APPLICATION OF ELECTRICAL RESISTIVITY AND HYDROGEOLOGY MODELLING METHODS TO MAP
AND FORECAST THE SALTWATER INTRUSION IN THÁI BÌNH PROVINCE

1NGUYỄN NHƯ TRUNG, 1TRỊ NH HOÀI THU, 2NGUYỄN VĂN NGHĨA

1Institute of Marine Geology and Geophysics, VAST;
2Division of Water Resource Management, MNRE

Abstract: This paper focuses on the application of electrical resistivity and hydrogeology modelling to map the freshwater/saltwater interface and forecast the saltwater intrusion caused by the freshwater withdraw. Based on VES data and pore water conductivity, measurements were carried out in Thái Bình coastal plain from 2006 to 2007, the total dissolved solids (TDS) and the spatial distribution of saltwater / freshwater interface in the Pleistocene aquifer were predicted and mapped.  Significant changes in pore water conductivity from the north to the south were detected by these methods. A low resistivity value of the Pleistocene aquifer (< 9 ohm) or high conductivity of Pleistocene pore water (> 0.24 S/m) implies the saltwater / freshwater boundary. The results of spatial distribution of TDS indicate the saltwater affected aquifer is confined to the southern part of the Thái Bình province (from the longitude 20030’ N to the south).

     Based on the modified Archie’s law and the correlation between the conductivity and TDS of the pore water suggested by the authors, two empirical relations between pore water resistivity of the Pleistocene beds and bulk resistivity and pore water resistivity and TDS are obtained. They can be used to convert the computed bulk resistivity and pore water conductivity to the TDS, and thus, leading to the confine of the groundwater contamination. Based on the TDS maps, the hydrological modelling method is conducted for forecasting a safety-pumping rate of the region and groundwater contamination from present time to the 2030 year.

I. INTRODUCTION

Thái Bình, a coastal plain province, is fast urbanized and becomes an agricultural and industrial centre in the Red River Delta. With the complicated hydrogeological conditions, although having abundant reserves of groundwater, but almost aquifers are confronted with saltwater intrusion. Although there is not a large-scale groundwater exploiting station, but thousands of UNICEF wells had been drilled uncontrollably to meet the freshwater demand, reducing the water table. It may be the main reason of saltwater intrusion. Actual state, rate, forecasting and protection of saltwater intrusion should be considered to exploit the groundwater resources substantively. The conventional method is used to investigate saltwater intrusion is chemical analysis of water sample. This method gives very highly precise results. However, it consumes many time and intensive and expensive labour. Another method has been widely and effectively used to investigate the saltwater intrusion is the electrical resistivity method [3, 4, 10-15]. Based on measuring apparent resistivity of the aquifer on the surface, the method allows us to determine saltwater/freshwater interface from the resistivity changes of the aquifer. Main advantages of the method are fast, cheap and reliable results.

In this paper, both the electrical resistivity and chemical analysis have been conducted in this study for investigating the saltwater intrusion in Pleistocene aquifer in Thái Bình Province. The aquifer structures and their bulk resistivity are determined by the inversion of the VES data. The pore-water conductivity and its TDS of selected samples are measured and analyzed in the field and in laboratory. The distribution of the TDS of the region is mapped for the 2006-2007 years. Finally, the distribution of saltwater intrusion is also forecasted for the 2015, 2020 and 2030 years by the hydro-geological modelling method.

II. GEOLOGICAL SETTING

The study area is located between the longitudes 106006’33” and 106037’35” E and the latitudes 20005’15” and 20043’57”N. It is about 1542 km2 large and consists of 8 districts and 1 city. The topography of the study area is quite flat. The average elevation is 1-2.5 m above the sea level. The Quaternary alluvial sediments, including mud clay, stiff clay, silty-clay, silt-loam, sand, and gravel cover it completely. The Quaternary sediments may be divided into 5 formations [6]: Lệ Chi Fm (QI lc), Hà Nội Fm (QII-III hn), Vĩnh Phúc Fm (QIII2 vp), Hải Hưng Fm (QIV1-2 hh), and Thái Bình Fm (QIV3 tb). There is unconformity separating Quaternary sediments and Paleozoic-Neogene sandstone, siltstone and limestone. There are three aquifers in Quaternary sediments namely: Upper Holocene aquifer (qh2), Lower Holocene aquifer (qh1), Pleistocene aquifer (qp) and unconfining Quaternary layers namely: Hải Hưng Upper Sub-Fm (QIV1-2 hh2) and Vĩnh Phúc Upper Sub-Fm (QIII2 vp2). According to the hydrogeological data, the main aquifer in Quaternary sediments are Pleistocene one. It is considered to be at the depth of 26-143 m and its average thickness is about 29-127 m. The water qualities of Pleistocene aquifer can be clearly divided into 2 parts: the salt water (TDS >1 g/l) in the southern part and the fresh water (TDS <1 g/l) in the northern part of the province.

