Statistical evaluation of chemical and biological

contamination in Yamchi Dam basin water

Romina Rasuli Asl and Hossein Saadati*

Department of Natural Resourses, Ardabil Branch, Islamic Azad University, Ardabil, Iran

ABSTRACT

Rivers are the most vulnerable water resources due to carry urban and industrial wastewater and agricultural wastes.

In this study, the trend of changes of water quality of main branches of Yamchi Dam and water of dam reservoir dur-

ing the time series were evaluated according to the Spatiality changes and the most sensitive branch was identi ed

due to produce contamination as well as the time and location of changes (amounts) of water quality parameters to

Yamchi Dam were evaluated. According to the statistical results in 2001 to 2015 related to the stations studied, it can

be concluded that Nir River among all the variables studied is as the most pollutant river from the three main riv-

ers of the upstream region of Yamchi Dam. According to the results, the river’s water quality parameters from 2001

onwards has been increasing process, variables of fecal coliform, bicarbonate, DO, BOD and sodium respectively, as

main factors have the greatest impact on water quality survey of the main branches of Yamchi Dam. Three main

branches of Yamchi Dam are different in terms of rate and amount of variables mentioned and have had signi cant

effect on water quality of Yamchi Dam. Variables studied except DO in all stations are considered pollutant in terms

of the drinking water standard

KEY WORDS: YAMCHI DAM, WATER QUALITY, STATISTICAL INDICATORS

ARTICLE INFORMATION:

*Corresponding Author: Ah.saadati@iauardabil.ac.ir

Received 27

Nov, 2016

Accepted after revision 20

March, 2017

BBRC Print ISSN: 0974-6455

Online ISSN: 2321-4007

Thomson Reuters ISI ESC and Crossref Indexed Journal

NAAS Journal Score 2017: 4.31 Cosmos IF : 4.006

reserved.

Online Contents Available at: http//www.bbrc.in/

INTRODUCTION

Today, surface water quality is one of the major concerns

and considers a health index for community (Charu and

Verma, 2008). Data that describes temporal and spa-

tial variations of water quality in river can be used to

identify the relative importance of natural and human

effects (Ramazani and Hashemi, 2011). Over the past

decade, monitoring the water quality of the river has

increased by measuring water quality parameters (Bu

et al, 2010). Therefore, evaluating water quality is very

important because it directly affects public health and

life of aquatic ecosystem (Dixon and Chysol, 1996). P.

et al (2009) used multivariate regression analysis includ-

ing principal component analysis (PCA) and cluster (CA)

to study the spatial and seasonal variations of surface

Biosci. Biotech. Res. Comm. Special Issue No 1:87-93 (2017)

88 STATISTICAL EVALUATION OF CHEMICAL AND BIOLOGICAL CONTAMINATION BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS

Romina Rasuli Asl and Hossein Saadati

FIGURE 1. Position of the study area

water quality of Haraz Watershed in Iran (during the

summer and fall 2007 and winter and spring 2008). To

assess the temporal and spatial variations of water of

Jajrud River in Iran, they investigated water samplesin

a three-year period for each month in the 18 stations

with PCA and CA statistical analysis. CA and PCA led to

similar results, box plots show that PCA can show the

temporal and spatial variations approximately (Razm-

khah et al 2010 ).

Vega and colleagues in evaluating the seasonal

changes and pollutants effects of river water quality of

Pisuerga Spain using statistical analysis investigated var-

iables of physicochemical of water during 2.5 years from

3 stations. The PCA showed that the mineral amount,

human pollution and the temperature has dropped.

