Вы находитесь на странице: 1из 5

J

O

U

R

N

A

L

O

F

C

O

A

S

T

A

L

S

C

I

E

N

C

E

S

J O U R N A L O F C O A S T A L
J O U R N A L O F C O A S T A L

JOURNAL OF COASTAL SCIENCES

Journal homepage: www.jcsonline.co.nr

Volume 1 Issue No. 1 - 2014

ISSN: 2348 – 6740

Pages 58-62

Statistical evaluation of groundwater geochemistry: a case study between Chinnakuppam and Kulathur, South Chennai, Tamil Nadu, India

Amadeius Abin a* , Sibin Raj a , T.B. Aghil b , N.S. Magesh b , S.G.D. Sridhar a

a Department of Applied Geology, University of Madras, Chennai, Tamil Nadu 600 025, India b Centre for Geotechnology, Manonmaniam Sundaranar University, Tirunelveli, Tamil Nadu 627 012, India

`

A B S T R A C T

A R T I C L E

I N F O

The groundwater quality in the southeast coast of India is severely affected by sea water intrusion and it is a growing concern to the masses. Lack of sufficient recharge with only one prevailing monsoon season, the quality and quantity of the groundwater aquifers in this region are deteriorating. A geochemical study was conducted on the southern part

Received 2 January 2014 Accepted

 

of Chennai metropolitan city in the state of Tamil Nadu, India along the coastal reaches on developing sub-urban

12

March 2014

population areas. Groundwater samples from different locations along the coast were collected using random

Available online

 

sampling and analyzed for various physico-chemical parameters. The results were tabulated and further statistical work was carried out. The statistical approach provides a means to estimate the distribution characteristics of the data

 

19

March 2014

 

thus provide a means for the development of hypothetical models. The primary objective of the analysis was to verify certain observation found in the earlier analysis about the distribution properties of geochemical data. Tests were

Keywords

 

conducted for the goodness of fit on log-normal distribution with 5% confidence interval. Eight parameters were

Groundwater

tested on the probability distribution plot to estimate the best fit. It has been confirmed that the geochemical values

Geostatistics

does follow the log-normal distribution. As such, after obtaining the required parameters of suitable sample size,

Geochemistry

hypothetical models can be developed using this observation. The analysis support the earlier assumptions, suggesting

Coastal aquifer

an effective method of evaluation and estimation of errors.

Distribution analysis

Modeling

 

*Corresponding author, E-mail address: scienceloving@gmail.com

Southeast coast

Phone: +91 8891545334 © 2014 – Journal of Coastal Sciences. All rights reserved

India

1. Introduction

The numerical analysis of geochemical data is still an ill ventured horizon in the geological society and it is often discarded for the surreptitious mathematical complex equations that often accompany these analyses. Apart from stifling mathematical equations, the scope for statistical examination is valid without indulging in the extreme complexities. The proximity to urban landscapes exposes the in-situ groundwater resources to rapid encroachment, depletion and contamination. The natural influx of the saline water to the groundwater aquifers is a gruesome factor that further deepens the growing concern. Improper planning and rapid industrialization of the suburbs have paid heavy lapse on the groundwater potentials instilling fresh flow of saline water. Hydro-geochemical processes such as dissolution, precipitation, ion-exchange, and the residence time along the flow path control the chemical composition of the groundwater in the shallow alluvial aquifers (Apodaca et al. 2002). Groundwater condition along the south-east coast is a great concern due to the declining water level and increasing risk of salinity. The occurrence of only one prevailing monsoon underlies these concerns. The ever increasing quench of the limited water resources crush the strategic aquifers and pose serious threat for future generation. Geochemical qualitative analysis

is a direct estimate of the contamination providing vital clues regarding the quality and saline intrusion in the coastal region. The critical information regarding the complexity and vitality of the groundwater can be derived from these values. The sample collection and laboratory analysis often display heterogeneity in error that etches the final observation. However, for quantitative measurement of this error and to validate the final observations there is a need to exist a mathematical model. The observed data figures often follow a standard distribution. It has been mentioned by several authors that geochemical data are approximately log-normally distributed (Finney 1941; Ahrens 1965; Allegre and Lewin 1995). The aim of the present work is to study the distribution of the data and test the goodness of fit of the cumulative frequency distribution to the standard distribution curves, with the help of present day computer applications.

