Distribution of ²²²Rn in Seawater Intrusion Area and Its Implications on Tracing Submarine Groundwater Discharge on the Upper Gulf of Thailand

Submarine groundwater discharge (SGD), which includes both fresh and saline groundwater outflows to seas, has been considered an important contributor to the water and chemical budgets of the world’s oceans [1, 2]. It is an important process for the marine biogeochemical cycles of elements, such as nutrients, trace metals, carbon, and rare earth elements [3–7]. SGD is recognized as an “invisible” process, driven by both marine and terrestrial forcing components below the water surface. Different types of groundwater can be mixed with the water through the coastal subsurface sediments to form subterranean estuaries, transporting nutrients in coastal ecosystems, which can affect water quality and lead to environmental deterioration in coastal areas [8]. Hence, understanding SGD is important for the management of coastal waters.

The natural radioisotopes radium (Ra) and radon (Rn) have a wide range of applications in investigating SGD fluxes [4, 9–12]. Indeed, natural radioisotopes show the advantage of a composite tracer signal when entering the ocean by a variety of routes in the aquifer due to their conservative behavior [13]. ²²²Rn, with a half-life of 3.8 days, was considered to be an excellent tracer for quantifying submarine groundwater discharge [10, 13]. The ²²²Rn mass balance model proposed by Burnett and Charette et al. was used to evaluate the SGD. The model assumed that the water body was well mixed and there was no water stratification in terms of ²²²Rn distribution. The estimation of mixing loss was a speculative value, which would lead to some errors in the results. Methods for estimating SGD using chemical tracers are also being improved; for example, Lamontagne and Webster described how to better estimate SGD from trends in the chemical fingerprint in seawater [14, 15]. Although quantifying SGD remains challenging because of its spatial and temporal variability, a classical mass balance model for SGD estimation has been applied or accepted by many researchers to estimate SGD [8, 16–19]. The ²²²Rn mass balance model of estimating SGD flux is based on converting ²²²Rn inventories to the flux of Rn derived from the SGD by considering both the Rn distributions in sea water and groundwater inflowing to the coastal zone and taking into account radioactive decay, tidal effects, river input, sediment diffusion, atmospheric escape, and mixing with low-concentration offshore waters [13]. In order to transform ²²²Rn fluxes to SGD fluxes, the radon end-member in the groundwater needs to be determined. Due to the high spatial variability of the radon isotope in groundwater, large uncertainty in the model calculation can be observed. However, most studies of SGD have been primarily focusing on estimating the fluxes of Rn without assessing the distribution of ²²²Rn in the groundwater end-member [20].

The coastline of Thailand’s coastal regions has changed due to natural and human influences, resulting in environmental degradation and hydraulic balance change [21]. Indeed, groundwater and coastal water pollution is the main problem encountered in Thailand’s coastal regions [22, 23]. However, groundwater discharge may be easily overlooked as a hidden channel, even though it carries significant amount of nutrients and pollutants.

In addition to reducing groundwater pollution, it is also important to effectively identify the exchange process between groundwater and seawater. As an inert gas, the geochemical behavior of Rn is relatively conservative in aquifers and becomes significantly enriched in SGD fluids compared to seawater [10]. Therefore, ²²²Rn is widely used to estimate SGD fluxes, allowing the identification of groundwater discharge hotspots and pollution sources, thus facilitating targeted treatment in the study area [23, 24].

Due to the low-lying area proximity to the sea and the deep well pumping [25], the sea-level rise, land subsidence, and seawater intrusion have occurred in the Gulf of Thailand [26]. Salinization in freshwater aquifers can cause water–rock interaction, increasing total dissolved solid (TDS) concentrations and the risk of contamination of water resources [27]. Indeed, the salinity of groundwater has long been considered to be the primary factor affecting the Ra activity in the SGD end-member. Cerdà-Domènech et al. [20] showed that the desorption of Ra increases significantly with the increase of salinity. The spatial variability in the distribution of ²²⁶Ra may lead to changes in the ²²²Rn concentrations. Indeed, the groundwater salinization process leads to spatial heterogeneity of radioisotope activity in aquifers [27], thus causing great uncertainty in the SGD assessment.

