**Application of Porous Concrete Infiltration Techniques to Street Stormwater Inlets That Simultaneously Mitigate against Non-Point Heavy Metal Pollution and Stormwater Runoff Reduction in Urban Areas: Catchment-Scale Evaluation of the Potential of Discrete and Small-Scale Techniques**

**Shigeki Harada**

Department of Agroenvironmental Sciences, Faculty of Food and Agricultural Sciences, Fukushima University, Fukushima 960-1296, Japan; harada.shigeki@agri.fukushima-u.ac.jp

**Abstract:** The expansion of pervious areas is an essential and common concept in mitigating nonpoint pollution runoff in urban areas. In this review, literature related to the expansion of pervious areas is introduced. In addition, the potential application of porous concrete as a medium for constructing the bottom and side walls of street stormwater inlets is investigated. The effectiveness of this medium in reducing (i) the stormwater runoff volume via porous concrete by exfiltrating from the bottom and the wall, and (ii) the heavy metal pollution runoff loads via infiltration through the porous concrete is assessed using data obtained by the author and published in the literature. The urban hydrological model Infoworks ICM (Innovyze) was used to estimate the exfiltration rates through the porous concrete plates set at the bottom and side walls of the street stormwater inlets. The exfiltration rates used in the pre-reported literature varied depending on the methods used. In the present study, sensitivity tests were performed by changing the exfiltration rates. The results of this study indicated that porous concrete used at only the bottom and side walls of the street stormwater inlets is suitable for reducing the runoff volume and removing any heavy metals from stormwater at a catchment scale.

**Keywords:** non-point pollution; stormwater drainage systems; infiltration technique; storage of runoff water; quantity and quality of stormwater; porous concrete; heavy metals; urbanized areas

#### **1. Introduction**

*1.1. Overview and Review Objectives*

The potential and realized effects of non-point sources of pollution originating from runoff from urban areas, such as stormwater drains adjacent to roadways, have received increased attention in recent years. Comprehensive reviews of state-of-art remediation techniques and their development have been reported [1,2]. The impacts of non-point sources of pollution have been demonstrated [3–5] and the runoff characteristics have been described [6,7]. Major non-point pollutants, including organic materials (chemical oxygen demand (COD), biochemical oxygen demand (BOD), total organic carbon (TOC)), nutrients (various forms of nitrogen and phosphorus), heavy metals, suspended solids, PAHs [8], microplastics [9–11], and water-soluble aerosols [12], have all been found in the runoff from urbanized areas and their surrounding areas. Of these pollutants, heavy metal pollutions from urbanized areas [13–24] are the primary focus of this review owing to its high toxicity [25–29]. In this review, studies on heavy metal pollution in road effluents [13–15] will be cited and the data therein will be compared to the author's own data.

Since non-point pollutants, including heavy metals, in runoff water can be captured using infiltration techniques (or infiltration unit processes) [30–35], the author focused on the application of porous concrete in stormwater drainage systems. The aim of the study

**Citation:** Harada, S. Application of Porous Concrete Infiltration Techniques to Street Stormwater Inlets That Simultaneously Mitigate against Non-Point Heavy Metal Pollution and Stormwater Runoff Reduction in Urban Areas: Catchment-Scale Evaluation of the Potential of Discrete and Small-Scale Techniques. *Water* **2023**, *15*, 1998. https://doi.org/10.3390/w15111998

Academic Editor: Domenico Cicchella

Received: 3 April 2023 Revised: 10 May 2023 Accepted: 18 May 2023 Published: 24 May 2023

**Copyright:** © 2023 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

was the reduction of both non-point source heavy metals in runoff and surface runoff in urbanized areas. The author focused on the application of porous concrete plates set at the bottom [36–40] and sides of street stormwater inlets. The author intended to demonstrate that the placement of relatively small porous concrete plates at the bottom and sides of street stormwater inlets is sufficient to capture the heavy metals within the runoff and facilitate the efficient drainage of the stormwater itself in impermeable urban catchment areas. Specifically, the author examines the following points:


#### *1.2. Features and Definitions of Non-Point Pollution Sources and the Countermeasures Required for Their Reduction in Runoff*

The information provided on the homepage of the Japan Society for Water Environment (JSWE) and the US Environmental Protection Agency (EPA) is discussed in order to explore the similarities and differences in the approaches adopted by the two agencies to expand pervious areas and control non-point sources of pollution.

The JSWE (https://www.jswe.or.jp/eng/index.html (accessed on 10 May 2023)) has based its description of non-point sources of pollution in urban areas on the features of these sources and the actions that need to be implemented (http://jswe-nonpoint.com/ 1/documents.html (accessed on 10 May 2023)). The definitions employed by the EPA are based on basic information, such as "non-point vs. point sources" and "what we can do first, etc.".

JSWE describes the actions that are required to remediate non-point pollution runoff by emphasizing the need to understand runoff behavior and develop pollution control measures.

The US EPA published a fact sheet entitled, "Protecting Water Quality from Urban Runoff" (https://nepis.epa.gov/Exe/ZyPDF.cgi/20004PP1.PDF?Dockey=20004PP1.PDF (accessed on 10 May 2023)), which shows "How Urbanized Areas Affect Water Quality" in terms of increasing runoff and pollutant loads. Of particular interest were the roles of porous and pervious areas in natural landscapes, such as forests, wetlands, and grasslands, at trapping rainwater and snowmelt and how they promote water filtration into the ground. The roles of pervious areas as countermeasures to non-point pollution, and the ways in which pervious areas can be expanded include the following [41]:


These could be included in the BMPs of water managers [42]. These infiltration measures are similar to those employed in the experimental sewer system (ESS) in Japan [43]. The system, which is described in [41], could be considered to include the whole catchment area, while the porous concrete plates at the bottoms and sides of the street stormwater inlets could be considered as small discrete points.

#### **2. Runoff Behavior from Non-Point Sources in Urban Areas**

*2.1. Road Runoff Water Quality Assessments in Sendai City, Miyagi Prefecture, Japan*

#### 2.1.1. Materials and Methods

In Wakabayashi ward, Sendai City in Miyagi Prefecture, the author deployed a water collection device in a street stormwater inlet. The device, which consisted of a glass bottle with a floating ball, can collect 1 L samples of stormwater runoff. When the ball rises to the top of the bottle, further inflow is prevented. Therefore, the collected samples correspond to the initial stormwater runoff from 1.5 mm rainfall events (Figures 1–3). This estimate is based on the assumption that the area of inflow into the inlet was 5 m by 10 m, the initial loss was 0.75 mm, the runoff ratio of the area was 0.8, and the proportion of water entering the sample collecting vessel was 0.05 (Figure 3).

