. Such an extreme possibility of rainfall occurrence in one spell of one, two or three days duration of design storm is not the characteristic of the nature of rainfall and results in higher design flood peaks. Broadly, it can be approximated to one or two bells per day but not over the full period of the long duration storms of high intensity. It has also been mentioned in Central Water Commission (1993) that it is reasonable to consider this aspect in design flood since it is closer to the meteorologic reality.

In this study, the Clark's model has been used for derivation of the I-hour unit hydrograph for a catchment. Design flood estimate has been obtained using the calibrated parameters of the Clark model together with the time area diagram corresponding to a 48-hour design storm which has been temporally distributed and critically sequenced as a single bell giving due consideration to the ordinates of the unit hydrograph. Sensitivity of the design flood estimate has also been studied considering the different scenarios viz. (i) varying the design unit hydrograph characteristics (ii) considering the design storm as one bell, two bells and four bells for the 48 hour storm, and(iii) varying the design loss rate. An objective methodology has been suggested and applied for estimation of design unit hydrograph parameters after increasing the peak of the average unit hydrograph to account for the hydraulic conditions of the channel and the catchment for the extreme flooding situations, which would occur due to the occurrence of the design flood. The design flood peak is not much sensitive to the change in the design loss rate. The sensitivity runs taken considering the various design storm patterns show that the design flood estimates are very much dependent on the temporal distribution and the critical sequencing of the PMP values.