(Bardvaj et al 2010) In investigating water quality of Kali

Gandaki River in India using PCA, they analyzed factors

such as shelves, poor drainage, ion exchange, intensity

of use of nutrients and household pollutions. The results

of this study showed that in some areas, due to increase

of alkalinity, water becomes hard and is not suitable for

drinking and irrigation purposes. To assess water qual-

ity of upstream basin of Yamchi Dam,  rst the quality

parameters related to the main branch of Nir, Lai and

Jurab were collected from the Department of environ-

ment and regional water organization of Ardabil prov-

ince from 2001-2002 to 2014-2015 which contains con-

centrations of salts (TDS), acidity (pH), dissolved oxygen

(DO), biological oxygen demand (BOD), chemical oxygen

demand (COD) and total hardness (TH).

MATERIAL AND METHODS

THE STUDY AREA

Yamchi Dam basin with area of 709.18 square kilom-

eters is placed in the geographic area 28” 46’ 47 º to

57” 05’ 48 º east longitude and 24” 01’ 38 º to 15” 09’

38 º north that limits from north with Meshkinshahr,

from North-East with Ardebil, from west and south to

Eastern Azarbaijan and from southeast to the basin of

Ghuri Chay. Yamchi Dam main branches included Nir-

chay, Laichay and Lamchay that after connecting to

each other make a great river that is called up Balqe-

lychay that is one of the important branches of Aras

River in the North West of Iran that  ows in South North

and collects the waters of the main branches of Yam-

chi Dam from East and South of the area. YamchiDam

main branches because of  owing from important resi-

dential areas such as Nir and villages in the region and

also placing some industrial and resort centers near the

river and because of the drinking and agriculture water

supply of area is one of the major dams of Azerbaijan

region.On the other hand, the area because of the rul-

ing of the above conditions and serious threats caused

by wastewater of industry and domestic sewage is con-

stantly threatened by pollution. Therefore, such a situa-

tion increases the vulnerability of the study area to the

issue of pollution. Figure 1 shows Yamchi Dam position

in the North West of Iran, West of Ardebil and also the

position of hydrometric stations studied.

RESEARCH METHODOLOGY

Compliance with statistical principles to prevent from

wasting time and increasing the accuracy,  rst the

variables studied were minimized. Since the number

of parameters mentioned above is large and it is not

possible to check all of them in the form of a master's

thesis therefore, it was acted to minimize the variables

using principal component analysis (PCA). After deter-

mining the main variables of all three stations (Nir,

Lai and Jurab) by adopting a common statistical base,

shared variables were selected from three stations. After

BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS STATISTICAL EVALUATION OF CHEMICAL AND BIOLOGICAL CONTAMINATION 89

Romina Rasuli Asl and Hossein Saadati

selecting the common parameters at the later stage, it

was acted to group comparisons based on common vari-

ables. So that it was acted to investigate the presence or

absence of signi cant difference between the three main

branches of Yamchi Dam ie Nir, Lai and Jurab using mul-

tivariate analysis of variance, at this stage the presence

or absence of difference between the three stations is

determined. After this stage, it was acted to group aver-

ages of the variables among three stations and it became

clear which of the variables of these three groups are

different. In the next step, it was acted to compare the

amount of shared variables with standard value using

One Sample T-test; the purpose of this step is to identify

the most polluted river of the main branches of Yam-

chi Dam. In this study, a sample T-test hypotheses are:

1. The null hypothesis: The value of investigated vari-

ables is in permitted range. 2. Hypothesis: The value of

investigated variables is not in permitted range. Finally,

to determine the trend of changes of data and identify

the type and time of it, it was acted to analyze the trend

of time series using Mann-Kendall non-parametric test,

and the presence or absence of changes was speci ed.

Kolmogorov-Smirnov test was used to assess the

normality of variables. Determining the appropriateness

of data was performed using the KMO index and

Bartlett test, in components analysis, maximum equity

(Eigenvalue) belongs to the  rst component (Pc1) and

gradually by increasing class of components, this

amount decreases. It should be noted in this method, each

component is independent of other components (Yao et

al., 2013). Each component is a linear combination of

variables that its equation can be shown as equation 3-1

(Jolliffe, 1986).

