Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

18
Intrinsic and specic vulnerability of groundwater in central Spain: the risk of nitrate pollution Juan J. Martínez-Bastida & Mercedes Arauzo & Maria Valladolid Abstract The intrinsic vulnerability of groundwater in the Comunidad de Madrid (central Spain) was evaluated using the DRASTIC and GOD indexes. Groundwater vulner- ability to nitrate pollution was also assessed using the composite DRASTIC (CD) and nitrate vulnerability (NV) indexes. The utility of these methods was tested by analyzing the spatial distribution of nitrate concentrations in the different aquifers located in the study area: the Tertiary Detrital Aquifer, the Moor Limestone Aquifer, the Cretaceous Limestone Aquifer and the Quaternary Aqui- fer. Vulnerability maps based on these four indexes showed very similar results, identifying the Quaternary Aquifer and the lower sub-unit of the Moor Limestone Aquifer as deposits subjected to a high risk of nitrate pollution due to intensive agriculture. As far as the spatial distribution of groundwater nitrate concentrations is concerned, the NV index showed the greatest statistical signicance (p <0.01). This new type of multiplicative model offers greater accuracy in estimations of specic vulnerability with respect to the real impact of each type of land use. The results of this study provide a basis on which to guide the designation of nitrate vulnerable zones in the Comunidad de Madrid, in line with European Union Directive 91/676/EEC. Keywords Nitrate . Groundwater vulnerability . Land use . Agriculture . Spain Introduction Diffuse nitrate pollution of groundwater is currently considered one of the major causes of deteriorating water quality (Knapp 2005). European Union Directive 91/676/ EEC (Council of the European Communities 1991), relating to the protection of waters against pollution caused by nitrate from agricultural sources, was developed in response to the European Unions concern about the environmental and health implications of this phenom- enon. This directive established a limit for nitrate concen- trations in water bodies (50 mg L 1 ) and requires that Member States designate nitrate vulnerable zones and develop action programmes for the restoration of polluted areas. During the last decade, the implementation of these European regulations has prompted numerous publica- tions on the evaluation of the vulnerability of groundwater (Secunda et al. 1998; Fritch et al. 2000; Gogu and Dassargues 2000; Al-Adamat et al. 2003; Naqa 2004; Worral and Kolpin 2004; Stigter et al. 2006). However, to date, no standard method has been adopted for evaluating the vulnerability of groundwater to nitrate pollution. Since the concept of groundwater vulnerability was rst introduced by Margat (1968), many other denitions have been incorporated and used. Vrba and Zaporozec (1994) dened intrinsic vulnerabilityas an intrinsic property of a groundwater system and one that depends on the sensitivity of that system to human or natural impacts, whereas specic vulnerabilitywas dened as the risk of pollution due to the potential impact of specic land uses and contaminants. Many methods have been proposed for mapping aquifer vulnerability but parametric methods represent the most utilized approach (Instituto Geológico y Minero; IGME 2004). Within these, the DRASTIC index (Aller et al. 1987) is one of the most common methods used internationally to evaluate the intrinsic vulnerability (Auge 2004). The most important assump- tions made when assessing vulnerability with DRASTIC are that the contaminant is introduced at the ground surface, is ushed into the groundwater by precipitation and has the mobility of water (Aller et al. 1987). This model uses seven media parameters (depth to water-table, net recharge, aquifer media, soil media, topography, impact of the vadose zone and hydraulic conductivity) through an additive formulation to estimate intrinsic Received: 24 April 2008 / Accepted: 19 October 2009 Published online: 26 November 2009 © Springer-Verlag 2009 J. J. Martínez-Bastida ()) : M. Arauzo Departamento de Contaminación Ambiental, Centro de Ciencias Medioambientales, CSIC, Serrano 115, 28006, Madrid, Spain e-mail: [email protected] Tel.: +34-91-7452500 Fax: +34-91-5640800 M. Valladolid Departamento de Biodiversidad y Biología Evolutiva, Museo Nacional de Ciencias Naturales, CSIC, José Gutierrez Abascal 2, 28006, Madrid, Spain Hydrogeology Journal (2010) 18: 681698 DOI 10.1007/s10040-009-0549-5

Transcript of Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

Page 1: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

Intrinsic and specific vulnerability of groundwater in central Spain:the risk of nitrate pollution

Juan J. Martínez-Bastida & Mercedes Arauzo &

Maria Valladolid

Abstract The intrinsic vulnerability of groundwater in theComunidad de Madrid (central Spain) was evaluated usingthe DRASTIC and GOD indexes. Groundwater vulner-ability to nitrate pollution was also assessed using thecomposite DRASTIC (CD) and nitrate vulnerability (NV)indexes. The utility of these methods was tested byanalyzing the spatial distribution of nitrate concentrationsin the different aquifers located in the study area: theTertiary Detrital Aquifer, the Moor Limestone Aquifer, theCretaceous Limestone Aquifer and the Quaternary Aqui-fer. Vulnerability maps based on these four indexesshowed very similar results, identifying the QuaternaryAquifer and the lower sub-unit of the Moor LimestoneAquifer as deposits subjected to a high risk of nitratepollution due to intensive agriculture. As far as the spatialdistribution of groundwater nitrate concentrations isconcerned, the NV index showed the greatest statisticalsignificance (p<0.01). This new type of multiplicativemodel offers greater accuracy in estimations of specificvulnerability with respect to the real impact of each typeof land use. The results of this study provide a basis onwhich to guide the designation of nitrate vulnerable zonesin the Comunidad de Madrid, in line with European UnionDirective 91/676/EEC.

Keywords Nitrate . Groundwater vulnerability .Land use . Agriculture . Spain

Introduction

Diffuse nitrate pollution of groundwater is currentlyconsidered one of the major causes of deteriorating waterquality (Knapp 2005). European Union Directive 91/676/EEC (Council of the European Communities 1991),relating to the protection of waters against pollutioncaused by nitrate from agricultural sources, was developedin response to the European Union’s concern about theenvironmental and health implications of this phenom-enon. This directive established a limit for nitrate concen-trations in water bodies (50 mg L−1) and requires thatMember States designate nitrate vulnerable zones anddevelop action programmes for the restoration of pollutedareas. During the last decade, the implementation of theseEuropean regulations has prompted numerous publica-tions on the evaluation of the vulnerability of groundwater(Secunda et al. 1998; Fritch et al. 2000; Gogu andDassargues 2000; Al-Adamat et al. 2003; Naqa 2004;Worral and Kolpin 2004; Stigter et al. 2006). However, todate, no standard method has been adopted for evaluatingthe vulnerability of groundwater to nitrate pollution.

Since the concept of groundwater vulnerability wasfirst introduced by Margat (1968), many other definitionshave been incorporated and used. Vrba and Zaporozec(1994) defined “intrinsic vulnerability” as an intrinsicproperty of a groundwater system and one that depends onthe sensitivity of that system to human or natural impacts,whereas “specific vulnerability” was defined as the risk ofpollution due to the potential impact of specific land usesand contaminants. Many methods have been proposed formapping aquifer vulnerability but parametric methodsrepresent the most utilized approach (Instituto Geológicoy Minero; IGME 2004). Within these, the DRASTICindex (Aller et al. 1987) is one of the most commonmethods used internationally to evaluate the intrinsicvulnerability (Auge 2004). The most important assump-tions made when assessing vulnerability with DRASTICare that the contaminant is introduced at the groundsurface, is flushed into the groundwater by precipitationand has the mobility of water (Aller et al. 1987). Thismodel uses seven media parameters (depth to water-table,net recharge, aquifer media, soil media, topography,impact of the vadose zone and hydraulic conductivity)through an additive formulation to estimate intrinsic

Received: 24 April 2008 /Accepted: 19 October 2009Published online: 26 November 2009

© Springer-Verlag 2009

J. J. Martínez-Bastida ()) :M. ArauzoDepartamento de Contaminación Ambiental,Centro de Ciencias Medioambientales, CSIC,Serrano 115, 28006, Madrid, Spaine-mail: [email protected].: +34-91-7452500Fax: +34-91-5640800

M. ValladolidDepartamento de Biodiversidad y Biología Evolutiva,Museo Nacional de Ciencias Naturales, CSIC,José Gutierrez Abascal 2, 28006, Madrid, Spain

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 2: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

vulnerability values. Another parametric method is theGOD index (Foster and Hirata 1991; Foster et al. 2002),which has often been applied in the UK, Spain and LatinAmerica (Auge 2004). The two basic factors considered todetermine intrinsic vulnerability are the level of hydraulicinaccessibility of the saturated zone of the aquifer and thecontaminant attenuation capacity of the strata overlyingthe saturated aquifer (Foster et al. 2002). Unlike DRAS-TIC, this method does not consider the effect of themovement of the pollutant through the saturated zone andonly uses three parameters (groundwater confinement,overlying strata and depth to groundwater) in a multi-plicative system. Therefore, regarding the concept ofvulnerability, each index represents a different pointof view: while the GOD index just considers the risk ofpollution based on the attenuation capacity of theunsaturated zone, the DRASTIC index also incorporatesthe risk associated with the characteristics of the saturatedzone. Because of the higher number of parametersincluded in the DRASTIC model, it could be expected toobtain a greater degree of accuracy in the vulnerabilitymaps. On the other hand, the DRASTIC model couldfavor the fact that some of its seven parameters were notreflected in the final vulnerability results (as a conse-quence of the relative weights). Moreover, not all the datarequired for the DRASTIC index are always available on aregional scale.

