In absence of sufficient data, the design storm may be used by the transposition
method. In case of large and important projects the probable maximum precipitation
is used as the design storm. If probable maximum precipitation is used in design
flood estimation, the resulting flood hydrograph gives the maximum probable flood.
The unit hydrograph method cannot be applied for large areas, more than 5000 km
2 .
In such cases the basin is subdivided into a number of subbasins corresponding to
each tributary. These flood hydrographs of subbasins are then routed through the
channel reaches and superimposed with a common time scale to obtain the flood
hydrograph at the point of interest. In case of large areas, it is assumed that the design
storm produces rainfall over the entire basin that is simultaneous. The unit
hydrograph not only produces flood peak but also the complete flood hydrograph,
which is essential in fixing the spillway capacity after incorporating the effort of
storage of the reservoir on flood peak through flood routing. If the area of basin is
large, it is better to use frequency analysis methods (Ojha et al. 2008).
Trend Analysis of Rainfall
The rainfall data are collected, and trend analysis of rainfall determines the existing
trends in the data (Ghosh et al. 2009) by parametric and nonparametric methods. In
nonparametric methods, the Mann–Kendall test is used to determine the rainfall
trend (Ahamad et al. 2015). In the nonparametric test, the data outliers can be evaded
easily. The next test is Sen’s slope estimator, commonly used to sense a linear trend.
As the M-K test, the time series are evaluated by their ranks, known as Spearman’s
rho test (Siegel and Castellan 1998). To check the statistical changes in rainfall trend,
the CUSUM test is used (Chowdhury and Beecham 2010).
Frequency Analysis of Rainfall
Rainfall frequency can be analyzed by parametric methods such as log-normal,
log-Pearson type III, generalized extreme value (GEV), and Gumbel (GUM) and
Weibull probability distributions. The probability distributions are used to evaluate
the best fit probability distribution from the rainfall distribution analysis. The
goodness of fit is evaluated by the chi-square test, the Kolmogorov–Smirnov test,
the Cramer von Mises test, and the probability plot correlation coefficient (PPCC)
test. This method is applied for testing the null hypothesis whether the maximum
daily rainfall data follow the specified distribution. Tests such as Anderson–Darling
and Kolmogorov–Smirnov were used with the chi-square test at the α ¼ 0.01 level of
significance for the selection of the best fit probability distribution. The Gumbel
distribution is known as the extreme value type I distribution used for extreme
analyses such as flood and drought. In 1967, the United States Water Resources
Council suggested that log-Pearson type III distribution should be adopted as the
standard flood frequency distribution by all U.S. federal government agencies (Ojha
et al. 2008).
3 Flood Risk Assessment for a Medium Size City Using Geospatial Techniques with. . .
45
method. In case of large and important projects the probable maximum precipitation
is used as the design storm. If probable maximum precipitation is used in design
flood estimation, the resulting flood hydrograph gives the maximum probable flood.
The unit hydrograph method cannot be applied for large areas, more than 5000 km
2 .
In such cases the basin is subdivided into a number of subbasins corresponding to
each tributary. These flood hydrographs of subbasins are then routed through the
channel reaches and superimposed with a common time scale to obtain the flood
hydrograph at the point of interest. In case of large areas, it is assumed that the design
storm produces rainfall over the entire basin that is simultaneous. The unit
hydrograph not only produces flood peak but also the complete flood hydrograph,
which is essential in fixing the spillway capacity after incorporating the effort of
storage of the reservoir on flood peak through flood routing. If the area of basin is
large, it is better to use frequency analysis methods (Ojha et al. 2008).
Trend Analysis of Rainfall
The rainfall data are collected, and trend analysis of rainfall determines the existing
trends in the data (Ghosh et al. 2009) by parametric and nonparametric methods. In
nonparametric methods, the Mann–Kendall test is used to determine the rainfall
trend (Ahamad et al. 2015). In the nonparametric test, the data outliers can be evaded
easily. The next test is Sen’s slope estimator, commonly used to sense a linear trend.
As the M-K test, the time series are evaluated by their ranks, known as Spearman’s
rho test (Siegel and Castellan 1998). To check the statistical changes in rainfall trend,
the CUSUM test is used (Chowdhury and Beecham 2010).
Frequency Analysis of Rainfall
Rainfall frequency can be analyzed by parametric methods such as log-normal,
log-Pearson type III, generalized extreme value (GEV), and Gumbel (GUM) and
Weibull probability distributions. The probability distributions are used to evaluate
the best fit probability distribution from the rainfall distribution analysis. The
goodness of fit is evaluated by the chi-square test, the Kolmogorov–Smirnov test,
the Cramer von Mises test, and the probability plot correlation coefficient (PPCC)
test. This method is applied for testing the null hypothesis whether the maximum
daily rainfall data follow the specified distribution. Tests such as Anderson–Darling
and Kolmogorov–Smirnov were used with the chi-square test at the α ¼ 0.01 level of
significance for the selection of the best fit probability distribution. The Gumbel
distribution is known as the extreme value type I distribution used for extreme
analyses such as flood and drought. In 1967, the United States Water Resources
Council suggested that log-Pearson type III distribution should be adopted as the
standard flood frequency distribution by all U.S. federal government agencies (Ojha
et al. 2008).
3 Flood Risk Assessment for a Medium Size City Using Geospatial Techniques with. . .
45