International Journal of Hydrology

Research Article Open Access

A neuro-fuzzy modeling for the hydrological time

series of floods of river indus of Pakistan
Abstract Volume 3 Issue 2 - 2019

Many forecasting applications of science and technology are used to predict future
Salman Bin Sami,1 Tanveer Ahmad Siddiqi,2
values. Today, the flood peak flow discharges are extensively used on annual basis
in risk assessment by collecting quantitative data from several sources. The famous
Muhammad Jawed Iqbal3
1
Management Science Department, SZABIST (Karachi Campus),
rivers of Pakistan i.e. River Jhelum, River Kabul, River Chenab, and the very famous,
Pakistan
considering upper area parts and the lower area parts of the river, the Indus River are 2
Department of Mathematics, University of Karachi, Pakistan
the prime sources of flooding. These aforesaid rivers are the prime tributaries of the 3
Department of Mathematics and Institute of Space and
Indus River System which is the one, from all, most notable rivers of the world and Planetary Astrophysics, University of Karachi, Pakistan
for Pakistan, it is a supreme river. River Indus is Pakistan’s longest river with seven
(7) different gauge stations, Dams and various barrages, and plays a significant role- Correspondence: Salman Bin Sami, Management Science
in irrigation and maximum generation of power in Pakistan. In the present research Department, SZABIST (Karachi Campus), Karachi-Pakistan, Tel
the flood risk in the Indus River has been calculated by utilizing the historical peak +92-3333175611, Email
streamflow discharges recorded data on the daily basis. Nowadays, Adaptive Neuro-
Fuzzy Inference System (ANFIS) model which is widely used to analyze these Received: February 20, 2019 | Published: April 04, 2019
hydrological time series data sets obtained from different gauge stations. Adaptive
Neuro-Fuzzy Inference Systems (ANFIS) merges the potentiality of Fuzzy Inference
Systems (FIS) and Artificial Neural Networks (ANN) to work out problems of different
kinds. For this, the data covered eleven years from 2002 to 2012 daily (6-months
each year) streamflow period. From our analysis, the root mean square error (RMSE)
shows that the ANFIS model generated more satisfactory results than other models
with minimum prediction errors. The ANFIS model is very pliable and has feasibilities
of integrating the essence of a fuzzy system for real world.

Keywords: neuro-fuzzy network, fuzzy logic, fuzzy inference system, hydrological

modeling, river indus, root mean square error, adaptive neuro-fuzzy inference systems

Introduction unfavorable, devastating and calamitous in the history of Pakistan.

This research is an attempt to scrutinize the consequences of floods
Today floods are very commonly occurring natural disaster in on habitat along the marked off flood areas of Indus River of Pakistan.
the world. Flood damages all kind of lives i.e. humans, animals and
vegetation, and destroys environments. Very heavy losses are sustained Figure 1 shows the neighborhood elevation of the flooded areas
to the economy during floods. Looking at the flash flood data for the 2 in Pakistan. The most significant reason of floods in Pakistan are the
or 3 decades, the speed of flood has increased due to global warming.1 monsoon rains in which there is heavy rainfall which is intensified
Therefore, engineers and planners want to get accurate time of these by the snow melt flow in river drainage. Land sliding also results in
events. Technology has extensively been used to study the severity of floods. These floods cause heavy losses including loss of human life
flood damages, i.e. constructively or non-constructively. Constructive and animals, structural damages including bridges, sewerage systems,
measures require huge amounts of money and time. This is due to roadways and canals, raising waterborne diseases and disintegration
the fact that things to consider include; making of dams, reservoirs of farmland resulting in shortage of food. Nevertheless, in the annual
and change to river flow. As far as the non-structural measures are report for the year 2010, FFC (Federal Flood Commission) described
concerned they include flood- forecasting to plan for floods and to that highest damage is observed in the lower part of Indus River
provide relief if floods occur. Time series analysis predicts future specially in Sindh where flow is at a peak in the neighboring areas.1
values by analysis of data taken from the past. Pakistan lies in the River Indus is one of the longest rivers of the world. It has seven (7)
Western region of the Asian Subcontinent where it stands 23 degrees dams or barrages mainly and has a length of 1800 (miles). Its total
to 37 degrees north and rom 60 degrees to 77 degrees east. Pakistan has discharged area is around 450000 (square miles) out of which 275000
five provinces and a tribal area; Punjab, Sindh, Khyber Pakhtunkhwa, (square miles) lies in semi-arid deserts of Pakistan while the rest
Balochistan & Gilgit-Baltistan. All of these places face different types comprises of mountains of Pakistan. The River Indus runs through the
of weather conditions. There are temperature variations in the four region of Ladakh which is in Jammu and Kashmir, Gilgit-Baltistan
seasons throughout the year. Different conditions including heavy and flows through Pakistan in southerly direction along the whole
rain, bring floods. The snow melts on the mountains canal so cause length of Pakistan to converge into the Arabian Sea near the port
floods. Pakistan has long history of floods in River Indus and it nearby city of Karachi in Sindh. Indus River is largely observed by its seven
areas. Pakistan has coped with many floods in 1950, 1956, 1973, 1976, gauge stations which are Tarbela Dam, Chashma Barrage, Kalabagh
1978, 1988, 1992 and 2010. Graphs or sketches of floods between the or Jinnah Barrage, Taunsa Barrage, Guddu Barrage, Sukkur Barrage
years 1922 and 2010 vary. The 2010, flood proved to be the most and Kotri Barrage as shown in Figure 2.