Both methods could prove to be a useful groundwatermanagement tool, at medium to small scales, for provid-ing a general overview (Auge 2004; Vias et al. 2005) andalso for designating new Nitrate Vulnerable Zones, asdefined in EU Directive 91/676/EEC (Stigter et al. 2006).It could therefore be interesting to compare the respectiveefficiencies of these methods, as it is noted in severalstudies (Gogu et al. 2003; Vias et al. 2005). On the otherhand, several authors have expressed doubts about thereliability of estimations of intrinsic vulnerability based onthese methods (Garrett et al. 1989; Rosen 1994; Rupert2001; Stigter et al. 2002, 2006; Gogu et al. 2003; Colmanet al. 2005) because of discrepancies observed in somecases between vulnerability maps and nitrate pollutionmaps. For this reason, modifications to the DRASTICindex have been proposed. These essentially relate tochanges in rating ranges, relative weights and in vulner-ability classes and to the elimination of redundantparameters (Fritch et al. 2000; Thirumalaivasan et al.2003; Babiker et al. 2005; Panagopoulos et al. 2006;Stigter et al. 2006; Antonakos and Lambrakis 2007;Denny et al. 2007). Other authors have incorporated anew land use parameter with the aim of estimating thespecific vulnerability to groundwater pollution by nitrateor other pollutants. One example of this is the compositeDRASTIC index (CD index) proposed by Secunda et al.(1998). The CD index attempts to evaluate the potentialeffect of extensive land use upon groundwater qualityresulting from alterations to the soil matrix and unsatu-rated zone media over time. This index was applied toproduce groundwater vulnerability and risk maps for theSharon region of Israel (Secunda et al. 1998) and a

basaltic aquifer of the Azraq basin of Jordan (Al-Adamatet al. 2003).

The resulting specific vulnerability indexes are basedon an additive model (Secunda et al. 1998; Ribeiro 2000;Al-Adamat et al. 2003; Thirumalaivasan et al. 2003;Panagopoulos et al. 2006; Guo et al. 2007; Hamza et al.2007). However, the additive formulation perhaps failedto reflect the protective effect of land uses that do not haveany adverse effects on groundwater quality. In order to testa new approach based on a multiplicative model andfocused on the problem of nitrate pollution process, a newindex, the nitrate vulnerability index (NV index), isproposed in the current study. Combining parametricmethods with geographic information systems (GIS) hasmade intrinsic and specific vulnerability mapping possible(Evans and Myers 1990; Adams and Foster 1992; Robinset al. 1994; Hiscock et al. 1995; Secunda et al. 1998;Piscopo 2001; Gogu et al. 2003; Vias et al. 2005).

The objectives of this investigation were: (1) togenerate and to compare intrinsic vulnerability maps ofgroundwater in central Spain using the DRASTIC andGOD indexes, (2) to generate and to compare mapsshowing specific vulnerability to nitrate pollution forgroundwater in central Spain using the composite DRAS-TIC index (CD index; Secunda et al. 1998) and the nitratevulnerability index (NV index; a new approach based on amultiplicative model), (3) to compare the intrinsic andspecific vulnerability maps with the distribution ofgroundwater nitrate pollution observed in central Spain(IGME 1985; Hernández-García and Custodio 2004;Arauzo et al. 2008), and (4) to analyze the reliabilityand the respective utilities of all these vulnerabilityindexes to identify nitrate vulnerable zones (Directive91/676/EEC).

Study area

The study area selected for this research project corre-sponded to the Comunidad de Madrid (central Spain,8,028 km2, Fig. 1), which contains four main aquifersystems that belong to the Tajo River Basin: the TertiaryDetrital Aquifer, the Moor Limestone Aquifer (both ofthem are Tertiary groundwater systems), the CretaceousLimestone Aquifer and the Quaternary Aquifer. Thisregion is characterized by heavy human impact, with ahigh concentration of urban and industrial activity andvery intensive agriculture on Quaternary alluvial deposits.As a result, there is a significant water demand, whichresults in problems such as overfishing and groundwaterpollution. Water supplies for the local population areusually obtained from surface water reserves, althoughabout 5% of local inhabitants depend on groundwatersupplies (Hernández-García and Custodio 2004). Duringperiods of drought, groundwater ensures supplies to urbancentres as a supplementary source of water constitutingalmost 25% of the water supplies (Alcolea and GarcíaAlvarado 2006).

682

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 3: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

ClimateThe study area has a Mediterranean-Continental climatecharacterized by cold winters, hot summers and consid-erable oscillations in temperature, particularly in thehighest areas (IGME 1999). Rainfall is irregularly dis-tributed, with a summer dry season corresponding to themonths of July and August (Alcolea and García Alvarado2006) and wet periods in autumn and spring. Annualmean precipitation for the whole area is around 500 mmyear−1, but it ranges between 800 and 1,000 mm year−1 inthe northwest part of the region (IGME 1985). Periods ofdrought are also common and potential evapotranspiration

exceeds precipitation throughout most of the year, withannual deficits reaching 300 mm. Runoff consequentlytends to be very low (Comunidad de Madrid 2001).

HydrogeologyGroundwater systems represent nearly one third of thetotal water resources of the Comunidad de Madrid. Ofthese, the Tertiary aquifers are the most important andinclude the following systems: the Tertiary DetritalAquifer and the Moor Limestone Aquifer. The othermajor aquifer in the study area is the Cretaceous Lime-

Fig. 1 a Hydrogeological map of central Spain. Main aquifers distribution and groundwater flow directions in each aquifer. Hydraulichead contours (m) are shown with a contour interval of 20 m. b Cross section A–A′ of the study area

683

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 4: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

stone Aquifer (Fig. 1). The Quaternary Aquifer of theComunidad de Madrid is considered of only local interest(because of its lower storage capacity). Even so, itreceives the highest overall contribution of nitrogen fromagricultural activities (Ministerio de Medio Ambiente2001), which justifies its inclusion in this study.

The Tertiary Detrital Aquifer is the most importantaquifer in the region (Fig. 1) with an average thickness of1500 meters, variable transmissivity (1 – 852 m2 day−1)and renewable resources range from 130 × 106 to 200 ×106 m3 year−1 (IGME 1985). This is a heterogeneous andanisotropic aquifer composed of detrital deposits from theMiocene age, which have formed a silt-clay system of lowpermeability surrounding small and permeable arkosicsand lenses (IGME 1981). Recharge is mainly associatedwith the infiltration of rainwater in the Tertiary interfluvearea (Llamas and López Vera 1975) and discharge occursat the valley bottoms (IGME 1981; Fig. 1). This aquifer isalso hydraulically connected to some parts of theQuaternary Aquifer associated with the rivers Jaramaand Henares (IGME 1981; Fig. 1).

The Moor Limestone Aquifer (Fig. 1) is a free andisolated aquifer with an average thickness of 100–190 m.Transmissivity ranges from 90 to 550 m2 day−1 and therenewable resources are 122 × 106 m3 year−1 (IGME1985). Recharge occurs directly from rainfall infiltrationand discharge takes place through numerous springs andthe Quaternary Aquifer, which is connected with theTajuña River and its tributaries (IGME 1981; Fig. 1). Dueto fluvial erosion, Tajuña River valley has divided thisaquifer into two separate and low-thickness sub-units(IGME 1981). These groundwater resources have tradi-tionally been used to supply local demands (Alcolea andGarcía Alvarado 2006).

The Cretaceous Limestone Aquifer (Fig. 1) is composedof compact limestone rocks with pores and cracks caused bythe solvent water effect. This has produced channels ofvarious sizes, caves and karst structures. Its average thick-ness is 130 m and transmissivity ranges from 800 to1,000 m2 day−1 (IGME 1985). Recharge is by raininfiltration, lateral groundwater flow from surroundingsystems, and water draining from the bottom of rivers andstreams. Discharge is through springs along the course of theTajuña River and its tributaries (Fig. 1) and also through theextraction of water to supply the local population (Comuni-dad de Madrid 2001). Renewable resources are estimated at13 × 106 m3 year−1 (IGME 1985).

The rest of the groundwater resources in the study areaare small aquifers with low transmissivity and storagecapacities. The most interesting of these is the QuaternaryAquifer associated with the Jarama River and its tribu-taries (Fig. 1). This aquifer is formed by alluvial deposits(in valleys and on the first terraces), and is generallycomposed of gravel and sand with a variable content ofsilt-clay particles (IGME 1981). Its thickness rarelyreaches 10 m, except along some parts of the JaramaRiver, where it exceeds 40 m (Alcolea and GarcíaAlvarado 2006). Its transmissivity is high (1,500–4,000 m2 day−1) and the average renewable resources are

111 × 106 m3 year−1 (IGME 1985). This system receiveshigh contributions of nitrogen compounds from agricul-tural activities (Ministerio de Medio Ambiente 2001) andis connected to the underlying Tertiary aquifers (IGME1981) and to surface waters by its associated rivers(Arauzo et al. 2008).

The low permeability of the Tertiary Evaporite System(marl and gypsum facies), located in the eastern part of theComunidad de Madrid, gives it aquiclude characteristics(Fig. 1). The natural quality of this groundwater resourceis not good and these waters cannot be used to supplyhuman needs due to the high soluble salt concentrationresulting from gypsum dissolution (Alcolea and GarcíaAlvarado 2006). Finally, in the western part of theComunidad de Madrid, there are aquifuges formed byfissured igneous and metamorphic formations from thePaleozoic age. These include granitic rocks interlayeredwith schist and mica rocks (Fig. 1).

Soil typesThe main types of soil in the study area are Fluvisol,Luvisol and Cambisol (Monturiol and Alcalá 1990).Fluvisols are usually located on alluvial deposits associ-ated with the Quaternary Aquifer (Monturiol and Alcalá1990). These are relatively young and highly permeablesoils and play an important role in agricultural activities.The typical textural composition of the dominant Fluvisolsin the study area corresponds to loam, sandy loam andsandy-clay loam soils. Luvisols are usually located onthe arkosic sands of the Tertiary Detrital Aquifer, on theterrace sediments of the alluvial systems and on thelimestone rocks of the Moor Limestone Aquifer. They areusually used for cereal cultivation and are characterized bya clay B horizon that provides them with a great water-holding capacity (Monturiol and Alcalá 1990). Theirtypical texture varies from clay loam, mainly correspond-ing to the alluvial terraces and the northeast of the arkosicsands, and sandy loam, in the southwest. Cambisols arelocated on the Palaeozoic rocks (granitic, schist, shale,slate, quartz and mica rocks) from mountainous areas andin zones of arkosic facies (Monturiol and Alcalá 1990).Their water-holding capacity is medium, with intermediateyields between sandy and clay soils (Monturiol and Alcalá1990). The dominant Cambisols in the study area areclassified as sandy loam soils with an elevated fraction ofgravel and stones. Finally, Gypsisols dominate the lime-stone and gypsum deposits in the southern part of theregion (Monturiol and Alcalá 1990). They are usually clayloam soils with a high content of gravel or stones in somecases.

