Science of the Total Environment 432 (2012) 216–226

Contents lists available at SciVerse ScienceDirect

Science of the Total Environment

j ourna l homepage: www.e lsev ie r .com/ locate /sc i totenv

Assessment of groundwater contamination risk using hazard quantification, amodified DRASTIC model and groundwater value, Beijing Plain, China

Junjie Wang, Jiangtao He ⁎, Honghan ChenBeijing Key Laboratory of Water Resources and Environmental Engineering, School of Water Resources and Environment, China University of Geosciences, Beijing, 100083, China

⁎ Corresponding author. Tel.: +86 10 82322080.E-mail addresses: wjj198635@163.com (J. Wang), jt

chenhh56@126.com (H. Chen).

0048-9697/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.scitotenv.2012.06.005

a b s t r a c t

a r t i c l e i n f o

Article history:Received 5 April 2012Received in revised form 4 June 2012Accepted 4 June 2012Available online xxxx

Keywords:Groundwater contamination riskHazardIntrinsic vulnerabilityGroundwater valueBeijing Plain

Groundwater contamination risk assessment is an effective tool for groundwater management. Most existingrisk assessment methods only consider the basic contamination process based upon evaluations of hazardsand aquifer vulnerability. In view of groundwater exploitation potentiality, including the value ofcontamination-threatened groundwater could provide relatively objective and targeted results to aid in de-cision making. This study describes a groundwater contamination risk assessment method that integrateshazards, intrinsic vulnerability and groundwater value. The hazard harmfulness was evaluated by quantifyingcontaminant properties and infiltrating contaminant load, the intrinsic aquifer vulnerability was evaluatedusing a modified DRASTIC model and the groundwater value was evaluated based on groundwater qualityand aquifer storage. Two groundwater contamination risk maps were produced by combining the abovefactors: a basic risk map and a value-weighted risk map. The basic risk map was produced by overlayingthe hazard map and the intrinsic vulnerability map. The value-weighted risk map was produced by overlay-ing the basic risk map and the groundwater value map. Relevant validation was completed by contaminantdistributions and site investigation. Using Beijing Plain, China, as an example, thematic maps of the three fac-tors and the two risks were generated. The thematic maps suggested that landfills, gas stations and oil depots,and industrial areas were the most harmful potential contamination sources. The western and northern partsof the plain were the most vulnerable areas and had the highest groundwater value. Additionally, both thebasic and value-weighted risk classes in the western and northern parts of the plain were the highest, indi-cating that these regions should deserve the priority of concern. Thematic maps should be updated regularlybecause of the dynamic characteristics of hazards. Subjectivity and validation means in assessing the vulner-ability result were also discussed. In addition, GIS technology was essential in completing the assessmentwork.

© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Groundwater contamination is a widespread problem in the pro-cess of urbanization. When aquifers become polluted, contaminationis persistent and difficult to remediate due to their large storage,long residence times and physical inaccessibility (Foster andChilton, 2003). Moreover, groundwater contamination is an imper-ceptible and irreversible process, and prohibitive costs and time re-quirements may limit efforts to improve the groundwater condition(Causape et al., 2006; Yu et al., 2010). Thus, effectively preventing or re-ducing groundwater contamination is an important component ingroundwater management. Groundwater contamination risk assessmentis a useful tool for groundwater management. These assessments couldhelp to screen out potentially harmful sources and areas threatened by

he@cugb.edu.cn (J. He),

rights reserved.

groundwater contamination, which could be an important basis for deci-sion making, such as land planning and groundwater monitoring.

The concept of assessing groundwater contamination risk is based onthe “origin-pathway-target”model. The risk of contamination of ground-water depends on three elements: (1) the hazard posed by a potentiallypolluting activity (equivalent to origin); (2) the intrinsic vulnerability ofgroundwater to contamination (equivalent to pathway); and (3) the po-tential consequences of a contamination event upon the groundwater(equivalent to target) (Nobre et al., 2007; Zwahlen, 2004). At present,many countries in Europe (Andreo et al., 2006; Zwahlen, 2004), the Ame-ricas (Foster and Hirata, 1988; Nobre et al., 2007), Asia (Awawdeh andJaradat, 2009; Baalousha, 2011; Fadlelmawla et al., 2011; Nguyet andGoldscheider, 2006), Africa (Boughriba et al., 2010; Saidi et al., 2011)have carried out the risk assessment working.

Potential surface contamination sources and intrinsic vulnerability arethe two fundamental factors in the risk assessmentmethod. However, themethods used for assessing the two fundamental factors may differ fromplace to place due to various potential contamination sources andhydrogeological conditions. Existing methods to assess groundwater

217J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

vulnerability can be classified into three categories: (1) overlay and indexmethods; (2) process-based methods that apply deterministic modelsbased on physical processes; and (3) statistical models (NationalResearch Council, 1993; Nobre et al., 2007). Compared to groundwatervulnerability assessment, there are fewer general methods used for po-tential contamination source assessment, which can be attributed to thediversity of contamination source types, scales and relevant contaminanttypes and the randomness of contamination incidents. In addition to thepotential surface contamination sources and intrinsic vulnerability, re-searchers have added other factors, such as groundwater value (Ducci,1999; Saidi et al., 2011; Zaporozec, 2002), well capture zone (Baalousha,2011; Nobre et al., 2007) and specific geological settings (Gemitzi et al.,2006) into the risk assessment methods based on disparate perspectives.Factor aggregation means used in the assessment process include matrixrelation combination (Johansson et al., 1999; Zaporozec, 2002), multipli-cation (Johansson et al., 1999; Zaporozec, 2002) and addition (Gemitzi etal., 2006; Saidi et al., 2010). Additionally, validation of the vulnerability as-sessment result is worth discussing.