**Figure 1.** Study site and the street stormwater inlet.

**Figure 2.** Inside the street stormwater inlet.

**Figure 3.** Water collection device.

The daily rainfall volumes during the periods 1985–1990 and 2015–2019 in Sendai obtained from the Japan Meteorological Agency showed daily rainfall events of <1.5 mm, which accounted for 34.1% and 31.2% of the rainy days during the 1985–1990 and 2015–2019 periods, respectively. The histograms show that daily rainfall events ranged between 0.5 and 10 mm during these two periods (Figures 4 and 5), while daily rainfall events of 0.5 mm were the highest in frequency. In addition, despite the changes in climatological conditions over these periods, the shapes of the histograms are generally similar. The figures show that the sample collecting system employed in this study is well suited to collecting runoff from the most frequent rainfall events and, also, for collecting samples for some of the larger rainfall events. Whether collecting runoff for the first <1.5 mm rainfall events can capture the first flush effects should be clarified by further monitoring [6,44], literature reviews [2,45,46], and analyses using EMC [47].

**Figure 4.** Frequency of the daily rainfall volume focusing on 0.5–10 mm events in Sendai (1985–1990).

**Figure 5.** Frequency of the daily rainfall volume focusing on 0.5–10 mm events in Sendai (2015–2019).

In 2019, the author collected 1 L samples at intervals of 1–2 weeks (Table 1).

**Table 1.** Features of the samples collected 18 May to 19 December in 2019 and qualities (SS, particulate and dissolved heavy metals).



**Table 1.** *Cont.*

The collected water samples were initially stored in a cool dark space before being transferred to a refrigerator (4 ◦C) until filtration.

The collected water was pre-filtered using a 2 mm mesh filter followed by filtration using a glass fiber filter (Whatman GF/B, 47ϕ, pore size: 1 μm). Suspended solids (SS, see also Figures 5 and 6) on the glass fiber filters were measured after drying the filter for two hours at 105 ◦C. The filtrates were used to measure heavy metals (Cr, Cu, Zn, Cd, and Pb) by ICP–MS. Separately, pre-filtrated waters were used to measure the total heavy metals (Cr, Cu, Zn, Cd, and Pb) by ICP–MS to determine the proportion of the particulate heavy metals.

#### 2.1.2. Results

A list of the water samples and their physicochemical parameters is shown in Table 1. High SS values were often observed, which were consistent with the visual observations (Figure 6).

SS was observed in the samples, even when the antecedent dry weather days were zero. The author observed road sediment residual, even after rainfall events, and the SS values obtained, when antecedent dry weather days were zero, were consistent with this observation. Table 1 shows similar trends among SS, particulate heavy metals, and dissolved heavy metals alongside antecedent dry days. However, the correlation factor between the antecedent dry days and SS was not high (0.475) suggesting that more precise analyses, which consider the build-up and wash-off mechanisms, are needed to quantitatively characterize the effects of the antecedent dry weather days on the variations in water quality.

Table 2 shows that the runoff contained high levels of heavy metals, which was consistent with the visible abundance of SS (Figure 6), and suggests an important potential role for porous concrete in the removal of particulate heavy metals during the passage of runoff water via porous concrete because the particles are expected to be trapped.

**Figure 6.** Collected water after shaking to resuspend the particulate matter.


**Table 2.** Proportions of particulate heavy metal to total heavy metal in the samples shown in the Table 1.

#### *2.2. Comparison of Dissolved Heavy Metal Concentrations with Previous Studies*

Compared to previous studies [13–15], the heavy metal concentrations in the collected samples, shown in Table 1, were high; the results of the three studies are reviewed in Murakami [48] (Table 2.11). Specifically, the concentration of Cr was similar to those in Sansalone and Buchberger [14] and Pitt et al. [15], while it was one to two orders higher than in Shinya et al. [13]. The concentration of Cu was one order lower than in Shinya et al. [13] and Sansalone and Buchberger [14] and similar to those found by Pitt et al. [15]. The concentration of Zn was one to two orders lower than in Shinya et al. [13] and Sansalone and Buchberger [14] and similar to those found by Pitt et al. [15]. The Cd concentration was higher than in Shinya et al. [13], one order lower than in Sansalone and Buchberger [14], and similar to those in Pitt et al. [15]. The Pb concentration was one order lower than those identified in the three aforementioned studies [13–15].

The results of these comparisons suggest that the heavy metals levels in all four studies (including this study) varied markedly, presumably reflecting the differences in the environmental conditions at the four sites. For Zn, the environmental standards in Japan are set at 0.03 mg/L for rivers and lakes; this level was exceeded only once at the study site (19 December 2019). The mean heavy metal concentrations reported by Shinya et al. [13] and Sansalone and Buchberger [14] were one order higher than the environmental standards, suggesting that there is a need to reduce non-point Zn in Japan, and presumably the other heavy metals in stormwater runoff in urban areas.

A study by Flores-Rodriguez et al. [49] measured Pb, Zn, and Cd concentrations in stormwater samples collected at eight sites. The Pb and Cd concentrations were one order higher than most of the values obtained in this study, while the Zn levels were more varied, yet tended to be one order lower than the values shown in Table 1. In a study by Mikkelsen et al. [50], the Pb, Zn, Cd, and Cu concentrations were measured in different types of urban runoff. They found that the Cd concentrations were higher and Cu concentrations were lower than the values shown in Table 1, while the concentrations of Pb and Zn were mostly similar. Numerous factors are considered to affect non-point sources of heavy metals, as demonstrated in the studies by Ozaki et al. [51,52].

#### **3. Control of Non-Point Sources of Pollution and Sewage Systems**

*3.1. Non-Point Source Pollution and Sewage Systems*

The present study focused on reducing non-point heavy metal runoff. The benefits of using infiltration techniques to decrease heavy metals in runoff are (i) to avoid constructing water treatment facilities; (ii) to facilitate multipurpose uses for the surfaces used for infiltration during fine weather (i.e., these surfaces could be used for other activities); (iii) to retain the water that can be recycled for uses other than drinking.

Owing to the high ratio of paved (i.e., impervious) to non-paved (i.e., pervious) areas in urban centers, high runoff volumes and peak discharge rates are observed. Thus, mitigation measures employing infiltration and storage are very important in urban areas because they prevent flooding and/or inundation.

Furthermore, infiltration techniques are both indirectly and directly effective for water quality control, as described below.

Direct methods of control include methods such as those that the author is attempting to implement. In these cases, the infiltration site (i.e., the bottom and the sides of the street stormwater inlets) could be said to achieve two aims: water volume control and quality control. These two aims could be met by water penetration and these sites could be regarded as high-performance facilities.

