Vulnerability of the karst area related to potentially toxic elements

Soil samples from 31 locations in the Una river spring catchment were subject to chemical extraction analyses. The data were presented as distribution maps of potentially toxic elements (Al, Cu, Mn, Pb and Zn) in the surface soil of the area. To evaluate the vulnerability of the immediate spring zone of the karst catchment, the vulnerability map was derived from the application of the PI methodology proposed by the European COST Action 620. The PI method used to produce the vulnerability map takes into account the protective cover (P) and the infi ltration conditions (I). It is based on the origin-pathway-target model. The π-factor (π = P ́ I) describes the vulnerability in the area, subdivided into 5 classes: π-factor in the range 0-1 implies an area of extreme vulnerability, while π-factor in the range 4-5 implies an area of very low vulnerability. The extraction procedure for the elements Al, Cu, Mn, Pb and Zn, has been applied in order to determine the potential mobility and redistribution of elements that could infl uence the groundwater and affect its quality. The applied extraction was the second step of the sequential procedure proposed by TESSIER et al. (1979), i.e. extraction with 1 mol dm-3 CH3COONa/CH3COOH buffer (pH 5). The results provide information on the potential mobility of the studied elements, indicating the possibility of their mobilization through changes in pH. Lead shows the greatest amount of mobility, with a mean of 9% (max. 16%) extracted under an acidic condition. Manganese follows with a mean of 5% (max. 11%) and zinc, copper and aluminium show less than 1% (mean) mobility. The vulnerability map of the karst area was produced in order to predict potential problem areas of karst aquifers. The Una spring catchment area presents generally low to moderate vulnerability; 8% of the studied area can be considered as extremely vulnerable according to the PI-methodology. Based on these data it was possible to delineate areas with a low protection cover i.e. combining the vulnerability map of the karst area with the distribution maps of potentially toxic elements, areas considered extremely vulnerable could be identifi ed.


INTRODUCTION
The study area lies in the Una River valley, east of the small town of Gračac, in the southern part of the Lika region.It is a part of a catchment of the Una spring and extends over 135 km 2 (Fig. 1).
The Una River spring surfaces near the village of Suvaja, 399 m above sea level (BOGNAR, 2005), and is protected as a hydrogeological heritage site of Croatia.Typical karst landforms are observed here, including dolines, sinkholes and karst springs (vauclusian type).Sinkholes are covered, mostly by silt loam, and overgrown by vegetation.
The area is part of a high karst belt of the Outer Dinarides, the genesis of which is connected with the Adriatic Carbonate Platform (AdCP).The deposit can be very thick, in some areas surpassing 8000 m.The age of this carbonate succession ranges from Middle Permian, or even Upper Carboniferous to Eocene (VLAHOVIĆ et al., 2005).The area of investigation is part of a large carbonate aquifer system which accounts for 88% of Croatian ground water reserves, and is an important drinking water source (BIONDIĆ, 2009).
Karst aquifers are exceptionally vulnerable to contamination due to their heterogeneity, which results in huge variations in permeability, and enables poorly fi ltered, concentrated recharge to take place (ZWAHLEN, 2004).Sin ce contaminants can easily reach the aquifer by bypassing fi ltra tion, it is important to know where such events might occur.
Predicting the path of possible pollutants in aquifers is more diffi cult in karst due to the fact that water often lacks a specifi c fl ow path.
The objective of this paper was to evaluate the vulnerability of the karst catchment immediate to the Una spring zone.To assess the vulnerability, a combined soil/geochemical approach with vulnerability mapping was applied.
Since the protection of groundwater is important not only from an ecological point of view, but also from an economic one, different approaches to ground water protection have been developed.Comparative studies have shown that the application of various methods often leads to different results (NGUYET & GOLDSCHEIDER, 2006;GOGU et al., 2003).In order to develop an approach that considers the specifi cs of karst formations, a special group, COST Action 620, was established.The group developed an approach to "vulnerability and risk mapping for the protection of carbonate (karst) aquifers" (ZWAHLEN, 2004;NGUYET & GOLDSCHEIDER, 2006), and proposed a new method of vulnerability mapping -the PI method -for mapping karst groundwater resources (ZWAHLEN, 2003;GOLDSCHEI-DER, 2005).This approach was used in this study.
Acidifi cation of the topsoil cover may cause the mobilization of certain potentially toxic elements that accumulate in the soil and could be transported into aquifers, thus affecting the water quality.Therefore, topsoil can be considered a source of contamination, especially if possible changes in environmental conditions would lead to leaching of pollutants.With analysis of the soil for total elements and their mobile fraction, it is possible to distinguish between anthropogenic and geochemical sources of heavy elements (DUBE et al., 2001;FILGUEIRAS et al., 2002).

