Controlling factors and driving mechanisms of nitrate contamination in groundwater system of Bandung Basin, Indonesia, deduced by combined use of stable isotope ratios, CFC age dating, and

(1)

Controlling factors and driving mechanisms of nitrate contamination in groundwater system of Bandung Basin, Indonesia, deduced by combined use of stable isotope ratios, CFC age dating, and

socioeconomic parameters

Ahmad Tau ﬁ q

^a^,^b^,^c^,^*

, Agus J. Effendi

^d

, Irwan Iskandar

^e

, Takahiro Hosono

^b^,^f

, Lambok M. Hutasoit

^a

aFaculty of Earth Science and Technology, Bandung Institute of Technology, Ganesha 10, Bandung, 40132, Indonesia

bGraduate School of Science and Technology, Kumamoto University, 2-39-1 Kurokami, Kumomoto, 860-8555, Japan

cResearch Center for Water Resources, Ministry of Public Work and Housing, Juanda 193, Bandung, 40135, Indonesia

dFaculty of Civil and Environmental Engineering, Bandung Institute of Technology, Ganesha 10, Bandung, 40132, Indonesia

eFaculty of Mining and Petroleum Engineering, Bandung Institute of Technology, Ganesha 10, Bandung, 40132, Indonesia

fPriority Organization for Innovation and Excellence, Kumamoto University, 2-39-1 Kurokami, Kumomoto, 860-8555, Japan

a r t i c l e i n f o

Article history:

Received 27 December 2017 Received in revised form 6 August 2018

Accepted 17 October 2018 Available online 23 October 2018

Keywords:

Groundwater Nitrate contamination CFC age dating

Principal component analysis Developing region Bandung Basin

a b s t r a c t

Number of populations, industry, and economic activities in Indonesia are growing rapidly and these impacts on natural environments raise awareness about water quality issue over the country. Bandung Basin, one of the most growing rapidity urban areas in Indonesia, was assessed for NO3contamination in groundwater systems, and its controlling factors and driving mechanisms were investigated with the aim to demonstrate novelty on the use of combination of parameters of stable isotope ratios in nitrate (d¹⁵^N andd¹⁸^{O in NO}3), groundwater age (using CFC-12 age tracer), and socioeconomic parameters (land-use, population, and economic database). Groundwater NO3concentrations at present time did not exceed HWO limit for all the analyzed samples (3.00 mg/L in average with maximum value of 20.69 mg/L, n¼102). Dual stable isotopic analysis together with CFC-12 groundwater age determination suggest that anthropogenic activities are the major causes for increasing NO3concentrations in groundwater. Those activities under respective land-use are industrial and domestic wastes for urban areas and chemical fertilizers for paddy and plantations areas. In general shallow unconfined aquifer is more vulnerable to NO3contamination compared with deep confined aquifer because denitrification partly occurs in deep anoxic aquifer and this led attenuation of NO3pollution in groundwaterflowing. However, it seems likely at groundwater depression cones in urban areas that more concentrated waters are transported from shallow aquifer into deep aquifer system through downward vertical fluxes due to excessive pumping. Principal component analysis (PCA) on NO3concentrations with socioeconomic parameters indicated that industrial and population growths are the main factors related to groundwater NO3 contamination. This result corresponds to CFC-age dating which shows younger (more recently recharged) groundwaters as being more contaminated than older ones do. Our study implies that NO3 contamination in this area may become more severe in future with a lack of necessary controls and treatment for human-induced nitrogen sources. Proposed approach is useful to understand how the NO3 contaminant behaves in large basin aquifer system under urban environments and might be applicable in other developing regions too because increasing populations may be associated with increasing nitrogen loadings.

1. Introduction

Increase of dissolved nitrate (NO3) in aquatic system is now a

*Corresponding author. Faculty of Earth Science and Technology, Bandung Institute of Technology, Ganesha 10, Bandung, 40132, Indonesia.

E-mail addresses:ahmad.tauﬁ[email protected],[email protected](A. Tauﬁq).

Contents lists available atScienceDirect

Water Research

j o u r n a l h o m e p a g e : w w w . e l s e v ie r . c o m / l o c a t e / w a t r e s

(2)

widespread concern and a particularly important ongoing environmental problem (Foster and Chlton, 2003;Rivett et al., 2008). It has been shown in several studies that human activities (such as nitrogen fertilizer and animal manure overuse, domestic and industrial wastewater discharges, and activities that cause increased precipitation of nitrogen compounds) are the primary causes of increased NO3concentrations in groundwater (e.g.Kendall, 1998;

Otero et al., 2009;Xue et al., 2009;Kaushal et al., 2011;Gu et al., 2012;Zhang et al., 2015). The NO3contamination of groundwater poses risks to human health (Steffen et al., 2015) and can also cause many ecological and environmental problems (e.g.Rockstrom et al., 2009;Li et al., 2010,2013). In this regard, the World Health Orga- nization has set an upper NO3concentration limit of 50 mg/L for drinking water (WHO, 2017), but limit of 10 mg/L is used in Indonesia where this study was performed (Permenkes, 2017).

Comprehensive studies on determination of NO3sources and pathways in subsurface environments have initially been performed in Europe and North America (referred to as Western countries) using multiple isotope tracers (d¹⁵^NeNO3- with other isotope ratios of the elements such asd¹¹B,d¹⁸OeNO3-,d¹³C-DIC, d³⁴^SeSO₄^2-, and d¹⁸^OeSO₄^2-) (B€ottcher et al., 1990; B€ohlke and Denver, 1995; Wassenaar, 1995; Aravena and Robertson, 1998;

Fukada et al., 2003; Widory et al., 2004,2005; Singleton et al., 2007; Otero et al., 2009) and successfully applied by followed studies in many part of the worlds including Asian region (Liu et al., 2006;Umezawa et al., 2008;Hosono et al., 2009a, 2010,2011a, 2011b,2013, 2014, and references cited previously). It has already argued in the previous review on groundwater NO3pollution status among Asian megacities (Umezawa et al., 2008;Hosono et al., 2011c) that presence of the nitrogen form and its attenuation processes depend particularly on in situ hydrogeological setting and controlled by redox condition within aquifer systems, not due to the differences in developing stage among the cities. However, the previous studies have not yet sought deeply the possibility of pollution status transition (Hara, 2006;Taniguchi et al., 2008) that may be able to identiﬁed within particular study site with a set of key information for comparison (groundwater age and socioeconomic archives) to achieve this attempt.