#### *3.2. Indirect and Direct Means of Reducing Non-Point Pollution Runoff Loads*

Indirect control methods promote penetration of the water at the surface of the infiltration stratum and the subsequent retention therein. Percolation of the water in the infiltration stratum is referred to as temporal retention, which is different from the water inside the retention ponds because it is exfiltrated into the natural base soil below the stratum and/or through drainage pipes [53–55]. This method reduces the volume of runoff and decreases the volume and frequency of combined sewage overflow (CSO), as well as the sweep flow inside stormwater pipes, gutters, etc. Originally, the idea of runoff volume reduction developed from the viewpoint of flood control. The frequent occurrence of heavy rain events around the world in recent years has required catchment managers to re-evaluate runoff volume reduction, i.e., to consider the mutual benefits of flood control and runoff load reduction.

Direct control methods to control non-point pollution runoff aim at controlling (i.e., trapping and adsorbing) the pollution in the stormwater inside the infiltration stratum. In the case of porous concrete, the mechanisms of pollution control were clarified based on laboratory experiments (see Section 5). The economic aspects of the infiltration technique have been previously analyzed [36]. We intended to develop direct methods for runoff reduction into separate sewage systems.

#### **4. Infiltration as a Direct Pollution Control Method in a Separate System**

#### *4.1. Porous Concrete as An Infiltration Medium*

Porous concrete is produced using large aggregates, which ensures that permeability is maintained with only a minor decrease in hardness compared to normal cement. The author used an aggregate called Gmax15, which is composed of gravel measuring less than 15 mm.

Porous concrete columns (Figure 7) and porous concrete cubes (Figure 8) were used in the laboratory experiments. Columns were prepared using a mixture of cement (0.3 kg), gravel (Gmax15, 1.55 kg), water (8.05 kg), and admixture (high-range water-reducing admixture, 0.003 kg) [38].

**Figure 7.** Experimental porous concrete column measuring 10 cm in diameter and 10 cm in depth.

**Figure 8.** Porous concrete cube with each side measuring 4 cm.

The saturated hydraulic conductivity of the column was measured using the constant water level method and estimated to be 1800 mm/h.

#### *4.2. Deployment of Porous Concrete Plate at the Bottom of Street Stormwater Inlets*

Using porous concrete on the bottom of the street and not at the ground surface of stormwater inlets (Figure 9) has been proposed by the author's research group as a means of reducing heavy metal runoff, via filtration and adsorption, and water runoff reduction, via exfiltration, into the natural base soil around the bottom of the street stormwater inlets [36–38]. The reasons why the bottoms of the street stormwater inlets were selected as sites to deploy the porous concrete plates were because (i) porous concrete has a structural weakness and cannot bear significant loads, (ii) the inflows of the stormwater during rainfall events will keep the porous composite unclogged, and (iii) the street stormwater inlets are located in the stormwater drainage networks.

**Figure 9.** Porous concrete plate at the bottom of the street stormwater inlet.

Figure 10 shows the behavior of the water passing through the permeable bottom plates in the stormwater inlets, although, in reality, the walls could also be permeable, as described below. Here, the behavior of the water at the bottom of the inlets is shown to illustrate the two proposed functions of the stormwater drains, i.e., the reduction of heavy metals in the runoff via filtration and adsorption reactions, and the overall reduction in water runoff, via the exfiltration into the natural base soil around the bottom of the street stormwater inlets [36–38].

**Figure 10.** Schematic diagram showing the behavior of water through the porous concrete when placed only at the bottoms of street stormwater inlets.

#### **5. Laboratory Experiments Examining the Potential Reduction in Heavy Metals in Porous Concrete Exposed to Runoff**

Using artificial rainfall (mixture of Zn, Cu, and Pb), the author conducted experiments using porous concrete columns (Figure 7; diameter and depth: 10 cm) and porous concrete cubes (Figure 8; 64 cm3). Specifically, the author investigated the adsorption rates of heavy metals using concrete columns and cubes under various conditions.

The porous concrete column samples (C series) consisted of a column prepared in February 2008 at the School of Food, Agricultural and Environmental Sciences at Miyagi University, alongside the column samples (N series) prepared outside the University in November 2016. Both columns were prepared using the same mixture, coefficient of permeability, and size.

We performed 12 experimental runs (Table 3). In each run, a single column was placed in a Petri dish and a 50 mL solution was added comprising either a mixture of the heavy metals Zn, Cu, and Pb, or Zn or Cu, or Pb individually. The concentrations used for Pb were higher than the level of dissolved Pb in Table 1, although Zn and Cu concentrations were of the same order of magnitude, while each was sprayed onto the top of the sample column using a 50 mL volumetric pipette (Figure 11). The experiment was left to run overnight. Then, the leachate that had accumulated in the glass Petri dish was collected, and the amount of the leachate and the concentrations of Pb, Zn, and Cu were measured. The runs for each experimental condition were repeated 2–13 times. The amount of leachate in each run varied from 25 to 45 mL. The adsorption rate was determined based on the relationship between the volume of leachate, the concentration of each heavy metal, the initial volume of solution added, and the concentration. In addition, the time taken for the 50 mL of the solution to flow out of the sample (about 25 s) was measured periodically to confirm that the spray intensity had not changed.


**Table 3.** Twelve experimental runs for the leachate experiments.

**Figure 11.** Leachate experiments.

The variations in the proportion of Pb, Zn, and Cu adsorbed by the concrete columns in the indoor artificial rainfall experiments are shown in Table 4. The findings revealed that (i) similar proportions of Pb, Zn, and Cu were absorbed when the leachates contained mixed heavy metal solutions and when they contained individual heavy metals (Run 1-1, Run 2-1 and Run 3-1 and Run 5; Run 1-2, Run 2-2, Run 3-2 and Run 5; Run 1-3, Run 2-3 and Run 3-3 and Run 5), (ii) similar proportions of the three heavy metals were absorbed by the

columns, even when the concentration of each heavy metal showed variations (Run 1-1, Run 2-1 and Run 3-1; Run 1-2, Run 2-2 and 3-2; Run 1-3, Run 2-3 and Run 3-3), (iii) similar adsorption proportions for Cu (Run 3-3 and Run 4-1) and for Zn and Pb (Run 5 and Run 6) were shown and (iv), smaller proportions were shown when the concentrations of the heavy metals were low (less than 10 ppb), such as in Runs 4-1 and 6, where a few ppb Cu elution from the column N series arose. The elution was confirmed separately by adding the ultrapure water (Millipore MQW) onto sample columns and again by the subsequent measurements of Pb, Zn, and Cu in the leachate. The results showed that a few ppb of each heavy metal leached from the recycled concrete used in the porous concrete columns. The small amount of heavy metal leaching decreased the apparent adsorption proportions. To resolve this problem, the author attempted to purify the porous concrete columns by submerging the columns in ca. 30 L pure water for approximately 2 weeks; this process was repeated three times after replacing the water with new pure water. Using this procedure, leaching from the columns decreased by 90% compared to the original concentrations of heavy metals in leachate and the porous concrete columns could be used to assess actual stormwater samples.