Submit Manuscript | http://medcraveonline.com Int J Hydro. 2019;3(2):138‒147. 138

© 2019 Sami et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which
permits unrestricted use, distribution, and build upon your work non-commercially.
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 139

characteristics. Similarly, Khan et. al. quantified the probabilities of

incidents of floods to evaluate River Indus flood risks using archived
information about maximum peak discharges in Pakistan.4
According to many authors various countries like USA, Korea,
Australia, Pakistan, etc., are using barometrical parameters to assess
flood risk. A technique for graphing risk of flood based on provincial
or local weather information and area has been presented by Kalma
and Laughlin.5 Khan used a more effective technique called GIS
technique to study flood risks in the affected areas near River Indus
in Pakistan. This technique included remote sensing, geological data
system and digital image processing.6 Khan used satellite information
yielding the importance of making dams to minimize flood risks.
Nawaz and Shafique did the same work on river Jhelum in 2003.7 For
Dams and Rivers, several forecasting techniques strived successfully
to use the Linear and non-linear regression’s techniques used by Burn
and McBean in 1985,8 El-Fandy, Karunanithi and Awwadin 1994.9–
11
These studies yielded good dams river flow forecasts. Selas and
Smith generated synthetic stream flows using hydrological time series
modeling.12 Stedingerand Taylor also assumed stream flow image by
developing five different models.13 Time series forecasting is a highly
demanding and rapidly growing area for researchers in various fields
Figure 1 The map of the Pakistan for flooded areas. of interest; in engineering, finance, physics and also in medicine.
Traditionally, researchers have used several modeling and forecasting
techniques for time series, such as, AR (Auto-regressive) method
discovered in 1970,14 ARMA (Auto-regressive moving average) and
ARIMA (Auto-regressive integrated moving average) disaggregation
models formulated by Valencia and Schaake in 1973,15 etc. The
process to predict weather from time series data is called time series
forecasting. These approaches are used only for time series data to
generate forecasts. The time series data used is the numeric data type
for months of heavy stream flows such as a day or several hours.
In the year 2010, Hassan and Ansari performed some nonlinear
methods to predict the continuing behavior of Indus River.16 Sudheer
presented an ANN model for the same purpose but realized that
further research on ANN model is required to shape up the peak data
flows efficiently.17 In this study the potential of the ANFIS (adaptive
neuro-fuzzy inference system) model has been investigated for the
Indus basin. Moreover, Nayaka and Sudheer, in India, investigated
a hydrological model of time series for Baitarani River’s basin
stream flow in Orissa state by ANFIS model.18 In recent years, the
models which are widely used for prediction; adaptive neuro-fuzzy
Figure 2 The Map of Pakistan for the locus plans of River Indus. inference system (ANFIS), artificial neural network (ANN) and
genetic algorithm are highly systematic tools for complex hydrologic
Different flood risk levels; medium, high or very high floods have modeling systems. Fuzzy logic procedure to describe the human-like
been observed to occur at these seven gauge stations. Very high-level expertise system and to make the decisions was first developed in 1965.
flood risk is observed at Sukkur Barrage, while Kotri and Guddu The ANFIS model keeps the complete capacity of the ANN approach,
Barrages are seen to be fluctuating between high and very high flood and simplicity of the procedure of model building. Adaptive neuro
risks. Chashma and Taunsa Barrages have high flood risks but Tarbela fuzzy inference system (ANFIS) is the fuzzy-rule based or mapping of
Dam and Kalabagh have medium flood risks. In this paper, we are fuzzy algorithm that was developed by Tagaki-Sugeno-Kang (TSK)
proposing to use ANFIS technique for constructing time data series and Yasukawain in 1993.19 For water resources predictions ANFIS has
model of river flow for the Basin of River Indus in Pakistan using proved to be used on a higher scale during the last decade by various
annual flood peak discharges from gauge sites. In 2005, Ologunorisa scientists. Nowadays it is used for many applications such as, database
and Abawa described that many approaches are used to estimate management, the forecasting of resources of water and planning.
flood risk like barometrical conditions and equipment, hydrological
equipment, socioeconomic elements and a blend of hydro- Neuro-fuzzy model
barometrical and socioeconomic elements in addition to geological
Neuro-fuzzy model introduces the process of utilizing various
data network.2 Smith3 states that to estimate flood risk we need to
learning techniques developed in the literature of neural system to
observe the probability of events that took place and its outcomes.
fuzzy modeling or FIS (fuzzy inference system). The fuzzy inference
Uncertainty in resources of water result in natural dissimilarity of
system or FIS structure basically consists of three elements:
geophysical developments and changes in complex socioeconomic