Using Beijing Plain as the example, this paper describes a ground-water contamination risk assessment method by integrating potentialcontamination sources, intrinsic vulnerability and groundwater value.This research aims to ascertain the present groundwater contamina-tion risk situation of Beijing Plain in a comprehensive and targetedway by testing the feasibility of the risk assessment method.

2. Methods

2.1. Hazard assessment

In the context of groundwater contamination, a hazard is defined asa potential source of contamination resulting fromhuman activities tak-ing place mainly at the land surface (Zwahlen, 2004). There are a widerange of sources that are likely to generate a certain number of contam-inant loads. Often, the groundwater contamination level is determinedby natural attenuation processes that occur within the zone betweenthe pollution source and the aquifer (Gogu and Dassargues, 2000).The occurrence of a groundwater contamination incident can be de-scribed as a certain quantity of contaminants released from a sourceon the ground that penetrates through the vadose zone and reachesthe aquifer (Foster and Hirata, 1988; Foster et al., 2002). It should benoted that a contamination source could release more than one con-taminant simultaneously and contain various contaminant types. Thecontamination process could be generalized as the behavior of a certainquantity of contaminants under specific hydrogeological conditions.The behavior is determined by the properties of the contaminants,which mainly include toxicity, mobility and degradation (Andreo etal., 2006; Arias-Estevez et al., 2008; Aronson et al., 2006; Nobre et al.,2007; Vahidnia et al., 2007). These properties represent the destructivepower, distribution scale and persistence of a contaminant in the envi-ronment during a contamination process. Based on the analysis above,the harmfulness of a hazard could be quantified using the followingequations:

HIj ¼ ∑ni¼0Cij � Qij ð1Þ

Cij ¼ TijWT þMijWM þ DijWD ð2Þ

Where HIj (Hazard Indexj) is the harmfulness quantification resultof the hazard j, Cij is the behavior quantification result of the contam-inant i that belongs to hazard j and Qij is the evaluated potential quan-tity of the contaminant i from hazard j that can infiltrate into theground and is called infiltrating contaminant load. Tij, Mij and Dij are

the quantified ratings of properties of contaminant i from hazard jand WT, WM and WD are the corresponding weights, which are usedfor distinguishing the different effects of the three properties in thecontamination process.

To assess a hazard using the above equations, the following proce-dure should be carried out. First, the representative contaminants of ahazard need to be selected to determine the contaminant inventory,and then the corresponding infiltrating contaminant load needs tobe estimated. Because it is not possible to collect all contaminants re-leased into the inventory by the hazard, the TDS (total dissolvedsolids) and COD (chemical oxygen demand), which are usually usedfor estimating water quality, are included in the inventory to standfor unknown or abandoned contaminants. A hazard's contaminant in-ventory should meet the following requirements: (1) contaminantsbelonging to corresponding hazards should reflect the hazard's con-tamination features, and (2) contaminants detected in the localgroundwater must be contained in accordance with the hazard if itssource is known. The infiltrating contaminant load is estimated bymultiplying the potential leakage quantity and the contaminant con-centration. For the purpose of attaining the most exact values of po-tential leakage quantity and contaminant concentration, advancedpreparation, including field investigation, sampling, background in-quiry, and data collection, should be carried out. When estimatingthe potential leakage quantity, elements such as the hazard's type,scale, penetration probability and the annual precipitation of thestudy locale should be considered.

In addition, the following three properties of the contaminantneed to be quantified and normalized: toxicity, mobility and degrada-tion. First, the toxicity, mobility and degradation of a contaminant inthe inventory are quantified according to different indicators. Second,each property is normalized to a rating from 1 to 5 according to thegiven value range of the indicator. Table 1 lists the relevant informa-tion used for property quantification and normalization. The toxicityof all selected contaminants refers to relevant standards for drinkingwater quality (MOH, 2006) and is graded depending on orders ofmagnitude of the corresponding standard values. The mobility anddegradation of selected organic contaminants refer to the LogKoc

values and the degradation rating and are graded by their relevantdescriptions, which can be obtained by the EPI Suit program(Aronson et al., 2006; EPA, 2011; Nobre et al., 2007). Exact indicatorsare not available for quantifying the mobility and degradation of TDS,COD and inorganic contaminants, so their relevant ratings are deter-mined by expert analysis.

The three weights in Eq. (2) are calculated according to AHP (Ana-lytical Hierarchy Process) (Saaty, 1987). In light of the damage togroundwater environments and water supply security, the toxicity ofa contaminant is thought to be moderate importance over its mobilityand degradation. Mobility and degradation are thought to be of equalimportance. Calculations showed that WT, WM and WD were 0.6, 0.2and 0.2, respectively.

2.2. Intrinsic vulnerability assessment

Intrinsic vulnerability assessment is an effective tool for describinginnate features of specific hydrogeological conditions that providesomemeasure of defense against external contamination. Intrinsic vul-nerability is independent of the nature of the specific contaminants andthe contamination scenario and takes into account the geological, hy-drological and hydrogeological characteristics of an area (Zwahlen,2004). The DRASTIC model (Aller et al., 1987), which is an overlay andindex method, is the most popular tool for assessing intrinsic vulnera-bility. The DRASTICmodel is easy to implement and provides a good as-sessment of groundwater vulnerability to contamination (Baalousha,2010). The model uses seven parameters to assess vulnerability:depth to water table, net recharge, aquifer media, soil media, topogra-phy, impact of vadose zone and hydraulic conductivity. The DRASTIC

Table 1Information used for property quantification and normalization.