**Table 4.** Proportion of heavy metals adsorbed by the columns.

#### **6. Effectiveness of the Porous Concrete Plates Placed at the Bottom of the Inlets Based on Calculations Using Infoworks ICM (Innovyze)**

The magnitude of infiltration for a 1 ha catchment at 20 discrete street stormwater inlets where the porous concrete plates were deployed was estimated using the author's own simulation results.

Firstly, the density of the stormwater inlets in the catchment needed to be estimated. In Japan, the density of the street stormwater inlets is 10–30/ha (20 in Harada and Komuro [37]); therefore, in this study, a value of 20 was used for the inlet density in 1 ha. Using the maximum adsorption capacity of Zn, by the cube (Figure 8) [37], the duration that the Zn runoff did not occur was estimated at about 41 years [37]. By obtaining the EMC values, a more precise duration could be calculated.

Assuming that the bottom plate was circular with a bottom thickness of 10 cm and side walls that were 10 cm thick with a 10 cm water level, while the average diameter of all stormwater inlets was set as 1.8 m, the area of the porous concrete at 1 inlet would correspond to 31,086 cm2. Thus, for the 1100 inlets in the Fukumuro catchment (Figure 12, the 9 inlets are the ones selected to conduct the passage proportion of water via porous concrete, as mentioned in Section 8), the area of the permeable media was estimated as

3419.5 m2. This area corresponds only to 0.04% of the catchment area (ca. 900 ha). Using Infoworks ICM (Innovyze) [56,57], Harada and Kim [36] showed that mitigation of the inundation occurrence happened following a 68.5 mm rainfall event in the Fukumuro catchment.

**Figure 12.** Fukumuro catchment and the nine sewage traps used in the study.

The analyses described here show that the ability to control stormwater runoff is quite high when small and discrete infiltration is used in a catchment.

#### **7. Verification of Exfiltration Coefficient Obtained Using Infoworks ICM**

The simulation described in Section 6 assumed that the exfiltration rate (i.e., threedimensional water seepage into the natural base soil below and around the inlets) of every street stormwater inlet was 2000 mm/h. The magnitude of the exfiltration coefficient was quite sensitive in those analyses, thus, the suitability should be examined. This examination is of the validity of the 3D to 1D conversion coefficient. Here, the author has introduced four concepts from previous studies to clarify the suitable exfiltration rate.

Firstly, based on the observed infiltration rate in the vicinity of the street stormwater inlet in the Fukumuro catchment where the natural base soil was exposed at the surface (3600 mm/h, the 3D to 1D conversion coefficient was 36 because of the saturated hydraulic conductivity of the natural base soil (sand and silt), which was 100 mm/h) [36].

Next, referring to Herath et al. [58] and Herath and Mushiake [59], the ratio of the exfiltration rate to the hydraulic conductivity from infiltration trench (q/k0 in the Fig.2 of Herath and Mushiake [59]) was considered according to the matric and gravity potential slopes and water pressure (the model domain used for the numerical simulation is shown in Fig.1 of Herath and Mushiake [59]). The Fig.1 of Herath and Mushiake [59] shows a symmetric analysis, thus, the width of the trench shown in the Fig.1 of Herath and Mushiake [59] could be replaced with the average radius of the inlets in the Fukumuro area, 0.9 m. Substituting the width of the trench as 0.9 m, and the water level, 10 cm (obtained via infoworks ICM calculations) into the Fig.2 of Herath and Mushiake [59], the author obtained the 3D to 1D conversion coefficient corresponds to ca. 100.

Blazejewski et al. (2018) [60] did not consider the matric potential slope and demonstrated a 3D to 1D conversion coefficient of 1.0–1.3 [60].

The author performed the 3D to 1D conversion coefficient sensitivity tests in Infoworks ICM by changing the coefficient to 0, 5, 10, and 20 [36]. The mitigation mentioned in Section 6 used a coefficient of 20.

The author assumed that a value of 10–20 was plausible. First, the coefficient should be larger than in Blazejewski et al. (2018) [60], thereby considering the matric and gravity potential slopes with the water head. Moreover, the coefficient should be smaller than that shown by the author's observed value of 36 (where the soil was dry) because the natural base soil around the street stormwater inlets should be wet, whereby the design runoff ratio of the Fukumuro area was 0.40 [36].

#### **8. Proportion of Water Passing through the Porous Concrete Plates at the Bottom of the Street Stormwater Inlets to the Total Volume of Inflow Based on Estimates Calculated Using Infoworks ICM (Innovyze)**

The author analyzed the proportion of water passage through the porous concrete at the bottom and the side walls of the street stormwater inlets, using alternative 3D to 1D conversion coefficients of 0, 5, 10, 15, and 20 and simultaneously changing the rainfall volume to 1, 3, and 5 mm. The proportions of water passage through the porous concrete column were calculated as the "volume of water exfiltrated into the soil from the street stormwater inlets" divided by "the volume of water that enters the street stormwater inlets" multiplied by 100 (%). The proportions at the 9 street stormwater inlets in the Fukumuro area (Figure 12) during 3 mm of rainfall, when the 3D to 1D conversion coefficient was 10 using Infoworks ICM (Innovyze) [56,57], is shown in Figure 13.

**Figure 13.** The proportion of the passage of the porous concrete at the bottom of the sewage traps in case the rainfall volume is 3mm and the exfiltration rate is 1500 mm/h.

Figure 13 shows the passage proportion of the water through the porous concrete at the bottom and side walls of the inlets. The proportions varied from about 10% to 88%, excluding the proportion at the inlet of 173, meaning that the calculations there showed that these numerical analyses were unsuitable. The proportions were inversely related to the increasing rainfall volume and related to the increasing 3D to 1D conversion coefficient. However, reaching 100% was uncommon, even though heavy metal removal was expected as the proportions increased. Thus, the author proposed using a hanging-type porous concrete plate, shown in Figure 14. This configuration of which has already been deployed in Sendai.

**Figure 14.** Hanging-type placement of the porous concrete at the same inlet as the one shown in Figure 1. The grating is open. Closing the grating results in the porous concrete board assuming a horizontal orientation.

#### **9. Conclusions and Future Work**

The present study reported the results of the author's monitoring of heavy metal concentrations in urban road runoffs. Indoor experiments were conducted to analyze the adsorption of heavy metals by porous concrete. In addition, the study also examined the amount of runoff that could be treated in stormwater drains fitted with porous concrete filters. While Section 2 presented some of the behavior of non-point heavy metals, there remains a need to explain them in terms of environmental factors using statistical analyses, as highlighted by Ozaki et al. [51,52]. It is recommended that the EMC values should be corrected, and improvements be made to the 3D/1D conversion methods.