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 140

a. A rulebase, in which we have fuzzy rules selection consequent parts of this FIS and can be determined by the method of
simple least square error method.
b. A database element, which designates the functions called
membership functions (mfs) for the rules of fuzzy; and Let’s us consider, for FIS, x and y are the two input variables and
the one output variable is z. The fuzzy model of Sugeno for the first
c. A reasoning mechanism element, in which we execute the
order has two rules that are if-then rules which are conveyed as,
inference strategy based on the defined rules to yield an output
results as can see in Figure 3. Rule I: Let x stands for A1and y stands for B1, then for f1

= p x + q y + r1 (1)
1 1

Rule II: Let x stands for A2 and y stands for B2, then for f2

= p x + q y + r2 (2)
2 2

Here, for the input variables x & y the MFs are A1, A2 & B1, B2
respectively, and the parameters for the output function are p1, q1, r1
and p2, q2, r2. and Figure 4(A) explains fuzzy reasoning technique
to obtain the output function i.e. f for this Sugeno model from the
Figure 3 Brisky output by Fuzzy inference system. provided input variables [x, y]. The related equivalent architecture of
ANFIS is described in Figure 4(B), where the same layer nodes have
From the input area to the output area, fuzzy inference system
corresponding functions. The following are procedures of ANFIS
executes nonlinear plotting. This plotting is carried out by a number
functioning:
of if-then rules of fuzzy, in which each rule reports the behavior of
the plotting locally. The parameters of the if-then rules (referred to as
antecedents or premises in fuzzy modeling) define a fuzzy region of
the input space, and the output parameters (also consequents in fuzzy
modeling) specify the corresponding output. Thus, this FIS (fuzzy
inference system) relies on the calculated parameters for its proficiency.
Nonetheless,20 for the fuzzy set (delineated by antecedents) the choice
of pattern leading to an input is not accompanied by any method.
But the fuzzy inference rule of system and its structure make it easy
to assimilate human proficiency for the system into the process of
modeling directly to settle on significant no. of membership functions
for every number of input, etc. For parameter calculation there will be
leading numerical data. Nowadays, perception of the adaptive network,
which is an abstract principle of the ordinary backpropagation neural
network, is applied to tackle the parameter identification problem in
a FIS. ANN (adaptive neural network) is the multi layered furnish
Figure 4(a) Shows the inference system of Fuzzy; (b) Shows an Equivalent
shape structure whose overall output behavior is determined by the
architecture of ANFIS.
value of a collection of modifiable parameters. More specifically, the
configuration of an adaptive network is composed of a set of nodes Layer I: Here, every node of layer 1 creates grades of membership for
connected through directional links, where each node is a process unit variable inputs. To define the node output OP1i we have:
that performs a static node function on its incoming signal to generate
a single node output. Every function of node consists of parameterized = OP1i µ= A ( x ) wherei 1, 2
1
(3)
value function for some malleable parameters. It is noticeable in the
adaptive network that the attachments, between nodes with no weights,= OP1i µ= B (y) wherei 3, 4 (4)
i−2
only show the flow direction of signals affiliated with these links. In
1994, some of the readers mentioned21 additional characteristics of Here, the variable x or y is defined as input node & Ai or Bi-2
adaptive networks. The Jang22 established an understanding strategy is associated as a fuzzy set, designated by the structure of the
and a novel architecture for FIS that uses an understanding algorithm membership functions which is any suitable continuous and piecewise
of neural network for building a collection of if-then rules from the differentiable function like Gaussian, generalized bell, triangular &
determined input–output couples for fuzzy with the convenient MFs.23 trapezoidal shaped in this node. Considering MF has the shape of
Using this framework for the ANN (adaptive neural networks) for the generalized bell function, then the output
strategy of flourishing a FIS is known as an adaptive neuro-fuzzy
inference system (ANFIS). result OP1i will be,