In this study, the hydrochemical characteristics of groundwater along the Upper Gulf of Thailand were assessed to reveal the impact of sea intrusion groundwater quality. In addition, comprehensive monitoring was carried out to assess the spatiotemporal distribution of ²²²Rn in seawater and coastal groundwater and to calculate the SGD accurately in the upper Gulf of Thailand. Moreover, the present study is aimed at estimating the SGD fluxes in the East coast and the entire Upper Gulf of Thailand based on the characteristics of ²²²Rn distribution results and at analyzing the spatiotemporal changes in groundwater in coastal areas under the influence of tides. The research results are of great significance for the protection of ecological environment and groundwater resources in coastal areas of Thailand.

This study was conducted in the Upper Gulf of Thailand and its coastal sections (Figure 1). The Gulf of Thailand covers an area of about 320,000 km² [21]. Malaysia, Thailand, Cambodia, and Vietnam share this semienclosed tropical sea in the South China Sea (Pacific Ocean). The terrain in Thailand is high in the north and low in the south. On the other hand, river networks are densely dispersed from North to South and run into the Gulf of Thailand. The Upper Gulf is U-shaped and forms the catchment basin of two rivers in the western coast and four major rivers in its Northern part (Bang Pakong, Chao Phraya, Tha Chin, and Mae Klong Rivers) [28].

The study area belongs to a part of the Chao Phraya Plain which is a fault-bounded basin developed in the Polio-Pleistocene epoch. The main aquifers in this area consist of fluvial deposits and marine sediments, and aquitard layers of clay separate the aquifers from one another [26]. There are eight different aquifers in the Chao Phraya-Tha Chin Basin less than 700 m deep from surface level (Figure 2). They are confined aquifers and are on average 50 m thick. These aquifers and the depth from ground surface level (m) can be summarized as follows [29]: (1) Bangkok 50; (2) Phra Pradeang 100; (3) Nakhon Luang 150; (4) Nonthaburi 200; (5) Sam Khok 300; (6) Phaya Thai 350; (7) Thon Buri 450; and (8) Pak Nam 550.

The Gulf of Thailand is a vital source of the Thai economy. Indeed, the fishing sector, tourism, and port activities in the Gulf region have brought great benefits to the residents. In recent decades, with increasing population, as well as industrial and economic development, several coastal development projects have been planned and implemented. However, these projects have impacted local coastal and estuarine formations, resulting in changes in coastal processes and shorelines. Furthermore, the offshore area has become severely polluted [21, 23]. The causes of coastal groundwater and seawater pollution in Thailand are surface runoff and drainage of effluents from port, urban, and industrial areas. Moreover, rivers and groundwater are highly polluted by urban sewage, industrial wastewater, and sediments.

The upper Gulf’s surface water quality is often poor, particularly around the downstream of the four major rivers and important tourist destinations along the coast [21]. Water quality in the upper Gulf region is deteriorating due to the increased use of nitrogen fertilizers, mariculture activity, and urban waste.

In total, 32 water samples of the Gulf of Thailand’s coastline were collected from May to November 2018 for hydrochemical analysis (Table 1). The samples consisted of 29 groundwater and 3 seawater samples collected from residential wells and estuarine seawater, respectively (Figure 1).

The RAD-7 continuous monitoring system with a submersible pump was installed in the offshore groundwater of the eastern coastal zone (100.90° E, 12.77° N) to continuously monitor (24 hours) ²²²Rn. The monitoring was carried out every 2 hours, from 15:15 on November 23 to 17:15 on November 24, 2018 (Figure 1). In addition, tidal data were collected to compare the effect of tidal height on SGD.