**Funding:** This research was funded by Japan Science and Technology Agency grant number [JP-MJTM20BK], Fukushima University Research Fund grant number [262q006], Miyagi University Research Funds and KC Miyagi Research Fund 2019.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The author declares no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

**Kei Nakagawa 1,\*, Hiroki Amano 2, Zhi-Qiang Yu <sup>3</sup> and Ronny Berndtsson <sup>4</sup>**


**Abstract:** Nitrate pollution in groundwater is a severe problem in Shimabara Peninsula, Nagasaki Prefecture, Japan. Previous studies have investigated water quality characteristics in the northern part of the peninsula and shown serious effects of nitrate pollution in the groundwater. The present study aimed to investigate the groundwater quality in the southern areas of the peninsula for improved understanding of the water quality status for the entire peninsula. Groundwater samples were collected at 56 locations in Minami-Shimabara City from 28 July to 4 August 2021. The spatial distribution of water quality constituents was assessed by Piper-trilinear and Stiff diagrams for major ion concentrations. One agricultural area in the western parts exceeded Japanese recommended standards for water. According to the Piper-trilinear diagram, 44 sampling sites (78.6%) were classified as alkaline earth carbonate type, nine sites (16.1%) as alkaline earth non-carbonate type, and three sites (5.3%) as alkaline carbonate type. Stiff diagrams displayed Ca-HCO3 water type for most of the sites. Na-HCO3 and Mg-HCO3 types were found in coastal areas. Principal component analyses showed that the first component corresponded to dissolved constituents in groundwater and denitrification, the second effects of ion exchange and low nitrate pollution, and the third effects of severe nitrate pollution. Hierarchical cluster analysis was used to classify the groundwater into five groups. The first group included sites with relatively high nitrate concentration. The second group had relatively low ion concentration, distributed from center to eastern parts. The third group included intermediate ion concentration, distributed at lower altitudes along the coastal line. The fourth and fifth groups had a higher ion concentration, especially characterized by high sodium and bicarbonate concentration.

**Keywords:** groundwater; principal component analysis; hierarchical cluster analysis

#### **1. Introduction**

Groundwater is a globally important resource used for domestic, industrial, and agricultural purposes. In 2018, groundwater dependency for these uses in Japan was 20.1%, 27.3%, and 5.4%, respectively [1]. In some areas, the dependence on groundwater for domestic use is 100%. Therefore, groundwater quality assessment is essential to protect residents' health and preserve their living environment. In Japan, assessments have continuously been carried out for 28 water quality parameters (e.g., cadmium, lead, arsenic, dichloromethane, carbon tetrachloride, benzene, nitrate+nitrite-nitrogen (NO3+NO2-N), fluoride, and boron) for which environmental standards of groundwater pollution are established under the Basic Environment Law. In 2020, an investigation was conducted for 3103 wells to assess the overall groundwater quality. In total, 10 parameters were found

**Citation:** Nakagawa, K.; Amano, H.; Yu, Z.-Q.; Berndtsson, R. Groundwater Quality and Potential Pollution in the Southern Shimabara Peninsula, Japan. *Water* **2022**, *14*, 4106. https://doi.org/10.3390/w14244106

Academic Editor: Cesar Andrade

Received: 9 November 2022 Accepted: 14 December 2022 Published: 16 December 2022

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

to exceed environmental standards [2]. NO3+NO2-N had the highest exceedance at 3.3%. This was followed by arsenic at 2.1% and fluoride at 0.8%. NO3+NO2-N has continued to present the highest exceedance rates since it was added to the criteria in 1999, and this is typified throughout Japan. Nitrate pollution in groundwater is the most prevalent type of anthropogenic pollution [3]. In many cases, nitrate pollution in groundwater is caused by fertilizer and livestock wastewater in agricultural areas [3]. Other pollutants, such as arsenic and fluoride stem from natural sources originally contained in rocks [3]. The release and fate of these elements in groundwater are controlled by water-rock interaction processes in hydrogeological paths of the water [4,5].

The Shimabara Peninsula in Nagasaki Prefecture, Japan, depends on groundwater for most of its water supply due to the small amounts of surface water [6]. However, the peninsula is experiencing severe nitrate pollution in groundwater. Water quality monitoring has shown that the number of tap water source (groundwater) wells above environmental standards for NO3+NO2-N ranged from 5.1% to 10.0% during 2005 to 2020 [7]. Although the general pollution level is slowly decreasing, some wells are still showing increasing concentrations and require continuous monitoring. The Shimabara Peninsula consists of three cities: Shimabara, Unzen, and Minami-Shimabara. In 2011, we started an investigation on groundwater quality in Shimabara City, where pollution is particularly severe. We found concentrations up to 26.6 mg/L NO3-N related to intensive agricultural activities (fertilizer and livestock production) [8]. Further investigation displayed pollution spread with depth in groundwater [9] and that there is a strong relationship between pollution in soil and groundwater [10]. In 2018, the study area was expanded to Unzen City. We found maximum NO3-N concentrations of 19.9 mg/L and that the pollution is related to agriculture as in Shimabara City [11]. Although the assessment of groundwater pollution has been established for Shimabara and Unzen Cities; groundwater chemistry including nitrate pollution has not been studied in the Minami-Shimabara City. For this reason, we evaluated spatial characteristics of groundwater chemistry in the Minami-Shimabara City for improved understanding of the status of nitrate pollution in this area and the hydrochemical characteristics of groundwater in vulnerable aquifers compared to the hydrochemistry of the Shimabara and Unzen Cities.

#### **2. Materials and Methods**

#### *2.1. Study Area*

Minami-Shimabara City is located in the south of Shimabara Peninsula, Nagasaki Prefecture (Figure 1) with an area of approximately 17,000 ha. In 2017, the population was 44,200 with a water supply coverage of 92.8% [12]. There are 5 surface reservoirs, a main river, 60 groundwater sources, and 2 springs supplying water to the city [12]. Groundwater has the highest ratio of water withdrawal, reaching 86.1% [12]. The land use map is shown in Figure 1a [13]. Forests cover 35.6% of the total area [14]. The agricultural land area is 4730 ha (paddy fields 1650 ha; upland fields 3070 ha) [14], equivalent to 27.8% of the city area. This ratio is higher than that of Nagasaki Prefecture as a whole and is characterized by agriculture use in upland fields. Mainly cultivated crops are rice, potatoes, fodder crops, and leaf tobacco. Nitrogen application by chemical fertilizer and manure for the 10 major crops is estimated at 1715 kg/day [7]. Livestock production is thriving, with 223 livestock facilities (49 dairies, 138 beef cattle, 14 pork, 5 egg, and 17 broiler production units) in 2019 [7]. This corresponds to 35,950 cattle, 79,000 pigs, and 2,900,000 chicken [7]. The resulting nitrogen generation from livestock waste is estimated at 6625 kg/day [7].