Nowadays, the population in Indonesia is growing signiﬁcantly and it will increase progressively in the next quarter century. While half the population was already living in urban areas in 2015, this is expected to reach two thirds by 2035. By 2035, 90 per cent of those living in Java island, will be urban dwellers. This massive urban population will concentrate in the twin mega-urban regions:

Jakarta and Bandung (Jones, 2010). It is therefore that the Bandung basin is one of ideal study target area to assess how such rapid urbanization effects on subsurface environments.

In order to reconstruct historical trend of the groundwater NO3

contamination groundwater, residence time (determining from CFCs and FS₆age tracers,Busenberg and Plummer, 1992,2000) can be used as a time marker together with NO3concentration of the same water, hypothesized that contaminants were infiltered in recharge time and transported along the groundwater flows in aquifer systems. In fact, this concept has successfully been applied in small-scale catchment (e.g.Gooddy et al., 2006) and to other target material transportations (Ako et al., 2013;Ide et al., 2018), although much applications have not seen for regional groundwaterflow systems. Moreover, age-constraint groundwater can be used to further discuss a tendency and driving force of contamination in comparison with socioeconomic parameters available.

Motivated from current observations on increase in nitrogen loadings (Singh et al., 1995;Kumar, 2013) associated with population increases in large Asian cities, it is worth to test if proposed approach will be applicable to our study area.

In Indonesia, there are several areas where groundwater NO3

concentrations exceed the World Health Organization limit (Umezawa et al., 2008;Kagabu et al., 2010;Hosono et al., 2009a, 2009b,2009c,2011a,b,c). On the shallow groundwater in Jakarta, groundwater samples collected near dry agricultural ﬁelds area displayed NO₃contamination that is sourced from the agricultural activities and were more largely attributed to wastewaters from domestic leakage near city area (Umezawa et al., 2008; Kagabu et al., 2010; Hosono et al., 2011a,b,c). In Jogjakarta, it was reported that the NO3 concentrations in groundwater range from 0.09 to 74.80 mg/L (n¼43) and identiﬁed the highest anomaly of the presence of fecal coli bacteria in shallow groundwater under the urban areas (Souvannachith et al., 2017). However, there are no comprehensive publications reporting groundwater NO3contam- ination in Bandung Basin.

This paper is the ﬁrst comprehensive study reporting NO3

contamination of groundwater in Bandung Basin that treads for both shallow (unconfined) and deep (confined) aquifers. First, the current NO₃contamination status will be assessed and compared with land-use type to put constraint on their source. Then, the source of contaminations will be further discussed in detail with combination dataset of dual isotope ratios (d¹⁵^NeNO₃^- and d¹⁸ÔeNO3-), multiple chemical components (major ions andfield measurements) and chlorofluorocarbon-12 (CFC-12) ages. Finally, the relationships between NO₃concentrations and socioeconomic parameters will be compared to test if there would detect the relationships between NO3concentrations in groundwater and socioeconomic development parameters. The results and proposed approach could be a base for understand the ways of mitigating NO3contamination in the near future.

2. Study area

The Bandung Basin is situated in the center of West Java, Indonesia (Fig. 1). Bandung is the capital of West Java Province, which is the most densely populated province in Indonesia and the center of the Indonesian textile industry. The Bandung Basin covers about 2300 km² and hasﬁve administrative areas and a current population of more than 7 million. The basin is an intramontane basin consisting of late Tertiary and Quaternary volcanic deposits, and is surrounded by mountains up to 2400 m high. The mean annual temperature in the basin is about 23.7C, the annual precipitation is from 1500 to 2500 mm/year and the potential evapo- transpiration is 1606 mm/year (Nurliana, 2009). This basin has two seasons: the dry season for during April to August and the wet season for during September to March with a peak rainfall usually occurs in January throughout the region (Iwaco-Waseco, 1990).

The surface land of the Bandung Basin is mostly in three land- use (Fig. 1), urban including industrial activities, paddyfields, and plantation areas. The urban areas have expanded rapidly since the 1970s as the population and industrial sector have grown (Wagner and Sukrisno, 1998). Land use patterns in the Bandung Basin have been changing for decades. For instance, many of the lands for cultivations, plantation areas, and paddies have shifted to urban and industrial areas (Permana and Wijaya, 2017). In 2015, pro- portions for each land-use in total basin are: paddy fields (33.11%)>urban areas (28.34%)>plantation areas (16.16%)>forests (12.36%)>reservoirs and rivers (10.02%). The three main land uses (paddy fields, urban areas, and plantation areas) and their surrounding environments will be studied here. These land uses are, from higher to lower elevations, plantation areas (dominated by tea and coffee plantation), urban areas (residential area, office buildings, and industrial complexes), and paddy (rice-growing) fields.

From a hydrostratigraphical viewpoint, the Bandung Basin can be divided into two groundwater systems, a shallow groundwater

(3)

(4)

system and a deep groundwater system. The most recent study of the groundwaterflow system of the Bandung Basin was done by Sunarwan (2014), using descriptions of cuttings, drilling well logs, resistivity logs, major ions, and stable isotopes. Hydrostratigrapi- cally, the Bandung Basin can be divided into two main systems of aquifer: an unconfined and confined aquifer. The unconfined aquifer is composed of three facieses: sandy tuff, clayey sand, and silty clay. In this study the groundwater in this aquifer is called

‘shallow groundwater’. This water can be accessed via dug wells with a depth of 10e20 m. Meanwhile, the conﬁned aquifer is alteration layer composed of tuffaceous sandstone and tuffaceous breccia which contains‘deep groundwater’. This water is obtained from borehole wells with a depth of 80e150 m. Spring waters are distributed in mountain slope.

3. Methodology

Spring and groundwater samples were collected over Bandung Basin (Fig. 1) during the dry season from May to July in 2015 under the assumption that samples are least affected by dilution

processes from precipitations and would sustain highest solute concentrations during a year. In total of 10 of spring waters were collected from the mountainous foots where plantation areas widely distribute (we obtained these samples as hypothesized these waters are affected by agricultural activities) (Fig. 1a). 28 shallow groundwater samples were collected from dug wells at different land-uses (urban areas and paddyﬁelds) as to be able to put a constraint regarding possible source of nitrate contaminations (Fig. 1a). The access to obtain shallow groundwater was limited in paddyﬁelds and we have collected only few represen- tative samples from dug wells (Appendix 1). 66 deep groundwater samples were obtained both from drilled wells (mostly from pri- vate wells) and observation wells (government monitoring well) in the Bandung Basin.

