Abstract:
In this paper, flood frequency analysis has been carried out for Sub-Himalayan region (Zone-7) using peak flood series data of ten small and medium catchments varying in size from 6 sq.kms. to
2072 sq.kms. The study presented, involves application of Extreme Value Type-I, General Extreme Value and Wakeby distributions using (i) at site data (ii) at site and regional data combined and (iii) regional data alone. Statistical test
based on U.S.G.S. method has been performed in order to test the
homogeneity of the region. The data of eight sites are
utilised for estimating the parameters, keeping the data of
remaining two sites for the purpose of testing the methodology
based on descriptive ability criteria. The predictive ability of
the different methods considered in the study is also tested
through Monte Carlo experiments; wherein synthetic flood
series love been generated using the regional EV1, GEV and
Wakeby parameters derived from the historical data. Generated
data sets of specific record lengths (same as the record
length of historical data for respective gauging sites) have
been considered for the eight sites. For the two independent
sites variable record lengths viz.1, 5, 10, 13 ,20, 30 and 40 one
at a time, have been considered.
The above methodology has been applied to the generated data
of different sample sizes for each population for the two
independent gauging sites. Performance of different methods has
been evaluated based on predictive ability criteria viz, bias,
and root mean square error. It is seen that the methods based
on EV1(PWM), GEV(PWM) and Wakeby (PWM) approaches using at
site and regional data in combined form provide estimates
of flood peaks with computationally less bias and
comparable root mean square error for the two test catchments.