III. MEASUREMENT AND INTERPRETATION METHODS

1. Electrical resistivity method

Vertical electrical sounding (VES) has been used for estimating the geoelectrical params of the structure and TDS of aquifer in the periods 1996 and 2006-2007. To determine the geoelectrical structure, 210 VES stations measured in 1996 and 79 VES stations measured at pumping wells in 2006-2007 were used [8]. Apart from VES data, 222 pore water samples at the pumping wells were collected for measuring electrical conductivity and analyzing chemistry [8].

a. Geoelectrical structure of Quaternary sediments: The VES data are interpreted by the forward-inversion methods, which constrain by the available well data. The bulk resistivity and thickness of layers at each VES station are gotten from this interpretation process. The structure of geoelectrical cross-section in the study area is determined by the combination of the interpretation results. It consists of 5 layers in Quaternary sediments corresponding to the 3 aquifers and 2 unconfined layers as mentioned above:

- First layer: lying on the top of geological cross-section. Its average thickness is about 9 m and resistivity value is in range of 8-70 Ωm.

- Second layer: average thickness - 18 m, resistivity value - 0.7-50 Ωm.

- Third layer: average thickness - 17 m and resistivity value - 3.2-16 Ωm.

- Fourth layer: average thickness - 18 m and resistivity value - 5.3-30 Ωm.

- Fifth layer: average thickness - 41 m and resistivity value - 2.2-14.3 Ωm.

- Sixth layer: average depth - 110 m and resistivity value - 5-60 Ωm.

As mentioned above, the main aquifer in Quaternary sediments is Pleistocene one (qp), so in this study we only investigate the saltwater intrusion in this aquifer.

b. Formation factor of Pleistocene aquifer: As you may know, the relationship between the bulk resistivity and pore water resistivity of the aquifer is presented by the modified Archie’s law: rbuk = Frw;   where F is the formation factor of aquifer; rbuk is the bulk resistivity of aquifer and rw - pore water resistivity [2].

According to Archie’s law, each aquifer is characterized by one formation factor. The change of bulk resistivity depends on the change of pore water resistivity [2]. Basing on the bulk resistivity of the qp (rbuk) and pore water resistivity (rw) at 25 pumping wells, we have a relationship between bulk resistivity and pore water resistivity as illustrated in Fig. 1. The regress formula can be obtained as:

rbuk = 2.20 rw                                       (1)

The formation factor of the qp is determined as: F = 2.2. Substituting the bulk resistivity for the equation (1), we can determine the pore water resistivity of the qp.

c. Determining TDS and chloride from pore water resistivity: The analytical result of TDS and content of chloride, bulk resistivity and pore water resistivity at 25 pumping wells are illustrated in the Fig. 2. Based on these distributions, the following regress equations can be obtained:   

       Y = 4099.52*X                            (2)

Where Y is TDS in water (mg/l); X - pore water resistivity (S/m);

And     Y = 2317.30* X-64.67           (3)

where Y is content of chloride (mg/l), X - pore water resistivity (S/m).

 

Figure 1. Correlation between bulk and pore water resistivity of the qp.

The empirical formulas (1), (2) and (3) allow us to calculate the TDS and content of chloride of qp from pore water resistivity or bulk resistivity.

2. Hydrogeological modelling method

Based on the aquifer structures and their distribution of TDS of aquifer, the forecasting problem of saltwater intrusion is developed from the solution of two following problems [1, 5, 7, 9]:

- Flow modelling problem, the argument is a distribution of water table in space;

- Transmitted material problem, the argument is a distribution of TDS in space.

So far, these two problems are solved fairly good by finite-difference and finite-element methods. Many hydrogeological laboratories in the world have been developing these problems into a package software to compute three-D groundwater flow and contaminant transport simulations. In this paper, the authors used the Visual Modflows software of Watertoo Company, Canada [9] for our calculations.


                                                                               

Figure 2. Relation between (a) TDS and pore water resistivity
and (b) content of chloride and pore water resistivity.

The general calculating steps are performed as follows:

- Assigning boundary condition: In the east side along the Bắc Bộ (Tonkin) Gulf, the boundary condition is imitated the first class and the water table is constant and equals H = 0 m, total solute chemitry in seatwer is 3,0 g/l. The river systems are imitated third class boundary condition and assume that they are not under the influence of saltwater intrusion.