Rating Toxicity Mobility Degradation

Drinking waterstandards range

LogKoc Description Calculateddegradation rating

Description assigned, expected totaldegradation time

1 >100 mg >4.5 Very strong sorption to soil and sediment, negligible migrationpotential to groundwater

>4.25–5 Hours and hours to days

2 >1–100 mg >3.5–4.5 Strong sorption to soil and sediment, negligible to slow migrationpotential to groundwater

>3.25–4.25 Days and days to weeks

3 >100–1000 μg >2.5–3.5 Moderate sorption to soil and sediment, slow migration potentialto groundwater

>2.25–3.25 Weeks and weeks to months

4 >10–100 μg >1.5–2.5 Low sorption to soil and sediment, moderate migration potentialto groundwater

>1.75–2.25 Months

5 ≤10 μg ≤1.5 Negligible sorption to soil and sediment, rapid migration potentialto groundwater

≤1.75 Recalcitrant

218 J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

model applies a linear combination of all factors (Babiker et al., 2005)according to the following equation:

VI ¼ DRDW þ RRRW þ ARAW þ SRSW þ TRTW þ IRIW þ CRCW ð3Þ

Where VI (vulnerability index) is the final assessment result, D, R,A, S, T, I, and C are acronyms of the seven parameters and the sub-scripts R and W indicate the corresponding ratings and weights.

Several shortcomings of this model, such as the subjectivity inher-ent in determining the rating scale and weighting coefficients and theselection and exclusion of specific parameters (Saidi et al., 2011;Zwahlen, 2004), have been pointed out. Still, its advantages cannotbe ignored. These advantages include the following: (1) some of thefactors, such as rainfall and depth to groundwater, can be appliedover large areas, which makes them suitable for region-scale assess-ments (Chitsazan and Akhtari, 2009; Thapinta and Hudak, 2003),and (2) there are fewer methodological uncertainties compared tomethods that require extensive data input (Fadlelmawla et al.,2011). Compared to other models, the DRASTIC model can assessthe intrinsic vulnerability using various types of hydrogeological pa-rameters and requires a relatively small amount of data.

To better represent the stratigraphic characteristics of the Quater-nary aquifer by DRASTIC, the China Geological Survey modified therating system of the original model as shown in Table 2, and theweights for the seven parameters were consistent with the original(CGS, 2006). The modified DRASTIC model was used in this study.

2.3. Groundwater value

Valuing groundwater is an important component of groundwaterresource protection schemes. It is important to recognize the valueeven when one cannot develop specific quantitative separations of thevarious components (National Research Council, 1997; Zaporozec,

Table 2Modified DRASTIC rating.

Rating Depth to watertable (m)

Net recharge(mm/year)

Aquifer media(media type)

Soil media(media type)

1 >30.5 0–51 Clay Non-shrinking or non-2 26.7–30.5 51–71.4 Loam Backfill3 22.9–26.7 71.4–91.8 Sandy loam Clay loam4 15.2–22.9 91.8–117.2 Silt Silty loam5 12.1–15.2 117.2–147.6 Fine sand Loam6 9.1–12.1 147.6–178 Sand Gravelly loam7 6.8–9.1 178–216 Medium sand Shrinking or aggregatin8 4.6–6.8 216–235 Coarse sand Peat9 1.5–4.6 235–254 Sand and gravel Sand and gravel10 0–1.5 >254 Gravel Gravel

2002). Different researchers have used various factors to quantify ground-water value, such as (1) total economic value as measured by extractivevalue and in situ value (National Research Council, 1997); (2) socio-economic value as measured by total extraction/groundwater recharge,variation in storage and groundwater quality (Saidi et al., 2011); and(3) estimated criteria measured by the current dependence on ground-water, the potential future dependence on groundwater and interactionwith environmental resources (Fadlelmawla et al., 2011).

The aim of valuing groundwater is to make decision makers realizethe present and future value of exploiting groundwater resources. Thevarious indicators used for valuing groundwater are based upon thegroundwater quantity and quality. A region with abundant groundwa-ter storage and high groundwater quality will have a higher groundwa-ter value. Accordingly, groundwater value is evaluated according to thefollowing Eq. (4):

Groundwater value ¼ groundwater storage class� groundwater quality class ð4Þ

2.4. Groundwater contamination risk

In this study, groundwater contamination risk is divided into twotypes: (1) basic contamination risk (Gogu and Dassargues, 2000;Zwahlen, 2004), which is composed of hazard and intrinsic vulnerability,and (2) value-weighted contamination risk (Karaouzas et al., 2009),which is composed of the basic contamination risk and groundwatervalue. These two assessments are based upon different considerationsand cannot be used interchangeably. The basic groundwater contamina-tion risk considers only the basic contamination process, whereas thevalue-weighted groundwater contamination risk considers both the con-tamination process and the groundwater exploitation potentiality. Theformer could depict the potential groundwater contamination class dis-tinctly, and the later could screen out areas of potential groundwater ex-ploitation that are threatened by contamination.