Laboratory analysis
Topsoil (0−2 cm) and subsoil (40−50 cm) samples were collected at 31 locations.Some of the topsoil of each sample was used for granulometric analysis, which was performed on 29 samples, by wet sieving using ASTM standard stainless steel sieves.Due to an insuffi cient amount of samples, granulometric analysis could not be performed for three samples.Hydraulic conductivity for the 29 samples was defi ned by grain size composition of sediments using the Hydraulic Conductivity Software, SizePerm.All samples were analyzed for total (HF-HClO 4 -HNO 3 -HCl) and mobile Fi gu re 1: Geographic location of the studied area with the sampling sites.
(CH 3 COONa / CH 3 COOH, pH 5) contents of the elements of interest (Al, Cu, Mn, Pb and Zn).Analyses were carried out on the fraction <0.063 mm.Inductively coupled plasma atomic emission spectrometry (ICP-AES) was used for determination of the total content of elements in soils, for which samples were dissolved using 4-acid digestion (HF, HClO 4 , HCl, HNO 3 ), by standard procedure (MIKO et al., 2000).
In order to determine the potential mobility and redistribution of the elements that could have an infl uence on the groundwater, the extraction procedure for elements Al, Cu, Mn, Pb, and Zn was also applied.The sequential extraction procedure provides more information, than the analysis of the total heavy metal content of soil, i.e. it is possible to determine preferential binding sites for observed elements, to evaluate the potential mobility of metals in the environment and to differentiate the lithological from the anthropogenic contributions of elements (TESSIER et al., 1979;PROHIĆ & KNIEWALD, 1987;AUBERT et al., 2004;KAASA LAI-NEN & YLI-HALLA, 2003).For the purpose of the paper, only the second step of fi ve-step sequential extraction procedures based upon TESSIER et al. (1979) was carried out, i.e. extraction with 1 mol dm -3 CH 3 COONa / CH 3 COOH buffer (pH 5), on the <0.063 mm size.The goal of this extraction was to observe the amount of the mobile fraction under specifi c pH conditions (5).

Intrinsic vulnerability mapping
The data for the vulnerability map were compiled from the digital topographic data (scale 1:25000), which included land-use data, sinkhole distribution data, a digital elevation model (DEM), soil map data, geological and hydrogeological digital maps, and a database of potential pollution sources (road infrastructure, settlements).The vulnerability map was the fi nal result of different analyses and a few intermediate steps, which are briefl y described in this paper.The vulnerability map is coloured according to π-factor, which is expressed by π = P ´ I (Table 1).The PI method takes into account the protective cover (P) and the infi ltration conditions (I).Its goal is to describe how vulnerable the groundwater is.Extreme vulnerability is presented as a red colour, while very low vulnerability is presented in a blue colour (Tab.1).In order to be able to compute the P factor (protective cover), the following information about the protective function was used: topsoil and subsoil thickness, precipitation regime, grain size, lithology, fi ssuring and karstifi cation.The P-factor ranges from 1 to 5, with 1 presenting a very low protec-tive function.The I-factor gives an insight into infi ltration conditions and the degree to which the protective cover is by-passed as a result of lateral surface and subsurface fl ows.In order to compute the I-factor, information about the hydraulic conductivity, depth to low permeability layers, slope gradient and vegetation was used.The I factor ranges between 0.0 and 1.0, with 0 being the most permeable (GOLD-SCHEIDER et al., 2000).Some of the information on vegetation and slope gradient were directly taken from digital data.
The vulnerability map was derived from application of the PI methodology proposed by the European COST Action 620; detailed information and a description of these methods can be found in ZWAHLEN (2004).According to this methodology, intrinsic vulnerability takes into account the geological, hydrological and hydrogeological characteristics of an area, but it is independent of the nature of the contaminants and the contamination scenario.
It is based on the origin-pathway-target model, and as fi nal result, areas considered extremely vulnerable could be highlighted.The origin is the term used to describe the location of a potential contaminant release.The pathway includes the passage of potential contaminants, from its origin to the target (receptor).The target may be the groundwater surface in the aquifer (GOLDSCHEIDER et al., 2000).