Most of data forfield measurement, major ion concentrations, CFC-12 concentrations, and stable isotope ratios of nitrogen (d¹⁵^NeNO3-) and oxygen (d¹⁸ÔeNO3-) in NO3were supplied partly in previous reported papers with analytical procedures (Taufiq et al., 2018a,2018b); summary of results and a set of complete data are provided in Table 1 and Appendix 1, respectively. Briefly, field Table 1

Statistic of dissolved nitrogen concentrations in each groundwater type.

a) Shallow groundwater

Category Concentrations for nitrite and ammonium Concentrations and stable isotopes ratios for nitrate ion

NO2 NH4þ NO3 d¹⁸OeNO3- d¹⁵NeNO3-

mg/L mmol mg/L mmol mg/L mmol (⁰/00) (⁰/00)

1.1 Plantation areas

Mean 0.03 0.6 1.85 102.3 3.39 54.6 3.05 5.68

Standard deviation 0.02 0.5 1.50 82.9 1.67 26.9 4.30 3.31

Maximum 0.06 1.3 4.70 260.6 5.99 96.6 15.07 14.44

Minimum 0.00 0.1 0.22 12.2 0.36 5.8 0.87 2.09

Number (N) 10 10 10 10

1.2 Urban areas

Mean 0.01 0.3 1.04 57.5 5.04 81.3 6.55 14.40

Maximum 0.07 1.5 3.60 199.6 20.69 333.7 33.10 39.14

Minimum 0.00 0.1 0.14 7.8 0.97 15.6 12.18 1.23

Number (N) 19 19 24 22

1.3 Paddyﬁelds

Mean 0.01 0.3 1.52 84.4 3.36 54.1 12.92 22.34

Maximum 0.05 1.0 2.35 130.3 5.75 92.7 18.16 47.01

Minimum 0.00 0.1 0.37 20.5 1.25 20.2 3.12 9.26

Number (N) 4 4 4 4

b) Deep groundwater

Category Concentrations for nitrite and ammonium Concentrations and stable isotopes ratios for nitrate ion

NO2 NH4þ NO3 d¹⁸ONO3 d¹⁵NNO3

mg/L mmol mg/L mmol mg/L mmol (⁰/00) (⁰/00)

2.1 Older groundwater

Mean 0.01 0.2 1.18 65.2 2.42 39.1 20.17 19.20

Maximum 0.05 1.1 2.75 152.5 6.59 106.3 37.97 38.28

Minimum 0.00 0.1 0.10 5.5 0.10 1.6 8.46 10.52

Number (N) 18 18 22 18

2.2 Younger groundwater

Mean 0.01 0.2 0.70 38.8 1.69 27.3 11.41 13.02

Maximum 0.03 0.7 2.77 153.6 5.65 91.1 34.64 40.84

Minimum 0.00 0.1 0.10 5.5 0.13 2.1 8.19 2.69

Number (N) 30 30 36 29

Fig. 1.Land-use map of the Bandung Basin (Bakosurtanal, 2009) showing groundwaterflow and the sampling points for (a) shallow and (b) deep groundwater. Groundwaters in shallow unconfined aquifer generallyflow from the peripheral to the center of the basin, while the groundwaters in deep confined aquifer isflowing towards three depression zones in the industrial area (Taufiq et al., 2018a). Hydrogeological cross-section in CeCʹline are shown inFig. 8. CMHI, DYHK and RCK inFig. 1b represent three depression areas, Cimahi, Dayeuhkolot, and Rancaekek, respectively. The numbers in each sampling location represent the sample ID. The identification of the well types (dug well, spring, drilled well, and observation well) are explained in the text (3. Methodology).

(5)

measurements were monitored continually using a portable meter (D-54, Horiba, Japan), and the concentrations of the major ions (Na^þ, K^þ, Ca^2þ, Mg^2þ, SO42, and Cl) and nitrogenous nutrients (NO3, NO2, and NH4þ) were determined using an ion chromatograph (Compact IC 761, Metrohm, Switzerland), respectively.

Chloroﬂuorocarbons (CFCs) concentrations were determined using a closed system purge-and-trap gas chromatograph (GC-2014;

Shimadzu, Japan), but in this study only the CFC-12 concentration was used as a marker of young groundwater to calculate apparent groundwater residence time (Appendix 1) because CFC-12 is relatively stable in the subsurface environment than other compounds (Plummer et al., 1998). Thed¹⁵^NeNO3- andd¹⁸^OeNO3- were determined applying the denitriﬁer method presented bySigman et al.

(2001) and Casciotti et al. (2002)followed laboratory routine in the Kumamoto University (Hosono et al., 2013,2014). The analytical precisions for d¹⁵^NeNO3- andd¹⁸^OeNO3- were better than ±0.2‰ and±0.3‰, respectively.

In this study, the NO₃ concentrations in spring and shallow ground-waters will be assessed taking the type of land-uses into accounts as has been previously demonstrated for its effectiveness (Lockhart et al., 2013;Qin et al., 2013). In contrast, direct contributions of pollutants from just above surface lands seem to be unlikely for deep groundwater with longer lateral flow paths considering regional groundwaterflow dynamics except in an area of depression zones (Taufiq et al., 2018a,2018b). In this regard, the deep groundwater could be classified mainly into two types:‘older groundwater’(the groundwater with CFC-12 concentrations un- detectable and recharge periods estimated as to be before 1950) and ‘younger groundwater’ (the groundwater with CFC-12 concentrations detectable and recharge periods estimated as to be after 1950) (Appendix 1).

Socioeconomic data for the basin for the period from 1953 to 1991 were compiled and used in the study. The annual data of the Bandung Basin were taken from Book of West Java Province (BPS, 2016) and were compiled inAppendix 2, such as permanent population (PP) in number of people, population density (PD) in people/km², average monthly gross domestic product (GDP) in Indonesian Rupiah (IDR), industrial growth (IG) in number of industries, built-up urban areas (BUA) in ha, green areas of agriculture, plantation and forests (GA) in ha, and amount of use of chemical fertilizers (CF) in ton. This period is somehow matching to groundwater age determined by CFC-12 concentration (from 1953 to 1991;Tauﬁq et al., 2018a, seeAppendix 1).

