Research papersRevised Dicken’s method for flood frequency estimation of Upper Ganga basin
Introduction
Design of any important developmental project, for example small hydropower projects or other hydraulic structures, requires data of flood magnitudes and their frequencies. Suitable methodology to be adopted for estimation of design flood for a project depends upon importance of the project, economic and safety consideration, data availability and computational facilities etc. An accurate assessment of the river flow is essential for estimating the design discharge for any hydraulic projects at the proposed site. Increasing trend in one or more flood characteristics have led researchers around the globe to find a way for more reliable flood prediction measures (Hall et al.,2014, Vormoor et al., 2016, Kundzewicz, 2019). However, for a number of potential sites, there is either insignificant data or no data which can be used to estimate the flow statistics. Regional flood frequency analysis in ungauged basins have been a subject of interest. Various methods like L-moments method and annual maximum series of mean daily streamflow has been used for reliable prediction of flood quantiles in various countries (Smithers and Schulze, 2001, Hosking and Wallis, 2005, Kumar and Chatterjee, 2005, Malekinezhad and Zare-Garizi, 2014, Hailegeorgis and Alfredsen, 2017). Empirical formulas are the only alternative method to provide an estimate of the design flood in ungauged catchment and as such various empirical formulae have been devised for different regions of the country (Acreman, 1985, Commission, 1980, Greenwood et al., 1979, Huq et al., 1986). Most of these empirical formulae involve one or two physical characteristics for estimation of the flood peaks. These formulae are essentially the regional formulae based on statistical correlation of the observed peaks and important catchments and climatologic characteristics. However, in the present form, these formulae are not capable of providing flood estimates for a desired return period (Commission, 1980).
One of the empirical formula adopted for moderate size North Indian catchments is Dickens formula, which is given aswhere Q is flood volume in cumecs, A is catchments area in Sq. Km., and C is the Dickens coefficient, whose values for various regions are given in Table 1.
The coefficient C is chosen from the given range for the specific regions. In spite of its simplicity and applicability, Dicken’s formula is not capable of yielding flood volume at different desired frequency.
In this paper, a methodology is being presented for the computation of revised coefficient of the Dickens formula, to estimate design flood of different return period, by relating the frequency curves developed for hydro meteorologically homogeneous regions by fitting probability weighted moments (PWM) based Index flood method, General extreme Value (GEV) and Wakeby distributions to the data of 16 gauging sites of Ganga basin in Himalayan region. The floods of different return periods are computed for four test catchments for validation of the revised Dickens formula. This revised Dicken’s empirical formula will enhance its practical utility for computation of flood volume for Upper Ganga basins.
Section snippets
Geoidentity and data availability
Based on hydrological response, India has been geographically categorized into seven major hydro- meteorological zones. These are further sub-divided into 26 hydro-meteorologicalsub zones. The Upper Ganga basin is situated between latitude 32–34° N & longitude 76–78° E. The present study area falls into hydro-meteorological zone 7 of India (See Fig. 1). Most of rivers and their tributaries of upper Ganga basin are snow fed while the normal annual rainfall in this region varies between 700 mm
Methodology
Available methods for carrying out flood frequency analysis are listed in Table 3 (Goswami, 1972, Greis and Wood, 1981, Greis NPaEFW, 1983, Jakhade et al., 1984, James et al., 1987). The USGS methods have been used for testing the regional homogeneity. The methods for development of regional flood frequency curves using probability weighted moments (PWM), Index flood method, General extreme value (GEV) and Wakeby distribution have been described in detail elsewhere (Kuczera, 1982, Kuczera, 1983a
Revision of Dicken’s coefficient
By indication, the Dicken’s formula for T -year return period can be written as
Dividing Eq. (8) by Eq. (7), the equation obtained is
Substituting for in the Eq. (10) from Eq. (2), (4), (6), the revised Dickens coefficient CT can be written as shown in Table 6.
Development of regional frequency curves
Observed annual maximum peak flood data were used for development of regional frequency curves for Upper Ganga basin. These curves are shown in Fig. 7, Fig.8, Fig. 9. An analysis of these curve reveal that for catchments draining up to 5000 Km2, numerical value of CT shows a sharp gradient. Further, CT value, ranges between 0.834 and 6.924. The CT –values are lower when calculated by Index Flood method while these are higher by Wakeby method. A plot of the growth factor in Fig. 5 of the
Model validation
The historical data pertaining to GD13, GD14, GD15 and GD16 have been used to verify the developed Flood Frequency Model for Upper Ganga basin. From the historical records for these test sites, the site parameters pertaining to Index Flood method, GEV and Wakeby distribution were computed.
From these computed parameters, the flood volumes at sites for different recurrence intervals have been computed using the regional flood model, developed in present study and tabulated along with their ADF in
Conclusion
Insignificant or no data for a number of potential sites present difficulty in estimating the flood volume and as such the flow statistics. Several empirical methods are available for this task but they are complex and computationally not efficient. The proposed method of flood frequency analysis is based on Dicken’s formula and it is very simple and widely used for predicting the flood discharge at these ungauged sites and is used by practicing field engineers for flood estimation. However the
CRediT authorship contribution statement
K.K. Pandey: Conceptualization, Data curation, Formal analysis, Funding acquisition, Methodology, Investigation, Validation, Project administration, Resources, Supervision, Visualization. Amiya Abhash: Methodology, Investigation, Software, Validation, Writing - original draft, Writing - review & editing. Ravi Prakash Tripathi: Writing - original draft, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (29)
- et al.
Regional flood frequency analysis and prediction in ungauged basins including estimation of major uncertainties for mid-Norway
J. Hydrol. Reg. Stud.
(2017) Effect of sampling uncertainty and spatial correlation on an empirical Bayes procedure for combining site and regional information
J. Hydrol.
(1983)- et al.
Regional frequency analysis of daily rainfall extremes using L-moments approach
Atmósfera
(2014) - et al.
A methodology for the estimation of short duration design storms in South Africa using a regional approach based on L-moments
J. Hydrol.
(2001) - et al.
Evidence for changes in the magnitude and frequency of observed rainfall vs. snowmelt driven floods in Norway
J. Hydrol.
(2016) Identification of homogeneous regions for flood frequency analysis
J. Hydrol.
(1986)Predicting the mean annual flood from basin characteristics in Scotland
Hydrol. Sci. J.
(1985)- Commission CW (1980) Flood Estimation Report For Lower Godavari Sub zone (Subzone-3f)-A Method Based on Unit Hydrograph...
- Goswami A (1972) Regional Flood Frequency Analysis of Brahmputra Basin in North-East India Institution of Engineers...
- et al.
Probability weighted moments: definition and relation to parameters of several distributions expressable in inverse form
Water Resour. Res.
(1979)