In this equation, PC is the main component, is the coef-

 cient or special vector (Eigen Vector) and is variable

considered (Jolliffe, 1986). In order to select the number

of effective components, components were selected that

their Eigenvalue was more than one. To interpret the

effective features in a component that controls the most

changes, selection criteria was used (Doran and Parkin,

1994).

In this equation, SC is selection criteria, PC is the main

component and Eigenvalue is the special value (Doran

and Parkin, 1994). After identifying the most in uen-

tial variables in each station by considering the common

base among the main station variables, variables that

were present in all three stations were selected as the

major variables in qualitative studies of upstream water

of Yamchi Dam. All statistical analyzes were conducted

in SPSS v.20 software.

In this study, the method of multivariate analysis of

variance was used to investigate the signi cant differ-

ence between the variables of factor analysis (including

 ve variables of bicarbonate, sodium, DO, BOD and fecal

coliform) in three main branches of Yamchi Dam (Nir,

Lai and Jurab).

Among the groups after the signi cant difference

among the main branches of Yamchi Dam was approved

then, using Duncan method from grouping methods of

averages, it was acted to identify the different variables

in the three branches.

In order to understand the trend of changes of data

and identify the type and time of it, it was acted to

analyze the time series using the Mann-Kendall non-

parametric test and the presence or absence of changes

was speci ed. In this study, the graphical method of

Mann - Kendall test was used. About the graphical

method, mentioning points about the statistic U and U

‘is necessary. If the sequence U and U ‘ based on i is

drawn as a graph, in the signi cance mode of the trend,

two graphs at the starting point of phenomena outside

the scope intersect each other and will move in opposite

direction of each other, this point of collision is called

mutation. While if there isn’t trend, two sequences U and

U ‘move almost in parallel or will act in several times of

collision, so that does not lead to change direction. U

graph to the year (axis X) is drawn and for that signi -

cance of trend and point of its mutation to be achieved,

the sequence U ‘is de ned.

RESULTS AND DISCUSSION

Principal components analysis (PCA) was used in

order to determine the minimum effective variables in

investigating qualitative changes of upstream water of

Yamchi Dam (main branches of Nir, Lay and Jurabchay).

First KMO factor was used in determining the suitability

of the data for principal components analysis to achieve

this goal. The value of this factor is always variable

between zero and one. If the value of this ratio is less

than 0.5, the data will not be suitable for factor analysis

and if its value is between 0.5 to 0.69, it can be analyzed

main components more carefully. If KMO coef cient is

larger than 0.7, principal components analysis in the

decrease of data will be effective (Jolliffe, 1986). As well

as to ensure correlation between input variables or inde-

pendent, Bartlett’s test was used. Based on the results,

KMO coef cients in three main branches of Nir, Lai and

Jurabchay was obtained 0.769, 0.782 and 0.724 respec-

tively that the amount con rms the correlation between

input variables for principal components analysis. Bar-

Romina Rasuli Asl and Hossein Saadati

90 STATISTICAL EVALUATION OF CHEMICAL AND BIOLOGICAL CONTAMINATION BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS

Table 1. Investigate KMO and Bartlett tests to

detect changes in upstream water quality of

Yamchi Dam

Station

Test

KMO

Bartlett

test

Signi cant

level

Nir0.76986.7740.000

Lai0.78289.3450.000

Jurab bridge0.72482.7610.000

Table 2. Number of main components,: EV (Eigenvalue),: SC (selection index) and: Cu (cumulative variance)

in assessing the changes of water quality of the main branches of Yamchi Dam

Station

First main

component

Second main

component

Third main

component

Fourth main

component

Fifth main

component

Nir

EV8.746EV6.898EV2.538EV2.017EV-

SC0.169SC0.19SC0.313SC0.352SC-

Cu41.648Cu74.495Cu6.581Cu96.184Cu-

Lai

EV6.967EV5.365EV4.33EV2.109EV1.228

SC0.189SC0.215SC0.24SC0.344SC0.451

Cu34.837Cu61.664Cu83.313Cu93.358Cu100

Jurab

EV11.078EV3.702EV2.314EV1.966EV-

SC0.15SC0.259SC0.328SC0.356SC-

Cu55.391Cu77.901Cu85.473Cu95.304Cu-

Table 3. The results of multivariate analysis of variance

Source

Dependent

Variable

Type III Sum

of Square

df Mean

Square

F Sig.