An understanding of the mechanisms controlling the current NO3groundwater contamination status was gained by performing several statistical analyses using SPSS version 23.0 software (SPSS, IL, USA). Differences in NO₃concentrations in groundwater from areas with different land uses were evaluated using the nonpara- metric KruskaleWallis test and the ManneWhitney U test. The degrees to which groundwater NO₃concentrations and socioeconomic parameters were associated were determined by performing regression analyses. The main mechanisms driving groundwater NO₃ concentrations were identiﬁed by performing principle component analysis (PCA).

4. Results

4.1. Concentrations for dissolved inorganic nitrogen

Among three different forms of dissolved inorganic nitrogen (NO3, NO2, and NH4þ), NO3and NH4þconsist of the major forms of nitrogen compounds in the studied groundwaters but NO2is less important as its concentration is very low ranging from ~1.52m^mol to ~1.09mmol for shallow and deep groundwater, respectively (Table 1andAppendix 1). Concentrations of NO3and NH4þvary

~333.71mmol and ~260.53mmol for shallow groundwater and

~106.29mmol and ~152.44mmol for deep groundwater, respectively. For shallow groundwater, NO3 concentrations tend to increase with decreasing of NH4þ concentrations and vice versa (Appendix 1), suggesting input sources of nitrogen partly stay as being ammonium ion form but not be nitrified probably due to the presence of anoxic conditions similar phenomenon as identified in other Asian aquifer systems (Umezawa et al., 2008). Such tendency is less obvious for deep groundwater (Appendix 1) probably because of the mixing of shallow and deep groundwater (Taufiq et al., 2018a,2018b) as partly discussed later.

Concentrations of NH4þ are comparable to those of NO3 and undoubtfully important to discuss for their sources and fates.

However, this paper focus on NO₃concentrations since nitrogen in nitrate form is threat for human health. There are more groundwaters in shallow aquifer with higher NO3concentrations than the deep groundwater (Figs. 2 and 3). The NO3concentrations are from detection limit to 20.69 mg/L for all studied groundwaters and do not exceed the World Health Organization limit of 50 mg/L in any sample but exceed the Indonesian standard (10 mg/L; S3, S6, S30) in some samples. Although the current NO₃concentrations seem to be acceptable, the concentrations are likely to increase with time (Fig. 4) and threaten drinking water supplies in the future. This will be discussed in section5.2.

The effects of human activities on groundwater contamination were investigated further by comparing the NO3concentrations in the shallow groundwater by different land uses. The mean of NO₃ concentrations in different land uses were different but the differences were not signiﬁcant (P>0.05), except in urban area.

Higher mean NO3concentration (5.04±4.68 mg/L) was observed for the urban areas compared for plantation areas (3.39±1.94 mg/

L) and paddyﬁelds (3.36±1.94 mg/L). The mean NO3concentra- tions for older and younger groundwater were 2.42±1.92 mg/L and 1.69±1.49 mg/L, respectively. These concentrations are lower than shallow groundwater.

4.2. Stable isotope ratios of nitrate

The d¹⁵^NeNO3- and d¹⁸^OeNO3- were determined to allow possible sources of dissolved nitrate (NO₃). Analytical results are summarized inTable 1(and all data are supplied inAppendix 1), and compositional relationships between d¹⁵^NeNO3- and NO3

concentrations and dual isotope ratios are plotted inFigs. 5 and 6 for shallow and deep groundwater, respectively. Overall, either shallow or deep groundwater show wide compositional ranges (Figs. 5b and 6b): from5‰^{to 50}‰^ford¹⁵NeNO3-and from10‰ to 40‰ ^for d¹⁸^OeNO₃^-, suggesting variabilities in their source characteristics and occurrences of microbiological reactions. Ma- jority of samples (both shallow and deep groundwaters) show dissolved oxygen concentrations more than 2.0 mg/L (Fig. 2c and d) and only 3 samples showed the value less than 2.0 mg/L. It is generally accepted that nitrate microbial reduction could occur under anoxic condition typically dissolved oxygen concentrations

<2.0 mg/L (Hosono et al., 2013 and references therein). In fact, decline of NO3 concentrations seems not to be associated with increasing ind¹⁵^NeNO₃^- (Fig. 5a and b), while NO₃concentrations rather tend to increase with increasing ind¹⁵^NeNO₃^-for the samples with its value<20‰, implying the samples may inherit dual isotopic compositions as indicating their source characteristics except ones have elevatedd¹⁵^NeNO₃^-values>20‰^.

For shallow groundwater, the samples from plantation areas tend to have lower d¹⁵^NeNO3- and d¹⁸^OeNO3- compositions compared to those from urban areas and paddy ﬁeld (Fig. 5b).

Source of NO3-N in groundwaters could be estimated from their isotopic compositions together with land-use constraint: chemical

(6)

Fig. 2.Distributions of the (a,b) NO3concentrations, (c,d) dissolved oxygen concentrations, and (e,f)d¹⁵NeNO3-values for groundwater samples collected from the (a,c,e) shallow and (b,d,f) deep groundwaters.

(7)

fertilizers for plantation areas, septic waste for urban areas, and manure and/or septic waste for paddyﬁelds in rural area (details are discussed in later part). For deep groundwater, majority of younger groundwater (recharge age after 1950) shows isotopic compositions (Fig. 5b) those indicate a source mixture among chemical fertilizers, septic waste, and manures, implying multiple pathways of nitrate transportations. On the contrary, older groundwater tend to show more elevatedd¹⁵^NeNO3-andd¹⁸^OeNO3-

compositions than younger groundwater (Fig. 5b), partly reflecting the occurrence of denitrification along groundwater flows with longer paths (for the samples withd¹⁵^NeNO3-values>20‰^{). Other} samples show distinct isotopic compositions from 10‰^{to 20}‰^for both isotope ratios (d¹⁵^NeNO3-andd¹⁸ÔeNO3-), perhaps indicating nitrogen source from septic waste (or application of manure in agriculture) but this is not clear because there are no favorite conditions that may support this scenario from other aspects such as sociological background.