Land usesFigure 2 shows the Comunidad de Madrid CORINE LandCover map 2000 (European Environmental Agency 2000).There is a high level of urban development in the centralarea, corresponding to the city of Madrid and itsmetropolitan area (which is in some areas corresponding

684

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 5: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

to the Tertiary Detrital Aquifer, the Quaternary Aquiferand the Tertiary Evaporite System). The implications ofhigh-population density are rapid urbanization, the large-scale development of industrial and commercial sectors,and a high volume of urban and industrial waste water andsewage (Alcolea and García Alvarado 2006). Agriculturalzones cover about 38% of the land surface of theComunidad de Madrid (305,201 ha), of which 54%corresponds to non-irrigated field crops and 10% toirrigated field crops (Observatorio de la Sostenibilidad enEspaña 2006). The rest of the agricultural surfacecorresponds to field crops mixed with natural areas andzones with a mixture of irrigated and non-irrigated fieldcrops. The main non-irrigated crops are cereals, vineyards,olive trees and aromatic plants. Most of the irrigated cropsin the region are present on the alluvial deposits of theQuaternary Aquifer, with corn as the dominant crop.Farmers in central Spain use a traditional irrigationschedule such as surface-furrow irrigation (Román et al.1999). Semi-natural areas, uncultivated land (shrubberyand agroforestry systems) and forests are present in thenorthwest of the area.

Materials and methods

Intrinsic vulnerability

The DRASTIC indexThe DRASTIC index (Aller et al. 1987) is based on theevaluation of seven hydrogeological parameters related to

groundwater vulnerability: depth to the water table (D),net aquifer recharge (R), aquifer media (A), soil media (S),topography slope (T), impact of the vadose zone (I) andhydraulic conductivity (C). Two values are assigned toeach parameter: a relative weight (w, on a scale from 1 to5; Table 1), according to the relative impact on potentialpollution in comparison with the rest of the parameters,and a rating (r, on a scale from 1 to 10), according to themagnitude of each parameter in the different zones of thestudy area (Aller et al. 1987).

The DRASTIC index ranges from 23 to 230 and iscalculated according to the following equation:

DRASTIC index ¼ Dw � Dr þ Rw � Rr þ Aw � Arþ Sw � Sr þ Tw � Tr þ Iw � Irþ Cw � Cr ð1Þ

Where,

Dw Relative weight of the depth to the water tableDr Rating of the depth to the water tableRw Relative weight of the net aquifer rechargeRr Rating of the net aquifer rechargeAw Relative weight of the aquifer mediaAr Rating of the aquifer mediaSw Relative weight of the soil mediaSr Rating of the soil mediaTw Relative weight of the topography slopeTr Rating of the topography slope

Fig. 2 Comunidad de Madrid CORINE Land Cover map 2000 (Source: European Environmental Agency 2000)

685

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 6: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

Iw Relative weight of the impact of the vadose zoneIr Rating of the impact of the vadose zoneCw Relative weight of the hydraulic conductivityCr Rating of the hydraulic conductivity

Aller et al. (1987) did not propose any classification fortheir DRASTIC results, so the vulnerability ranges of theDRASTIC index used in this study correspond to the mostcommonly used references in the literature (Civita and DeRegibus 1995; Corniello et al. 1997; Table 2). Thisclassification is the result of dividing the final index intovulnerability classes such as low, moderate, high, and veryhigh potential. Civita and De Regibus (1995) applied it forthe ISIS index, a hybrid method developed taking intoaccount the rating and weighting systems of DRASTICand SINTACS (Civita 1994) methods, and the GODmethod for the general structure design.

Seven raster maps were generated for each DRASTICindex parameter and were combined by overlayingaccording to the index equation. Geographic informationsystem ArcGIS 9.2 (Environmental Systems ResearchInstitute 2006) was used for this task, employing a meshcomposed of 142 rows and 130 columns and a 1,000 ×1,000-m pixel resolution. Depth to the water table in theTertiary aquifers and in the Cretaceous Limestone Aquiferwas obtained from average values of long-term data(1985–2003) associated with the piezometric samplingnetwork of the Confederación Hidrográfica del Tajo (CHT,Tajo River Water Authority, unpublished data, 2004) witha total of 29 sampling points throughout the area. Depth tothe water table in the Quaternary Aquifer was obtainedfrom Arauzo et al. (2008), using average values from twosampling events performed during March and August2005, with information corresponding to 17 samplingpoints over the alluvial deposits. ArcGIS 9.2 was used asthe interpolation tool to perform parameter estimationthroughout the study area. Topographic information wasobtained from the digital elevation model of the Comuni-dad de Madrid using ArcGIS 9.2 to calculate slopes. Datarelating to lithology were obtained from IGME (1988),while the source of the soil data was Monturiol and Alcalá(1990). The hydraulic conductivity and net aquiferrecharge (annual renewable resources) for each aquiferwere obtained from IGME (1985, 1993). Net rechargevalues for irrigated field crops were corrected in order to

incorporate return flows which were estimated using theaverage volume of irrigation applied in the Comunidad deMadrid (7,000 m3 ha−1; IGME 1985) and the percentageof irrigation water that drains to the aquifer with theconventional farming methods used in the central Jaramabasin (20% of applied irrigation water; Román et al.1996). Table 3 shows the ranges and ratings assigned toeach parameter for the DRASTIC index.

The GOD indexThe God index (Foster and Hirata 1991; Foster et al.2002) is calculated by assigning ratings from 0 to 1 for thefollowing three parameters: groundwater confinement (G),overlying strata (O) and depth to groundwater (D). Thisindex can vary from 1 (greatest vulnerability) to 0 (leastvulnerability) according to the following equation:

GOD index ¼ G � O � D ð2Þ

The vulnerability ranges of the GOD index (Table 4)and the ranges and ratings for each parameter in eachaquifer (Table 5) were designed according to Foster et al.(2002). Vulnerability mapping of the GOD index wasperformed according to the same methodology used in theDRASTIC index calculation (overlying of thematic mapsusing ArcGIS 9.2). The data sources were the same asthose used with the DRASTIC index.

Specific vulnerability to nitrate pollution

The composite DRASTIC index (additive method)The Composite DRASTIC index (CD index; Secunda etal. 1998) is an adaptation of the DRASTIC index based onthe addition of a new parameter defining the potential riskassociated with land use (L). The objective of thisapproach is to evaluate the potential effect of extensiveland use upon groundwater quality resulting from alter-ation over time of the soil matrix and unsaturated zonemedia. This was applied by Secunda et al. (1998) to assessthe potential level of groundwater vulnerability to pollu-tion in Israel’s Sharon region. Al-Adamat et al. (2003)also applied this model to assess groundwater vulner-ability and produce risk maps for an aquifer of the Azraqbasin of Jordan.

Table 1 Relative weights given to the DRASTIC parameters (Source:Aller et al. 1987)

Parameters Relative weight

Depth to the water table 5Impact of the vadose zone 5Net aquifer recharge 4Aquifer media 3Hydraulic conductivity of the aquifer 3Soil media 2Topography slope 1

Table 2 Vulnerability ranges corresponding to the DRASTIC index(Sources: Civita and De Regibus 1995; Corniello et al. 1997)

Vulnerability Ranges (DRASTIC index)

Very low <80Low 80–120Moderate 120–160High 160–200Very high ≥200

686

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 7: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

The specific vulnerability to nitrate pollution rangesfrom 28 to 280 according to this index and is calculatedusing the following equation:

CD index ¼ Dw � Dr þ Rw � Rr þ Aw � Ar þ Sw � Srþ Tw � Tr þ Iw � Ir þ Cw � Cr þ Lw � Lr ð3Þ

Where,

Lw Relative weight of the potential risk associated withland use

Lr Rating of the potential risk associated with land use

The rest of the parameters are the same as in Eq. (1).The map of the potential risk associated with land use

(CORINE Land Cover 2000: European EnvironmentalAgency 2000) was produced using the same methodologyapplied to the rest of the DRASTIC index parameters.Table 6 shows the ratings assigned to the potential riskassociated with land use (L) in the CD index, according toSecunda et al. (1998). These ratings characterize extensiveland uses as potential sources of groundwater pollution.The greatest impacts corresponded to irrigated field cropsand urban areas, with scores twice as high as for non-irrigated field crops. The lowest impact corresponded tonatural areas such as forests. The intrinsic vulnerabilitymap based on the DRASTIC index and the potential riskassociated with land use map were combined usingArcGIS 9.2 to generate a new map showing specific

vulnerability to nitrate pollution according to the CDindex. The vulnerability ranges for the CD index areshown in Table 7.

