9 Frequency Analysis+

Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 4

Frequency Analysis

Introduction
Among the ways devised for estimating storm rainfall or flood flows for engineering
designs is statistical analysis based upon observed records of adequate length. In the design of
hydraulic structures the maximum rainfall or the transformed from maximum rainfall peak
flow that can be expected with an assigned frequency is of primary importance for adequate
design of bridges, culverts, waterways, spillways and dams. Unfortunately, however, recorded
information is seldom sufficiently extensive to identify critical magnitudes directly from
experience. Therefore, information must be generalized to arrive at rational extrapolations for
the frequency or recurrence interval of design storms or the magnitude of flows of design
frequency.
In certain regions of the world increase of flood levels are predictable as a result of climate
change or to the impact of other changes on the catchment. If due to climate change flood
risks should increase, evaluation of this risk can be part of the design standards and the
growing risk should be seriously considered. A global re-evaluation of flood probabilities is
initiated, and flood control measures should be enhanced.
Obviously, it is prime importance for these particular regions to maintain a convenient
weather monitoring system to aware floods. And that must be together with the extension of
raingauge network and weather data record improvement.
Point-Rainfall Frequency
Point rainfall; also known as station rainfall refers to the rainfall data of a station.
Depending upon the need, data can be listed as daily, weekly, monthly, seasonal or annual
values for various periods.
In many Hydraulic-engineering applications such as those concerned with floods, the
probability of occurrence of a particular extreme rainfall, will be of importance, such
information is obtained by the frequency analysis of the point-rainfall data. The rainfall at a
place is a random hydrologic process and the rainfall data at a place when arranged in
chronological order constitutes a time series. The probability of occurrence of an event in this
series is studied by frequency analysis of this data series.
Flood Frequency
Hydrologic processes such as floods are exceedingly complex natural events, and are
therefore very difficult to model analytically. For example, the flood in a catchment depends
upon the characteristics of the catchment; rainfall and antecedent conditions, each of these
factors in tern depend upon a host of constituent parameters. This makes the estimation of the
flood peak a very complex problem leading to many different approaches. The empirical
formulae and the unit hydrograph are some of them. Another approaches to predict floods,
rainfalls, and also applicable to another hydrologic processes, are the statistical methods of
frequency analysis.

Plotting Position
The value of the annual maximum flood from a given catchment area for large number of
successive years constitutes a hydrological data series called annual series. The data is then
arranged in decreasing order of magnitude and the probability P of each event being equaled
to or exceeded (plotting position) is calculated by plotting position formula,

Where, m = order number of the events,


N = total number of events in data observed.
The recurrence interval T (is also called the return period or frequency) is calculated as:

A plot of rainfall magnitude RF or flood magnitude Q vs. return period T yields the
probability distribution. For small return period (i.e. for interpolation) or limited extrapolation
is required, a simple best fitting curve through plotted points can be used as the probability
distribution, as shown in the figure below,
A logarithmic scale for T is often advantageous. However, when larger extrapolation of T
are involved, theoretical probability distribution have to be used. In frequency analysis of
floods or storms that producing floods the usual problem is to predict the extreme flood
events. Towards this, specific extreme value distributions (functions) are assumed and the
required statistical parameter calculated from the available data. Using these the flood
magnitude for a specific return period is estimated or, vice versa, the return period
(frequency) of a specific flood magnitude is estimated.

Frequency-Distribution Function
General equation of hydrologic frequency analysis, which expresses most of the frequency
- distribution functions applicable in hydrologic studies, proposed by Ven Te Chow as;
XT = X + K……………………………..(3)

Where, XT = value of the variate X of a random hydrologic series with a return


period T,
X = mean of the variate,
 = Standard deviation of the variate
K = frequency factor which depends upon the return period T and the assumed
frequency distribution.
Some of the commonly used frequency distribution functions for predication of extreme
flood values are:
1. Gumbel’s Extreme-Value distribution (GEV),
2. Log - Pearson Type III distribution (LP3)
3. Log Normal distribution (LN)

Log Pearson Type III Distribution


The Pearson type III is a special case of gamma distribution and all the more useful. This
distribution has been widely adopted as a standard method for flood frequency analysis in a
form known as the log-Pearson type III in which the transform Z=logX is used to reduce
skewness. Although, all the three moments: second, third and fourth moments of the variance
about the mean are required to fit the distribution. Yet it is extremely flexible, in the sense
that a zero skew will reduce log-Pearson type III distribution to a log-normal and a Pearson
type III to a normal.
The probability distribution function for Gamma is,
Where, α and β are constant and they may be estimated by the method of moments or by
the method of maximum likelihood
Г(α +1) is a Gamma function. and Г(α +1) = α!
Log Pearson Type III distribution is extensively used in hydrology. In this the variate is
first transformed into logarithmic form (base 10) and the transformed data is then analyzed. If
X is the variate of a random hydrologic series, the series of Z are first obtained, where,
Z = log (X) ……………………….…………(5)
For this Z series, for any recurrence interval T, equation (3) gives,
ZT = Z + KZ …………………….………. (6)
Where, KZ = a frequency factor which is a function of recurrence interval T and the
coefficient of Skew CS, i.e.; KZ = f (CS, T)

Z = standard deviation of the Z variate sample,

CS = Coefficient of Skew of variate Z,

Where, Z = mean of the Z values


The variation of KZ = f (CS , T) is given in special tables for use in LP3 distribution in most
of the available references in hydrology.
Further, the coefficient of skew CS is adjusted to account for the size of the sample by using
the following relation proposed by Hazen A.,

Where, Ĉs = adjusted coefficient of skew.


N = sample size = number of years of record.

You might also like