4.3. Nitrate concentrations and socioeconomic development As explained earlier deep groundwater could be divided into two, younger groundwater (recharged after the year 1950) and older groundwater (recharged before the year 1950). Therefore,

NO3 contaminations of age-dated younger groundwater can compare with time serious of the socioeconomic parameters (Appendix 2). Similarly, the CFC-12 concentrations in the shallow groundwater samples covered a wide range (from 12.2 to 540.8 pptv) and recharge ages were estimated for these groundwaters that range from 1953 to 1991 (Appendix 1). The NO₃contaminations of these shallow groundwaters can also be used for comparison with socioeconomic parameters (Appendix 2). By obtaining time series of both groundwater recharge age and socioeconomic parameters, groundwater NO3contaminations at known recharge time can be compared with socioeconomic data of the same year (Fig. 7). Then, the relationships between the groundwater NO₃ concentrations and annual data of several socioeconomic parameters were analyzed by performing regression analyses. The increases of NO₃concentrations in both shallow and deep younger groundwaters are weakly correlated (P<0.01) with increase of permanent population in number of people (PP), population density in people/km²(PD), average monthly gross domestic product in IDR (GDP), industrial growth in number of industries (IG), built-up urban areas in ha (BUA), and amount of use of chemical fertilizers in ton (CF), but with decrease of green areas of agriculture, plantation and forests in ha (GA).

Fig. 3.Comparison of NO3concentrations between (a) shallow and (b) deep groundwater from the Bandung Basin.

Fig. 4.Increasing tendency of NO3concentrations with groundwater residence time determined by CFC-12 concentration (Tauﬁq et al., 2018a, seeAppendix 1).

(8)

5. Discussion

At present, NO3contamination in groundwater does not appear to threaten human health in the study area. However, it is still essential to gain an understanding of the factors that control and the mechanisms that drive the NO3concentrations in groundwater to allow measures to be taken to prevent contamination in the area becoming more severe.

5.1. Factors controlling nitrate contamination

5.1.1. Source of nitrate contamination for shallow unconﬁned groundwater

It has been previously investigated that nitrogen loads from the surface of particular land-uses imprint their sources in groundwaters (Lockhart et al., 2013;Qin et al., 2013;Zhang et al., 2015).

Thus, descriptions for land-uses are important to study the NO3

contamination in groundwater system. The controlling factors of NO₃contaminations in groundwaters were identiﬁed by assessing isotopic ﬁngerprinting together with land-uses analysis because NO3 enters subsurface system through several pathways that

depend strongly on the type of land-uses.

Elevated NO3concentrations in groundwaters (spring waters, triangle symbol inFig. 2a) from plantation areas would be transported from vegetable and livestock farms, which are distributed around the periphery of the basin. These concentrations tend to decrease with increase ind¹⁵NeNO3-(Fig. 5a) and this tendency may be explained by the occurrence of denitriﬁcation (Kendall, 1998).

However, all groundwater samples from the plantation areas showed dissolved oxygen concentrations >6 mg/L (Fig. 2b). It is commonly assumed that denitrification will not occur at such dissolved oxygen concentrations (Spalding and Exner, 1993;McCallum et al., 2008;Hosono et al., 2013), and thus, occurrence of denitrification is unlikely for nitrate dissolved in these waters. From isotopic fingerprinting evidence (Fig. 5b), it can be concluded that NO3-N in groundwaters from plantation areas is attributed to applications of chemical fertilizers (B€ohlke and Denver, 1995).

It is commonly understood that leakage of sewage waters from sewage pits or pipes are important factors affecting groundwater recharge and nutrient contamination in urban areas (Lerner, 2002;

Wakida and Lerner, 2005). This may contribute to shallow groundwater in urban areas of the Bandung Basin as well as partly Fig. 5.(a)d¹⁵NeNO3-vs. [NO3] and (b)d¹⁵NeNO3-vs.d¹⁸OeNO3-diagram for the samples from shallow aquifer. Denitriﬁcation trend (straight line) and compositionalﬁelds of possible source materials (dotted line) are afterKendall (1998).

(9)

suggested by previous report (Nurliana, 2009). Most parts of the urban areas are underlain by poorly porousfine clay in old lake deposits and overlain by asphalt. This may lead subsurface condition to be anoxic, favoring denitrification (relative to the subsurface in the volcanic fan). This scenario could partly be applied for the groundwaters with high d¹⁵^NeNO3- >20‰but low to moderate NO₃concentrations (Fig. 5a). However, the effect of denitrification looks sporadic because there is no clear negative trend between the concentrations and d¹⁵^NeNO3- values (Fig. 5a) and most of the groundwaters are characterized by DO>2.0 mg/L. It is therefore reasonable from isotopic comparison (Fig. 5b) to think that NO₃in shallow groundwater in urban areas is the dominant nutrient released from domestic wastewater and that denitrification may remove NO₃but just too some extent.

It is interest to further describe that the groundwaters where near to sewage treatment plants in industrial complexes (such as at points S3 and S6 in the CMHI area and S30 in the RCK area, see Fig. 1) contain relatively high NO3(>10 mg/L). Treated wastewaters at these points may directly inﬁltrate the shallow groundwater from the leakage pathways. It has recently been reported that

inappropriate on-site sanitation is becoming an issue for an environmental degradation at all big cities of Indonesia (Saﬁtri et al., 2010). For instance, fecal coli bacteria are detected in shallow groundwater in Jogjakarta (Souvannachith et al., 2017). It seems to be reasonable to conclude that the NO3in groundwater in this area is partly supplied by wastewater derived from sewage and partly by industrial wastewater.

The concentrations of nitrogen compounds at the atmosphere in Southeast Asia, including Indonesia, have increased because of human activities that could supply NO3 in rainwater that could enter the shallow groundwater in urban areas (Boring et al., 1988;

Streets et al., 2001,2003). Some of the water samples from the urban areas (e.g., at site S12, see Fig. 5) had particularly high d¹⁸^OeNO₃^- value (33.10‰), suggesting the contribution of atmospheric NO3. The contributions of atmospheric deposition may partly elevate NO3concentrations in surface and subsurface water (Campbell et al., 2002) in limited areas in the Bandung Basin.