- Assigning the aquifer and aquitard: All structural data of aquifer and aquitard and the TDS of Pleistocene are used for the above interpretated results. All other properties and relevant params, such as permeability, storage coefficients, discharge and recharge rate ..., of the aquifer and aquitard are referred from hydrogeological reports [6].

- Adjusting the parameters of the model: All params of the model, for example, permeability, storage coefficients, discharge and recharge rate ..., will be adjusted by the stable inversion and unstable inversion problems. Apart from adjusting, the hydrogeology param, the TDS in 2006-2007, the statistic actual pumping rate are also used for adjusting the material spreading params. The TDS in 2006-2007 is used as solution constrained condition during performing the inverse procedure. Fig. 3 is the TDS measured in 2006-2007 (Fig. 3a) and the TDS for the year 2006 is calculated  from the adjusted input params (Fig. 3b). The measured TDS and calculated TDS are similar to the same show that the adjusted params for the modelling can be accepted.

III. SALTWATER INTRUSION BOUNDARY AND FORECASTING RESERVES

1. Saltwater intrusion boundary in the Pleistocene aquifer (qp)

The above measurement and interpretation data allow us to map the spatial distribution of chloride and TDS of the qp in 2006-2007 (Fig. 4). The Fig. 5a shows the content of chloride changing from 0.2 g/l in the north to 12.5 g/l in the south. Most northern part of study area the content of chloride is less than 0.5 g/l (account for 40 % area). In the southern part, the content of chloride is very high, changing from 0.5 to 12.5 g/l.

In the Fig. 4b, the spatial distribution of TDS is the same as that of chloride. the TDS also increases from the north to the south of the study area. The highest TDS is 21 g/l located in the southern part. If we take the 1 g/l TDS contour is the boundary of the salt/fresh water (the red line in Fig. 4), all fresh-water area is distributed in the north of the study area and accounts for about 605 km2.

Figure 3. Model of params adjusted by inversion problem based on TDS measured
 in 2006-2007: (a) TDS measured in 2006-2007; (b) TDS calculated for 2006
 from the adjusted params of the model.

Figure 4. Spatial distribution of content of chloride (a) and TDS in 2006-2007
(b) in Thái Bình Province.

2. Estimating the fresh-water exploited reserves of Pleistocene aquifer

To estimate the exploited reserves of the freshwater area, we set up every 2 km a pumping well for all freshwater area (see Fig. 5a). We will have 215 pumping wells and total pumping rate is 118250 m3/day (550 m3/day per well). The calculated results of the reduction of water table and saltwater intrusion for the year 2010, 2015, 2020 and 2030 are as follows:   


Figure 5. Water level and TDS of the Pleistocene aquifer calculated for the 2006 (b), 2010 (c), 2015 (d); 2020 (e) and 2030 (f) years.

The water level in the qh2 and qh1 aquifers change in the range of -1 to 4 m. Since 2015, the eastern part of  Hưng Hà District the qh2 and qh1 aquifers is being to take shape the drawdown funnel having water level of 0,5 m in 2020 and -1 m in 2030. The salt/fresh water boundary of the qh2 aquifer changes lightly by time. The TDS varies from 2-3 g/l in the saltwater area and reduces in the freshwater area. The evidence for that is the occurrence of the freshwater spots in the qh2 aquifers.

The qh1 aquifer has saltwater intrusion completely at 2006 with the TDS of 2-3 g/l. By the time, the water exploiting activities making the freshwater penetrated into qh1 from qh2 desalts the northern part of the region. The freshwater area in this aquifer is spreading by the time.

The forecasting results of the water level reduction and saltwater intrusion of the qp are shown in Fig. 5b, 5c, 5d, 5e, and 6f. The lowest elevation of the forecasted water levels is -8 m in 2010, -12 m in 2015, -14 m in 2020 and -16 m in 2030. The water flow direction is moving from the south to the north. The centre of the drawdown funnel is located in the centre of the Thái Thụy District. The salt/freshwater boundary moves up strongest to the north direction. The Thái Thụy District is salted completely in the year 2030. The salt/freshwater boundary is spreading to the center of Đông Hưng, Hưng Hà and Quỳnh Phụ districts. The calculated results also show that the horizontal salt rate is predominated. The evidence for that no isolated saltwater zone is occurred in the freshwater area.    

The forecasted exploiting reserves are determined by the way of accounting at each time there are how many wells still lie on the freshwater area. The result of the exploiting reserves of the Pleistocene aquifer is as follows:

- 2006 year: total freshwater exploiting wells are 215 and total pumping rate is 11825 m3/day.