Architecture of ANFIS 1
OP1i µ=
= A ( x) (5)
1
(x −ci )bi
The common construction for ANFIS is reported in the Figure
3. When designing ANFIS, the main choice for FIS is a model of Here, the shape of MF can be changed by the parameter set {ai, bi,
particular objective system. Many categories of FIS are presented ci} from maximum 1 to minimum 0.
in the literature.24–26 Each example is portrayed by part of their
consequent parameters. Present research utilizes the fuzzy model of Layer II: Here each node multiplies the approaching signals,
Sugeno.26,27 Therefore, the linear equation and parameters are the most symbolized by Π, and the rule of firing strength is illustrated by the

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 141

2
output OP i2 which can be calculated as: A LSE for X can be required that reduces the squared error AX − B
. Figure 4 shows the architecture of ANFIS and shows that for the
=
OP 2
i w= µ A (=
x ) µ B ( y ), i = 1, 2 (6) provided parametric values of the premise part, the total and final
i i i
output is conveyed by a linear combination of the parameters of the
Layer III: For this layer the ith node is symbolized by N, which consequent part. Formally, the final output f can be changed into:
calculates the normalized firing strengths as:
wi f =w f + w f ==( w x ) p + ( w y ) q + ( w ) r + ( w x ) p + ( w 2 y ) q + ( w ) r
OP i3 = wi = i = 1, 2 (7) 1 1 2 2 1 1 1 1 1 1 2 2 2 2 2
w1 + w2
(14)
Layer IV: In layer 4 the node i calculates the contribution of the Which is linear for parameters (p1, q1, r1, p2, q2 and r2) of the
ith rule towards the output of the model, with the following node consequent part. In an ANFIS, the overall no. of parameters (S), as a
function: result, can be classified into two sets in such a way that the set of premise
OP i4= wi fi = wi (px i + qy i + ri ) (8) part parameters is S1 and the set of consequent part parameters is S2.
Accordingly, the hybrid-learning algorithm combines the methods
Here {pi, qi, ri} represents the parameter set and layer 3 output is of least squares and back propagation gradient descent method,
represented by w . which is applied for a fruitful or a powerful search of the excellent
i

Layer V: In layer 5, the total output of ANFIS is calculated by a parameters of ANFIS. More indicatively, the algorithm of hybrid
single node as: learning in the forward pass, the node output goes forward until layer
4 and parameters of the consequent part are recognized by method
∑ wifi of least squares. The propagation of the error signal backwards in
OP i5
= = ∑=
Total Output wi f i i
(9) the backward pass and the premise parameters are corrected by the
∑ wi
i gradient descent technique. As stated earlier, the consequent part
i
parameters thus recognized are excellent under condition that the
Computation of parameters (Hybrid algorithm) premise part parameters are confirmed. Consequently, the approach
of hybrid algorithm converges much faster since in the original back-
Accession parameters are the premise parameters {ai, bi, ci} in propagation method, it minimizes the dimension of the search space.
ANFIS, which reports the structure of MFs, and {pi, qi, ri} consequent A comprehensive and complete description can be found for this
parameters which report the final output of this ANFIS system. In algorithm.29
1986, backpropagation algorithm by Rumelhart, the basic learning
rule of an adaptive network based on the gradient descent rule, can The defuzzification
be successfully applied to calculate these parameters.28 Nevertheless,
Jang insisted that the rule of gradient descent is normally slow and Basic ANFIS takes the fuzzy inputs or the crisp inputs, but the total
gets trapped in local minima.22 Jang suggested a much faster learning and final outputs are sets of fuzzy. Consequently, a defuzzification
algorithm, which merges the technique of least squares estimate approach is required to convert a fuzzy set to a value of crisp. By21
(LSE) and gradient descent to locate parameters as presented below: the crisp output is usually secured using different defuzzification
approaches. The approach of Takagi-Sugeno is applied for the present
The adaptive network has only one output and is supposed to be: inspection.26 Xiong, et al.30 do not have a direct defuzzification
technique or rather they combine two techniques, the decision of
output = F ( I , S ) (10)
logic and the procedures of defuzzification into the one composite
Here, the input variables set is I and the parameters set is S. If procedure.
the existence of function H is there in such a way that the composite
function H◦F in rare elements of S is linear, then the method of least Model configuration
squares is used to recognize these elements. Precisely, parameters of The course and broad facts regarding applied models are discussed
the set S are divided into two sets: in this section.
S S1 ⊕ S 2
= (11) Non-linear time series forecasting
Here ⊕ sign shows the direct summation in such a way that S2 These forecasting are to be made with the help of Neuro-fuzzy.
elements possess linear H◦F. Then H is applied to Eq.10, to get: For this purpose, we have observed and evaluated the flow forecast of
the three gauge stations (Tarbela Dam in FATA, Chashma Barrage in
H ( output ) = H  F ( I , S ) (12) Punjab and Sukkur Barrage in Sindh). Using Neuro-fuzzy system, we
have selected readings of ten years daily stream flow from 2002-2011.
Which is linear for S2 elements; For the provided values of S1 Each year contains data of 6-months (from 1st April to 30th September
elements, the training data P can be plugged into Eq.12 in order to peak months). The daily stream flow data of six peak months of the
attain the equation of matrix; foresaid stations in 2012 is also chosen to test and verify the model.
AX = B (13) This data is then given to the system network.