Nitrate vulnerability index (multiplicative model)This is a new specific vulnerability index which isproposed as another adaptation of the DRASTIC indexand has been developed with the objective of achievinggreater accuracy in the estimation of specific vulnerabilityto nitrate pollution, based on the real impact of each landuse. The model attempts to integrate the risks of ground-water pollution by nitrate related to land uses (as apotential source of nitrogen). It incorporates both thenegative impacts, over time, of some of these uses andalso the protective effects that others may have uponaquifer media (uses that do not contribute significantquantities of nitrate and do not enhance leaching, such asthe protected natural areas). This is possible because it isbased on a multiplicative model, involving the addition of

Table 3 Ranges and ratings of the DRASTIC parameters (Source: Aller et al. 1987)

Depth to the watertable (m)

Net recharge (mm) Topography slope (%) Hydraulic conductivity (m day–1)

Range Rating Range Rating Range Rating Range Rating

0.0–1.5 10 0–51 1 0–2 10 0–4.1 11.5–4.6 9 51–102 3 2–6 9 4.1–12.2 24.6–9.1 7 102–178 6 6–12 5 12.2–28.5 49.1–15.2 5 178–254 8 12–18 3 28.5–40.7 615.2–22.9 3 >254 9 >18 1 40.7–81.5 822.9–30.5 2>30.5 1Soil media Aquifer media Impact of the vadose zoneRange Rating Range Ratinga Range Ratinga

Thin or absent 10 Massive shale 1–3 (2) Confining layer 1Gravel 10 Metamorphic/igneous 2–5 (3) Silt/clay 2–6 (3)Sand 9 Weathered

metamorphic/igneous3–5 (4) Shale 2–6 (3)

Peat 8 Glacial till 4–6 (5) Limestone 2–5 (3)Shrinking and/oraggregated clay

7 Bedded sandstone,limestone and shalesequences

5–9 (6) Sandstone 2–7 (6)

Loam 5 Massive sandstone 4–9 (6) Bedded limestone, sandstone and shale 4–8 (6)Silty loam 4 Massive limestone 4–9 (8) Sand and gravel with significant silt and clay 4–8 (6)Clay loam 3 Sand and gravel 4–9 (8) Sand and gravel 4–8 (8)Muck 2 Basalt 2–10 (9) Basalt 2–10 (9)Non-shrinking andnon-aggregated clay

1 Karst limestone 9–10 (10) Karst limestone 8–10 (10)

a Typical ratings in parentheses according to Aller et al. (1987)

Table 4 Vulnerability ranges corresponding to the GOD index(Source: Foster et al. 2002)

Vulnerability Ranges (GOD index)

Negligible 0–0.1Low 0.1–0.3Moderate 0.3–0.5High 0.5–0.7Extreme ≥0.7

687

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 8: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

a new parameter called the “potential risk associated withland use” (LU), which is calculated according to thefollowing equation:

NV index ¼ ðDw � Dr þ Rw � Rr þ Aw � Ar þ Sw � SrþTw � Tr þ Iw � Ir þ Cw � CrÞ � LU

ð4Þ

Where,

NV index Nitrate vulnerability indexLU Potential risk associated with land use

The rest of the parameters are the same as in Eq. (1).The NV index ranges from 23 to 230. The ratings

applied to the potential risk associated with land use(LU) are shown in Table 8. LU ranges from 0.1 to 1.0,with the lowest values corresponding to areas with landuses that “protect” water resources from nitrate pollu-tion and the highest values associated with areas withland uses that contribute significant amounts of nitrate.It can be observed that natural areas, which do notusually have an impact on aquifer quality, receive alow rating (0.2). It should, however, be noted that thisrating does not correspond to the lowest possible valueof the LU parameter due to the presence of nitrogeninputs such as atmospheric deposition (Heuer et al.1999; Lin et al. 2000). The lowest value (0.1) would beassigned to large natural areas (that are free for anyanthropic influences in their surrounding areas) andwhich hardly receive any nitrogen contribution. Theratings applied to the potential risk associated withagricultural land uses were estimated in a directly

proportional relationship to the amount of excessnitrogen (in kg N ha−1) associated with the calculatednitrate balance for different agricultural activities inSpanish agricultural systems in 2005 (Ministerio deAgricultura, Pesca y Alimentación 2007; Table 8). Thegreatest amount of excess nitrogen was associated withirrigated field crops, with almost 83 kg N ha−1 (whichis consistent with available information about the roleof intensive agriculture as the main source of nitratepollution; Guimera 1993; Cabrera et al. 1995; Arrateet al. 1997; Sanchez-Pérez et al. 2003). Therefore, thehighest rating was assigned (1.0) to this land use. In thecase of non-irrigated field crops, levels of excessnitrogen were around 39 kg N ha−1; about half of thetotal contribution from irrigated land. The quantity ofexcess nitrogen from uncultivated land and semi-naturalareas, including agroforestry systems and grazing areas,was around 10 kg N ha−1. For these reasons, in the LUparameter, the different levels of excess nitrogencorrespond to proportional ratings of 0.6 and 0.3,respectively. The rating applied to urban areas dependson population density and on the possible existence ofdischarges and leakage from sewer networks servingmetropolitan areas. Regional studies indicate that urbanuses are the other main nitrate source in the studyarea (Hernández-García and Custodio 2004; IGME 1985;Comunidad de Madrid 1995, 2001). The existence ofimportant leakages from sewer networks serving the

Table 5 Ranges and ratings of the GOD parameters (Source: Fosteret al. 2002)

Groundwater confinement Depth to groundwaterRange Rating Range Rating

None 0 All depthsa 0.9Overflowing 0 <5 m 0.9Confined 0.2 5–20 m 0.8Semi-confined 0.4 20–50 m 0.7Uncovered (confined) 0.6 >50 m 0.6Unconfined 0.7–1.0Overlying strataRange RatingEstuarine clays <0.4Residual soils 0.4Alluvial silt, loess, glacial till 0.5Mudstones 0.5Shales 0.5Aeolian sands 0.6Siltstones 0.6Igneous/metamorphic formations and older volcanics 0.6Volcanic tuffs 0.6–0.7Alluvial and fluvio-glacial sands 0.7Alluvial fan gravels 0.8Sandstones 0.7–0.8Recent volcanic lavas 0.8Chalky limestone calcarenites 0.9Calcretes + karst limestones 0.9–1.0

a Karst limestones

Table 6 Ranges and ratings applied to the potential risk associatedwith land use (L) according to the CD index (Source: Secunda et al.1998)

Potential risk associated with land use (L)Range Rating

Site-specific land usageToxic-waste disposal 9Oil spillage 7Industries 6Solid-waste disposal (regional) 6Domestic-waste disposal (local) 5Effluent irrigated fields 4Effluent reservoirs 3Extensive land usageCotton 10Urban areasa 8Irrigated field cropsa 8Greenhouses/tomatoes 8Citrus orchards 7Orchards of other fruit 6Pasture or other land unsuitable for agricultural usea 5Uncultivated landa 5Temporarily uncultivated landa 5Vineyards 5Olives 5Quarries 5Non-irrigated field cropsa 4Avocados 2Forestsa 1Natural areas or reservesa 1Dune sands–Open areas 1

aMain land uses observed in the Comunidad de Madrid

688

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 9: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

metropolitan area of Madrid (IGME 1993) generatesunquantified pollution events. Dilution processes fromleakages from supply networks (not quantified as well)should also be considered. As there were not enoughavailable data to make an accurate estimation, therating assigned (0.8) was an intermediate valuebetween the highest rating (for irrigated field crops)and the moderate rating assigned to non-irrigated fieldcrops. The specific vulnerability ranges established forthe NV index are shown in Table 7.

Nitrate pollution mapThe map showing nitrate pollution of groundwater in theComunidad de Madrid was generated from a long-termdata set (1985–2005, although data were not complete forall the years) showing nitrate concentrations in theTertiary aquifers and in the Cretaceous Limestone Aquiferand from data for the year 2005 for the QuaternaryAquifer (because there was no other informationavailable). In the first case, the data corresponded tothe Groundwater Quality Sampling Network of theTertiary aquifers (CHT, Tajo River Water Authority,unpublished data, 2005) and related to a total of 29sampling points (wells and boreholes) located abovethese aquifers (Fig. 3). In the second case, the datawere obtained from the Groundwater Quality SamplingNetwork of the Comunidad de Madrid QuaternaryAquifer (Arauzo et al. 2008), relating to 17 points(Fig. 3). For the Tertiary aquifers and the CretaceousLimestone Aquifer, the sampling points were mainlyboreholes, with depths ranging from 10 to 200 m(depending on the thickness of the aquifers); however,in the case of the Moor Limestone Aquifer, most of thesampling points were springs. For the QuaternaryAquifer, the sampling points were shallow wellsscattered across the alluvial area and mainly used foragricultural irrigation. This aquifer was sampled inMarch and August 2005. For the Tertiary aquifers andthe Cretaceous Limestone Aquifer, data were availablefrom two sampling events per year in most cases. Thesesamples were mainly taken in April or May and inNovember or December. In all cases (except in someboreholes of the Tertiary Detrital Aquifer), sampleswere taken from the upper layer of the aquifer inquestion.

The nitrate pollution map was based on average valuesfor all of the measurements of nitrate concentration

recorded for each sampling point (as there were nosignificant intra-annual variations). ArcGIS 9.2 was usedto interpolate the nitrate concentrations used to generatethe nitrate pollution map. Only one of the sampling pointsfor the Cretaceous Limestone Aquifer was located withinthe administrative boundaries of the Comunidad deMadrid, so three additional sampling points were alsoused, located in the Comunidad de Castilla La Mancha,when making interpolations. Nitrate concentration classeswere defined according to the official limits established byEuropean regulations (a guide level of 25 mg L−1 forEU Directive 80/778/EEC (Council of the EuropeanCommunities 1980), relating to the quality of waterintended for human consumption, and a maximumadmissible concentration of 50 mg L−1 for EU Directive91/676/EEC).

Statistical analysisA Pearson correlation matrix was generated in order tocompare the four vulnerability indexes and to evaluate theconsistency of each index with respect to the spatialdistribution of nitrate pollution. A total of 37 samplingpoints were used to construct the correlation matrix.Sampling points removed from the matrix correspondedto boreholes of the Tertiary aquifers located over thesurface of the Quaternary Aquifer. This is because of theproblem for assigning to them vulnerability valuesthrough the maps.