**Figure 1.** Study area and groundwater sampling locations in the Minami-Shimabara City.

Figure 2b shows the geology of the study area [15]. In the southern part of the area, the Kuchinotsu formation, which consists mainly of freshwater sediment and shallowmarine sediment of middle Pliocene to early Pleistocene age [16], is exposed. Pre-Unzen volcanic rocks (4 Ma–500 ka) are partially distributed above the Kuchinotu formation. Pre-Unzen volcanic rocks and Kuchinotsu formation are overlain by the Unzen volcanic rocks (500 ka–present), which are widely distributed in the Shimabara Peninsula [16]. Unzen volcanic rocks are observed near the southern border and in the center of the peninsula. Alluvial deposits are found at lower elevation areas in the north. Pre-unzen volcanic rocks are mainly composed of olivine basalt and amphibole andesite, while Unzen volcanic rocks contain hornblende andesite to dacite rich in plagioclase, hornblende, and biotite mottling [16]. Figure 2c shows the hydrogeological map of the study area [17]. Groundwater levels are only evident in a part of the study area. There are many faults in this area, and the groundwater levels differ greatly due to these. Especially, in the south and east, groundwater levels are complex due to faults.

**Figure 2.** Land use and geology, (**a**) vegetation, (**b**) geology, (**c**) hydrogeology; (Pyfl) Pyroclastic flow. Vegetation map is based on data collected by Biodiversity Center of Japan [13]. Geological map 1/200,000 scale is based on data collected by Geological Survey of Japan [15]. Hydrogeological map is based on the groundwater investigation by Murakami [17].

The climate is humid subtropical. Mean temperature and mean annual precipitation are 17.3 ◦C and 1836 mm, respectively (1992 to 2021; data from Kuchinotsu observatory, 32◦36.7 N, 130◦11.6 E) [18].

#### *2.2. Sampling and Analyses*

Groundwater samples were taken from 56 municipal wells (Figure 1) during July to August 2021. All samples were stored in pre-washed bottles. Stagnant water in the well pipes was removed before water sampling. pH, electrical conductivity (EC), oxidant redox potential (ORP), and dissolved oxygen (DO) were measured on-site by using hand-held instruments (HORIBA D-51 and D-54, and HACH HQ30d). Bicarbonate ion (HCO3 −) was determined by the titration method with 0.1 N HCl on-site. All water samples were filtered by 0.45 μm membrane filter on-site. Major anions (Cl−, NO3 −, SO4 <sup>2</sup>−) and cations (Na+, K+, Mg2+, Ca2+) were measured by ion-chromatography (DKK-TOA, IA-300) in the laboratory. The validity of the analytical results was confirmed by converting ion concentrations from mg/L to mmolc/L and then calculating the charge balance errors (CBE) using:

$$\text{CBE} = \left(\frac{\sum \text{cations} - \sum \text{anions}}{\sum \text{cations} + \sum \text{anions}}\right) \times 100,\tag{1}$$

According to Rahman et al. [19], CBE < ±5 is good, and CBE < ±10 is acceptable. In this study, the CBEs of the samples ranged from −3.6 to 7.3, with <±5 in 41 samples and <±10 in 15 samples.

Principal component analysis (PCA) and hierarchical cluster analysis (HCA) for the hydrochemical parameters were used to interpret processes controlling water chemistry (mineral dissolution, anthropogenic input, sea water intrusion, ion exchange, evapotranspiration, etc.) and grouping the water samples e.g., [20–24]. These methods were also used previously for the Shimabara and Unzen groundwater [8,11]. PCA and HAC were performed based on major ion concentrations (Cl−, NO3 −, SO4 <sup>2</sup>−, HCO3 <sup>−</sup>, Na+, K+, Mg2+, and Ca2+). The principal components were extracted based on the Kaiser criterion to only retain components with eigenvalues greater than 1. The HCA was performed based on

Ward's method. All analyses were performed by using the statistical software JMP Pro 13 (SAS Institute Inc., 100 SAS Campus Drive, Cary, NC 27513-2414, USA).

#### **3. Results and Discussion**

#### *3.1. Water Chemistry*

The analysis results of hydrochemical parameters are shown in Table 1 and Figure 3. pH ranged from 6.3 to 8.7, indicating that groundwater in the study is weakly acidic to weakly alkaline. EC ranged from 74 to 563 μS/cm. The range of pH and EC was generally close to the Shimabara and Unzen groundwater [8,11]. ORPs ranged from −21 to 738 mV, meaning that groundwater is oxic or anoxic. The latter was observed at two sampling points (MW51 and MW55). The redox conditions are strong indicators of contaminants that might be present at elevated concentrations [25]. The nitrate concentration is more likely to exceed the recommended limits for oxic conditions [25]. On the other hand, microbially driven reduction of nitrate to nitrogen gas occurs only under anoxic conditions [25]. Gillham and Cherry [26] reported that denitrification processes can occur when the DO in groundwater is less than 2 mg/L. In our study, DO ranged from 1.7 to 10.2 mg/L, with two sampling points (MW44 and MW50) showing less than 2.0 mg/L. Thus, at sampling points (MW44, MW50, MW51, and MW55) with low DO or ORP, NO3 − concentrations between 0.1 to 1.9 mg/L indicate that denitrification is likely to occur. These conditions were not observed in Shimabara and Unzen [8,11]. The total dissolved solids (TDS) were estimated from EC by the following equation [27].

$$\text{TDS} = 640 \times \text{EC}\_{\prime} \tag{2}$$

where TDS and EC are in mg/L and dS/m, respectively. TDS ranged from 47.4 to 360.3 mg/L with a mean of 134.4 mg/L. The samples were characterized by HCO3 − followed by SO4 <sup>2</sup>−, Cl−, and NO3 − for anions (Figure 3). Cations were characterized by Ca2+ and Na<sup>+</sup> followed by Mg2+ and K+.

**Table 1.** Descriptive statistics of hydrochemical components of groundwater samples.


Note: <sup>1</sup> S.D. = standard deviation.

**Figure 3.** Box plots of hydrochemical components of groundwater samples. A circle is a far outlier. Starred point is a "far out" outlier.

Groundwater samples were plotted using trilinear Piper diagram for easy visual understanding the characteristics of water chemistry (Figure 4). Forty-four sampling points (78.6%) were classified as alkaline earth carbonate type, which is common in shallow groundwater in Japan. Nine points (16.1%) showed alkaline earth non-carbonate type, which is classified as groundwater with high nitrate concentration in Shimabara and Unzen [8,11]. Three samples (5.3%) were classified as alkaline carbonate type. This type of water is commonly found in deep groundwater with long residence time and was not identified in Shimabara and Unzen [8,11]. To consider salinity, Total Ionic Salinity (TIS) is shown in Figure 5 [28]. According to the figure groundwater in Minami-Shimabara has low TIS (<10 mmolc/L).