Total dissolved solids (TDS) and pH were measured using a portable multiparameter water quality analyzer (YSI Pro Plus, American YSI). To purify the samples and prevent air from entering, the water sample was first filtered via a 0.45 μm membrane filter and then placed in a bubble-free polyethylene bottle and sealed with an adhesive stripe. Finally, all samples were stored strictly in cold storage before being sent to the laboratory.

The K⁺, Na⁺, Ca²⁺, Mg²⁺, Br^-, and ²³⁸U were analyzed using an inductively coupled plasma-atomic emission spectrometer (ICP-MS, American Thermo Fisher). Cl^- and $SO42‐$ were determined by an ion chromatograph (ICS-3000, American DIONEX). A titration method with phenolphthalein solution and standard HCl solution was used to analyze the $HCO3‐$ [30, 31].

To assess the activity of ²²²Rn, the water samples were slowly injected into a 250 mL glass bottle, in which ²²²Rn was analyzed using a RAD-7 radon monitor and its RAD H₂O water accessory. The ²²²Rn activity was determined continuously in coastal groundwater using radon automatic monitoring systems [32]. The radon automatic monitoring systems worked by providing a constant stream of water (driven by a submersible pump) to be analyzed to the RAD-7. Continuous activity of ²²²Rn in groundwater can be obtained by setting the pumping time [10].

The hydrochemical results (Table 2) showed that Cl^- was the dominating anion in most groundwater samples, followed by $HCO3‐$⁠, while Na⁺ was the dominating cation, followed by Ca²⁺ and Mg²⁺. In addition, the ranges of TDS, Cl^-, $HCO3‐$⁠, Na⁺, Ca²⁺, and Mg²⁺ concentrations in groundwater samples were 96.43-5092.52, 17.88-3788.00, 69.78-652.10, 17.70-533.10, 14.69-299.10, and 2.71-157.40 mg/L, respectively.

Based on the location and chemical compositions of groundwater samples (Figure 1), the study area was divided into three main areas, namely, the southwestern area (TW1-TW11), the northwestern area (TW12-TW24), and the eastern area (TE1-TE5). The hydrochemical type of groundwater on the southwestern area is mainly Na-HCO₃ and Na-Cl. The hydrochemical type of groundwater on the northwestern area is Na-Cl. The hydrochemical type of groundwater on the eastern area is mainly Na-HCO₃ and Ca-HCO₃.

The pH of groundwater in the southwestern area ranged from 6.76 to 7.98, with an average value of 7.36. The range of TDS was 386.01-1417.01 mg/L, with an average of 792.93 mg/L. The results showed that Na⁺ was the dominant cation in the groundwater of the southwestern area, ranging from 87.94 to 442.40 mg/L, with an average of 190.19 mg/L, while the dominant anion was $HCO3‐$⁠, ranging from 341.27 to 652.10 mg/L, with an average value of 465.43 mg/L. Regarding the anions, the contents of $HCO3‐$ and Cl^- raised alternately, while the contents of $SO42‐$ were very low in Figure 3. The hydrochemical types of groundwater were mainly Na-HCO₃ and Na-Cl, suggesting a slight intrusion of seawater in this area, which leads to the increase in Cl^- and Na⁺.

The pH of groundwater in the northwestern area varied from 6.67 to 7.74, with an average of 7.19. The variation range of TDS was relatively large, ranging from 683.54 to 5092.52 mg/L, with an average of 1853.15 mg/L. The Na⁺ was the dominant cation in the groundwater in the northwest area, ranging from 100.30 to 533.10 mg/L, with an average of 269.59 mg/L, while the Cl^- was the dominant anion, ranging from 213.00 to 3788.00 mg/L, with an average of 1067.20 mg/L. Indeed, the Cl^- and Na⁺ contents in the whole area were high, which is consistent with the Na-Cl hydrochemical type, indicating that this area was severely affected by seawater intrusion.