Here, the unknown vector X has the parameters of S2 elements. Recruited of the neuro-Fuzzy network
Suppose lS2l = M, then P x M, M x 1 and P x 1 are the dimensions The managerial fuzzy interference network was chosen to be
of A, X and B. Therefore, for the training data the no. of pairs (P) executed by Sugeno model accompanied by the product operator. Two
are generally larger than the no. of linear parameters (M). This is an categories of Neuro-fuzzy system network have also been implemented
over-determined problem and usually Eq.13 has no accurate solution. that have different membership functions. Flow discharge is taken at

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 142

two values, the present time dt and the past time dt–1 for one network of three previously mentioned gauge stations reveals that the surveyed
and flow discharge is taken at three values that are, the present time data is highly varied and disordered. Flow behavior of Tarbela Dam,
dt, the past time dt–1 and one step before the past time dt – 2. Five Chashma and Sukkur Barrages, from 2010 to 2012, are displayed in
different ANFIS models are considered as 2, 3, 4, 6 and 8 membership Figure 5. Parameters which are interrelated to three foregoing stations
functions. For all model networks the membership functions are are provided in Table 1. From the Table 1, it is clear that standard
predicted as Gaussian and Triangular shaped once. Hence the first deviation and the difference between the maxima and minima are
system network uses dt and dt –1 as an input for two variables x and large, so that gauge station modeling will be complex. In this regard,
y and the second network system uses dt, dt –1 and dt–2 as an input for for fuzzy network, we consider input data for ten years from year
three variables x, y and z for the executed fuzzy rule base. Then in 2002 to year 2011 of the daily stream flow to check different models
accordance with the relation the yielding results are generated as the as training data. Moreover, different models were also tested using
final outcomes in defuzzification. We can see that ten different system testing data for only 1 year i.e. 2012, the daily stream flows (6 peak
networks are applied and the outcomes are determined and discussed months’data) for each gauge station. Entered the Input data into the
in the coming sections. neuro-fuzzy network and results are shown in Tables below.

Work explanation
Various categories of traditional statistic can be referred to as
statistical work explanation. But here we used root mean square error
(RMSE). This estimation test can also be estimated by using the
below equation.
p 2
∑in=1 ( di − di )
o
RMSE = (15)
n
o p
Where, d is the observed flow of stream at any time t, and d is
i i

the predicted flow of stream at any time t.

Result testing and review

For the present study, we have collected the data from Federal
Flood Commission (FFC), situated in Islamabad, which is comprised Figure 5 Stream flows of Tarbela, Chashma and Sukkur Barrage through 2010
of eleven years recording at three gauge stations. Data investigation – 2012.
Table 1 Estimated parameters corresponding to Dam and Barrages for 10 years & 1 year alone (2002–2011 & 2012)

Estimated parameters Tarbela Dam Chashma Barrage Sukkhur Barrage

2002–2011 2012 2002–2011 2012 2002–2011 2012

Average (fps) 141643.2 123941 177492.7 158845.2 116962.8 81611.91

Standard deviation (fps) 88743.25 76086.45 101397.2 79063.51 124973.6 47645.55

Minimum amount (fps) 18800 26000 23493 26169 16405 15630

Maximum amount (fps) 557100 284000 957309 276745 1130995 214780

Ratio of average to standard deviation (fps) 1.6 1.63 1.75 2.01 0.94 1.71

Outcomes and analysis stations. It is of great importance to note that the mentioned outcomes
are informed for Gaussian membership functions only because its
For obtaining outcomes, we have processed the network in various calculations and outcomes are better than the Triangular membership
approaches and RMSE is used to figure out the outcomes. Tables functions with error tolerance 0.001.
2-4 are the calculated results for the three previously mentioned
Table 2A Error evaluation for Tarbela dam in 2012 of flow prediction as testing data

Serial no. Different input variations No. of membership functions RMSE

1 dt, dt – 1 2 113.787

2 dt, dt – 1 3 108.474

3 dt, dt – 1 4 105.902

4 dt, dt – 1 6 104.87

5 dt, dt – 1 8 97.42

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 143

Table Continued....