RESULTS
Analyses for the total and mobile content of observed elements, (Al, Pb, Cu, Mn and Zn) are presented in Table 2.The analyses were used to display the percentage of mobile content for all topsoil samples.Results of soil texture classifi cation, based on a granulometric analysis are also presented in Table 2.
Information about the hydraulic conductivity was needed to compute the I-factor of the PI method.This was calculated from granulometric analysis, with the software EasyPerm, and for all samples,values fall in the range of 10 -8 cm/s to 10 -9 cm/s.
From the geochemical analysis, colour scale maps were generated for the mobile fraction content of the topsoil cover in relation to the deeper layer (Figs.2a-e).Maps indicate which part of the examined area is mostly exposed to migration of the mobile fraction of elements.Lead shows the highest probability of mobility, with a mean of 9% (max.16%) extracted under acidic conditions.Manganese follows with a mean of 5% (max.11%), and zinc, copper and aluminium   show less than 1% (mean) of mobile content.It is important to note that the same colour on these maps, for different elements, do not hold the same value.The colour scale was made separately for each map according to its minimum and maximum values.For instance, orange colour on the map represents 1.46% mobile content for aluminium, but 11.68% for manganese.
The vulnerability map of the karst area was compiled by combining a digital topographic map (scale 1:25000) and values of P and I factor (Fig. 2f).It was made to predict potential problem zones where the karst aquifer might be easily contaminated.The Una spring catchment area has low to moderate vulnerability in most parts, but 8% of the studied area could be considered as extremely vulnerable according to the PI-methodology.The steps preceding the creation of the vulnerability map are not displayed here, but details needed to produce a vulnerability map can be found in the Cost Action 620 Report (ZWAHLEN, 2004).

DISCUSSION AND CONCLUSION
Two types of maps, the PI-map (or vulnerability map), and maps of the mobile fraction content of the topsoil cover in relation to the deeper layer, were combined in order to assess the potential hazard to groundwater.Maps of the mobile fraction content for observed elements (Al, Mn, Pb, Cu, Zn) provide information about their potential mobility, indicating the possibility of mobilization through changes in pH.On these maps, the area with the highest content of mobile fraction is shown in red.It is important to be aware of wellconnected environmental processes, in which acidifi cation of the terrain as a result of climate change or anthropogenic impact may lead to increased levels of metals in the aquifer (DUBE et al., 2001).Extraction analysis, used for making these maps, is a common procedure when the goal is to indicate potential risk of toxic species entering the groundwater (RAO et al., 2008).Depending on the type of soil, and the pH used in the extraction procedure, heavy metals can be retained in a soil sample, but some amount can be observed as mobile content (LAFUENTE et al., 2008;RAO et al., 2008;FILGUEIRAS et al., 2002).Not all the mobile content will dissolve in water.Heavy metals may interact, chemically or physically, with the natural compounds found in the water and, in general, may react with particular species, change oxidation states and precipitate out (DUBE et al., 2001).However, one part will dissolve in the water, while the other could be found in colloidal fractions or associated with colloids (JENSEN et al., 1999;CHRISTEN-SEN et al., 2001).In the same aquatic environmental condition, dissolution of mobile elements will increase the higher the content of mobile elements.Further, due to features of ka rst hydrogeology, contaminants can quickly appear in areas distant from the source, and affect the quality of the aquifer.
If the protective cover is by-passed due to karst features, potential pollution can reach the aquifer in a short time.In order to prevent such scenarios, it is important to know which area is the most vulnerable.The PI-map provides such information.The red-coloured areas of the PI-map represent the zones with the lowest protection function of the aquifer.According to the PI-methodology, as shown in results, 8% of the studied area can be considered extremely vulnerable (Fig. 2f).By combining the aforementioned maps, it is possible to discuss potential scenarios that could lead to contamination of groundwater, and use the composite map as a decision-making tool in a sanitation or pollution-prevention process.The worst-case scenario would be if the red areas of both types of maps overlap.That would imply the absence of a suitable effi cient barrier between the possible contaminant and the groundwater, and should lead to protective measures.Luckily, in the investigated area presented in this paper, red zones from the maps do not entirely coincide.As the red zones do not overlap, the question arises of which type of map should have preference as decision-making tool, (the PI map, or the mobile content map).Each situation should be considered separately, depending on the observed conditions.By considering both variables, it is possible to identify areas that must be prioritized in terms of protection, or monitoring and restriction of use (NOBRE et al., 2007).In such assessment, the areas of the highest proportion of mobile elements and those of extreme vulnerability are to be considered as important guidelines.For instance, if a red area of the PI-map (extremely vulnerable) overlapped with a low proportion of mobile elements (blue), then that area would have a lower priority than a yellow area of the PI-map (medium vulnerability) overlapping with an extremely high proportion of mobile elements (red).Such conclusions could be a helpful guide for investors, e.g. a water bottling branch.The main information that can be obtained from the maps in the paper is that the groundwater would not be seriously affected if environmental conditions changed to pH 5.

Fi gu re 2 :
The maps of the content of mobile fraction of the topsoil cover in relation to the deeper layer for aluminium (a), copper (b), manganese (c), lead (d), zinc (e) and the PI vulnerability map (f).

Table 3 :
The highest measured content of the mobile fraction for observed area and the maximum allowed concentration of elements of interest in drinking water.