Topography(%)

Impact of vadose zone(media type)

Hydraulic conductivity(m/day)

aggregated clay >18 Clay 0–4.117–18 Loam 4.1–12.215–17 Sandy loam 12.2–20.313–15 Silt 20.3–28.511–13 Fine sand 28.5–34.69–11 Sand 34.6–40.7

g clay 7–9 Medium sand 40.7–61.14–7 Coarse sand 61.1–71.52–4 Sand and gravel 71.5–81.50–2 Gravel >81.5

219J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

According to Eqs. (5) and (6) below, the basic and value-weightedcontamination risk evaluations are calculated and then classifiedaccording to the basic groundwater contamination risk map and thevalue-weighted groundwater contamination risk map.

Basic contamination risk index ¼ Nh � Ni ð5Þ

Value�weighted contamination risk index ¼ Nb � Nv ð6Þ

Where Nh is the quantified number of the hazard class, Ni is thequantified number of the intrinsic vulnerability class, Nb is the quan-tified number of the basic contamination risk class and Nv is the quan-tified number of the groundwater value class.

2.5. Overlay and visualization methods

Much of our original data processing and results visualization wasaccomplished using ArcGIS10. First, a basemap of the study areawas di-vided into small cells using the fishnet tool according to the fit unit size.Considering the amount of data to be calculated and the requirementsneeded for its visual expression, the unit size in this study area was1 km*1 km. Second, various information regarding the hazards,hydrogeological parameters, groundwater storage and groundwaterquality were reflected on the base map using the spatial join tool.Third, different thematicmapswere produced based on built-in overlayrules using attribute table calculation and classified using Jenks NaturalBreaks.

Fig. 1. Location of Beijing Plain and potential grou

2.6. Validation methods

COST Action 620 outlined some ways to validate the performance ofthe risk assessment model for the karst region. Validation methods caninclude: the analysis of hydrographs and chemographs, isotopic chemis-try, use of artificial tracers and the use of analytical and numericalmodels(Zwahlen, 2004). Many other researchers also pointed out some similarvalidationmeans, such as nitrate concentrations in groundwater, ground-water quality analysis, groundwater flowmodels, statistics methods andsensitivity analysis (Babiker et al., 2005; Boughriba et al., 2010; Chitsazanand Akhtari, 2009; Nobre et al., 2007; Saidi et al., 2010; Saidi et al., 2011).In this research, the method performance can be validated through con-taminant distributions in the local groundwater and site investigation.

3. Applications

3.1. Description of the area

Beijing Plain, which covers an area of about 6400 km2 and accountsfor 39% of Beijing, is located in northeastern China (Fig. 1). It consists ofalluvial–proluvial fan as well as alluvial–proluvial plain and is broad andflat in its overall geomorphology. It has a semi-arid, semi-humid conti-nental monsoon climate with a mean annual temperature of 11.7 °C.The average annual precipitation and evaporation are 542.7 mm and1800 mm, respectively.

The Quaternary pore water system of Beijing Plain is the mainsource of groundwater used for surface activities. According to itshydrogeological conditions and the present exploitation situation,the Quaternary pore water of Beijing Plain is divided vertically into

ndwater contamination source distribution.

220 J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

three main aquifer groups: the shallow aquifers, the middle aquifersand the deep aquifers. The shallow aquifers are the target of the riskassessment and are composed predominantly of unconfined ground-water and some shallow-confined groundwater in confined areas. Itcan be recharged directly from natural rain infiltration, surfacewater body leakage and irrigation return water.

Beijing Plain is the most important base for local industry, agricultureand habitation in Beijing. Intensive anthropogenic activities consume alarge quantity of the groundwater and lead to the deterioration ofgroundwater quality. The groundwater consumption of Beijing in 2009was 1.97×1010 m3 and accounted for approximately 55% of the totalwater used in the year (Beijing Statistical Yearbook, 2010). Field investi-gation and sampling showed that in some areas of Beijing Plain, ground-water contamination has changed the original groundwater chemicaltype and increased the concentration of some contaminants, such as ni-trate and ammonium.

3.2. Hazard map

Six types of hazards were assessed in this study: gas stations and oildepots, polluted rivers, landfills, industrial areas, residential areas, andagricultural areas (Fig. 1). Landfills and agricultural areas were sub-divided into two types depending on anti-seepage treatment andwastewater irrigation history, respectively. Subcategories of industrialareas and polluted rivers were based upon industry type and waterquality differences, respectively. Table 3 lists the contaminant invento-ry and the corresponding property rating. Selection of the hazard'scontaminant inventory was based on the requirements stated inSection 2.1. Assignments of Qij for the six types of hazards are de-scribed in the supplementary material.

The above hazards were assessed using Eqs. (1) and (2). As shown inFig. 2, the harmfulness is divided into five classes: very low, low, moder-ate, high, and very high. The approximate order of the hazards based ontheir potential harmfulness from low to high is polluted rivers, agricultur-al areas, residential areas, industrial areas, gas stations and oil depots, andlandfills. Although the area occupied by landfills and gas stations and oildepots is small, they have a higher potential harmfulness class than theother hazards. In contrast, agricultural and residential areas occupy

Table 3Groundwater contaminant inventory of Beijing Plain and the corresponding property ratin

Contaminants Relevant hazard CAS number Drinking wstandards

TDS Industry, residents, landfills,polluted rivers, agriculture

NA 1000

COD Industry, residents, landfills,polluted rivers, agriculture

NA 5

Cr6+ Industry, landfills, polluted rivers,agriculture

7440-47-3 0.05

NH4-N Industry, residents, landfills,polluted rivers, agriculture

14798-03-9 0.5

NO3-N Industry, residents, polluted rivers 14797-55-8 20Benzo(a)pyrene Industry, residents, polluted rivers,

agriculture50-32-8 0.0000

Heptachlor Agriculture 76-44-8 0.0004Carbontetrachloride

Industry, polluted rivers, agriculture 56-23-5 0.002

Benzene Industry, polluted rivers, agriculture,gas stations and oil depots

71-43-2 0.01

Dichloromethane Industry, polluted rivers, agriculture 75-09-2 0.02Methyl tertiarybutylether