Three froups

of Nir, Lai,

Jurab

Bicarbonate 59233.637 2 29616.818 180.443 .000

Sodium 1921.608 2 960.804 69.644 .000

DO 1.264 2 4.632 1.509 .000

BOD 5.621 2 2.810 8.002 .000

Fecal

coliform

5411.641 2 2705.820 4.072 .000

tlett’s test was signi cant for all three rivers in 0.01 so

the data are suitable for PCA.

The main components analysis was used to assess

the water quality changes of main branches of Yamchi

Dam. Therefore, variables that at least one of the coef-

 cients of them is used to form the component have

relatively high amount. According to Table 2, Nir, Lai

and Jurab stations had 4, 5 and 4 main components and

in total according to the considered criteria, it was speci-

 ed among 10 variables, Nir, Lai and Jurab stations have

respectively 9, 3 and 5 main variable. Because of the

variety in number and type of main variables in each

station, considering common base between main vari-

ables related to each of the three stations, three variable

of DO, BOD and fecal coliform were selected as the main

variables (effective factors). So the variables of fecal

coliform and DO and BOD as the main factors have the

greatest impact and other variables have the least impact

in investigating water quality of the main branches of

Yamchi Dam.

According to the results of analysis of variance,

there is a signi cant difference between the three main

branches. Values Sig. that for stations studied is less than

0.05, and then by 95% con dence level, there is a sig-

ni cant difference between the main branches of Yam-

chi Dam based on  ve variables (bicarbonate, sodium,

DO, BOD and fecal coliform). In other words, three main

branches of Yamchi Dam in terms of amount of vari-

ables mentioned are different with each other and have

signi cant effect on water quality of Yamchi Dam.

It was speci ed in the results of multivariate analysis

of variance that there is a signi cant difference between

three main branches, but it is not clear which of the

variables (bicarbonate, sodium, DO, BOD and fecal coli-

form) are different. Comparisons of the average of vari-

ables in groups surveyed by Duncan method showed

bicarbonate amount in rivers of all three with Sig. equal

one in each three main branches are different, sodium

amount among two branches of Nir and Jurab with Sig.

equal to 0.763, there is no signi cant difference but the

amount of sodium in the Lai branch with sodium in two

branches of Nir and Jurab is quite different. The number

of fecal coliform in the water of branches of Lai and

Jurab is very similar but the number of fecal coliform

Romina Rasuli Asl and Hossein Saadati

BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS STATISTICAL EVALUATION OF CHEMICAL AND BIOLOGICAL CONTAMINATION 91

in Nir branch is different with two other branches (Table

4).The amount of BOD in water, branches of Lai and

Jurab not have much difference with each other but the

amount of this variable in Nir branch is very different

from the other two branches. DO variable situation is

similar to variable BOD.

According to the results of a sample t test, the

signi cance level (Sig) is larger than 0.05, then, the null

hypothesis is con rmed with con dence of 95 percent.

As the value of Sig of one sample t-test in all parameters

except DO is less than 0.05 so the variables considered

in all stations except DO in terms of the standard of the

drinking water are considered pollutant. DO variable in

two stations of Lai and Jurab is equal to the threshold

of pollution and is not considered statistically pollutant.

According to the results of Mann - Kendall non-par-

ametric test, all stations follow the process of changing

the station of Nir that is why only the results of Station

Nir is provided. In the graph of the annual average DO

of station, a signi cant mutation with positive trend in

Table 4. A sample t test results with constant value of main three branches variables

One-Sample Test

Station

Common

variables

Degree of

freedom

Sig.