Results

Intrinsic vulnerability of groundwater accordingto the DRASTIC and GOD indexesFigures 4 and 5 show intrinsic vulnerability maps obtainedby application of the DRASTIC and GOD indexes,respectively. A great similarity can be observed in thedistribution of the vulnerable zones recognized by bothindexes. In addition, there was a positive correlation witha high statistical significance (p<0.001) between theDRASTC index and the GOD index (Table 9). TheQuaternary Aquifer was the system with the highestintrinsic vulnerability values, corresponding to high

Table 7 Vulnerability ranges corresponding to the CD index(Source: Secunda et al. 1998) and the NV index

Vulnerability Ranges (CD index) Ranges (NV index)

Very low <100 <70Low 100–145 70–110Moderate 145–190 110–150High 190–235 150–190Very high ≥235 ≥190

Table 8 Ranges and ratings applied to the potential risk associatedwith land use (LU) as a source of nitrate pollution for the NV index.Sources used to assign the ratings are shown

Potential risk associated with land use (LU)Range Rating Source

Irrigated field crops 1.0 MAPA (2007)Urban areas 0.8 Arauzo et al. (2008);

IGME (1993)Non-irrigated field crops 0.6 MAPA (2007)Uncultivated land,semi-natural areas

0.3 MAPA (2007)

Forests, natural areas 0.2 Heuer et al. (1999) ;Lin et al. (2000)

689

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 10: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

Fig. 3 Location of the sampling points in the main aquifers of central Spain (Sources: CHT 2005; Arauzo et al. 2008). The aquifers aredescribed as in Fig. 1

Fig. 4 Thematic maps corresponding to each DRASTIC parameter and intrinsic vulnerability map for groundwater in central Spain,according to the DRASTIC method. The parameters used for the DRASTIC index are: depth to the water table (D), net aquifer recharge (R),aquifer media (A), soil media (S), topography slope (T), impact of the vadose zone (I), hydraulic conductivity of the aquifer (C). Thevulnerability classes of the DRASTIC index were designed according to Civita and De Regibus (1995)

690

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 11: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

vulnerability levels according to both indexes. The maindifferences between the two indexes are observed inkarstic aquifers (Cretaceous Limestone Aquifer and MoorLimestone Aquifer), which were classified as exhibitingmoderate vulnerability according to DRASTIC and highvulnerability according to GOD. On the other hand, theTertiary Detrital Aquifer showed low vulnerability accord-ing to DRASTIC and moderate vulnerability according toGOD. Intrinsic vulnerability in the Paleozoic formationsand the Tertiary Evaporite System was very low accordingto DRASTIC and low, according to GOD.

Specific vulnerability of groundwater to nitratepollution according to the CD and NV indexesFigures 6 and 7 show the maps for specific vulnerability tonitrate pollution according to the CD and NV indexes,respectively. There was a positive correlation with a high

statistical significance (p<0.001) between the two indexes(Table 9). There was also significance (p<0.001) in thepositive correlation coefficient obtained between the twoindexes and the intrinsic vulnerability indexes (Table 9).

According to the CD index, the areas with the highestrisk of nitrate pollution of groundwater (high and veryhigh) were situated in the Quaternary Aquifer (Fig. 6),where irrigated field crops and urbanization are the mainland uses (see Fig. 2). Specific vulnerability was usuallylow in the Tertiary Detrital Aquifer, although there werealso some moderate risk zones corresponding to urbanterritories and their respective areas of influence (e.g.metropolitan area of Madrid), whose unsaturated zones areless thick (see Fig. 4). Specific vulnerability was mainlylow in the Cretaceous Limestone Aquifer, while it rangedfrom low to moderate in the Moor Limestone Aquifer,with the highest vulnerability zones corresponding mainlyto the lower sub-unit of the aquifer, where intensiveagriculture covers most of the surface (see Fig. 2). Specificvulnerability to nitrate pollution in Paleozoic formationsand in the Tertiary Evaporite System ranged from low tovery low according to the CD index.

The NV index (Fig. 7), unlike the CD index, allowedimproved accuracy in estimations of the vulnerability ofthe Quaternary Aquifer, according to the real impact ofeach land use over its total area. According to the NVindex, the zones with the highest risk of nitrate pollution(high and very high) corresponded to the middle andlower parts of the Quaternary Aquifer (the Jarama,Henares, Tajuña, Manzanares and Tajo river basins),where irrigated agriculture is the main land use. Nitrogenpollution from urban sources also affects the groundwater

Fig. 5 Thematic maps corresponding to each GOD parameter and intrinsic vulnerability map for groundwater in central Spain, accordingto the GOD method. The parameters used for the GOD index are: groundwater confinement (G), overlying strata (O), depth to groundwater(D). The vulnerability classes of the GOD index were designed according to Foster et al. (2002). Specific vulnerability of groundwater tonitrate pollution according to the CD and NV indexes

Table 9 Pearson correlation matrix using the mean nitrate concen-tration of groundwater in the Comunidad de Madrid (Spain), DR-ASTIC index, GOD index, CD index and NV index. Number ofdata points is equal to 37

[NO3−] DRASTIC GOD CD

DRASTIC 0.47a

GOD 0.46a 0.74b

CD 0.51a 0.99b 0.73b

NV 0.66b 0.77b 0.55b 0.83b

a Statistical significance at p<0.01b Statistical significance at p<0.001

691

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 12: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

in these middle and lower parts of the Jarama, Henaresand Manzanares rivers alluvial deposits (see Fig. 2).However, the upper parts of the Quaternary Aquifer(corresponding to the alluvial deposits of the Jarama andManzanares rivers), were classified as exhibiting lowvulnerability zone according to the NV index (seeFig. 2), because the dominant land uses are non-irrigatedfield crops, uncultivated land and semi-natural areas. Thespecific vulnerability in the area corresponding to the

Cretaceous Limestone Aquifer and the Tertiary DetritalAquifer ranged usually from low to very low, presenting aslightly higher risk (moderate) in zones of urban influence.Specific vulnerability in the area corresponding to theMoor Limestone Aquifer generally ranged from low (inthe northern part) to moderate (in the southern part), thisbeing largely determined by the distribution of irrigatedland (see Fig. 2). Specific vulnerability to nitrate pollutionwas very low in the areas corresponding to Paleozoic

Fig. 7 Map of specific vulnerability to nitrate pollution for groundwater in central Spain according to the NV index; a thematic map of thepotential risk associated with land use (LU) is also shown

Fig. 6 Map of specific vulnerability to nitrate pollution for groundwater in central Spain according to the CD index (Secunda et al. 1998);a thematic map of the potential risk associated with land use (L) is also shown

692

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 13: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

formations and to the Tertiary Evaporite System, accord-ing to the NV index.

Groundwater nitrate pollution and vulnerabilityindexesTable 10 shows average values and standard deviations fornitrate concentrations at each sampling point on theaquifers in the Comunidad de Madrid. The standarddeviations show only a low level of variation at most ofthe sampling points. Furthermore, some of the 12

sampling points that showed high standard deviationscorresponded to highly polluted areas in which variationsdid not signify changes in their status as polluted sites.Figure 8 shows the groundwater nitrate pollution map forthe Comunidad de Madrid. The most polluted areas can beobserved in the middle and lower parts of the QuaternaryAquifer (Jarama, Henares, Tajuña, Manzanares and Tajoriver basins), with nitrate concentrations in some areas thatwere much higher than the limit established by the EUNitrate Directive 91/676/EEC. In the Tertiary aquifers, thehighest values corresponded to the Moor LimestoneAquifer and to isolated areas at the northern and southernends of the Tertiary Detrital Aquifer. The lowest valuescorresponded to the Cretaceous Limestone Aquifer, to thenorthern sector of the Tertiary Detrital Aquifer, and to theupper parts of the Quaternary Aquifer.

Positive coefficient correlations were obtained betweengroundwater nitrate concentration in the study area andDRASTIC, GOD and CD index values, with statisticalsignificance at p<0.01. The NV index proposed by theauthors, based on a multiplicative model, showed betterstatistical significance (p<0.001) than the other vulner-ability indexes in relation to nitrate distribution (Table 9).

Discussion

The application of vulnerability indexes on a regionalscale in central Spain has permitted the identification anddescription of vulnerable areas in the main aquiferslocated in Comunidad de Madrid (Figs. 4–7; Table 9).The results of this study highlight the usefulness of theseintrinsic (DRASTIC and GOD) and specific (CD and NV)vulnerability indexes in studies to evaluate groundwatervulnerability and are consistent with the observations ofSecunda et al. (1998), McLay et al. (2001), Al-Adamat etal. (2003) and Debernardi et al. (2007). The greatsimilarity observed between maps based on the GODand DRASTIC indexes and the high correlation betweenthese indexes indicate that these two methods could beequally useful for evaluating intrinsic vulnerability on aregional scale. This similarity could be explained by thefact that, on a regional scale, the vulnerability indexeswere over-influenced by the lithological characteristic.Thus, the two parameters of the DRASTIC index withhigher ratings (impact of the vadose zone and depth towater table) represent the major influence in the finalvulnerability value, which correspond to two of the threeparameters of the GOD index (overlying strata and depthto groundwater). Moreover, recharge, a parameter incor-porated in the DRASTIC index, shows problems with itsestimation on a regional scale, once all anthropogenicactivities change the aquifer recharge. These are the mainreasons why the simplicity of the GOD method probablymakes it more attractive to use on a regional scale.However, the DRASTIC index could be more accurate instudies developed on a local scale or when more detaileddata are available (Gogu and Dassargues 2000).