2. Study area

The study area is situated along the northeastern coast of Tamil Nadu that includes Chinnakkuppam, Vilambur, Tembakkam, Marakanam, and Kolathur from North to South. These sites fall in Kanchipuram district along the coast that stretches for about 30 km. Geographically, it is located between 12⁰20' and 12⁰52’30’’ N latitude

58

O R I G I N A L

A R T I C L E

J

O

U

R

N

A

L

O

F

C

O

A

S

T

A

L

S

C

I

E

N

C

E

S

and from 79⁰20’ to 80⁰17’30’’ E longitudes. The area is characterized by sand dunes of quaternary and recent period, Cuddalore sandstone of mio-pliocene age, shales and sandstone of upper gondwannas and charnockites of archaean era (Chidambaram et al. 2011). The greater part of the area mainly consists of metamorphic crystalline rocks of archaean age that belongs to the charnockite and khondalite groups (CGWB 2001). The upper most formation is coastal alluvium which is few meters thick and is underlined by thick sequences of crystalline rock. The quaternary/recent sediments weathered and fractured crystalline charnockite function as an unconfined aquifer system. The groundwater head follows the topography, i.e., it flows towards east and west from the central part of the area. The groundwater requirement of this area is mostly met by wells penetrating up to upper quaternary/recent sediments. During the study it is noticed that most of the dug well are tapping ground water from recent alluvium at a depth ranging from 0.5 m to 9.8 m. The study area map is shown in the figure1.

pH were measured on the field. According to APHA (1995) the groundwater samples were analyzed in the geochemical laboratory to find out anions and cations. Pre-cleaned polyethylene bottles were used for collecting water samples from open wells and bore wells. Groundwater samples were collected using simple random sampling method. Electrical Conductivity (EC) and pH were measured using conductivity and pH meters. TDS was computed from EC multiplied by 0.64. Na + and K + ions were determined by flame photometer. Total hardness (TH) as CaCO3, total alkalinity (TA) as CaCO3, calcium, carbonate, bicarbonate and chloride were analyzed by volumetric methods. Mg was calculated from TH and Ca contents. SO4 were estimated by Nephelometer techniques. Fluoride (F) was determined using the SPADNS method. All concentrations are expressed in milligrams per liter (mg/l). The results were tabled and statistical parameters were calculated. Means and standard deviations were estimated using the robust log-probability (MR) method, percentile concentrations and inter-quartile ranges (IQR) were estimated by

Fig.1 Location map of the study area

J O U R N A L O F C O A S T A L

The coastal area of Kanchipuram district is characterized by several strand lines, lagoons, mangroves, salt marsh, estuaries, creeks, barrier dunes, spits, beach terraces. The major water bearing formation being the coastal sands extending 30-40 feet, and the sand thickness is increasing towards the coast. In many places, the coastal water is brackish in nature due to the presence of Buckingham canal running parallel to the coast. The major zone of groundwater occurrence is the Palar mouth where the Palar river bed acts as a potential formation. The important mineral found in this stretch is silica sands along the Mamallapuram coast.

the maximum likelihood estimation (MLE) method, as described by Helsel (1990) and Helsel and Hirsch (1992). Computer application Minitab was used to test the goodness of fit with 5% confidence intervals.

4. Results and Discussions

Data represented in table 1 demonstrate each statistical parameters used in the present study. Analysis was primarily concentrated on the parameters mean, SD, variance and distribution parameters:

3. Methodology

A total of 30 water samples were collected from Chinnakuppam in the north to Kolathur in the south of the study area during pre- monsoon (June, 2011) and post-monsoon (January, 2012) from bore wells and dug wells to assess the quality of groundwater. The water samples were numbered starting from 1 to 30 that are collected from Chinnakuppam to Kolathur along East Coast Road (ECR). The EC and

skewness and kurtosis. Highest mean values are calculated for the EC values with median at 490. TDS similarly has a higher value; however in the case of elemental concentration sodium (Na) and bicarbonate (HCO3) have major peaks. TDS also display a high variation with SD of 276.95 while EC has highest standard deviation among variable analyzed. The increase in TDS in some well may be attributed by the nature of overlying rock material. In area having more clay content the TDS generally increases with recharge due to rainfall. It is also possible that rising level in the backwater contributes to the increase