**Figure 4.** Trilinear Piper diagram of groundwater in Minami-Shimabara City.

**Figure 5.** Total Ionic Salinity of groundwater in Minami-Shimabara City.

The concentration of major dissolved ions is visualized using Stiff diagram in Figure 6. Most of the samples represent the Ca-HCO3 type. The Ca-(SO4+NO3), Ca-SO4, Na-HCO3, and Mg-HCO3 types were observed in a small number of wells. Na-HCO3 and Mg-HCO3 were found in the coastal area and thus, have larger arrows than other samples. To decipher mechanisms controlling groundwater chemistry, Gibbs diagram [29–31] is shown in Figure 7. Gibbs diagram is described by the ratio of cation and anion endmembers (Na+/(Na+ + Ca2+)) and (Cl−/(Cl<sup>−</sup> + HCO3 −)), respectively, as a function of TDS. This indicates the functional sources of dissolved chemical constituents, such as precipitation, rock weathering, or evaporation dominance. The diagram shows that most groundwater samples are affected by rock weathering processes, suggesting that longer residence time, large contribution of mineral dissolution, and the influence of ion exchange form water chemistry in the study area. A few samples were influenced by precipitation. The difference of these factors appears as the difference of ion concentrations and water type. The water samples in precipitation dominance have lower concentration. Higher concentrations and water types of Na-HCO3 and Mg-HCO3 are dominated by rock weathering processes. Influence of seawater was mainly found in the south part of Minami-Shimabara City [17], but no Na-Cl type samples were identified. Figure 8 shows the distribution of groundwater samples in the Hydrochemical Facies Evolution (HFE) diagram [32] by using Excel Macro [33]. HFE-diagram reveals that 70% of the groundwater samples correspond to freshening phase and 30% correspond to intrusion phase. Since the study area faces to the sea, it is assumed that groundwater samples, which is characterized by high concentration and ratio of Na+, such as MW42, MW43, and MW44 in the left upper corner of freshening phase, are affected by seawater intrusion.

**Figure 6.** Major ion variation in Minami-Shimabara City groundwater using Stiff diagrams.

**Figure 7.** Gibbs diagram of groundwater in Minami-Shimabara City.

**Figure 8.** Hydrochemical Facies Evolution (HFE) of groundwater in Minami-Shimabara City.

The southern part of the city has little surface water and relies on groundwater for irrigation. However, as mentioned above, groundwater with high Na+ concentrations were observed. Therefore, the suitability of groundwater for irrigation was evaluated by plotting a USSL diagram (Figure 9). The USSL diagram is drawn based on the salinity (EC) and sodium hazard (Sodium Adsorption Ratio: SAR) [34]. The SAR is calculated by [34]:

$$\text{SAR} = \frac{\text{Na}^+}{\sqrt{\left(\text{Ca}^{2+} + \text{Mg}^{2+}\right)/2}} \text{} \tag{3}$$

where the concentrations of Na+, Ca2+, and Mg2+ are all expressed in mmolc/L. Most of the groundwater samples belong to C1-S1 (low salinity hazard and low sodium hazard) class or C2-S1 (medium salinity hazard and low sodium hazard). Only two samples (MW42 and MW43) were classified in the C2-S2 (medium salinity hazard and medium sodium hazard) class. The following is a summary of each class [34]. Low salinity water (C1) can be used for irrigation with most crops on most soils with little likelihood that soil salinity will develop. Some leaching is required, but this occurs under normal irrigation practices except in soils of extremely low permeability. Medium salinity water (C2) can be used if a moderate amount of leaching occurs. Plants with moderate salt tolerance can be grown in most cases without species practices for salinity control. Low salinity water (S1) can be used for irrigation on almost all soils with little danger of the development of harmful levels of exchangeable sodium. However, sodium sensitive crops, such as stone-fruit trees and avocados may accumulate harmful concentrations of sodium. Medium sodium water (S2) will present an appreciable sodium hazard in fine textured soils having high cation exchange capacity, especially under low leaching conditions, unless gypsum is present in the soil. Consequently, Figure 9 shows that groundwater in the study area can be used for irrigation, but attention must be paid to the use of groundwater belonging to the C2-S2 class. Further evaluation of water quality for irrigation was conducted based on Food and Agriculture Organization (FAO) guidelines [35] (Table 2). According to the FAO guidelines, although most of the groundwater samples has no potential irrigation problem, some samples have elevated EC, SAR, Na, NO3-N, and HCO3 or even potentially severe problem in view of SAR. In addition, some of the samples have a pH somewhat out of range. Therefore, caution should be exercised in some places when using groundwater. Since most tree crops and woody plants are sensitive to sodium [35], groundwater having high SAR should be avoided. These results are generally consistent with the caution given from the USSL diagram.

**Figure 9.** USSL diagram based on EC and SAR.

**Table 2.** Guidelines for interpretations of water quality for irrigation.


#### *3.2. Nitrate Pollution*

The nitrate (NO3-N) concentration ranged between 0.02 and 12.6 mg/L, with an average of 1.8 mg/L. The maximum concentration was found at MW39, the only point where NO3-N concentration exceeded the Japanese drinking water standard (10 mg/L). The drinking water standard of World Health Organization [36] for NO3 − concentration (50 mg/L) was exceeded in one point. The maximum nitrate concentration in Shimabara City was 26.6 mg/L [8] and that in Unzen City 19.9 mg/L [11]. Compared to these values, the pollution level in Minami-Shimabara City is lower. Only one sampling point exceeded the standard limits for NO3-N concentration. This is lower than the 38% in Shimabara City [8] and the 12% in Unzen City [11]. Even if the standard limits were not exceeded, relatively high NO3-N concentrations were observed, e.g., 8.4 mg/L at MW49, 7.3 mg/L at MW53, and 6.5 mg/L at MW15. Nitrate can occur naturally, but nitrate

concentrations greater than 1 mg/L are likely to indicate effects of human activities [37]. Under this criterion, 21 sampling points (38%: except for the smallest blue circle in Figure 10) can be considered polluted by nitrate. As mentioned above, denitrification may occur in the groundwater in this area, and it is considered that at some sampling points the nitrate concentration has decreased below 1 mg/L due to denitrification. Sampling points with nitrate concentration exceeding the criteria of 1.0 mg/L [37] were observed in the southwestern and eastern parts of the study area (Figure 10), and the land use here is upland agriculture (Figure 2). This suggests that nitrate pollution of groundwater in this area is related to agricultural activities.