The paddyfields are widely distributed in the center of the basin (Fig. 2). The DO concentrations in this area are no less than 2 mg/L (Fig. 2b), at which condition denitrification will not commonly Fig. 6.(a)d¹⁵NeNO3-vs. [NO3] and (b)d¹⁵NeNO3-vs.d¹⁸OeNO3-diagram for the samples from deep aquifer. Denitrification trend (straight line) and compositionalfields of possible source materials (dotted line) are afterKendall (1998).

(10)

occur (Spalding and Exner, 1993;Cey et al., 1999;McCallum et al., 2008;Hosono et al., 2013). However, very highd¹⁵^NeNO3- value observed for one sample (site S24, seeFig. 5) strongly suggest the occurrence of denitriﬁcation. In fact, the paddyﬁelds are in the

lowlands in the basin center, and are often water-saturated, so the groundwater will be in a reducing environment that will favor denitrification. The denitrification may, therefore, be able to occur in the paddyfields, but the number of the samples (4 samples, see Fig. 7.Relationships between the NO3contaminations and socioeconomic parameters of permanent population in number of people (PP), population density in people/km²(PD), average monthly gross domestic product in IDR (GDP), industrial growth in number of industries (IG), built-up urban areas in ha (BUA), green areas of agriculture, plantation and forests in ha (GA), and amount of use of chemical fertilizers in ton (CF) for the shallow and deep younger groundwaters from the Bandung Basin.

(11)

Fig. 5) is too small expanding this aspect as for general conclusion.

As mentioned in Result section, rest of 3 samples show isoto- pically distinct as these sources corresponding to manures and septic wastes but not as originated of chemical fertilizers (Fig. 5b).

Therefore, it could be possible that nitrogen was derived from wastewater infiltering from wastewater pits in rural houses (Umezawa et al., 2008;Hosono et al., 2011b). Alternatively, an effect of application of manure to paddy should be considered; in fact, paddyfields in Indonesia using organic management system (i.e., only manure being applied) cover only 0.09% of whole paddyfields in 2003 (Takada et al., 2004;Umezawa et al., 2008) and synthetic fertilizers are major applied fertilizers. Thus, the possibility of NO3

contaminations due to the application of manure is unlikely and leaking of wastewaters are the most probable cause of elevated NO3 concentrations, although the effects of the use of chemical fertilizers (and subsequent denitriﬁcation) could not completely be

eliminated. Nevertheless, difference in d¹⁵^NeNO3- and d¹⁸^OeNO3-

compositions between the groundwaters from plantation areas and paddy ﬁelds may suggest the difference in their source characteristics.

5.1.2. Source of nitrate and its attenuation processes for deep conﬁned groundwater

The source of NO₃-N contaminations and their attenuation processes including deep confined aquifer system are assessed along the groundwaterflow systems in Bandung Basin (Taufiq et al., 2018a) using hydrogeological cross-section CeC’ (Fig. 8). In recharge area (described as zone A inFig. 8) the NO3concentrations in shallow groundwaters (spring waters) are somewhat elevated due to agricultural activities in plantation areas (we have discussed it in previous section). The NO3concentrations decreased in both shallow and deep groundwaters without increase ind¹⁵^NeNO3-and Fig. 8.Changes in the nitrate concentrations, dissolved oxygen concentrations,d¹⁵NeNO3-values, andd¹⁸OeNO3-along a groundwaterflow as shown in hydrogeologic cross section C-Cʹ(seeFig. 1for its location). The sampling sites are the same as inFig. 2. The horizontal axis indicates the distance from recharge zone in cross section C-Cʹ. Some indication lines of attenuation processes of NO3were expressed in this section which explained in chapter 5.1.2.

(12)

d¹⁸OeNO3-values at the entrance of basin aquifer system (zone B in Fig. 8), suggesting it might be caused of dispersion and/or dilution by inﬁltrating water in the recharge area, but not through denitriﬁcation.

The NO₃concentrations then increased as groundwaterflows under CMHI groundwater depression area (Taufiq et al., 2018a,b) for both deep and shallow aquifers with simultaneous increase in d¹⁵^NeNO₃^- andd¹⁸ÔeNO₃^- values especially for deep groundwater (Fig. 8). Isotopic compositions for this elevated d¹⁵^NeNO3- and d¹⁸ÔeNO3-values are corresponding to those for manure and septic wastes (Fig. 6). Redox condition under CMHI area is not anoxic as denitrification favorably occurs (Fig. 8). Previous studies assessing groundwater mixing between shallow and deep groundwater revealed that shallow waters with contaminant materials are increasingly mixed into deep unpolluted groundwater with time, enhancing groundwater contaminations due to excess pumping by the rapid urbanization (Taufiq et al., 2018a,2018b). From these observations withflow-constrained by previous studies it could be reasonably concludes that nitrate concentrations in deep groundwater increased under CMHI area by increasing mixing of surface contaminants such as urban and septic wastes.

The NO3concentration in the deep groundwater sustains in the same level after the groundwaterflows out of CMHI depression area but the redox condition turned to somewhat anoxic condition as documented previously (Fig. 8; Taufiq et al., 2018a,b). The d¹⁵^NeNO3-andd¹⁸ÔeNO3-values for the deep groundwater tended to increase over 20‰ in the end of topographical slope (in the entrance of basin plain) (Fig. 8), indicating the occurrence of denitrification particularly at such specific area. This tendency can be confirmed in dual isotopic composition diagram (Fig. 6b) as some of deep groundwater samples are plotted along denitrification trend.

In the end ofﬂow the deep groundwaters come under another depression area, DHYK area (Fig. 8). Interestingly, the redox condition turned to more oxic even under stagnantﬂow regime as anoxic condition might apparently be presented. The d¹⁵^NeNO3-

andd¹⁸^OeNO₃^-values under DHYK area showed similar compositional range found in groundwater under CMHI depression area (Fig. 8). In fact, the occurrence of similar groundwater mixing as stated for CMHI area has been argued in the previous studies for DHYK depression area too (Tauﬁq et al., 2018a,2018b). Thus, similar scenario (enhancing NO3contaminations from surface contaminants of urban and septic wastes by groundwater mixing due to excess pumping) could be applied for explanation of observed NO₃ contaminations under this depression area.

5.2. Mechanisms driving nitrate contamination of groundwater 5.2.1. Nitrate concentrations versus regional socioeconomic development

It is important to identify the mechanisms driving NO3

contamination to help develop ways of regulating groundwater NO₃contamination. The relationships between the NO₃concentrations and socioeconomic development parameters and groundwater age were assessed. Here we will discuss more deeply for previously mentioned results, positive correlation (P<0.01) between NO₃concentrations and PD, PP, BUA, GDP, IG, and CF, and negative correlation (P<0.01) between same concentrations and GA (Fig. 7).