Table 10 Mean nitrate concentration and standard deviations foreach sampling point. Average values from long-term data (1985–2005) were used for the Tertiary aquifers and the Cretaceous Lim-estone Aquifer, and average values of data for 2005 were used forthe Quaternary Aquifer. n number of available data for each samp-ling point

Sampling point Aquifer n Nitrate(mg L−1)

1 Quaternary Aquifer 1 02 Quaternary Aquifer 2 89±33 Quaternary Aquifer 1 14 Quaternary Aquifer 2 10±135 Quaternary Aquifer 2 32±36 Quaternary Aquifer 2 50±27 Quaternary Aquifer 2 58±218 Quaternary Aquifer 2 70±139 Quaternary Aquifer 2 31±3410 Quaternary Aquifer 2 64±311 Quaternary Aquifer 2 13±512 Quaternary Aquifer 2 85±2613 Quaternary Aquifer 2 21±814 Quaternary Aquifer 2 30±315 Quaternary Aquifer 2 16±2116 Quaternary Aquifer 2 68±117 Quaternary Aquifer 2 103±2218 Cretaceous Limestone Aquifer 23 14±719 Tertiary Detrital Aquifer 28 14±420 Tertiary Detrital Aquifer 21 8±421 Tertiary Detrital Aquifer 24 64±2822 Tertiary Detrital Aquifer 24 14±523 Tertiary Detrital Aquifer 25 23±1024 Tertiary Detrital Aquifer 27 40±1925 Tertiary Detrital Aquifer 20 41±1626 Tertiary Detrital Aquifer 26 13±2127 Tertiary Detrital Aquifer 27 60±2228 Tertiary Detrital Aquifer 27 54±3529 Tertiary Detrital Aquifer 28 23±630 Tertiary Detrital Aquifer 28 11±331 Tertiary Detrital Aquifer 26 5±532 Tertiary Detrital Aquifer 26 25±833 Tertiary Detrital Aquifer 26 58±1834 Tertiary Detrital Aquifer 26 34±735 Tertiary Detrital Aquifer 25 23±4536 Tertiary Detrital Aquifer 26 20±837 Tertiary Detrital Aquifer 27 27±638 Tertiary Detrital Aquifer 25 39±1739 Tertiary Detrital Aquifer 26 77±2940 Moor Limestone Aquifer 26 33±341 Moor Limestone Aquifer 23 31±442 Moor Limestone Aquifer 22 40±543 Moor Limestone Aquifer 24 26±344 Moor Limestone Aquifer 26 31±345 Moor Limestone Aquifer 25 89±2846 Moor Limestone Aquifer 22 55±5

693

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 14: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

As expected, specific vulnerability indexes producedhigher coefficients of correlation for nitrate concentrationsin groundwater because of the incorporation of the“potential risk associated with land use” parameter thatconsiders potential inputs of nitrogen associated withdifferent land uses. The NV index was more consistentthan the CD index with respect to the spatial distributionof nitrate in the study area (Figs. 6 and 7; Table 9). Thisresult could be explained by the improvement achieved inestimating the specific vulnerability according to theimpact of each specific land use by applying the multi-plicative model of the NV index (Eq. 4). The intrinsicvulnerability values tend to be smoothed using this model,in areas in which land uses do not have any adverseeffects and therefore “protect” water resources. In the caseof the Quaternary Aquifer, whose alluvial nature results inhigh intrinsic vulnerability on all of its extension, it iseasier to appreciate this effect (Figs. 4 and 5). Applyingthe NV index allowed one to distinguish low specificvulnerability zones in the upper parts of the alluvialsystem (traditionally associated with uses with a low riskof pollution; Figs. 2 and 7) from zones in the middle-lower parts of the same system with the highest levels ofspecific vulnerability (associated with irrigated land uses;Figs. 2 and 7). The CD index, based on an additive model,does not achieve that level of accuracy which explains thelower statistical significance obtained from its correlationwith nitrate concentration in the study area (Table 9).

Both maps of intrinsic and specific vulnerability for theComunidad de Madrid (Figs. 4 and 5) show the zones ofthe Quaternary Aquifer (medium-lower part, according tothe NV index) and the lower sub-unit of the MoorLimestone Aquifer as the most vulnerable areas. Theseresults are consistent with the distribution of the mainpolluted areas (nitrate concentrations higher than 50 mgL−1) in the nitrate pollution map (Fig. 8). Nevertheless, itis necessary to analyse these results carefully as contam-ination from surface sources regularly generate stratifica-tions and production wells with long screen sectionsfrequently take water from different levels and mix them.Anyway, the nitrate pollution map (Fig. 8) also showsconsistency with previous information about nitrate dis-tribution in aquifers of the Comunidad de Madrid (IGME1985; Consejería de la Comunidad de Madrid 2001; CHT2005; Martínez-Bastida et al. 2006; Arauzo et al. 2008).The significant proportion of irrigated field crops corre-sponding to the middle and lower parts of the QuaternaryAquifer results in a high risk of diffuse nitrate pollutionaffecting the underlying aquifer. This is the result ofnitrogen fertilization, poorly optimized irrigation techni-ques (Arauzo et al. 2008), the high permeability of thealluvial sediments and the shallow depth to water table—all of which encourage the development of nitrate leachingprocesses. Furthermore, the risk of nitrate pollutionincreases with irrigation return flows that proceed fromfluvial waters used in irrigation systems that contain high

Fig. 8 Map of nitrate pollution of groundwater in the Comunidad de Madrid. Average values obtained from long-term data (1985–2005)were used for the Tertiary aquifers and the Cretaceous Limestone Aquifer and average values of data for 2005 were used for the QuaternaryAquifer

694

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 15: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

levels of nitrate from urban sources (Arauzo et al 2008).There is also an important development in irrigated fieldcrops in the lower sub-unit of the Moor LimestoneAquifer (Fig. 2), which is consistent with the vulnerabilityestimations obtained.

In the Tertiary Detrital Aquifer the risk of nitratepollution is low, although there are small zones withmoderate specific vulnerability (Figs. 6 and 7), mainlyassociated with urban and industrial activities (Fig. 2), thatare consistent with the highest nitrate concentration areas(Fig. 8). There is, however, a considerable discrepancy inthe southern part of the Tertiary Detrital Aquifer, where apolluted area can be observed (Fig. 8) that is notconsistent with the degree of vulnerability projected byany of the vulnerability maps (Figs. 4–7). There are twopossible reasons for this: one is the lack of complete landuse information for the area corresponding to the southernpart of the Tertiary Detrital Aquifer (resulting in lessaccurate mapping of potential risk associated with landuse, which was used to estimate specific vulnerability inthis area); and the other, which seems more likely,involves the possible movement of nitrate, carried bygroundwater flows, from the highest risk areas (such asthe metropolitan area of Madrid) to stagnant zones, whichcould result in an accumulation of pollutants.

It should also be highlighted that none of theparameters in the DRASTIC and GOD intrinsic vulner-ability indexes consider the influence of groundwater flowdirection within vulnerable areas. The groundwater flowdirection implies that some points of a given aquiferreceive groundwater (and its corresponding pollutants)from a larger area than other points within the sameaquifer. This tends to favour an increase in the concen-tration of these pollutants in stagnant areas, where thereare convergences of groundwater flows (accumulationprocess), or a reduction in the concentration of thesepollutants (dilution process), depending on the quality ofthe groundwater received. This also explains the origins ofpollution plumes, which are phenomena that should alsobe considered when constructing vulnerability indexes.Although some of the parameters included in theDRASTIC index have a direct relation with saturatedzone properties (the groundwater flow velocity within theaquifer), they do not take into account the influence offlow direction. Furthermore, ratings assigned to thehydraulic conductivity parameter of the DRASTIC indexincrease with the velocity of the groundwater flow,reflecting the potential capacity of the system to extendthe pollution process. These criteria do not, however,consider the effect of a high flow rate on preventing theaccumulation of pollutants at specific points within theaquifer. These limitations could explain the recent generaltrend to incorporate flow models into intrinsic and specificvulnerability methods (Al-Adamat et al. 2003; Gogu et al.2003; Debernardi et al. 2007; Nobre et al. 2007) andpoints to the need to replace qualitative models withquantitative models (Conell and den Daele 2003; Gogu etal. 2003; Debernardi et al. 2007). Another weakness of theDRASTIC method is the difficulty to achieve an accurate

estimation of certain parameters such as hydraulic conduc-tivity and net aquifer recharge (Fritch et al. 2000; Stigter etal. 2006). As a result, the same values could be assigned tolarge and non-homogeneous areas. Stigter et al. (2006) alsopoint out another limitation of the DRASTIC index—theexcessive emphasis on the attenuation capacity of theunsaturated zone in the case of mobile pollutants such asnitrate. As a result, high unsaturated thicknesses, of the typefound in some areas of the Tertiary Detrital Aquifer, do notattenuate the pollutant process but rather delay the arrival ofnitrate to the aquifer (Foster 1987). Along the same lines,Aller et al. (1987) stressed the risk of error, in the case ofstable and mobile pollutants such as nitrate, when dilutionprocesses are more important than attenuation processes. Forall of the above reasons, and as noted by several otherauthors, it could sometimes be necessary to modify some ofthe ranges and weightings of the different parameters used tocalculate the DRASTIC index. It is also necessary to adaptthe ratings of the LU and L parameters in all the vulnerabilityindexes so that they can consider the specific circumstancesof each setting, particularly with respect to urban land uses.

Despite these limitations, the results of this studysupport the great utility of vulnerability indexes and theirstatus as a very useful tool for decision making to promotethe sustainable management of different land uses and theidentification of nitrate vulnerable zones at the regionalscale. The global perspective of the hydrogeologicalcharacteristics of the study region provided by thesemethods favours a better understanding and interpretationof nitrate pollution processes. Given that the distributionof areas subject to nitrate pollution is quite consistent withthat of vulnerable zones identified by the vulnerabilityindexes for the Quaternary Aquifer and the lower sub-unitof the Moor Limestone Aquifer, both areas should beanalysed in greater detail in order to consider thepossibility of designating them as nitrate vulnerable zoneswithin the Comunidad de Madrid (in compliance with theEU Directive 91/676/EEC). Whatever the case, comple-mentary studies are necessary at the local scale and theseshould specifically focus on heterogeneous aquifers inareas with specific problems of nitrate pollution.