Gas stations and oil depots 1634-04-4 0.02

Trichloroethylene Industry, polluted rivers, agriculture 79-01-6 0.07Naphthalene Industry, polluted rivers, agriculture 91-20-3 0.3Toluene Gas stations and oil depots 108-88-3 0.7Xylene Gas stations and oil depots 1330-20-7 0.5Ethylbenzene Gas stations and oil depots 100-41-4 0.3Volatile phenol Industry, polluted rivers, agriculture 108-95-2 0.002Anionicsurfactant

Industry, residents, polluted rivers,agriculture

27176-87-0 0.3

large tracts of land but have a much lower potential harmfulness class.The distribution of harmfulness classes within industrial areas rangesfrom high to very low because of the variety of industrial subcategories.It should be noted that there are some cells that overlay more than onehazard, such as gas stations and agricultural areas. This will lead to therise of their corresponding harmfulness classes.

3.3. Intrinsic vulnerability map

The intrinsic vulnerability map of Beijing Plain (Fig. 3), which wascreated using the modified DRASTIC model, was acquired through ma-terial collection by the Beijing Institute of Geological Survey. Similar tothe hazardmap, the vulnerability indexwas divided into 5 classes rang-ing from very low to very high. The western and northern parts of theplain show the highest vulnerability class, indicating that they are themost vulnerable to external contamination. This is due to either theshallow groundwater table or the media of both the aquifer and vadosezone, which is mainly sand and gravel, gravel. In the southern centralarea of the plain, following along the upper, middle and lower part ofthe alluvial–proluvial plain, the vulnerability class shows a decreasingtrend from west to east. The “very low” areas are somewhat scattered.In the central and southeastern areas of the plain, the very low classifi-cation is either due to the fact that backfill is the major soil type or thatthe aquifer and vadose zone are mainly composed of sandy loam orloam. In the northern and western areas, the very low classification ismainly due to the deep groundwater table.

A hydrogeological section (its location was plotted in Fig. 3) wasshown to compare with the assessed intrinsic vulnerability result.As Fig. 4 demonstrates, along the section from northwest to south-east, namely from top of the alluvial–proluvial fan to its bottom andthe alluvial–proluvial plain, aquifer media changes from coarse tofine and aquifer structure transits from one single layer to multi-layers. Also, the groundwater level changes from deep to shallow.The section can be divided into 5 parts from A to E and detailed aqui-fer medium features of each part are listed in Table 4. Generally, theintrinsic vulnerability map by modified DRASTIC model is in accor-dance with the revealed local hydrogeological information.

g.

ater(mg/l)

Toxicityrating

LogKoc Mobilityrating

Calculated degradationvalue

Degradationrating

1 5 5

2 3 2

4 4 4

3 4 3

2 5 41 5 5.95 1 1.84 4

5 4.38 2 0.53 55 1.85 4 1.95 4

5 1.75 4 2.44 3

4 1.44 5 2.67 34 1.35 5 2.78 3

4 2 4 2.39 33 2.96 3 2.33 33 2.07 4 2.94 33 2.25 4 2.81 33 2.23 4 2.91 35 1.9 4 3.07 33 3.57 2 2.84 3

Fig. 2. Hazards map.

221J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

3.4. Groundwater value map

Throughmaterial collection, the groundwater quality and quantity as-sessment outcomes of Beijing Plain (Figs. 5 and 6) were obtained. Valuesfrom1 (verypoor) to 4 (very good)were assigned to groundwater quality

Fig. 3. Intrinsic vulnerability map

(Fig. 5) and values from 1 (very poor) to 4 (very rich) were assigned togroundwater storage (Fig. 6). The groundwater valuewas then calculatedaccording to Eq. (4) and classified as shown in Fig. 7. The northern andsoutheastern areas of the plain have the highest groundwater quality(Fig. 5), and thewestern andnorthern areashave the richest groundwater

by modified DRASTIC model.

Fig. 4. Hydrogeological section of Beijing Plain.

222 J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

storage (Fig. 6). Compared to the southern andeastern areas, thenorthernand western regions of the plain show relatively higher groundwatervalues (Fig. 7).

3.5. Groundwater contamination risk map

Both the hazard harmfulness and the intrinsic vulnerability wereassigned values from 1 to 5 based on classes from “very low” to “veryhigh”. The basic groundwater contamination risk map (Fig. 8) was gen-erated by spatial overlay of the hazard map and the vulnerability map,and the results were reclassified. Taking into account the hazards andintrinsic vulnerability simultaneously, the area with the highest con-tamination risk class is located in the southwestern region of the centralplain. Hazards in this area include untreated landfills, gas stations and asteel industry zone and the vulnerability class is “very high”. The otherscattered areas with the highest risk class are mainly caused byuntreated landfills and high vulnerability class. It can be found that anabundance of potential contamination factors exist in the most vulner-able areas of the plain. On the whole, Fig. 8 could reflect the differentclass distributions of the hazard map (Fig. 2) and the vulnerabilitymap (Fig. 3).

Similar to the production of the basic risk map, the value-weightedgroundwater contamination risk map was generated using the ground-water value map and the basic risk map and calculated according toEq. (6). Areas containing higher groundwater value and basic risk classgenerate a higher value-weighted groundwater contamination riskclass. Fig. 9 shows that the western and northern areas of Beijing Plaincontain groundwater with the highest value and contamination risk,

Table 4Descriptions of the hydrogeological section.