59

O R I G I N A L

A R T I C L E

 

J

O

U

R

N

A

L

O

F

C

O

A

S

T

A

L

S

C

I

E

N

C

E

S

Parameter

Mean

StDev1

Variance1

CVariation1

Q1

Median1

Q3

 

Sum1

 

SSQ1

 

Skewness1

 

Kurtosis1

 

TDS

438.40

276.95

76701.90

63.17

257.50

338.00

636.00

13152.00

 

7990192.00

 

1.05

 

1.48

 

EC

631.60

400.42

160338.80

63.40

369.00

490.00

919.50

18948.00

16617382.00

 

1.08

 

1.58

 

pH

6.81

0.53

0.28

7.74

6.54

6.95

7.20

 

204.22

 

1398.25

 

-1.06

 

0.99

 

HCO3

147.30

110.51

12212.08

75.02

64.00

126.00

190.50

 

4419.00

 

1005069.00

 

1.46

 

2.43

 

TH

137.00

75.29

5668.90

54.96

80.25

118.50

196.00

 

4110.00

 

727468.00

 

0.48

 

-0.49

 

Ca

33.30

18.64

347.32

55.97

19.00

30.00

52.00

 

999.00

 

43339.00

 

0.41

 

-0.59

 

Mg

13.00

7.28

53.03

56.02

7.00

11.00

19.00

 

390.00

 

6608.00

 

0.60

 

-0.19

 

Na

71.53

59.21

3506.05

82.78

23.50

54.00

116.00

 

2146.00

 

255186.00

 

1.14

 

0.84

 

SO4

14.20

21.73

472.23

153.03

2.00

4.00

20.50

 

426.00

 

19744.00

 

2.12

 

4.18

 

Table 1. Statistical variation of different physico-chemical parameters in the study area

of ionic concentration. The high values of sodium and bi-carbonate are observed in the distribution analysis with sodium having a mean of 71 and bi-carbonate with 147. The SD is also having a higher value in the range of 59 and 110 which is in lieu with observed mean values. The median values are 54 and 126 for these ions.

Probability Plot for pH Goodness of Fit Test Lognormal - 95% CI Lognormal 99 AD =
Probability Plot for pH
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.979
P-Value = 0.012
95
80
50
20
5
1
5
6
7
8
9
pH
Probability Plot for EC
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.899
P-Value = 0.019
95
80
50
20
5
1
100
1000
10000
EC
Percent
Percent

Relatively lower concentrations are observed for the other elements namely magnesium, sulphate and calcium. The mean values for these elements are 33 for calcium, 13 for magnesium, and 14.2 for sulphate. Estimated mean concentrations are often seen to exceed the median values and this reflects the right-skewed cumulative distribution (Gilbert 1987; Helsel 1990). Probability graphs for TDS, pH, TH, Ca, Mg, Cl, Na, HCO3, SO4 are plotted with 95% confidence interval (Fig. 2). The general observation is all the data values tend to follow the log-normal distribution. The probability plot for EC,

display sufficient variation with lower p<0.05, suggesting the log normal distribution may not be a best fit for the values. Step like frequency percentiles are observed for the values of sulfates and sodium suggesting a recurrence in data values. The probability plots for electrical conductivity and pH show marked

Probability Plot for TDS Goodness of Fit Test Lognormal - 95% CI Lognormal 99 AD =
Probability Plot for TDS
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.945
P-Value = 0.014
95
80
50
20
5
1
10
100
1000
10000
TDS
Probability Plot for TH
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.549
P-Value = 0.145
95
80
50
20
5
1
10
100
1000
TH
Percent
Percent

minor deviations from the log-normal plots. The increase in the sodium and carbonate content and deviations in the electrical conductivity may be attributed to the regions of active saline intrusion along the coast. Provided with ample samples the values could be plotted along log-normal distribution to produce a hypothetical model .The possibility of larger sample space could be used to test the validity of central limit theorem, which is presently beyond the scope of this work. Probabilistic data analysis techniques often require very large sample populations.