Conclusions

The results of this study confirm the utility of intrinsicvulnerability indexes (DRASTIC and GOD) and specificvulnerability to nitrate pollution indexes (CD and NV) forevaluating the vulnerability of the groundwater in centralSpain on a regional scale. The great degree of similarityobserved between maps based on the GOD and DRASTICindexes indicates the usefulness of applying a simplemodel such as the GOD method, to evaluate intrinsicvulnerability on a regional scale. The map of specificvulnerability based on the NV index proved to be the mostconsistent with respect to the real nitrate distribution ingroundwater within the study area. This model, which wasproposed as an adaptation based on the DRASTIC index(although results indicate that it could also be used as an

695

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 16: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

adaptation of the GOD index), has been developed with theobjective of achieving greater accuracy in the estimation ofspecific vulnerability to nitrate pollution. It is based on amultiplicative model that integrates the risk of groundwaterpollution by nitrate related to different land uses andconsiders both the negative impacts, over time, of some ofthese uses on aquifer media and also the protective effects ofothers. Both intrinsic and specific vulnerability maps showedareas corresponding to the Quaternary Aquifer (the medium-lower part according to the NV index) and the lower sub-unitof the Moor Limestone Aquifer, where intensive agricultureis the dominant land use, to be the most vulnerable areas inthe Comunidad de Madrid. These results are consistent withavailable information about the nitrate pollution in theseaquifers. In the southern part of the Tertiary Detrital Aquifer,there is a polluted area that is not consistent with the degree ofvulnerability assigned to it by any of the vulnerability maps.This result could be explained by incomplete land useinformation for this area or by movements of nitrate due togroundwater flows from the areas of highest risk to stagnantzones. In this respect, the DRASTIC and GOD intrinsicvulnerability indexes show certain limitations that should beimproved, related to a lack of parameters that consider theeffects of groundwater flow direction on the distribution ofvulnerable zones. The results of this study provide a basis forconsidering the Quaternary Aquifer and the lower sub-unit ofthe Moor Limestone Aquifer as possible Nitrate VulnerableZones in the Comunidad de Madrid, as defined by EUDirective 91/676/EEC.

Acknowledgements This research was funded by the Comunidadde Madrid and the European Social Fund (GR/AMB/0745/2004).The Confederación Hidrográfica del Tajo and Instituto Geológico yMinero provided hydrogeological data relating to piezometric andquality sampling of the groundwater networks in the study area.

References

Adams B, Foster S (1992) Land surface zoning for groundwaterprotection. J Inst Water Environ Manage 6:312–320

Al-Adamat RAN, Foster IDL, Baban SMJ (2003) Groundwatervulnerability and risk mapping for the Basaltic aquifer of theAzraq basin of Jordan using GIS, remote sensing andDRASTIC. Appl Geogr 23:303–324

Alcolea MA, García Alvarado JM (2006) El agua en la Comunidadde Madrid [Water resource in the Comunidad de Madrid]. ObsMedioambient 9:63–96

Aller L, Bennet T, Lehr JH, Petty RJ (1987) DRASTIC: stand-ardized system for evaluating groundwater pollution potencialusing hydrogeologic settings. US EPA Report 600/2-87-035 USEPA, Ada, OK, 622 pp

Antonakos AK, Lambrakis NJ (2007) Development and testing ofthree hybrid methods for the assessment of aquifer vulnerabilityto nitrates, based on the drastic model: an example from NEKorinthia, Greece. J Hydrol 333:288–304

Arauzo M, Martínez-Bastida JJ, Valladolid M (2008) Contamina-ción por nitrógeno en el sistema “río-acuífero aluvial” de lacuenca del Jarama (Comunidad de Madrid, España): ¿origenagrícola o urbano? [Nitrogen pollution of the “river-aquifersystem” in the Jarama basin (Comunidad de Madrid, Spain): Isit from agricultural or urban sources? Limnetica 27:195–210

Arrate I, Sánchez-Pérez JM, Antiguedad I, Vallecillo MA, Iribar V,Ruiz M (1997) Groundwater pollution in quaternary aquifer of

the Vitoria-Gasteiz (Basque Country, Spain). Environ Geol30:257–265

Auge M (2004) Vulnerabilidad de Acuíferos [Aquifer vulnerability].Rev Latino-Am Hidrogeol 4:85–103

Babiker IS, Mohamed MAA, Hiyama T, Kato K (2005) A GIS-based DRASTIC model for assessing aquifer vulnerability inKakamigahara Heights, Gifu Prefecture, central Japan. Sci TotalEnviron 345:127–140

Cabrera F, Fernandez-Bohy E, Aparicio MG, Murillo JM, Moreno F(1995) Leaching of nitrate from a sandy loam soil under cornand two N-fertilizations. Fres Environ Bull 4:250–255

Civita M (1994) Le Carte délia Vulnerabilitâdegli Acquiferiall’In-qulnamento [Maps of aquifer vulnerability to pollution]. Teoria iPratica. Pitagora, Bologna, Italy

Civita M, De Regibus C (1995) Sperimentazione di alcunemetodologie per la valutazione della vulnerabilita` degli aquiferi[Testing methodologies for assessing aquifer vulnerability].Quad Geol Appl 3:63–71

Colman IP, Palmer RC, Bellamy PH, Hollis JM (2005) Validation ofan intrinsic groundwater pollution vulnerability methodologyusing a national nitrate database. Hydrogeol J 13:665–674

Comunidad de Madrid (1995) Plan de saneamiento y depuración delas aguas residuales de la Comunidad de Madrid 1995–2005[Action programme for the treatment of the Comunidad deMadrid waste waters 1995–2005]. Madrid, 81 pp

Comunidad de Madrid (2001) Plan regional de actuaciones enmateria de suelos contaminados de la Comunidad de Madrid2001–2006 [Action program on a regional scale for polluted-soils restoration in the Comunidad de Madrid]. Madrid, 184 pp

Conell LD, den Daele GV (2003) A quantitative approach to aquifervulnerability mapping. J Hydrol 276:71–88

Confederación Hidrográfica del Tajo (2005) Informe Anual de AguasSubterráneas de la Confederación Hidrográfica del Tajo año 2005[Annual report of the Confederación Hidrográfica del Tajo for theyear 2005 on groundwater]. Available on the Internet. http://www.chtajo.es/redes/cantidad/aguas_subte.htm. January 2008

Corniello A, Ducci D, Napolitano P (1997) Comparison betweenparametric methods to evaluate aquifer pollution vulnerabilityusing GIS: an example in the “Piana Campana”, southern Italy.In: Marinos PG, Koukis GC, Tsiambaos GC, Stournaras GC(eds) Engineering geology and the environment. Balkema,Rotterdam, The Netherlands, pp 1721–1726

Council of the European Communities (1980) Directive 80/778/EEC relating to the quality of water intended for humanconsumption. 15 July 1980, Council of the European Commun-ities, Brussels, Belgium

Council of the European Communities (1991) Directive 91/676/EEC concerning the protection of waters against pollutioncaused by nitrates from agricultural sources. 12 December 1991,Council of the European Communities, Brussels, Belgium

Debernardi L, De Luca DA, Lasagna M (2007) Correlation betweennitrate concentration in groundwater and parameters affectingaquifer intrinsic vulnerability. Environ Geol. doi:10.1007/s00254-007-1006-1

Denny SC, Allen DM, Journeay JM (2007) DRASTIC-Fm: amodified vulnerability mapping method for structurally con-trolled aquifers in the southern Gulf Islands, British Columbia,Canada. Hydrogeol J 15:483–493

Environmental Systems Research Institute (2006) ESRI ArcGIS 9.2,Environmental Systems Research Institute, Redlands, CA

European Environment Agency (2000) Mapa de ocupación delsuelo para España del año 2000 [CORINE Land Covermap2000; for Spain]. Available on the Internet. http://www.ign.es/ign_iberpix/index.htm. February 2008

Evans BM, Myers WL (1990) A GIS-based approach to evaluatingregional groundwater pollution potential with DRASTIC. J SoilWater Conserv 45:242–245

Foster SSD (1987) Fundamental concepts in aquifer vulnerability,pollution risk and protection strategy. In: van Duijvenbooden W,van Waegeningh HG (eds) Vulnerability of soil and ground-water to pollutants. Proceedings and Information no. 38, TNOCommittee on Hydrological Research, The Hague, pp 69–86

696

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 17: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

Foster SSD, Hirata R (1991) Determinación del riesgo decontaminación de aguas subterráneas, una metodología basadaen datos existentes [Groundwater pollution risk assessment: amethodology using available data]. CEPIS, Brussels, 81 pp

Foster S, Hirata R, Gómez D, D’Elia M, Paris M (2002) Groundwater quality protection: a guide for water utilities, municipalauthorities, and environment agencies. The World Bank,Washington, DC, 104 pp

Fritch TG, Mcknight CL, Yelderman JC, Arnold JG (2000) Anaquifer vulnerability assessment of the Paluxy aquifer, centralTexas, USA, using GIS and a modified DRASTIC approach.Environ Manage 25:337–345

Garrett P, Williams JS, Rossoll CF, Tolman AL (1989) Are groundwater vulnerability classification systems workable? In: Proc.Focus Conference on Eastern Regional Ground Water Issues,NWWA, Dublin, Ohio, USA, pp 329–343

Gogu RC, Dassargues A (2000) Sensitivity analysis for the EPIKmethod of vulnerability assessment in a small karstic aquifer,southern Belgium. Hydrogeol J 8:337–345

Gogu RD, Hallet V, Dassargues A (2003) Comparison of aquifervulnerability assessment techniques: application to the Néblonriver basin (Belgium). Environ Geol 44:881–892

Guimera J (1993) Metodologia de estudio de la contaminacion deacuiferos por nitratos. Aplicacion al acuifero del Maresme.[Methodology for studying aquifer pollution by nitrates:applicability in the Maresme aquifer]. In: Candela L, Varela M(eds) La zona no saturada y la contaminacion de las aguassubterraneas [Unsaturated zone and groundwaters pollution].CIMNE, Barcelona, pp 197–219