Higher urbanization levels (indicated by PP and PD) appear to be causing large amounts of domestic wastewater into the aquatic environment, resulting in increase in nitrogen concentrations in groundwater.Bilsborough and Mann (2006) stated that the per capita supply of nitrogen in human excreta will be stable because the basic protein requirement of humans is stable. It was therefore

expected that NO₃concentration in groundwater would strongly correlate with the PD if the nitrogen removal reactions could be neglected such as denitriﬁcation. The positive correlations (P<0.01) were found between the NO3concentration in groundwaters and the BUA, but the negative correlations were found between the NO3concentration and the GA for both shallow and deep aquifers. The BUA and GA are indices that reﬂect the degree of urbanization. The larger the built-up area (BUA) or the less of the green area (GA) link to the larger population. Therefore, it could be accommodated that the more potential for these domestic wastewater productions enhances NO₃contamination in groundwater.

A similar positive correlations (P<0.01) were found between the NO3concentration in groundwater and the GDP and IG too. This is quite reasonable because the economy of the Bandung Basin area is somehow reliant on industrial activity. Relatively large amounts of industrial wastewater may be released from industrial plants, and this may cause the NO3concentrations in groundwater to be relatively high. It has been argued that shallow groundwater is heavily contaminated close to industrial complexes and argued that huge amounts of untreated industrial wastewater will have directly drained into nearby rivers and may have infiltrated into the groundwater (Rosadi, 2004). The argument was also supported by rivers close to industrial complexes being strongly contaminated with industrial wastewater (Tohir, 2004) that can infiltrate shallow groundwater because of the influent relationships between rivers and shallow aquifers (Lubis and Puradimaja, 2005). These observations correspond to our results showing the NO₃concentrations in the groundwater were higher in the industrial areas than else- where in the study area.

The NO3concentration in groundwater are correlated (P<0.01) with the CF as well, indicating that increasing in leaching of nitrogen from chemical fertilizers in plantation areas as a factor of the amounts of applicated fertilizers. Agricultural activities in plantation areas (and possibly in paddy ﬁelds) may act as non-point sources of NO3to groundwater, in addition to domestic and industrial wastewater point sources.

Overall, it can be summarized that all relationships for both shallow and deep younger groundwater somehow followed the same trends. This may support the scenario that the deep groundwater is affected by the shallow groundwater because of verticalfluxes from the shallow to the deep groundwater (Taufiq et al., 2018a,2018b). Observed correlations were not high probably because of the several complex reasons such as lacking of the considerations for denitrification effects and influence of other nitrogen compounds (especially for ammonium). Nevertheless, our trial can shed a light on the effective use of socioeconomic parameters combined with hydrochemistry data of the regional groundwater.

5.2.2. Driving force for nitrate contamination of the groundwater The main factors driving the groundwater NO3concentrations were identiﬁed using PCA to further detect key independent factors from seven variables parameters. The key independent factor is a set of values that called principle components (PC) converted from a set of observation in orthogonal transformation of statistical procedure. They are wo principal components (PC1 and PC2) with eigen values>1 explained 70.19% of the total variance for the shallow groundwater data and explained 75.38% for the deep groundwater data as shown inTable 2. The PC1 had strong positive correlation (>0.80) between IG, GDP, PD, PP, and BUA, while weak positive (<0.30) and strong negative (<-0.80) correlations between CF and GA, respectively. The IG, GDP, PD, PP, BUA and GA had strong correlations indicating that primary industrial and population growth was the primary driving force. The IG and GDP indices reﬂect industrial development, meanwhile the PD, PP, BUA, and GA

(13)

indices reﬂect urbanization or population growth. PC1 therefore indicated that anthropogenic activities are the main driving forces for groundwater NO3contamination in the study area. This result is similar with the PCA that was examined byZhang et al. (2015)in southern China that the main driving force is the population growth.

PC2 explained 19.01% of the total variance for the shallow groundwater and 14.18% for the deep groundwater show strong positive (>0.80) correlation between only the CF. The CF reﬂects chemical fertilizer use. PC2 therefore indicated that is the secondary driving force of groundwater NO3contamination. Notably, the PCA results for shallow and deep younger groundwater were similar, meaning that the same driving forces are very important for both shallow and deep groundwater. Addressing these driving forces should be important to develop means controlling the current NO3contamination in the basin.

6. Conclusions

With the aim to understand the status and primal causes of NO3

contamination of the groundwater in the Bandung Basin isotopic fingerprinting (d¹⁵^NeNO3- and d¹⁸ÔeNO3-) method was applied with combined dataset of groundwater residence time and socioeconomic parameter archive. The dissolved NO₃concentrations in the Bandung Basin do not exceed the maximum concentrations specified by the World Health Organization for both shallow unconfined and deep confined aquifers. Our results, however, demonstrate that these concentrations are increasing with time due primary to wastewater leakage in urban areas closely linked with populations and industrial activities increase, and secondary, to increasing application of chemical fertilizers in rural area. These contaminants are transported from shallow aquifers toward‘un- contaminated’ deep confined aquifer triggered by vertical groundwater depression accelerated by excess pumping in the urban center. It is clear that rapid urbanization in Bandung city enhances NO3 concentrations in multiple step and this study illustrates how urban groundwater deteriorates its water quality.

We are lacking the presentation of ammonium loading in this study which should be important to further address; nevertheless, our

trial can propose the importance on the use of hydrochemistry data in combination with groundwater residence time and socioeconomic parameters. Information from our study is important as a base for protections and preservations of groundwater contaminations.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.watres.2018.10.049.

References

Ako, A.A., Shimada, J., Hosono, T., Kagabu, M., Richard, A., Nkeng, G.E., Tongwa, A.F., Ono, M., Eyong, G.E.T., Tandia, B.K., Mouncherou, O.F., 2013. Flow dynamics and age of groundwater within a humid equatorial active volcano (Mount Cameroon) deduced bydD,d¹⁸O,³H and chloroﬂuorocarbons (CFCs). J. Hydrol.

502, 156e176.