- 2010 year: total freshwater exploiting wells are 194 and total pumping rate is 106700 m3/day.

- 2020 year: total freshwater exploiting wells are 175 and total pumping rate is 96250 m3/day.

- 2030 year: total freshwater exploiting wells are 98 and total pumping rate is 53900 m3/day.

 3. Estimating the substainable pumping rate

The substainable pumping rate is how volume of water exploited so that the salt/freshwater boundary has no change by the time. Assuming the pumping wells are located in the centre of every commune of the freshwater area of the qp. The pumping rate at earch well is adjusted and operations repeated until a total acceptable pumping rate doesn’t make change of the salt/freshwater boundary. After many interactive computing, we have determined the substainable pumping rate for the qp is 27557 m3/day. With the pumping rate of 27557 m3/day, the salt/freshwater boundary of the qp has no change for the 2030 year.  

CONCLUSIONS

1. The interpretation results allow us to reconstruct the spatial distribution of the TDS, content of chloride and salt/freshwater boundary of the Pleistocene aquifer in Thái Bình Province for the year 2006-2007. The freshwater area of the Pleistocene aquifer account for 605 km2, distributed mainly in the northern part of the province including the Hưng Hà, Đông Hưng, Quỳnh Phụ and Thái Thụy districts. The TDS maps and the structure of aquifer and confines are useful data resources for applying numerical hydrogeology modelling to forecast the saltwater intrusion and suitable pumping rate in the study area.

2. According to the modelling calculation, if we maintain the pumping rate of 118,250 m3/day, the freshwater reserves of the Pleistocene aquifer may remain only about 80 %, equivalent to the pumping rate of 96250 m3/day for the year 2020 and about 45 %, equivalent to pumping rate of 53900 m3/day for the year 2030. To remain the current salt/freshwater boundary of the Pleistocene aquifer, the pumping rate of the Pleistocene aquifer should be 27557 m3/day and the well net is distributed evenly in local communes.    

REFERENCES

1. Anderson M.P., W.W. Woesseer, 1992. Applied groundwater modelling. Acad. Press, Inc., New York.

2. Archie G.E., 1944. The electrical resistivity log as an aid in determining some reservoir characteristics. Am. Inst. Min. Metallurg Petr. Eng. Tech. Paper, 1422 pp..

3. Ardau F., Barbiere G., 1994. Evolution of phenomena of salt water intrusion in the coastal plain of Muravera (southeastern Sardinia): 13th saltwater intrusion Mtg. Exp. abstr., 305-312.

4. Chieh-Hou Y., Lun-Tao Tong, Ching-Feng Huang, 1999. Combined application of DC and TEM to sea-water intrusion mapping. Geophysics, 64/2 : 417-427.

5. Herbert F.W., W.W. Woesseer, 1982. Introduction to groundwater modelling. Acad. Press Inc., New York.

6. Lại Đức Hùng (Ed.), 1996. Final report on drawing hydrogeological map, scale 1:50.000. Arch. of Dept of Geology, pp. 105.

7. Mary P.A., W.W. Woesseer, 1992. Applied groundwater modelling. Acad. Press Inc., New York.

8. Nguyễn Như Trung (Ed.), 2007. Evaluating the actual state and forecasting the saltwater intrusion of groundwater aquifer of Thái Bình coastal plain. Final report. VN Acad. of Sci. and Techn., 150 pp.

9. Nilson G., Thomas F., 2004. Visual mudflow guideline, Version 4.0. Waterflow Hydrogeology Software, Toronto.

10. Paul M.B., 2003. Groundwater in freshwater-saltwater environments of the Atlantics coast. Circular 1262, US Geol. Surv., 113 pp..  

11. Roberto B., E. Gavaudo, Fr. Ardau, G. Ghiglieri, 2003. Geophysical approach to the environmental study of the coastal plain. Geophysics, 68/5 : 1446-1459.

12. van Overmeeren, 1989. Aquifer boundaries explored by geoelectrical measure-ments in the coastal plain of Yemen: Case of equivalence. Geophysics, 54 : 38-48.

13. Waheidi M.M., F. Merlanti, M. Pavan, 1992. Geoelectrical resistivity survey of the central part of Azraq basin (Jordan) for identifying saltwater/freshwater interface. J. of Applied Geoph..

14. Wang H.F., W.W. Woesseer, 1982. Introduction to groundwater modelling. Acad. Press Inc., New York.

15. Zhdanov M., Keller G., 1994. The geoelectrical methods in geophysical exploration. Methods in geoch. and geoph., 31, Elsevier, Amsterdam-London-New York - Tokyo, 873 pp..