The pH values of groundwater in the eastern area ranged from 5.11 to 6.94, with an average of 6.30. In addition, the range of TDS was relatively small, ranging from 96.43 to 626.78 mg/L, with an average of 352.64 mg/L. The dominant cations in the groundwater were Na⁺ and Ca²⁺, with average values of 66.29 and 46.78 mg/L, respectively. On the other hand, the dominant anion was $HCO3‐$⁠, ranging from 69.78 to 265.15 mg/L, with an average of 154.65 mg/L. It can be seen from the Piper diagram that Na-HCO₃ and Ca-HCO₃ were the main hydrochemical facies in the eastern area, indicating that this area was the least affected by seawater intrusion.

In terms of the average ion concentrations, the three areas were not evenly distributed, reflecting the different spatial magnitude of the impact of seawater intrusion. The seawater intrusion in the western area was more serious, especially in the northwestern area, while that observed in the eastern area was relatively slight. The groundwater chemical type in the entire study area was first Ca(Mg)-HCO₃, then transformed into Na(Ca/Mg)-HCO₃ and Na(Ca/Mg)-Cl, and finally transformed into Na-Cl as a result of seawater intrusion.

The average of the field survey data showed a great spatial variation in the ²²²Rn activity. Indeed, the spatial amplitude of variation was about 50 times (Figure 4). The ²²²Rn activity ranged from 987.96 to 49858.91 Bq/m³, with an average value of 15560.06 Bq/m³. The average activity values of ²²²Rn were 20417.67, 10525.17, and 17964.33 Bq/m³ in the southwestern, northwestern, and eastern areas, respectively.

Generally, the activity of ²²²Rn in groundwater is fundamentally determined by the type and abundance of uranium-bearing minerals in the parent rock. Various uranium-bearing minerals in the strata decay continuously to produce ²²²Rn, resulting in the high activity of ²²²Rn in groundwater, especially in confined aquifers. Influenced by geological background, the activity of ²²²Rn in magmatic rocks is higher than that in sedimentary and metamorphic rocks [33]. The sampling sites in the southwestern and eastern areas are mainly located in the magmatic rock area. Therefore, the average activity of ²²²Rn in these areas is higher than that of the northwestern area, which is characterized by the presence of Quaternary loose sediments. Compared with the distribution of ²²²Rn isotope in the four watersheds around Jiaozhou Bay [34], the average activity of ²²²Rn in the Dagu River Basin with loose Quaternary sediments was 4399 Bq/m³, which is significantly lower than those in Licun River, Moshui-Baisha River, and Yang River. In the three other watersheds where the sampling sites are mostly located in the Mesozoic acidic magmatic rock area, the average activity of ²²²Rn ranged from 13342 to 17234 Bq/m³.

The contents of uranium and radium in magmatic rocks were relatively high. The parent of ²²²Rn is ²²⁶Ra, both of which belong to the ²³⁸U series. The activity of ²³⁸U can directly determine the activity of ²²²Rn [34]. Indeed, the results revealed a positive correlation between ²²²Rn and ²³⁸U (Figure 5).

However, low ²³⁸U and high ²²²Rn values were observed in TW1, TW2, TW5, TE1, and TE5. Br^- is stable in nature and is mainly found in the ocean. Thus, Br^- can be used to identify the source of salt in coastal groundwater [35]. The Br/Cl ratios in these samples were quite different from seawater (Figure 6), suggesting that these groundwaters have little interaction with seawater. Therefore, the sampling site is relatively closed, with a long water residence time and a significant water-rock interaction, leading to high radon levels.

The results showed that the activity of ²²²Rn in groundwater is mainly controlled by the geological background. Indeed, the geological background controls its activity mainly by the lithology of the aquifer. Moreover, other factors may affect the geochemical behavior of ²²²Rn in groundwater, such as groundwater sources, contact with air, recharge conditions, rock fissures and soil voids, water retention time, and water-rock interaction time [33, 34].