Serial no. Different input variations No. of membership functions RMSE

6 dt, dt – 1, dt - 2 2 101.913

7 dt, dt – 1, dt - 2 3 90.465

8 dt, dt – 1, dt - 2 4 85.161

9 dt, dt – 1, dt - 2 6 68.037

10 dt, dt – 1, dt - 2 8 51.114

Table 2B Error evaluation for Tarbela dam in 2002-2011 of flow prediction Table 3B Error evaluation for Chashma Barrage in 2002-2011 of flow
as training data prediction as training data

Different input No. of membership Different input No. of membership

Serial no. RMSE Serial No. RMSE
variations functions variations functions

1 dt, dt – 1 2 155.639 1 dt, dt – 1 2 186.708

2 dt, dt – 1 3 134.878 2 dt, dt – 1 3 169.368

3 dt, dt – 1 4 129.603 3 dt, dt – 1 4 162.853

4 dt, dt – 1 6 125.749 4 dt, dt – 1 6 156.526

5 dt, dt – 1 8 124.374 5 dt, dt – 1 8 156.105

6 dt, dt – 1, dt - 2 2 129.476 6 dt, dt – 1, dt - 2 2 158.58

7 dt, dt – 1, dt - 2 3 117.745 7 dt, dt – 1, dt - 2 3 155.167

8 dt, dt – 1, dt - 2 4 114.935 8 dt, dt – 1, dt - 2 4 148.86

9 dt, dt – 1, dt - 2 6 110.249 9 dt, dt – 1, dt - 2 6 141.91

10 dt, dt – 1, dt - 2 8 107.461 10 dt, dt – 1, dt - 2 8 139.753

Table 3A Error evaluation for Chashma Barrage in 2012 of flow prediction Table 4A Error evaluation for Sukkur Barrage in 2012 of flow prediction as
as testing data testing data

Different input No. of membership Different input No. of membership

Serial no. RMSE Serial no. RMSE
variations functions variations functions

1 dt, dt – 1 2 117.188 1 dt, dt – 1 2 86.433

2 dt, dt – 1 3 109.666 2 dt, dt – 1 3 80.395

3 dt, dt – 1 4 107.837 3 dt, dt – 1 4 75.492

4 dt, dt – 1 6 101.392 4 dt, dt – 1 6 68.732

5 dt, dt – 1 8 95.455 5 dt, dt – 1 8 63.513

6 dt, dt – 1, dt - 2 2 114.15 6 dt, dt – 1, dt - 2 2 77.205

7 dt, dt – 1, dt - 2 3 100.935 7 dt, dt – 1, dt - 2 3 69.544

8 dt, dt – 1, dt - 2 4 97.08 8 dt, dt – 1, dt - 2 4 60.337

9 dt, dt – 1, dt - 2 6 77.261 9 dt, dt – 1, dt - 2 6 48.093

10 dt, dt – 1, dt - 2 8 69.699 10 dt, dt – 1, dt - 2 8 44.276

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 144

Table 4B Error evaluation for Sukkur Barrage in 2002-2011 of flow prediction respectively for different inputs and different membership functions.
as training data Curves in Figure 6A, indicate the predicted values and surface area for
Tarbela dam using observed data and applied neuro-fuzzy technique
Different input No. of membership
Serial No. RMSE for the year 2012. In the graph above of predicted values in Figure
variations functions
6A, the output data on y- axis means stream data values in fps and on
1 dt, dt – 1 2 170.088 x-axis index means no. of days i.e. 183 peak days for the year 2012
and also Figure 6B indicates the predicted values and surface area
2 dt, dt – 1 3 149.33 for the years 2002 to 2011. In the graph above of predicted values in
Figure 6B, the output data on y- axis means stream data values in fps
3 dt, dt – 1 4 138.725 and on x-axis index means no. of days i.e. 1830 peak days for above
mentioned years.
4 dt, dt – 1 6 118.698
Chashma barrage results
5 dt, dt – 1 8 112.94 Same as previous calculations for Tarbela Dam, we can see the
results of Chashma barrage for flow prediction in year 2012 and
6 dt, dt – 1, dt - 2 2 135.046 in years 2002 - 2011 are presented in Table 3A and 3B. Curves in
Figure 7A, indicate the predicted values and surface area for Chashma
7 dt, dt – 1, dt - 2 3 114.976 barrage using the observed data and applied neuro-fuzzy technique
for the year 2012 and Figure 7B shows the same for the years 2002
8 dt, dt – 1, dt - 2 4 106.599 to 2011.

