Arsenic In A Hard Rock Aquifer

32ND INTERNATIONAL GEOLOGICAL CONGRESS FLORENCE, ITALY 20/08/04 to 28/08/04

Arsenic in a hard rock aquifer: multivariate optimisation of the groundwater monitoring network L.F. Molerio León CESIGMA, S.A, P.O..Box [email protected]

6219,

CP

10600,

Habana

6,

Ciudad

de

La

Habana,

Cuba,

R. Toujague de la Rosa Instituto de Geofísica y Astronomía, Calle 232 No. 19206, La Lisa, La Coronela, Ciudad de La Habana, Cuba, [email protected]

ABSTRACT: This paper describes the results of a geomathematical analysis of natural and artificially induced arsenic sources in groundwater in a hard rock aquifer. Data on chemical composition, ground water levels, and other physical variables were collected systematically but discontinuously between 1974 and 2001. A special survey was carried out in 2002 to clarify several geologic settings. Geomathematical tools involved factor and cluster analysis, kriging mapping techniques and information theory resources. Results allowed to assess the actual informativity of the monitoring network and optimise its spatial distribution, operation frequency and data collection structure, the improvement of the transit time of arsenic in groundwater, and to separate artificial sources from natural sources of arsenic occurrence. Surface waters appears to be contaminated by arsenic coming from lateral flow migration and leaching of concentrates in zones close to the surface water divides.

INTRODUCTION Water quality control and aquifer remediation in mining regions contaminated by toxic metals as As, Pb, Hg, Cd, among others, particularly important were no other sources of water are available for domestic consumption. Historically, several regions have been affected by contamination from geologic sources, like Taiwan (Williams, 1997), Córdova, Argentina (Thor nton et to the, 1997), Mexico (Cebrian et al., 1983) as well as west Bengal, India. This has been considered a case of world calamity due to aquifer contamination by arsenic (As) (Badal et al., 1996). In these cases the contamination is associated to the pyrite oxidation. Delita, a Cuban region linked to gold mineralization has shown problems of As contamination in ground and surface waters. The drinkable water is brought from other towns and the identif ication and characterization of the generating sources of this contamination is not very clear. Delita mines has been exploited for gold since the XVI century. Secondary geochemical anomalies of Ag, As, Pb among other elements contributes to the poor local water quality. Delita is part of the metamorphic belt of the Cuban Isle of Youth. Its mineralized body is composed by mineralized quartz veins associated to apoterrigenous quartz-chlorine-sericít icgraphitic schists of the formation Cañada (K2 ). Veins contain gold-arsenopirite, sulfoantimoniets and disseminated sulphurs, corresponding to the Au-Ag-As-Sb geochemical type. The precocious generation of gold is associated to arsenopirite and the late one to Pb sulfoantimoniets Fig. 1). According to EPA (2000), the arsenic in rocks is likely to occurr in its reduced form mineral veins and as well as an oxide. The arsenate (AsO4 -3 ) appears in some minerals replacing phos-

1

phates in apatites. The arsenate species H2 AsO 4 - and HAsO4 -2 are the balanced forms in the pH range of most of the natural waters. The HAsO4 -2 prevails with pH of 7,2 and H2 AsO4 - for lower pH. In ambient reducing environments the form HasO2 (aq) can be present. Arsenite (As3+) is present in aerobic conditions and arsenate in anaerobic conditions. Arsenic could also incorporate to the drinkable waters from wastes and from certain insecticides and herbicides. The balanced arsenate ions concentration in water depends on the concentrations of the cation and on the solubility of the different arsenates that could precipitate. The solubility products of the arsenates indicate that many of these compounds are not very soluble in terms of the anion AsO4 -3 . This form only prevails at pH 11,5 and, in the normal range of natural waters, calcium or of magnesium arsenate solubility is enough to allow that several dozens of milligrams of As can be retained in solution. Other metals like Cu can limit the solubility from the As to few dozens of mg/L. Probably the arsenate sorption in the precipitated ferric hydroxide or other active surfaces is an important factor restrictive in the solubility of As in natural waters. Local and regional flow systems were described by Arellano et al, (1993, 1995) and Molerio (1993, 1994).

Fig. 1. Location map MULTIVARIATE ANALYSIS FOR GROUNDWATERS For the uni- and multivariate statistical analysis of ground waters an observational matrix composed by a discrete and discontinuous series of 67 cases was available for the following data: Sampling depth (m) PROF_ Water temperature (ºC) TEMP pH PH Specific Electric Conductivity (MS/cm) SPC Total iron concentration of (mg/l) FE Arsenic concentration (mg/l) AS Factorial decomposition methods were applied with extraction by Principal Components (ACP) and Varimax Rotation of the observational matrix. The extraction of two factors, applying Kaiser´s Approach allowed explaining 72% of the total variance of the available time and spatial series. This maximation of the correlations suggests that the variables involved in the

