CIV 5306: Water Resources Engineering

Pre-requisite: CIV4213: Hydrology

CIV4217: Hydraulics
Statistics and Probability methods in Hydrology: Hydrological Forecasting, Reservoir and Flood
routing, Design of urban storm drainage systems, Groundwater Hydraulics and Groundwater
exploration.
Hydraulic Structures: Functional design of service and impounding reservoirs. Sedimentation
control in reservoirs, The design of earth dams, rock fill dams, concrete dams and associated
components e.g. spillways, stilling basing, outlet works, sluice ways etc.
Water Resources Management: National legislation in water resources, Integrated planning of
river basins, Economic analysis of water resources projects, Environmental impact of water
resources projects: Social Ecologic, Health aspects.
STATISTICS & PROBABILITY
Statistics and probability methods are widely used in hydrology for hydrological forecasting.
These methods involve the analysis of historical data to predict future water levels, streamflow,
and other hydrological variables.
Some commonly used statistical methods in hydrology include regression analysis, time series
analysis, and correlation analysis. These methods are used to identify relationships between
different hydrological variables and to develop models that can be used to forecast future
hydrological conditions.

Probability methods, such as probability distribution functions and stochastic modeling, are
also used in hydrology for forecasting. These methods involve the analysis of probability
distributions of hydrological variables to estimate the likelihood of different outcomes.

Probability distribution functions (PDFs: PDFs provide a mathematical representation of the

probability distribution of a random variable, allowing for the estimation of various statistical
properties and the quantification of uncertainties.
PDFs are commonly used in hydrological forecasting to model and analyze the behavior of
hydrological variables, such as streamflow, precipitation, and water levels.

Review of Statistical Methods

A probability function defines the relationship between the outcome of a random process and the
probability of occurrence of that outcome.

Observed values of x (variate) for a finite number of years is known as ‘sample’ of x. Say annual
flood peaks or annual rainfall for 75 years gives the sample; On the other hand, the population
consists of the values of annual flood peaks from time immemorial to eternity.
The ‘population parameters’ can be estimated by means of parameters obtained from the sample,
known as ‘sample parameters’.
If the value of one variate is independent of any other, the variable in question is a ‘random
variable’. Hydrological processes are mostly random and hence the respective variables are
equally random.
  If the sample space, S, contains discrete elements, then the sample space is called a
discrete sample space (e.g toss of a coin; or casting of a die); Discrete probability
distributions describe the probability of outcomes in discrete sample spaces,
 If S is a continuum, the sample space is called a continuous sample space (e.g amount of
rainfall that accumulates during a certain rainy day; 10-day low flow value for a river
reach, each year); Continuous probability distributions describe the probability of
outcomes in continuous sample spaces.
 The sample space of a random variable is commonly denoted by an upper-case letter
(e.g., X), and the corresponding lower-case letter denotes an element of the sample space
(e.g., x).

Some of the most commonly used probability distribution functions in hydrology:

1. Normal (Gaussian) Distribution: The normal distribution is widely employed in
hydrology due to its simplicity and applicability to a wide range of variables. It is
characterized by its mean (μ) and standard deviation (σ). In hydrological forecasting, the
normal distribution is often used to represent variables with symmetric and bell-shaped
distributions, such as daily or monthly streamflow data.
 2. Log-Normal Distribution: The log-normal distribution is frequently used to model
hydrological variables that are positively skewed and have a multiplicative behavior. It is
characterized by its mean (μ) and standard deviation (σ) of the logarithm of the variable.
For instance, the log-normal distribution is often applied to represent annual maximum
flood peaks or precipitation amounts. By transforming the data to the logarithmic scale,
the log-normal distribution can be used to estimate the probability of extreme events,
such as floods exceeding a certain threshold.
3. Generalized Extreme Value (GEV) Distribution: The GEV distribution is employed in
hydrology to model the extreme values of hydrological variables, such as flood peaks
and drought durations. It is characterized by three parameters: location (μ), scale (σ), and
shape (ξ). The GEV distribution encompasses three types of distributions: Gumbel (ξ =
0), Frechet (ξ > 0), and Weibull (ξ < 0). The choice of the GEV type depends on the
nature of the data. For example, the Gumbel distribution is often used to model flood
peaks, while the Frechet distribution is suitable for representing high flows in rivers with
persistent runoff sources.
The GEV distribution allows for estimating return periods and quantifying the probability of
extreme events beyond the observed data.
These three probability distribution functions, along with others such as the gamma distribution,
Pearson distribution, and others, provide hydrologists with flexible tools to characterize the
statistical properties of hydrological variables and make probabilistic forecasts.
By selecting an appropriate distribution function and estimating its parameters from
historical data, hydrological forecasts can be generated, aiding in water resource
management, flood risk assessment, and infrastructure planning.