Hydrogeologicalsection part

Features

A Top of the alluvial–proluvial fan, main recharge area ofBeijing Plain, single aquifer layer consisted by sand andgravel, gravel

B Upper region of the alluvial–proluvial fan, 2–3 aquifer layersconsisted by sand and gravel, gravel, interbedded with clay

C Lower part of the alluvial–proluvial fan and alluvial–proluvial plain, aquifer layers transiting to multiplegradually, consisted by sand and gravel mainly, gravel less

D Alluvial–proluvial plain, multiple aquifer layers with mediachanging from sand and gravel to coarse sand, mediumsand, sandy, fine sand gradually

E Alluvial–proluvial plain, multiple aquifer layers consisted bysandy media of small particle size, and sand and graveloccasionally

and should be given the most attention. Overall, the value-weighted riskclass of Beijing Plain is gradually reduced from west to east and north tosouth.

3.6. Validation

The assessment results can be validated by collected materials tosome extent. Relevant information indicates that there is a certain de-gree of similarity between the intrinsic vulnerability map and nitratecontamination distributions of shallow groundwater. The southwesternregion of the central plain, where vulnerability classes are “very high”and “high”, suffers the most serious nitrate contamination. In the mostof other “very high” and “high” regions of the intrinsic vulnerabilitymap, nitrate contamination also exists. In the southwestern region ofthe central plain of the basic risk map, where the basic risk class is“very high” and “high”, organic contaminants, such as trichloro-methane, 1,2-dichloropropane, have been constantly detected in localgroundwater. Actually, the steel plant and its ancillary industriesthere, have been moved out. Also, the untreated landfill there wasclosed and relevant environmental remediation has started. It shouldbenoted that the groundwater qualitymap (Fig. 5) is not suitable to val-idate the results, because the groundwater quality evaluation containsmany natural elements. The validation demonstrated that the centralsouthwestern region of Beijing Plain is the most risky area.

4. Discussion

It is very useful for groundwater management to carry out contam-ination risk assessments and create relevant thematic maps. Differentthematic maps represent divergent aspects to be considered whenmaking decisions about groundwater management, and the maps can-not be used interchangeably. The hazard map could be used for dis-tinguishing the harmfulness class of different contamination sources,and the intrinsic vulnerability map could be used for identifying theareaswith specific hydrogeological conditions that are vulnerable to ex-ternal contamination, which could be used as a basis for land planning.The rationality of existing contamination sources' location could bejudged according to the basic groundwater contamination risk map.The groundwater value map shows the exploitation potential with re-spect to considerations of aquifer storage and groundwater quality.Compared to the basic groundwater contamination risk map, thevalue-weighted groundwater contamination risk map is more compre-hensive and targeted by taking into account issues of essential contam-ination and extrinsic groundwater exploitation. Maps and summaryinformation could be made available to the public and stakeholders toraise awareness of the threats to their groundwater resources (Mimi

Fig. 5. Groundwater quality map.

223J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

and Assi, 2009). And such schemes should be kept as simple as possibleand require a minimum of data (Nguyet and Goldscheider, 2006).

The risk of pollution is determined both by the intrinsic characteris-tics of the aquifer,which are relatively static, and the existence of poten-tially polluting activities, which are dynamic and can be changed andcontrolled (Saidi et al., 2010, 2011). The present work is based on thecurrent status of various factors; thus, the results are temporary. It isnecessary to do as much research as possible to find the variations or

Fig. 6. Groundwate

trends when factors change in status. Additionally, the risk map has tobe updated using the latest information to facilitate continued ground-water management.

While subjectivity is practically unavoidable when consideringregional-scale schemes, the authors believe that amore comprehensiveapplication of the risk principles will enable more defendable land-surface zoning schemes (Fadlelmawla et al., 2011). Amore comprehen-sive consideration of different perspectives, such as contamination

r storage map.

Fig. 7. Groundwater value map.

224 J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

source, hydrogeology, socio-economic factors, and policies (Roselló etal., 2009), will generate a more applicative and objective mapping.Compared to the defined matrix relation combination, it is also more ob-jective to aggregate maps by the method of multiplication.

The DRASTIC model has received much criticism due to its inherentlack of proper validation (Andreo et al., 2006;Nobre et al., 2007). Distribu-tions of specific contaminants, such as nitrates (Al-Adamat et al., 2003;Saidi et al., 2010), have been adopted to validate the intrinsic vulnerabilitymap in this research. Furthermore, statistics about contaminant types and

Fig. 8. Basic groundwater co

concentration differences detected in the local groundwater should alsobe considered if the data are supported (Moratalla et al., 2011). After all,contaminant types and quantities generated by different sources show agreat diversity in such a scale. Groundwater flow models could also beused for validating the vulnerability map (Chitsazan and Akhtari, 2009).Some research has applied sensitivity analyses to compare “effective”weights with original “theoretical” weights and identified different con-tributions of the seven factors to the total vulnerability index (Babikeret al., 2005; Napolitano and Fabbri, 1996; Saidi et al., 2011). However,

ntamination risk map.

Fig. 9. Value-weighted groundwater contamination risk map.

225J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

sensitivity analyses only help to identify how the composition of factorsand weights work. It is the final vulnerability index that is needed for ef-fective decision making.