SGD is an important channel for transporting terrestrial materials to the ocean [1], and it is one of the main causes of nearshore hypoxia and acidification [18, 36]. SGD might have a role in the coastal pollution that has been reported to be a tool to reveal offshore pollution from the large industrial complex in the Gulf of Thailand [22].

The hydrochemical types of groundwater along the Gulf of Thailand, as well as the activity, distribution, and influencing factors of ²²²Rn isotope, were first determined to better use ²²²Rn as a tracer to study various oceanographic processes. A ²²²Rn mass balance model was built to estimate the submarine groundwater discharge in the Upper Gulf of Thailand.

The ²²²Rn mass balance model proposed by Burnett and Dulaiova and Charette et al. was used to evaluate the SGD in the East coast of the Gulf of Thailand (Figure 7) and speculate the changes in SGD with tides [10, 37].

This paper refers to the data of the activity of ²²²Rn in the radon inventory used by Burnett et al. [23] to estimate SGD in the East coast of the Gulf of Thailand in July 2018. The average value of the activity of ²²²Rn in the radon inventory was 173.00 Bq/m².

By considering the value of the bottom temperature in Equation (5), $Do$ is $1.48×10−5$ cm²/s (⁠$1.28×10−4$ m²/d). Therefore, $Ds$ is equal to $7.94×10−5$ m²/d. $Fsed$ is equal to 22.14 Bq/(m²·d).

The contribution of rivers to SGD is very limited. Indeed, the rivers that flow into the bay are Mae Klong River, Tha Chin River, Chao Phraya River, and Bang Pakong River. The average flow of the four rivers recorded in November was 800 m/s [40], equivalent to $6.91×107$ m/d. The activity of radon in rivers was very low, and the difference of the activity between all rivers was small. Field measurements of radon activity were made in two river water samples from the lower reaches of Mae Klong River. The radon activity of the two groups of river water samples is 158.03 and 224.65 Bq/m³, respectively. So the average activity of ²²²Rn in river water was 191.34 Bq/m³. The area of the upper Gulf of Thailand is approximately $1×1010$ m². The $Friv$ is, therefore, equal to 1.32 Bq/(m²·d).

$Fmix$ denotes the loss from mixing with low radon water from the open sea. When sediment diffusion, river input, and atmospheric escape loss fluxes were calculated, the ²²²Rn activity in the radon inventory was only affected by SGD and mixed loss flux [18]. The maximum negative value of variation in radon inventory was chosen as the mixing loss [10, 41]. This method underestimates the mixed loss, since each period may have a specific mixed loss and SGD input and the mixing loss value should be larger to offset the radon flux carried by SGD. The underestimated estimate of mixing loss leads to the underestimated estimate of SGD according to Equation (1). [43]. The results showed a $Fmix$ value of 282.00 Bq/(m²·d).

Based on the flux values mentioned above and Equation (1), an $FSGD$ value of 305.64 Bq/(m²·d) was obtained.

The activity of ²²²Rn in groundwater at different times was determined using the continuous monitoring data of groundwater. The following results are the $Q$ values obtained, considering the $FSGD$ value into the ²²²Rn activity in different periods in Table 3.

The average, maximum, and minimum groundwater flux values observed on the East coast of the Gulf of Thailand were 0.0922, 0.1097, and 0.0811 m/d, respectively. Although the $FSGD$ estimate is relatively conservative, the ²²²Rn activity at the continuous monitoring sites was relatively low compared to that in the entire Gulf of Thailand coast. Thus, the average groundwater flux estimated for the East coast of the Gulf of Thailand was relatively high.

The tidal heights observed during the period from 6:00 on November 23, 2018, to 17:00 on November 24, 2018, are reported in Table 4. It can be seen that the first low tide height value (2.4 m) was observed at 12:00 on November 23, while the first high tide height value (2.7 m) was at 14:00-15:00. On the other hand, the second low tide height value (1.2 m) was observed at 22:00-23:00 on November 23, while the second high tide height value (3.1 m) was recorded at 7:00 the next day (Figure 8). The average tidal height value during this period was 2.43 m. The first and second tidal ranges were 0.3 and 1.9 m, respectively. In addition, the second tidal range was larger and the time span was longer.