Difference

of average

Con dence level 95%

Less More

Nir

Sodium -98.84 8 0.000 -175.11 -179.2 -171.03

bicarbonate -14.31 8 0.000 -84.46 -98.07 -70.85

fecal coliform 4.8 8 0.000 47.61 24.75 70.47

DO -3.05 8 0.02 -0.46 -0.81 -0.11

BOD -51.93 8 0.000 -12.67 -13.24 -12.11

Lai

Sodium -2618.4 8 0.000 -193.26 -193.4 -193.09

bicarbonate -132.2 8 0.000 -155.68 -158.4 -152.9

fecal coliform 3.907 8 0.05 24.39 9.9 38.7

DO 0.206 8 0.862 0.056 -0.568 0.68

BOD -69.4 8 0.000 -13.015 -13.447 -12.582

Jurab

Sodium -145.6 8 0.000 -175.64 -178.4 -172.8

bicarbonate -9.809 8 0.000 -42.18 -52.097 -32.26

fecal coliform 6.353 8 0.000 58.31 37.146 79.48

DO -0.467 8 0.653 -0.098 -0.579 0.384

BOD -79.762 8 0.000 -11.92 -12.267 -11.57

FIGURE 2. Mann - Kendall non-parametric test, Nir Station

Romina Rasuli Asl and Hossein Saadati

92 STATISTICAL EVALUATION OF CHEMICAL AND BIOLOGICAL CONTAMINATION BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS

2007 can be seen that is presented as graph. In the graph

of annual average bicarbonate of station Nir, two sig-

ni cant mutations with positive and negative trend in

2003 and 2007 can be seen as in 2007 has undergone a

sudden change and governance of its trend is negative. It

is concluded from investigating the behavior of changes

U and U ‘related to annual average BOD of station Nir

that, without signi cant change and mutation has expe-

rienced the normal process of maximum Sodium. In the

graph of average annual fecal coliform of station Nir,

a signi cant mutation with the positive trend in years

2005 is seen. As in 2007, it has experienced a mutation

and its trend is declining.

According to the results of the previous sections and

referring to the multi-year average related to the sta-

tions studied (Fig. 3), it can be concluded that Nir river

in all the variables studied is as the most pollutant river

among the three main rivers of the upstream region of

Yamchi Dam.

CONCLUSION

According to the results of this study, it can be said

evaluating water quality conducted in this study has

integrated and comprehensive expertise knowledge. As

it can be acted to better understanding, manage and

reduce its pollution by the selection of the main fac-

tors in water quality. Based on the results obtained,

features such as fecal coliform, bicarbonate, DO, BOD

and sodium were selected as effective factors, which

are a combination of chemical and biological param-

eters of water. Research about selecting effective factors,

determining the minimum or the most optimal effective

characteristics on water quality in different regions can

be guidance for proper assessment of water quality by

spending a minimum cost and time. Reviews related to

multivariate analysis of variance indicate high accuracy

and ability to detect the presence or absence of differ-

ence between study stations.

FIGURE 3. Graph of average of variables investigated

in the main branches of Yamchi Dam

The use of this method in the analysis of different

groups with different variables is very effective and

provides signi cant results. About the identi cation of

variables that shows signi cant difference among the

stations studied, the Duncan method due to advantages

compared to other methods such as LSD, Dunnett and

can be compared groups (Nir, Lai and Jurab stations)

mutually and to be ensured the signi cant difference

between the variables of the stations. So it is concluded

Duncan method has a highly accurate in identifying dif-

ferent variables between stations. A sample T test as a

standard method to examine the value of variables with

standard value of them has a high ef ciency and the

results obtained of it indicate pollution of parameters

investigated in available stations. As Nir station has the

highest pollution among the stations studied. The rea-

son for this is the  owing Nir River through urban area

of Nir and adjacent villages as well as industries in the

region, which always production waste in these indus-

tries and urban sewage directly enter the river and is

a serious threat to aquatic and Yamchi Dam that is as

producer reservoir of drinking water for Ardabil. Of the

results obtained of time series of Mann-Kendall can be

clearly observed that parameters investigated during 13

years, have maintained its upward trend and are rising.