GIS technology has fundamentally changed the way we thinkabout and manage natural resources, especially water resources (Jhaet al., 2007). It provided an effective platform for this research, simpli-fied the data management in the assessment process and facilitatedthe mapping work. Different thematic maps were produced withhigher efficiency and quality compared to traditional methods. Withappropriate cartography, one can improve communication and thusbridge the gap between experts and users (Lahr and Kooistra, 2010).

5. Conclusions

This study built a method for assessing groundwater contamina-tion risk through the integration of hazards, intrinsic vulnerabilityand groundwater value. In this method, different types of hazardswere assessed in a unified formby quantifying the contaminant's proper-ties andpotential infiltrating load, amodifiedDRASTICmodelwas appliedto generate intrinsic vulnerability assessments of the studied aquifers,and the groundwater value was evaluated by combining groundwaterquality and quantity. Two types of risk mapswere produced, and the val-idation was done by nitrate distributions and site investigation. Thismethod considered both the contamination process, which was basedon the “origin-pathway-target”model, and the groundwater exploitationpotentiality, which was based on current groundwater conditions.

The final results demonstrated that the western and northernparts of Beijing Plain not only contained the highest groundwatervalue but were also the most vulnerable to groundwater contamina-tion due to inappropriate land use and special hydrogeological condi-tions. Applying groundwater contamination risk assessment techniquesis an excellent decision-making tool for land planning and groundwatermanagement. Thematic maps could reveal current inappropriate landusage and help distinguish the different groundwater contaminationrisk classes. Relevant thematic maps should be updated regularly due tothe dynamic nature of hazards. Although subjectivity was unavoidablein the study, some measures, such as the aggregation and classification

of different factors in a unified form, could overcome this disadvantageto some extent. If the data are available, statistics for different contami-nants and distributions and groundwater flow models could be used tovalidate the vulnerability assessment results. In addition, theGIS platformis a convenient tool for overlayingdifferent elements andexpressing eval-uation results.

Acknowledgments

The study was conducted under the “Potential groundwater haz-ard identification and contamination prevention zoning technologyresearch” project (Number: 200909038-1), which was funded bythe Ministry of Environmental Protection, the People's Republic ofChina. We acknowledge the assistance of the Beijing Institute of Geo-logical Survey, which provided the necessary materials to completethis research.

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.scitotenv.2012.06.005.

References

Al-Adamat RAN, Foster IDL, Baban SMJ. Groundwater vulnerability and risk mappingfor the Basaltic aquifer of the Azraq basin of Jordan using GIS, Remote sensingand DRASTIC. Appl Geogr 2003;23:303–24.

Aller L, Bennet T, Lehr JH, Petty RJ, Hackett G. DRASTIC: a standardized system for eval-uating ground water pollution potential using hydrogeologic settings. Dublin,Ohio: National Water Well Association; 1987. 266 pp.

Andreo B, Goldscheider N, Vadillo I, Vias JM, Neukum C, Sinreich M, et al. Karst ground-water protection: first application of a Pan-European approach to vulnerability,hazard and risk mapping in the Sierra de Libar (Southern Spain). Sci Total Environ2006;357:54–73.

Arias-Estevez M, Lopez-Periago E, Martinez-Carballo E, Simal-Gandara J, Mejuto JC,Garcia-Rio L. The mobility and degradation of pesticides in soils and the pollutionof groundwater resources. Agric Ecosyst Environ 2008;123:247–60.

Aronson D, Boethling R, Howard P, Stiteler W. Estimating biodegradation half-lives foruse in chemical screening. Chemosphere 2006;63:1953–60.

Awawdeh MM, Jaradat RA. Evaluation of aquifers vulnerability to contamination in theYarmouk River basin, Jordan, based on DRASTIC method. Arab J Geosci 2009;3:273–82.

226 J. Wang et al. / Science of the Total Environment 432 (2012) 216–226

Baalousha H. Assessment of a groundwater quality monitoring network using vulnera-bility mapping and geostatistics: a case study from Heretaunga Plains, NewZealand. Agric Water Manage 2010;97:240–6.

Baalousha HM. Mapping groundwater contamination risk using GIS and groundwatermodelling. A case study from the Gaza Strip, Palestine. Arab J Geosci 2011;4:483–94.

Babiker IS, Mohamed MAA, Hiyama T, Kato K. A GIS-based DRASTIC model for assessingaquifer vulnerability in Kakamigahara Heights, Gifu Prefecture, central Japan. SciTotal Environ 2005;345:127–40.

Beijing Statistical Yearbook. Beijing Municipal Bureau of Statistics, NBS (National Bu-reau of Statistics) Survey Office in Beijing. Beijing: China Statistics Press; 2010.

Boughriba M, Barkaoui A-E, Zarhloule Y, Lahmer Z, El Houadi B, Verdoya M. Groundwa-ter vulnerability and risk mapping of the Angad transboundary aquifer usingDRASTIC index method in GIS environment. Arab J Geosci 2010;3:207–20.

Causape J, Quilez D, Aragues R. Groundwater quality in CR-V irrigation district (BardenasI, Spain): alternative scenarios to reduce off-site salt and nitrate contamination. AgricWater Manage 2006;84:281–9.

CGS (China Geological Survey). Technical requirements for assessing groundwater vul-nerability. GWI-D3; 2006 (In Chinese).

Chitsazan M, Akhtari Y. A GIS-based DRASTIC model for assessing aquifer vulnerabilityin Kherran Plain, Khuzestan, Iran. Water Resour Manage 2009;23:1137–55.

Ducci D. GIS techniques for mapping groundwater contamination risk. Nat Hazards1999;20:279–94.