**Figure 10.** Distribution of nitrate concentration in groundwater.

Correlation among the eight major dissolved ion components is shown in Table 3. In the investigation of Shimabara and Unzen [8,11], NO3 − showed a high positive correlation with Cl−, SO4 <sup>2</sup>−, and K+, suggesting that nitrate is derived from chemical fertilizers ((NH4)2SO4), manure, and livestock waste. However, in Minami-Shimabara, NO3 − is not correlated with Cl−, SO4 <sup>2</sup>−, and K+. Focusing on Cl−, SO4 <sup>2</sup>−, and K+ concentrations in MW39, where NO3-N concentration exceeded the Japanese drinking water standard, the relationship is as follows: Cl<sup>−</sup> 16.9 mg/L > K<sup>+</sup> 5.1 mg/L > SO4 <sup>2</sup><sup>−</sup> 4.2 mg/L. For MW49 with NO3-N of 8.4 mg/L, which has the next highest NO3-N concentration, a relationship of Cl− 12.3 mg/L > K<sup>+</sup> 6.4 mg/L > SO4 <sup>2</sup><sup>−</sup> 2.7 mg/L was observed. The SO4 <sup>2</sup><sup>−</sup> concentration at these points is similar to or lower than those at other points with low NO3-N concentration (Figure 6). At the sampling points where such a relationship between ion concentrations is found, SO4 <sup>2</sup><sup>−</sup> concentration is low, and manure and/or livestock waste are supposed to be a dominant nitrate source. Analysis of stable nitrogen oxygen isotopes NO3 − is required for a more valid assessment of the nitrate source.

**Table 3.** Correlation matrix for eight dissolved ions (*n* = 56).


#### *3.3. Multivariate Analysis*

A principal component analysis summarized the 12 hydrochemical parameters into three principal components based on the Kaiser index (Table 4). The eigenvalues of principal component 1 (PC1), 2 (PC2), and 3 (PC3) were 5.64, 1.92, and 1.33, respectively. These explained 46.9%, 16.0%, and 11.1%, respectively, of the total variance. PC1 had positive loadings for EC and all ions except NO3 <sup>−</sup> and K+, indicating that it is related to dissolved constituents controlled by rock weathering and precipitation. The negative loadings of ORP and DO in PC1 are related to denitrification. PC2 had positive loadings for Ca2+ and negative loadings for Na+, implying ion exchange. The positive loading of NO3 − in PC2 represents lower nitrate pollution exceeding the criteria of 1.0 mg/L indicating effects of human activities. PC3 had positive loadings for Cl− and NO3 −, indicating relatively severe nitrate pollution including the exceedance of standard limits.


**Table 4.** Relationship between extracted principal components (PCs) and ions.

Results of the HCA are shown in Table 5 and Figure 11. The 56 groundwater samples were classified into five groups, with the number of samples in each group ranging from 3 to 26. Figure 12 shows the scatter plot of the principal components related to groups. PC1 effectively separated groups, meaning that dissolved constituents are different for each group. Group 1 shows relatively low scores for PC1, indicating lower dissolved constituents. Mainly the sampling points indicated by smaller Stiff diagrams in Figure 6 are contained in this group. Group 1 is distributed from the center to east in the study area. Group 2 and 3 has similar scores of PC1 and PC2, representing intermediate ion concentrations. However these groups are be distinguished by PC3, meaning that nitrate pollution level are different for these groups. As shown in Table 5, water samples having higher NO3 − concentrations were classified into Group 2. The difference between Group 2 and 3 is characterized by difference in locality (Figure 13). Group 2 is located in the western part of study area, while Group 3 is mainly located at lower altitude along the coastal line. PC1 and PC3 of Group 4 is similar to that of Group 5, but these groups are clearly distinguished by PC2. In other words, Group 4 has relatively high negative scores for PC2, indicating negative effects of ion exchange, and anthropogenic activities. Group 4 is associated with deep groundwater with high concentrations of sodium and bicarbonate ions (Table 5). The five groups can be generally summarized in three categories.PC1 and Figure 11 show that Group 4 and 5, which are characterized by high ion concentration especially Na+ and HCO3 −, are different from the other groups. Based on overall ion concentration including nitrate, the remaining groups can be categorized by either intermediate ion concentration (Group 3) or lower ion concentration (Group 1 and 2).


**Table 5.** Hydrochemical component depending on each group.

**Figure 11.** Dendrogram for groundwater samples divided into five groups.

**Figure 12.** Scatter plots for PC1, PC2, and PC3.

**Figure 13.** Spatial location of each group.

#### **4. Conclusions**

To improve the understanding of cause and effects of nitrate pollution and hydrochemical characteristics of groundwater in the Shimabara Peninsula, groundwater samples were

collected from 56 municipal wells in Minami-Shimabara City. Major dissolved ions and pH, EC, ORP, and DO were analyzed. The ORP and DO values suggested that denitrification may be responsible for decreasing nitrate concentrations. The major groundwater composition was Ca-HCO3 type. In addition, Na-HCO3, Mg-HCO3, and Ca-(SO4+NO3) types were also observed at a few locations. This water chemistry is formed by rock weathering, precipitation, and mixing with saltwater from seawater intrusion. NO3-N concentrations exceeded Japanese drinking water standards (10 mg/L) at one location. The pollution is related to agricultural land use. The high Cl− and low SO4 2+ concentrations at this point suggest that the pollution source is manure and/or livestock waste. PCA showed that processes controlling water chemistry are explained by three principal components: PC1 corresponds to dissolved constituents in groundwater and denitrification, PC2 represents ion exchange and low nitrate pollution, and PC3 represents severe nitrate pollution. HCA classified the 56 water samples 5 five groups. These can be broadly divided into 3 categories: the first characterized by high ion concentration especially in Na<sup>+</sup> and HCO3 − (Group 4 and 5); the second representing the intermediate ion concentration group (Group 3), and the third with low ion concentration (Group 1 and 2).

The study revealed that the extent of nitrate pollution in Minami-Shimabara City is small. In other words, NO3-N concentrations are lower than in Shimabara and Unzen Cities, and the percentage of NO3-N exceeding the standard limits is also small. The sampling campaign was between July and August when rainfall amount was large. In the investigations conducted during this period, a decrease in nitrate concentration was observed in Shimabara City due to dilution caused by rainfall [8,10]. Therefore, future surveys are needed at different seasons. It is also necessary to introduce analysis of nitrogen oxygen stable isotopes of NO3 − to clarify the nitrate source.

Since this study found high Na<sup>+</sup> concentrations in some wells, we additionally evaluated the suitability of groundwater for agricultural use. Groundwater was evaluated using the USSL diagram and FAO guideline. In some places, due to high EC, SAR, Na, NO3-N, HCO3, and pH, caution is necessary when using groundwater for irrigation. In addition, groundwater with high SAR should not be used for some types of crop species. In Minami-Shimabara City, agricultural wells have been installed in addition to the municipal wells that were investigated in this study. Since groundwater is affected by seawater in some areas [17], it is necessary to collect groundwater samples for agricultural wells to evaluate their suitability for agricultural production.

**Author Contributions:** Conceptualization, methodology, supervision, K.N.; investigation, K.N. and Z.-Q.Y.; formal analysis, data curation, H.A., K.N. and Z.-Q.Y.; writing—original draft preparation, K.N. and H.A.; writing—review and editing, K.N. and R.B.; funding acquisition, K.N. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by JSPS KAKENHI Grant Number JP20K12209.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors thank the Water Works Bureau of Minami-Shimabara City for support of the water sampling.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