Guo QH, Wang YX, Gao XB, Ma T (2007) A new model(DRARCH) for assessing groundwater vulnerability to arseniccontamination at basin scale: a case study in Taiyuan basin,northern China. Environ Geol 52:923–932

Hamza MH, Added A, Frances A, Rodriguez R (2007) Validity ofthe vulnerability methods DRASTIC, SINTACS and SI appliedto the study of nitrate pollution in the phreatic aquifer ofMetline-Ras Jebel-Raf Raf. CR Geosci 339:493–505

Hernández-García MA, Custodio E (2004) Natural baseline qualityof Madrid Tertiary Detrital Aquifer groundwater (Spain): a basisfor aquifer management. Environ Geol 46:173–188

Heuer K, Brooks P, Tonnessen KA (1999) Nitrogen dynamics intwo high elevation catchments during spring snowmelt 1996,Rocky Mountains, Colorado. Hydrol Process 13:2203–2214

Hiscock KM, Lovett AA, Brainerd JS, Parfitt JP (1995) Ground-water vulnerability assessment: two case studies using GISmethodology. Q J Eng Geol 28:179–194

Instituto Geológico y Minero de España (IGME)(1981) PlanNacional de Investigación de Aguas Subterráneas (PNIAS):Proyecto de investigación hidrogeológica de la cuenca del Tajo.Informes técnicos 1–5 [National Plan for Groundwater Research(PNIAS). Research Project on the hydrogeology of the Tajobasin. Technical reports 1–5]. IGME, Madrid

Instituto Geológico y Minero de España (IGME) (1985) Calidad ycontaminación de las aguas subterráneas en España [Ground-water quality and groundwater contamination in Spain]. Informede síntesis, Tomo II, Anejos, IGME, Madrid, 385 pp. Availableon the Internet. http://aguas.igme.es/igme/publica/libro43/lib43.htm. March 2008

Instituto Geológico y Minero de España (IGME) (1988) MapaGeológico de la Comunidad de Madrid [Geological map of theComunidad de Madrid]. Scale 1:200000. IGME, Maddrid

Instituto Geológico y Minero de España (IGME) (1993) Las aguassubterráneas en España [Groundwater in Spain]. IGME, Madrid,531 pp. Available on the Internet. http://aguas.igme.es/igme/publica/libro20/lib20.htm. March 2008

Instituto Geológico y Minero de España (IGME) (1999) Programade actualización del inventario hidrogeológico (PAIH) [Programto update the hydrogeological inventory]. IGME, Madrid, 183pp. Available on the Internet. http://aguas.igme.es/igme/publica/libros3_RHS/paih/lib32.htm. January 2008

Instituto Geológico y Minero de España (IGME) (2004) Protecciónde las aguas subterráneas frente a vertidos directos e indirectos

[Groundwater protection from direct and indirect discharges].In: Loreto FR, Sánchez F, López Geta JA, Antonio J (eds)Hidrogeología y aguas subterráneas [Hydrogeology and ground-water], No. 13. IGME, Madrid, 352 pp. Available on theInternet. http://aguas.igme.es/igme/publica/lib120/lib120.htm.January 2008

Knapp MF (2005) Diffuse pollution threats to groundwater: a UKwater company perspective. Q J Eng Geol Hydrogeol 38:39–51

Lin B, Sakoda A, Shibasaki R, Goto N, Suzuki M (2000) Modellinga global biogeochemical nitrogen cycle in terrestrial ecosystems.Ecol Model 135:89–110

Llamas MR, López Vera CF (1975) Estudio sobre los recursoshidráulicos subterráneos del área metropolitana de Madrid y suzona de influencia: avance de las características hidrogeológicasdel Terciario Detrítico de la Cuenca del Jarama [Hydraulicresources of the metropolitan area of Madrid and its area ofinfluence: hydrogeological characteristics study of the TertiaryDetrital Aquifer of the Jarama basin]. Agua 88:36–55

Margat (1968) Vulnérabilité des nappes d´eau souterraines á lapollution [Groundwater vulnerability to pollution]. Bases de lacartographie. BRGM 68. SLG198 HYD, BRGM, Orléans, France

Martínez-Bastida JJ, Arauzo M, Valladolid M (2006) Contamina-ción por nitrato en el acuífero aluvial [Nitrate contamination inthe alluvial aquifer]. In: Camargo J (ed) Ecología y conserva-ción del río Henares y sus tributarios [Ecology and conservationof the River Henares and its tributaries]. Universidad de Alcaláde Henares, Ediciones Cersa Madrid, pp 23–34

McLay CDA, Dragden R, Sparling G, Selvarajah N (2001)Predicting groundwater nitrate concentrations in a region ofmixed agricultural land use: a comparison of three approaches.Environ Pollut 115:191–204

Ministerio de Agricultura, Pesca y Alimentación (2007) Balance delnitrógeno en la agricultura Española durante 2005 [Nitratebalance in Spanish agricultural systems in 2005]. SecretaríaGeneral de Agricultura y Alimentación. Dirección General deAgricultura, Madrid, 51 pp

Ministerio de Medio Ambiente (2001) Caracterización de las fuentesagrarias de contaminación de las aguas por nitratos [Description ofagricultural sources that cause water pollution by nitrate].Secretaría de estado de Aguas y Costas, Dirección General deObras Hidráulicas y Calidad de Aguas, Madrid, 151 pp

Monturiol F, Alcalá L (1990) Mapa de asociación de suelos de laComunidad de Madrid [Soil type map of the Comunidad deMadrid]. Scale 1:200.000, Consejería de Agricultura y Cooper-ación, Comunidad de Madrid y Consejo Superior de Inves-tigaciones Científicas, Madrid, 71 pp

Naqa AE (2004) Aquifer vulnerability assessment using theDRASTIC model at Russeifa landfill, northeast Jordan. EnvironGeol 47:51–62

Nobre RCM, Rotunno OC, Mansur WJ, Nobre MMM, CosenzaCAN (2007) Groundwater vulnerability and risk mapping usingGIS, modeling and a fuzzy logic tool. J Contam Hydrol 94:277–292

Observatorio de la Sostenibilidad en España (2006) Cambios deocupación del suelo en España: implicaciones para la sosteni-bilidad [Land use changes in Spain: implications for sustain-ability]. Ministerio de Medio Ambiente, Univ. Alcalá deHenares, Madrid, 428 pp

Panagopoulos GP, Antonakos AK, Lambrakis NJ (2006) Optimiza-tion of the DRASTIC method for groundwater vulnerabilityassessment via the use of simple statistical methods and GIS.Hydrogeol J 14:894–911

Piscopo G (2001) Groundwater vulnerability map explanatorynotes: MacIntyre Catchment. NSW Department of Land andwater Conservation, Sydney, 13 pp

Ribeiro L (2000) IS: um novo indice de susceptibilidade deaquiferos á contaminaçao agrícola [SI: a new index of aquifersusceptibility to agricultural pollution]. Internal report, ERSHA/CVRM, Instituto Superior Técnico, Lisbon, 12 pp

Robins N, Adams B, Foster S, Palmer R (1994) Groundwatervulnerability mapping: the British perspective. Hydrogéologie3:35–42

697

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5

Page 18: Intrinsic and Specific Vulnerability of Groundwater in Central Spain-The Risk of Nitrate Pollution

Román R, Caballero R, Bustos A, Díez JA, Cartagena MC, VallejoA, Caballero A (1996) Water and solute movement underconventional corn in central Spain. I. Water balance. Soil SciSoc Am J 60:1530–1536

Román R, Caballero R, Bustos A (1999) Field water drainage undertraditional and improved irrigation schedules for corn in centralSpain. Soil Sci Soc Am J 63:1811–1817

Rosen L (1994) A study of the DRASTIC methodology withemphasis on Swedish conditions. Ground Water 32:278–285

Rupert MG (2001) Calibration of the DRASTIC ground watervulnerability mapping method. Ground Water 39:625–630

Sanchez-Pérez JM, Antigüedad I, Arrate I, García-Linares C,Morell I (2003) The influence of nitrate leaching throughunsaturated soil on groundwater pollution in an agriculturalarea of the Basque country: a case of study. Sci Total Environ317:173–187

Secunda S, Collin ML, Melloul AJ (1998) Groundwater vulner-ability assessment using a composite model combining DRAS-TIC with extensive agricultural land use in Israel’s Sharonregion. J Environ Manage 54:39–57

Stigter T, Almeida P, Carvalho Dill A, Ribeiro L (2002) Influence ofirrigation on groundwater nitrate concentrations in areas

considered to have low vulnerability to contamination. In: Proc.XXXII IAH and VI ALHSUD Congress, Mar del Plata,Argentina, Oct. 2002 (CD-ROM)

Stigter TY, Ribeiro L, Carvalho Dill AMM (2006) Evaluation of anintrinsic and a specific vulnerability assessment method incomparison with groundwater salinisation and nitrate contami-nation levels in two agricultural regions in the south of Portugal.Hydrogeol J 14:79–99

Thirumalaivasan D, Karmegam M, Venugopal K (2003) AHP-DRASTIC: software for specific aquifer vulnerability assess-ment using DRASTIC model and GIS. Environ Model Softw18:645–656

Vias JM, Andreo B, Perales MJ, Carrasco F (2005) A comparativestudy of four schemes for groundwater vulnerability mapping ina diffuse flow carbonate aquifer under Mediterranean climaticconditions. Environ Geol 47:586–595

Vrba J, Zaporozec A (1994) Guidebook on mapping groundwatervulnerability. IAH Int Contrib to Hydrogeol 16, Heise, Hannover,Germany, 131 pp

Worral F, Kolpin DW (2004) Aquifer vulnerability to pesticidepollution: combining soil, land-use and aquifer properties withmolecular descriptors. J Hydrol 293:191–204

698

Hydrogeology Journal (2010) 18: 681–698 DOI 10.1007/s10040-009-0549-5