APPLICATION OF THE PDF IN HYDROLOGICAL FORECASTING

Common application of probability theory in water-resources engineering involves the
assignment of an exceedance probability, Pe, to hydrologic events. A system designed to handle
a particular hydrologic event can be expected to fail with a probability equal to the probability of
exceedance, Pe, of the design event. This probability of system failure is called the risk of
failure, and the probability that the system will not fail, 1−Pe, is called the reliability of the
system.
Exceedance probabilities of hydrologic events are usually stated in terms of the probability that a
given event is exceeded in any given year. The average number of years between exceedances is
called the return period, T, which is related to the annual exceedance probability, Pe, by

The return period, T, is sometimes called the average recurrence interval (ARI) and is usually a
criterion for selecting design events in water-resource systems.
If a water-resource system is designed for a hydrologic event with exceedance probability, Pe,
and the system has a design life of N years, then the risk that the system will fail within its
design life, Pf , is given by

This formulation is particularly useful in assessing the cost of failure, since

the expected
damage cost is equal to the probability of failure multiplied by the cost
associated with the
failure.
EXAMPLE 8.1
Awater-resource system is designed such that the probability, f (xi), that the
system capacity is exceeded xi times during the 50-year design life is given
by the following discrete probability distribution:
xi f (xi)
0 0.13
1 0.27
2 0.28
3 0.18
4 0.09
5 0.03
6 0.02
>6 0.00
What is the mean number of system failures expected in 50 years? What are
the variance and skewness of the number of failures?
Solution The mean number of failures, μx, is defined by the following
equation, where
Binomial

Normal
Log Normal
The Gumbel distribution
The Gumbel distribution is commonly used in hydrological analysis to model extreme events
such as floods or droughts. It is a type of probability distribution that characterizes the maximum
or minimum of a number of random variables. Here's an explanation of how to implement the
Gumbel distribution in hydrological analysis:
The Gumbel distribution
The Gumbel distribution is commonly used in hydrological analysis to model extreme events
such as floods. It is a type of probability distribution that characterizes the maximum or
minimum of a number of random variables.
Frequency analysis using the Gumbel distribution can be implemented as below;
1. Collect the data: Start by collecting the historical data for the variable of interest, such as
annual maximum or minimum flows.
2. Order the data: Arrange the data in ascending order, from the smallest to the largest
values. This step is important because the Gumbel distribution focuses on the extreme
values.
3. Calculate the return periods: Compute the return periods for the ordered data. It is often
expressed in years. The formula for return period T= (n + 1) / m, where n is the total
number of data points and m is the rank of the data point.
4. Standardize the data: Convert the ordered data into standardized values using the formula
z = (x - μ) / σ, where z is the standardized value, x is the original data point, μ is the
mean, and σ is the standard deviation of the data.
5. Estimate the Gumbel parameters: Use the standardized data to estimate the parameters of
the Gumbel distribution.
6. Fit the Gumbel distribution: Once you have estimated the parameters, you can fit the
Gumbel distribution to the data. This involves generating the cumulative distribution
function (CDF) of the Gumbel distribution using the estimated parameters and comparing
it to the empirical CDF of the data. The fit can be visually assessed through a probability
plot or statistically evaluated using goodness-of-fit tests such as the Anderson-Darling
test or the Kolmogorov-Smirnov test.
7. Calculate design values: Once you have a good fit of the Gumbel distribution, you can
use it to estimate design values for specific return periods. For a given return period, you
can calculate the corresponding quantile (probability) using the Gumbel distribution's
inverse CDF. Then, convert the quantile back to the original scale using the inverse
standardization formula.
8. Validate the model: It is important to validate the Gumbel distribution model by
comparing the calculated design values with independent data or other well-established
methods. This step helps assess the reliability and accuracy of the Gumbel distribution in
representing extreme events.
   1 
y  In   In  1   
  T 

A 0.779 x

B  X  .4500 x

X B  Ay

  1 
y  In   In  1   
  T 

A 0.779 x

B  X  .4500 x

X B  Ay
 [ ( )]
y=−¿ −¿ 1−
1
Tr

QT =Qav + σ ( 0.78 y−0.45 ) σ=

√[ n
n−1 ( ∑ Q2 −Qav2
n
❑ )]
 Assignment:
For the following events, estimate the 100 yr flood using the Gumbel distribution
Q Q
Year (m3/s) Year (m3/s)
1981 56.25 1988 48.11
1982 65.57 1989 59.03
1983 45.72 1990 60.59
1984 44.78 1991 47.2
1985 48.02 1992 38.16
1986 54.48 1993 45.72
1987 51.53 1994 54.62

Qav =¿

Soln
Year Rank Q (m3/s) Tr Pr % ݊ ൅ ͳ
݉
Return Period Tr=
1982 1 65.57 15.0 6.67
1990 2 60.59 7.5 13.33

ܶ‫ݎ‬
1989 3 59.03 5.0 20.00 ͳͲͲ
Probability Pr=
1981 4 56.25 3.8 26.67
1994 5 54.62 3.0 33.33
1986 6 54.48 2.5 40.00
(ii)
1987 7 51.53 2.1 46.67
Gumbel T3

ܳ ൌ ܳ ൅ ߪ ͲǤ
ͺ‫ ݕ‬െ ͲǤ
1988 8 48.11 1.9 53.33
1985 9 48.02 1.7 60.00 ் ௔௩ Ͷͷ

‫ݕ‬ൌ െ ‫ ݊ܫ‬െ ‫ͳ ݊ܫ‬െ

1991 10 47.2 1.5 66.67

ܶ‫ݎ‬
ͳ
1983 11 45.72 1.4 73.33 Given Tr= 100
1993 12 45.72 1.3 80.00
1984 13 44.78 1.2 86.67 y= 4.6
1992 14 38.16 1.1 93.33 Q100 =
3
74.61 m /s

ܳ ௔௩ ൌ 51.41
σ= 7.39

σܳ ݊ σܳ ଶ
ܳ ൌ ߪൌ െ ‫ݒ‬
where
ଶ
௔௩
݊ ݊െ ͳ ݊ ᇹ
and