9 dt, dt – 1, dt - 2 6 93.434 Sukkur barrage results

The predicted results of Sukkur barrage of daily flow for different
10 dt, dt – 1, dt - 2 8 87.921 inputs and MFs in years 2012 and in years 2002 - 2011 are presented
in Tables 4A&4B respectively. Curves in Figure 8A, indicate the
Tarbela dam results predicted values and surface area for Sukkur barrage using the
observed data and applied neuro-fuzzy technique for the year 2012
Results of Tarbela dam for daily stream flow prediction for the
and Figure 8B is for the years 2002 to 2011.
year 2012 and in years 2002-2011 are presented in Table 2A and 2B

Figure 6A The predicted values and surface area for Tarbela Dam in year 2012.

Figure 6B The predicted values and surface area for Tarbela Dam in year 2002 to 2011.

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 145

Figure 7A The predicted values and surface area for Chashma Barrage in year 2012.

Figure 7B The predicted values and surface area for Chashma Barrage in year 2002 to 2011.

Figure 8A The predicted values and surface area for Sukkur Barrage in year 2012.

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 146

Figure 8B The predicted values and surface area for Sukkur Barrage in year 2002 to 2011.

Discussion on flood analysis results

We can see the modeled results in Table 2A–Table 4B. It is apparent
that the results are much better and faster by the application of the
ANFIS model. Comparing the results of these three stations shows
from the ratio of average to standard deviation and by calculating the
error RMSE, a discrete reduction in the errors in all by increasing the
inputs than MFs is shown in this Neuro-fuzzy model. It is surprising
that the RMSE error is slowly decreasing with an increase in the
membership functions, while increasing the number of inputs makes
efficiency decrease in the calculated error. So one thing should be
apparent that suitable number of membership functions should be
interpreted and for this the rise in inputs is a better option for predicted
values.
Furthermore, we find that the results of the three stations indicate
that the best results can be attained by the structure of neuro-fuzzy
model and the best outcomes with minimum errors are for the year
2012, for Tarbela Dam using data inputs dt, dt – 1, dt – 2 and eight
membership functions (MFs). Table 2A shows these results. From
Table 3A, it can be seen that for Chashma barrage the structure with
Figure 9A Structure for 2-inputs dt & dt–1.
minimum error using data inputs dt, dt – 1, dt – 2 and eight membership
functions (MFs) was obtained for the year 2012. But for Sukkur
barrage where the flood risk is very high, Table 4b shows an enormous
decrease or drop in the error i.e. 87.92 for large amount of data
i.e. 10 years data used on same structure with dt, dt–1, dt –2 inputs and
eight membership functions (MFs) while the error in the structure
with dt, dt–1 inputs and two membership functions is calculated as
170.087. In Table 1, very low of ratio of average to standard deviation
for the 10 years’ data is shown for Sukkur barrage.
Figure 9A & 9B shows that the performance evaluation for the
three-gauge station relies on the no. of membership functions for two-
input structure and three-input structure. It is visible that better results
might appear by increasing the input data {dt, dt – 1, dt – 2 .......dt – n} to
the network instead of increasing the membership functions.

Conclusion
An Adaptive Neuro-Fuzzy Inference System (ANFIS) has been
Figure 9B Structure for 3-inputs dt, dt–1&dt–2. executed at three different gauge stations to presume cyclic behavior of

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175
 Copyright:
A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan ©2019 Sami et al. 147