60

O R I G I N A L

A R T I C L E

Probability Plot for HCO3 Goodness of Fit Test Lognormal - 95% CI Lognormal 99 AD =
Probability Plot for HCO3
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.290
P-Value = 0.588
95
80
50
20
5
1
10
100
1000
HCO3
Probability Plot for Ca
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.636
P-Value = 0.088
95
80
50
20
5
1
1
10
100
Ca
Probability Plot for Mg
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.536
P-Value = 0.156
95
80
50
20
5
1
1
10
100
Mg
Percent
Percent
Percent

However, a drawback of using probabilistic methods may include the time and expense required to gather a sufficient number of data to accurately represent the modeled system.

J O U R N A L O F C O A S T A L
J
O
U
R
N
A
L
O
F
C
O
A
S
T
A
L
S
C
I
E
N
C
E
S
Probability Plot for Na
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 0.538
P-Value = 0.154
95
80
50
20
5
1
1
10
100
1000
Na
Probability Plot for SO4
Goodness of Fit Test
Lognormal - 95% CI
Lognormal
99
AD = 1.048
P-Value = 0.008
95
80
50
20
5
1
0.1
1
10
100
1000
SO4
Percent
Percent

Fig. 2 Probability plot of various physico-chemical parameters of

groundwater in the study area.

5. Conclusion

Geochemical data represent the overall quality of the groundwater

resources and suggest the degree of contamination the aquifers are

subjected into. However, the absence of well-developed statistical

techniques in the interpretation often leads to anomalous

conclusions and errors in the results. Very small sample populations

and the absence of supporting data often limit statistical analysis and

hinder the final outcomes. The result of this test confirms to the

earlier assumptions of the geochemical data following the log-normal

distribution. The increasing concentrations of sodium and bi-

carbonate suggest zones of saline intrusion and contaminated

groundwater aquifers. However, development of hypothetical model

need to be supported with a large number of samples but will provide an excellent tool in tackling the issues of sustainable development along the coastal regions of the southeastern India.

References

Ahrens, L.H., 1965. Distribution of the elements in Our Planet. McGraw-Hill. Allegre, C.J., Lewin, E., 1995. Scaling laws and geochemical distributions. Earth and Planetary Science Letters 132, 1–13.

61

O R I G I N A L

A R T I C L E

APHA. 1995. Standard methods for the examination of water and waste water (19 th edn.). Washington D.C: American Public Health Association. Apodaca, L.E. Jeffrey, B.B. and Michelle, C.S. 2002. Water quality in shallow alluvial aquifers, Upper Colorado River Basin, Colorado. Journal of American Water Research Association 38(1), 133–143. CGWB. 2001. Groundwater year book of Kancheepuram district. Central Ground Water Board, Ministry of Water Resources, Govt. of India. Chidambaram, S. Karmegam, U. Prasanna, M.V. Sasidhar, P. and Vasanthavigar, M. 2011. A study on hydrochemical elucidation of coastal groundwater in and around Kalpakkam region, Southern India. Environmental Earth

Science 64, 1419–1431 Finney, D.J., 1941, On the distribution of a variate whose logarithm is normally distributed. Journal of the Royal Statistical Society (suppl) 7, 155–

161.

Gilbert, R.O., 1987. Statistical Methods for Environmental Pollution Monitoring. Van Nostrand Reinhold, New York. Helsel, D.R., 1990. Less than obvious: statistical treatment of data below the detection limit. Environmetal Science and Technology 24, 1766–1774. Helsel, D.R., Hirsch, R.M., 1992. Studies in Environmental Science 49. Statistical Methods in Water Resources. Elsevier,New York.

J

O

U

R

N

A

L

O

F

C

O

A

S

T

A

L

S

C

I

E

N

C

E

S

62

O R I G I N A L

A R T I C L E