This implies the increase of human-produced pollutants

that without attention to the rights of posterity and due

to mismanagement is destroying environment.

RECOMMENDATIONS

Due to the sensitive situation of the region and the

expansion of residential and industrial areas adjacent

to rivers in the region, it is recommended to be used

drainage system suitable for agricultural lands

overlooking the river. In order to prevent the entry of

industrial wastewater and domestic sewage to the rivers

of area, sewage system is reconstructed in area and in

sections that is deprived of sewage, proper system of

sewage to be implemented. Lands use changes, espe-

cially deforestation and gardens and convert them to

residential areas in riverbanks to be prevented and in

accordance with the river privacy to be prevented the

entry of more pollution to groundwater and surface

water sources.

It is suggested to be achieved necessary information

from potential of land for various uses before use the

land in the area, with land capability assessment for dif-

ferent uses. According to the capabilities of GIS in land

planning, it is suggested to be implemented the logistic

plan in the region by taking advantage of the capabili-

ties of this system. It is suggested to be investigated the

effect of other parameters such as the amount of soil

Romina Rasuli Asl and Hossein Saadati

BIOSCIENCE BIOTECHNOLOGY RESEARCH COMMUNICATIONS STATISTICAL EVALUATION OF CHEMICAL AND BIOLOGICAL CONTAMINATION 93

erosion, rainfall and  ooding also with more details on

the amount of dissolution time.

REFERENCES

Bardwaj, V., D.S. Singh and A.K. Singh. (2010). Water quality

of the Chhoti Gandak River using principal component analy-

sis, Ganga Plain, India. Journal of Earth System Science, 119:

117-127.

Bu, H., Tan, X., Li, S. and Zhang, Q. (2010). Temporal and

spatial variations of water quality in the Jinshui River of the

South Qinling Mts, China. Ecotoxicology and Environmental

Safety. 73: 907-913 pp.

Charu, P. and Verma, N. (2008). Mulitivariate analysis of drink-

ing water quality in Bhopal, India. Envirionmental Monitoring

and Assessment. 140: 119-122 pp.

Dixon, W. and Chiswell, B. (2012). Review of aquatic monitor-

ing program design. Water Research. 30:1935-1984 pp.

Doran, J. W., and Parkin, T. B. (1986). De ning and assessing

soil quality. Journal of Soil Sciences Pages 3-21

Jolliffe, I. T. (1986) Principal Component Analysis Springer-

Verlag, NewYork.

Pezhman, A.H., Nabi Bidhendi, G.R., Karbasi, A.R., Mehradadi,

N., Esmaeili Bidhendi, M.(2009). Evaluation of spatial and sea-

sonal variations in surface water quality using multivariate

statistical techniques. International Journal of Environmental

Science and Technology (IJEST). 6:467-476 pp.

Ramezani, S. and Hashemi, H. (2011). Analysis of water qual-

ity of Zarrine Rud River with statistical technique of main

component analysis of Fourth Conference on Water Resources

Management in Iran. Amirkabir University of Technology,

Tehran.

Razmkhah, H., Abrishamchi, A. and Torkian, A. (2010). Evalu-

ation of spatial and temporal variation in water quality by

pattern recognition techniques: A case study on Jajrood River

(Tehran, Iran). Journal of Environmental Management. 91:

852-860 pp.

W. Doran and P Parkin (1996) (eds.) De ning soil quality for a

sustainable environment. il Science Society of America Special

Publication no. 35, Madison, WI.

Yao, H., Thornton, B., Paterson, E., (2012). Incorporation of

13C-labelled rice rhizodeposition carbon into soil microbial

communities under different water status. Soil Biology & Bio-

chemistry 53, 72e77.