Fadlelmawla AA, Fayad M, El-Gamily H, Rashid T, Mukhopadhyay A, Kotwicki V. A landsurface zoning approach based on three-component risk criteria for groundwaterquality protection. Water Resour Manage 2011;25:1677–97.

Foster SSD, Chilton PJ. Groundwater: the processes and global significance of aquiferdegradation. Philos Trans R Soc Lond B 2003;358:1957–72.

Foster SSD, Hirata R. Groundwater pollution risk assessment: a methodology usingavailable data. Peru, Lima: Pan American Centre for Sanitary Engineering and Envi-ronmental Sciences (CEPIS); 1988. 79 pp.

Foster SSD, Hirata R, Gomes D, D'Elia M, Paris M. Groundwater quality protection: aguide for water utilities, municipal authorities, and environment agencies. 2thed. Washington, DC, USA: World Bank; 2002.

Gemitzi A, Petalas C, Tsihrintzis VA, Pisinaras V. Assessment of groundwater vulnera-bility to pollution: a combination of GIS, fuzzy logic and decision making tech-niques. Environ Geol 2006;49:653–73.

Gogu RC, Dassargues A. Current trends and future challenges in groundwater vulnera-bility assessment using overlay and index methods. Environ Geol 2000;39:549–59.

Jha MK, Chowdhury A, Chowdary VM, Peiffer S. Groundwater management and devel-opment by integrated remote sensing and geographic information systems: pros-pects and constraints. Water Resour Manage 2007;21:427–67.

Johansson PO, Scharp C, Alveteg T, Choza A. Framework for ground water protection—the Managua ground water system as an example. GroundWater 1999;37:204–13.

Karaouzas I, Dimitriou E, Skoulikidis N, Gritzalis K, Colombari E. Linking hydrogeological andecological tools for an integrated river catchment assessment. Environ Model Assess2009;14:677–89.

Lahr J, Kooistra L. Environmental risk mapping of pollutants: state of the art and com-munication aspects. Sci Total Environ 2010;408:3899–907.

Mimi ZA, Assi A. Intrinsic vulnerability, hazard and risk mapping for karst aquifers: acase study. J Hydrol 2009;364:298–310.

MOH (Ministry of Health of the People's Republic of China). Standards for drinkingwater quality. GB 5749 - 2006; 2006 (In Chinese).

Moratalla A, Gomez-Alday JJ, Sanz D, Castano S, De Las Heras J. Evaluation of aGIS-based integrated vulnerability risk assessment for the Mancha Oriental System(SE Spain). Water Resour Manage 2011;25:3677–97.

Napolitano P, Fabbri AG. Single-parameter sensitivity analysis for aquifer vulnerabilityassessment using DRASTIC and SINTACS. HydroGIS 96: application of geographicinformation systems in hydrology and water resources management. Proceedingsof Vienna ConferenceIAHS Pub; 1996. p. 559–66.

National Research Council. Ground water vulnerability assessment: predicting relativecontamination potential under conditions of uncertainty. 1th ed. Washington, DC,USA: The National Academies Press; 1993.

National Research Council. Valuing ground water: economic concepts and approaches.1th ed. Washington, DC, USA: The National Academies Press; 1997.

Nguyet VTM, Goldscheider N. A simplified methodology for mapping groundwater vul-nerability and contamination risk, and its first application in a tropical karst area,Vietnam. Hydrogeol J 2006;14:1666–75.

Nobre RCM, Rotunno OC, Mansur WJ, Nobre MMM, Cosenza CAN. Groundwater vulner-ability and risk mapping using GIS, modeling and a fuzzy logic tool. J ContamHydrol 2007;94:277–92.

Roselló MJP, Martinez JMV, Navarro BA. Vulnerability of human environment to risk:case of groundwater contamination risk. Environ Int 2009;35:325–35.

Saaty RW. The Analytic Hierarchy Process—what it is and how it is used. Math Model1987;9:161–76.

Saidi S, Bouri S, Ben Dhia H. Groundwater vulnerability and risk mapping of theHajeb-jelma aquifer (Central Tunisia) using a GIS-based DRASTIC model. EnvironEarth Sci 2010;59:1579–88.

Saidi S, Bouri S, Ben Dhia H, Anselme B. Assessment of groundwater risk using intrinsicvulnerability and hazard mapping: application to Souassi aquifer, Tunisian Sahel.Agric Water Manage 2011;98:1671–82.

Thapinta A, Hudak PF. Use of geographic information systems for assessing groundwaterpollution potential by pesticides in Central Thailand. Environ Int 2003;29:87–93.

US EPA (United States Environmental Protection Agency). Estimation Programs Inter-face Suite™ for Microsoft® Windows, v 4.10; 2011. Washington, DC, USA.

Vahidnia A, Van der Voet GB, de Wolf FA. Arsenic neurotoxicity—a review. Hum ExpToxicol 2007;26:823–32.

Yu C, Yao YY, Hayes G, Zhang BX, Zheng CM. Quantitative assessment of groundwatervulnerability using index system and transport simulation, Huangshuihe catch-ment, China. Sci Total Environ 2010;408:6108–16.

Zaporozec A. Groundwater contamination inventory: a methodological guide. IHP-VI,Series on groundwater No. 2. Paris, France: UNESCO; 2002. 160 pp.

Zwahlen F. Vulnerability and risk mapping for the protection of carbonate (karst) aqui-fers, final report (COST action 620). European Commission, Directorate XII Science,Research and Development, Report EUR 20912, Brussels; 2004. 297 pp.