river flow. Input variables and membership functions were applied 11. Karunanithi N, Grenney WJ, Whitley D, et al. Neural networks for river
using two different neuro-fuzzy systems performed 5-times with flow prediction. J Comput Civ Eng, ASCE. 1994;8(2):10–220.
different membership functions; 2, 3, 4, 6, 8 with 2 and 3 inputs data. 12. Salas JD, Smith RA. Physical basics of stochastic models of annual flows.
To do so, we have accumulated ten years’ stream flow data of three Water Resour Res. 1981;41:428–430.
gauge stations of River Indus flow and have used it as training data.
13. Stedinger JR, Taylor MR. Synthetic stream–flow generation model
One-year stream flow behavior is executed as testing data. The system
verification and validation. Water Resource Res. 1982;18(4):909–918.
is executed in different levels. For this system, it is apparent that better
results may be obtained by raising the no. of inputs to the network 14. Box GEP, Jenkins GM. Time Series Analysis: Forecasting and Control.
instead of raising the no. of membership functions. Comparison Holden–Day, Oakland, CA. 1970.
between the observed and the predicted data values and this analysis of 15. Valencia RD, Schaake JC. Disaggregation processes in stochastic
qualitative data achievement of the model stipulates that ANFIS can be hydrology. Water Resour Res. 1973;9(3):58–585.
utilized fruitfully for time series hydrological data. The ANFIS model
16. Daniyal Hassan, Rakhshinda Bano, Kamran Ansari, et al. Hydrology of
is very adaptable and has possibilities of integrating the fuzzy nature
Upper Indus Basin. 2010.
of real world systems.
17. Sudheer KP, Nayak PC, Rangan DM. Rainfall runoff modeling using
Acknowledgment artificial neural network technique, Report No. CS/AR–16/1999–2000.
National Institute of Hydrology, Roorkee, India. 2000.
None.
18. Nayak PC, Sudheer DM, Rangan, et al. A neuro–fuzzy computing
Conflicts of interest technique for modeling hydrological time series. J Hydrol. 2004;291(1–
2):52–66.
The authors declare that there is no conflict of interest.
19. Sugeno M, Yasukawa T. A fuzzy–logic based approach to qualitative
modeling. IEEE Transactions on Fuzzy Systems. 1993;1(1):7–31.
References
20. Ojala T. Neuro–Fuzzy systems in control. M Sc. Thesis, TempereUniversity
1. Annual Flood report. From Federal Flood Commission. Department of
of Technology, Tampere, Finland. 1995.
Ministry of Water and Power, and Government of Pakistan–Islamabad.
2010. 21. Brown M, Harris C. Neuro–fuzzy Adaptive Modeling and Control.
Prentice Hall. 1994.
2. Ologunorisa TE, Abawua MJ. Flood risk assessment: A review. App Sci
Environ Mgt J. 2005;9(1):57–63. 22. Jang JSR. Rule extraction using generalized neural networks. In
Proceedings of the fourth IFSA World Congress. Volume for Artificial
3. Smith K. Environmental hazards. Rutledge, London. 1996.
Intelligence. 1991;4:82–86.
4. Khan B, Iqbal MJ, Yosufzai MAK. Flood risk assessment of River Indus
23. Jang JSR. ANFIS: Adaptive network based fuzzy inference systems.
of Pakistan. Arab Journal of Geoscience. 2011;4(1–2):115–122.
IEEE Trans. Syst., Man Cybernet. 1993;23(3):665–685.
5. Kalma JD, Laughlin GP. Flood risk mapping for landscape planning:
24. Mamdani EH, Assilian S. An experiment in linguistic synthesis with a
a methodology. International Journal of Theoretical and Applied
fuzzy logic controller. Int J Man Mach Stud. 1975;7(1):1–13.
Climatology. 1990;42:41–51.
25. Tsukamoto Y. An Approach to Fuzzy Reasoning Method. In: Gupta M,
6. Khan G. Flood hazard assessment and mitigation along River Indus
et al. Advances in Fuzzy Set Theory, North–Holland, and Amsterdam.
from Chashma Barrage to Sukkur Barrage using satellite image. M.Phil.
1979.
thesis, Institute of Space and Planetary Astrophysics, University of
Karachi. 2007. 26. Takagi T, Sugeno M. Fuzzy identification of systems and its application to
modeling and control. IEEE Trans Syst Man Cybernet. 1985;15(1):116–
7. Nawaz F, Shafique M. Data integration for flood risk analysis by using
132.
GIS/RS as tools. In proceedings of MAP Asia, Second Annual Asian
Conference and Exhibition in the field of GIS, GPS, Aerial Photography 27. Sugeno M, Kang GT. Structure identification of fuzzy model. Fuzzy Sets
and Remote Sensing. 2003. Syst. 1988;28(1):15–33.
8. Burn D, McBean E. River flow forecasting model for the Sturgeon 28. Rumelhart DE, Hinton and GE, Williams RJ. Learning representations
River. J Hydraulic Eng. 1985;111(2):316–333. by back–propagating errors. Nature. 1986;323:533–536.
9. Awwad H. Valdes J, Restrepo P. Streamflow forecasting for Han River 29. Jang JSR, Sun CT. Neuro–fuzzy modeling and control. Proceedings
Basin, Korea. J Water Resour Planning Manage. 1994;120(5):651–673. IEEE. 1995;83(3):378–406.
10. El–Fandy M, Ashour Z, Taiel S. Time series models adoptable for 30. Xiong LH, Shamseldin AY, O’Connor KM. A nonlinear combination of
forecasting Nile floods and Ethiopian rainfalls. Bull Am Meteorol Soc. the forecasts of rainfall–runoff models by the first order Takagi–Sugeno
1994;75(1):1–12. fuzzy system. Journal of Hydrology. 2001;245(1–4):196–217.

Citation: Sami SB, Siddiqi TA, Iqbal MJ. A neuro-fuzzy modeling for the hydrological time series of floods of river indus of Pakistan. Int J Hydro.
2019;3(2):138‒147. DOI: 10.15406/ijh.2019.03.00175