2

considered factors belong to the series but, however, it is necessary to include other complementary variables to explain the behaviour of the arsenic in the system. Table 1 shows the eigenvalues and accumulated variance of the series of superficial waters. Table 2 shows the factorial loads, considering significant those = 0,6. Table 1. Eigenvalues Factors 1 2

Eigenvalues 2,313 1,293

% total variance 46,262 25,863

Cu mulative eigenvalue 2,313 3,606

% Cumulative variance 46,262 72,126

Table 2. Factorial load PROF_ TEMP PH SPC AS Expl.Var Prp.Totl

Factor 1 0,12449277 0,03044848 0,90609084 -0,68627405 -0,94722028 2,2056245 0,4411249

Factor 2 -0,66307455 0,85446284 0,19418257 0,40934475 0,1601521 1,40069331 0,28013866

The standardized lineal components (SLC) has the following structure: Factor 1 = -0,95 AS + 0,91 PH 87 - 0,68 SPC Factor 2 = 0,85 TEMP + 0,66 PROF The first factor explains 45% of the total variance by three variable of which, the higher weight is the own As, closely associated to the pH, but in the same plane of the SPC (Fig. 2). For SPC role a fully satisfactory explanation should account for a mixture of waters of different origins in the sample. The control of temperature is more clearly because of its effect on the solubility patterns and the oxidation reduction potential of the system, in particular, to the development of aerobic or anaerobic conditions in the sampling station. The observed communa lities and the scores are presented in the Tables 3 and 4. Table 3. Observed communalities PROF_ TEMP PH SPC AS

From Factor1 0,01549845 0,00092711 0,82100061 0,47097207 0,89722626

From 2 Factor 0,45516632 0,73103386 0,85870748 0,63853519 0,92287495

Multiple- R2 0,1853549 0,20762012 0,7191284 0,49564815 0,80444712

Table 4. Factorial Scores (Varimax normalized rotation) Bodega89 SMina90 PArroyo89 PRuben90 PMalva89 V-1389 Cpipas89 PEsperanza

Factor 1 0,2495568 -2,8535261 0,34720938 0,39549939 -0,32094627 0,75079285 0,11522884 -0,09909391

Factor 2 -0,34984548 0,48556295 -0,17369291 -0,51957945 -0,08376767 0,2199185 -1,74158065 -0,81171584

3

MijailI89 MijailII8 S-789

0,78955606 0,24511565 0,38060732

1,80317804 1,49588604 -0,32436353

To identify the associations among the ground waters sampling stations, numerical classific ation based on non-supervised patterns was applied. An object (cases) dendrogram was built by simple connection with Ward´s measure of distance (Fig. 3). Factor Loadings, Factor 1 Rotation: Varimax normalized Extraction: Principal components 1

TEMP

0.8 0.6 SPC

Factor 2

0.4 0.2

PH

AS

0 -0.2 -0.4 PROF_

-0.6 -0.8 -1.2

-0.8

-0.4

0

0.4

0.8

1.2

Factor 1

Fig. 2. Factorial plane Four perfectly differentiated groups could be identified. One includes Mikhail's stations, Bodega 89 and V13 89; another PRubén 90, S 7 89, CPipas and PEsperan. A third group to SMIna and PMalva89 and a fourth group, independent and exclusive, integrated by the station PArroyo. In all these cases, waters of different origin and geochemical evolution are present and, therefore, should be evaluated in a different way in all the cases. Mapped scores are presented in Fig. 3 showing an interesting distribution of the orthogonally crossed factors indicating those points where controlling factors coincide. The relationship among the source points of Factors 1 and 2, SMina and Mikha il, indicates the strong contribution of the discharge points to local contamination. The point of minimum convergence of scores is associated to the station PArroyo 89 which deserves special attention for the inflection and the abrupt change of slope of the isolines.