Aravena, R., Robertson, W.D., 1998. Use of multiple isotope tracers to evaluate denitriﬁcation in ground water: study of nitrate from a large-ﬂux septic system plume. Ground Water 36, 975e982.

Bakosurtanal, 2009. Land Use Map, Sub Bandung Region scale 1: 50.000.

Bilsborough, S., Mann, N., 2006. A review of issues of dietary protein intake in humans. Int. J. Sport Nutr. Exerc. Metabol. 16, 129.

Boring, L.R., Swank, W.T., Waide, J.B., Henderson, G.S., 1988. Sources, fate, and impacts of nitrogen inputs to terrestrial ecosystems: review and synthesis. Bio- geochem. J. 6, 119e159.

B€ohlke, J.K., Denver, J.M., 1995. Combined use of groundwater dating, chemical, and isotopic analyses to resolve the history and fate of nitrate contamination in two agricultural watersheds, Atlantic coastal plain, Maryland. Water Resour. Res. 31, 2319e2340.

B€ottcher, J., Strebel, O., Voerkelius, S., Schmidt, H., 1990. Using isotope fractionation of nitrate-nitrogen and nitrate-oxygen for evaluation of microbial denitriﬁca- tion in a sandy aquifer. J. Hydrol. 114, 413e424.

BPS, 2016. West Java in Figure. Biro Pusat Statistik, Indonesia (In Indonesia).

Busenberg, E., Plummer, L.N., 1992. Use of chloroﬂuorocarbons (CCl3F and CCl2F2) as hydrologic tracers and age-dating tools: the alluvium and terrace system of central Oklahoma. Water Resour. Res. 28e9, 2257e2283. https://doi.org/

10.1029/92WR01263.

Busenberg, E., Plummer, L.N., 2000. Dating young groundwater with sulfur hexa- ﬂuoride: natural and anthropogenic sources of sulfur hexaﬂuoride. Water Resour. Res. 36e10, 3011e3030.https://doi.org/10.1029/2000WR900151.

Campbell, D.H., Kendall, C., Chang, C.C.Y., Silva, S.R., Tonnessen, K.A., 2002. Pathways for nitrate release from an alpine watershed: determination usingd¹⁵N and d¹⁸O. Water Resour. Res.https://doi.org/10.1029/2001WR000294.

Casciotti, K.L., Sigman, D.M., Galanter Hastings, M., B€ohlke, J.K., Hilkert, A., 2002.

Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitriﬁer method. Anal. Chem. 74, 4905e4912.

Cey, E.E., David, L.R., Ramon, A., Gary, P., 1999. Role of the riparian zone in controlling the distribution and fate of agricultural nitrogen near a small stream in Southern Ontario. J. Contam. Hydrol. 37, 45e67.

Foster, S.S.D., Chlton, P.J., 2003. Groundwater: the process and global signiﬁcance of aquifer degradation. Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 358, 1957e1972.

Fukada, T., Hiscock, K.M., Dennis, P.F., Grischek, T., 2003. A dual isotope approach to identify denitriﬁcation in groundwater at a river-bank inﬁltration site. Water Res. 37, 3070e3078.

Gooddy, D.C., Darling, W.G., Abesser, C., Lapworth, D.J., 2006. Using chloroﬂuoro- carbons (CFCs) and sulphur hexaﬂuoride (SF6) to characterise groundwater movement and residence time in a lowland Chalk catchment. J. Hydrol. 330, 44e52.

Gu, B., Dong, X., Peng, C., Luo, W., Chang, J., Ge, Y., 2012. The long-term impact of urbanization on nitrogen patterns and dynamics in Shanghai, China. Environ.

Pollut. 171, 30e37.

Hara, K., 2006. Groundwater contamination and quality management policy in Asia.

Int. Rev. Environ. Strat. 6 (2), 291e306.

Hosono, T., Ikawa, R., Shimada, J., Nakano, T., Saito, M., Onodera, S., Lee, K.-K., Taniguchi, M., 2009a. Human impacts on groundwaterﬂow and contamination deduced by multiple isotopes in Seoul City, South Korea. Sci. Total Environ. 407, 3189e3197.

Hosono, T., Umezawa, Y., Onodera, S., Wang, C.-H., Siringan, F., Buapeng, S., Delinom, R., Nakano, T., Taniguchi, M., 2009b. Comparative study on water quality among Asian megacities based on major ion concentrations. In:

Taniguchi, M., Burnttt, W.C., Fukushima, Y., Haigh, M., Umezawa, Y. (Eds.), From Headwaters to the Ocean: Hydrological Changes and Watershed Management.

Taylor and Francis, London, UK, pp. 295e300.

Hosono, T., Delinom, R., Onodera, S., Umezawa, Y., Nakano, T., Taniguchi, M., 2009c.

Cause of groundwater contamination in Jakarta alluvium volcanic fan deduced by sulfate and strontium isotope ratios. IAHS Publ. 329, 201e206.

Hosono, T., Siringan, F., Yamanaka, T., Umezawa, Y., Onodera, S., Nakano, T., Taniguchi, M., 2010. Application of multi-isotope ratios to study the source and quality of urban groundwater in Metro Manila, Philippines. Appl. Geochem. 25, Table 2

Mechanisms driving groundwater NO3contamination from the principal component analysis results.

Parameter Principal

components

PC1 PC2

a) Shallow Groundwater

PP (permanent population in number of people) 0.86 0.11 PD (population density in people/km²) 0.83 0.12 GDP (average monthly gross domestic product in IDR) 0.90 0.22 IG (industrial growth in number of industries) 0.98 0.05

BUA (built-up urban areas in ha) 0.97 0.09

GA (green areas of agriculture, plantation and forests in ha) 0.99 0.08 CF (amount of use of chemical fertilizers in ton) 0.30 0.84

Eigenvalue 5.62 1.52

%total variance 70.19 19.01

b) Deep younger groundwater

PP (permanent population in number of people) 0.97 0.24 PD (population density in people/km²) 0.97 0.24 GDP (average monthly gross domestic product in IDR) 0.99 0.03 IG (industrial growth in number of industries) 0.99 0.09

BUA (built-up urban areas in ha) 0.96 0.24

GA (green areas of agriculture, plantation and forests in ha) 0.97 0.24 CF (amount of use of chemical fertilizers in ton) 0.23 0.71

Eigenvalue 6.03 1.14

%total variance 75.38 14.18