Electrical resistivity tomography was used to monitor the influence range of tidal action (provided in Supporting-Information). The results (Fig. S1) showed that the formation resistivity near the continuous monitoring well changed with tides and the salt-freshwater interface moved constantly, which indicated that this site was affected by tidal action and the exchange of groundwater and seawater took place. Radon activity changed with tidal height. We compared tides with SGD, as the SGD was the result of changes in radon activity.

By comparing the SGD with the tidal data (Figure 8), we found that the groundwater flux and the tide were negatively correlated, and the response of the groundwater flux to the tidal height change showed a certain hysteresis. This may be due to the intrusion of seawater into the aquifer as a result of rising tides, which is not conducive to the discharge of submarine groundwater. However, when tides fall, groundwater can be discharged from the aquifer, which is conducive to submarine groundwater discharge.

At the first high tide, which was 2.9 m, the SGD was 0.0811 m/d; at the first low tide, which was 2.4 m, the SGD was 0.0961 m/d; at the second high tide, which was 2.7 m, the SGD was 0.0835 m/d; at the second low tide, which was 1.2 m, the SGD was 0.1097 m/d. The SGD observed during low tide was about 1.25 times higher than that observed during high tide.

The $FSGD$ was estimated to assess the groundwater discharge fluxes across the Gulf of Thailand from the East coast of the Gulf of Thailand to the entire Gulf of Thailand. The average activity values of ²²²Rn observed in the southwestern, northwestern, and eastern areas were 20417.67, 10525.17, and 17964.33 Bq/m³, respectively. Therefore, based on Equation (2), the estimated groundwater discharge flux values in the southwestern, northwestern, and eastern areas were 0.0150, 0.0290, and 0.0170 m/d, respectively, while the average groundwater discharge flux of the entire Gulf of Thailand was 0.0203 m/d.

Few studies on the SGD fluxes in the Gulf of Thailand were carried out. The results of the current study were compared with results obtained in the Gulf of Thailand and other regions worldwide. The results are reported in Table 5. The results of the current study were in line with those reported by other authors. However, the difference in the results of the current study may be due to the ignorance of some parameters with little influence in the model used and the estimation of some parameters. Furthermore, the difference in time scale and space scale, as well as measurement errors, may also affect the final results.

The activity of radon isotope in groundwater along the Gulf of Thailand showed a great spatiotemporal variation, but it was not affected by seawater intrusion. The relationship between radon isotope and Br/Cl ratio suggested an impact of the residence time and salinization mechanism of groundwater on the change in radon activity. The correlation between radon isotope and ²³⁸U suggested that the distribution of radon isotope in the aquifer is controlled by the geological environment, especially the content of ²²⁶Ra (which is the parent of ²²²Rn).

The $FSGD$ value of the East coast of the Gulf of Thailand was 305.64 Bq/(m²·d). Based on the radon activity in groundwater, the average submarine groundwater flux on the East coast of the top of the Gulf of Thailand was estimated to be 0.0922 m/d, while that on the submarine groundwater flux of the entire Gulf of Thailand was 0.0203 m/d. The tide was the main factor affecting the submarine groundwater discharge. The results showed that the SGD on the East coast of the Gulf of Thailand was negatively correlated with the tidal height. The SGD observed during low tide was about 1.25 times higher than that observed during high tide.

The data are listed in the Supplementary Materials.

The authors declare that they have no conflicts of interest.

We thank staff from Marine and Coastal Resources Research and Development Institute and Phuket Marine Biological Center. This work was supported by the National Natural Science Foundation of China-Shandong Joint Fund for Marine Science Research Centers (Nos. U2106203, U1806212), National Natural Science Foundation of China (41706067), and China-Thailand Cooperation Project “Research on Vulnerability of Coastal Zones.”