4

Fig. 3. Dendrogram

5

Fig 4. Factorial scores

SCORES FACTORIALES PARA LOS DOS PRIMEROS FACTORES EN LAS AGUAS SUBTERRANEAS

PRuben90

Bodega89

216500 SMina90

216000

215500

PArroyo89

215000

214500

214000

213500

213000 MijailI89 MijailII8 291000

0

Factor 1

291500

292000

1000

292500

293000

2000

Factor 2

6

IDENTIFICATION OF THE CONTAMINATION SOURCES AND THEIR AREA OF INFLUENCE The solution of the transport equation was expressed in the construction of a combined map of interpolation by means of kriging sustained on spherical variograms of the historical maximum and minimum concentration (Fig. 4. shows the map for maximum concentration). The isolines allows to identify the area of influence of the contamination and, consequently, the reach of risk scenarios defined, in this case, by the 0,05 mg/l isoline that represents the Acceptable Maximum Concentration of arsenic in the drinkable waters according to Cuban Standards. Maps show that the contributing sources are the same for surface and groundwater and are concentrated on the mine’s area. The shape of the isolines forms a lobe with its center at PMina sampling station and derives in South-North direction toward Mijail Station marking the histor ical limit of the mine’s influence area on ground waters. The zero isoline marks the contour of a risk scenario of danger, approximately limited by a rectangle of about 21 km2 . Towards the East, in the centre of the triangle formed by the points PRubén, Mikhail and PAmer, second scenario appears, and a plateau that is not conserved in the map of min imum values constituting a local effect, maybe associated to uncontrolled and non-monitored discharges. A third risk danger scenario is associated to arsenic transport in surface, as a result of migration and discharge of the ground waters and by discharges in the surface waters. Such a scenario becomes defined by the 0,01 mg/l isoline and covers the whole area between the mine and the Los Indios Dam, establishing an influence area of about 42 km2 . The eventual influence area toward the West could not be defined with the available information. CONCLUSIONS ?

? ? ?

The Monitoring network currently in operation shows a very low informativity due to its inadequate composition, surface distribution, monitoring frequency and selected data acquisition during monitoring. The information that it returns is of very low efficiency (8% for the Superficial Net and 60% for the underground one) The arsenic concentration shows tendency to increase in ground waters except in Cala Itabo. The reason of this is not clear, as long as during decades mine has not been exploited, but it can be associated to the low transit time of ground waters.. The source of the most important arsenic contamination of waters comes from the mining industry. All the identified focus, in surface and ground waters are associated to the mining facilities. The area of influence of the contamination, at the West of the mine, could not be identified, for lack of sampling points. To the North, South and East, it could be identified until a radius of 7 kilometres from the mine.

The surface waters can appear contaminated by two causes, by the underground migration of pollutants, by means of lateral flow, and by lixiviation of deposits concentrated close to recharge areas of aquifer or by surface drainage from mine wastes.

7

Fig. 5. Isolines of maximum recorded As concentration in terrestrial waters .

CONCENTRACIONES MAXIMAS DE As (mg/l) EN LAS AGUAS TERRESTRES Blan00

219000 Eliseo86

218000

MinaPozo79 Bodega89

Geodesia74 217000

PRuben74 PAlberto74 Terra 274

PMina79 Pmina75 SMina86 Mina74 216000 PMaestro74 Manantia01 PaPlanta79 PArroyo89 PArroyo74 215000

214000

Indios01 Indios00 Indios74 V-1279 P1274

213000

CIndios86

Mijail Mijail86 II89 PAmer89

212000 Terra 174 289000

290000

291000

0

292000

1000

As en las aguas subterráneas

293000

2000

294000

3000

295000

4000

296000

297000

5000

As en las aguas superficiales

8

REFERENCES Arellano Acosta, Daniela M.; B. Degournay; J. GutiérrezDíaz; L.F. Molerio León; O. Ascanio & A. Santos (1993): Isotope Hydrogeochemistry in the Study of Saline Aquifers. Case of Study-Isle of Youth, Cuba. Symp. Isotope Tech. in the Study of Past and Current Environmental Changes in the Hydrosphere and the Atmosphere, IAEA, Vienna, Paper IAEA -SM-329/27P,:514-516 Arellano Acosta, D. M.; L.F. Mole rio León & A. Santos (1995): Dinámica del Flujo Regional en el Macizo Metamórfico de la Isla de la Juventud, Cuba. IAEA TEC-DOC -835: Estudios de Hidrología Isotópica en América Latina 1994, Org. Internac. Energía Atómica, Viena :139-174 Badal K. 1996: Arsenic in groundwater in seven district of West Bengal, India – The biggest arsenic calamity in the world. Current Science, vol.70. No. 11, June.pp:976-986. Cebrian M. E. 1983: Human Toxicol., 2, 121-133. EPA 2000: Federal Register. Part II. National Primary Drinking Water Regulations: Arsenic and Clarifications to Compliance and New Source Contaminants Proposed Rule. Fed. Reg. Vol 65 (121), June 22, 2000, 97: Molerio León, L.F. 1993: Dinámica del Flujo Regional en el Macizo Metamórfico de la Isla de la Juventud. Taller sobre Aplicación de Técnicas Isotópicas en el Estudio de los Recursos y la Contaminación de las Aguas, OIEA, Maracaibo, Venezuela, 14: Molerio León, L.F. 1994: Isotopic and Geochemical Regionalization of a Tropical Karst Aquifer. Internatl. Symp. isotopes in Water Resources Management; OIEA, Vienna, Austria, Paper IAEA -SM336/88P, 6: Thornton, I, M. Farago (1997): The Geochemistry of arsenic. In Albernathy, C. O.; Calderon, R.L.; Chappell, W. R (eds.) Arsenic exposure and health effects. Chapman & Hall, London, pp: 1-16 Williams, M. (1997): Mining - related srsenic hazards: Thailand case-study Summary Report. Key worth, Nottingham, UK. British Geological Survey, 36 .

9