Multi-method assessment of nitrate and pesticide contamination in shallow alluvial groundwater as a function of hydrogeological setting and land use
Introduction
The leaching of large amounts of nitrate and pesticides from agricultural fields is a serious problem in many countries (Appelo and Postma, 2005). The occurrence in groundwater of high nitrate values (e.g. Hamilton and Helsel, 1995, Causapé et al., 2004, Jalali, 2005, Stigter et al., 2006a, Stigter et al., 2006b, Stigter et al., 2006c, Lorite-Herrera and Jiménez-Espinosa, 2008), high pesticide values (e.g. Chilton et al., 1998, Batista et al., 2002, Cerejeira et al., 2003, Papadopoulou-Mourkidou et al., 2004a, Papadopoulou-Mourkidou et al., 2004b, Guzzella et al., 2006, Shomar et al., 2006, Hildebrandt et al., 2008) or both (e.g. Cerejeira et al., 2000, van Maanen et al., 2001, Bouman et al., 2002, Silva et al., 2006) is described in many regions around the world. The relation between nitrate and/or pesticide concentrations in groundwater and agricultural activities is referred in most of these studies. In Europe, pesticide data are often very limited; however, groundwater contamination by nitrates and pesticides is a serious problem. The high nitrate levels are largely caused by man, particularly related to the use of nitrogen fertilizers and manure, which constitutes a diffuse source (EEA, 2000); nitrate in groundwater can also result from point sources, including cattle feed lots, septic tank, sewage discharge and the oxidation of organically bound nitrogen in soils (Appelo and Postma, 2005). At a large scale, agriculture contributes to 50–80% of the total nitrogen load on the aquatic environment in Europe (EEA, 2005).
The problem of groundwater contamination by agricultural practices was considered by the European Environment Agency (EEA, 2003) as an area of no progress in the European Union (EU), despite the adoption of European Council Directive 91/676/EEC in 1991. Known as the Nitrates Directive, its aim was to reduce the nutrient load of agriculture on surface and groundwater, but its implementation by the member states has proven extremely difficult, due to a number of factors, such as: i) the power of the agricultural lobby (e.g. Goodchild, 1998); ii) the lack of cooperation between environmental and agricultural agencies; iii) the complexity of the contamination process, involving other sources of nitrate and iv) the lack of comprehensive monitoring networks (Stigter et al., 2006b).
Nitrate (NO3−) is a very soluble anion that does not bind to soils and has a high leaching potential, and can only be removed from groundwater through reduction (Appelo and Postma, 2005). A drinking water guideline value of 50 mg/l is considered by the EU (EC, 1998) and the World Health Organization (WHO, 2006), whereas the Environmental Protection Agency of the USA considers a maximum contaminant level of 10 mg/l of nitrogen N (44 mg/l NO3−) as primary drinking water standard (EPA, 2003). Nitrate in drinking water is generally considered to constitute a health hazard for babies and young infants, its toxicity mainly attributable to the reduction to nitrite and associated to methaemoglobinemia. However, several recent studies question the importance of nitrate in drinking water as a risk factor for methaemoglobinemia (e.g. L’hirondel and L’hirondel, 2001, Addiscott, 2006). More consensus exists regarding the environmental problems that stem from enrichment of surface fresh- and saltwaters by nitrates, causing an accelerated growth of algae and ultimately leading to eutrophication (EC, 1991). Elevated levels of nitrate are particularly significant contributors to eutrophication in marine and coastal areas, where nitrogen is the limiting nutrient (Howarth and Marino, 2006).
Groundwater contamination by pesticides can result from incorrect or poor control on their use and application (EEA, 1999), from leaching after application due to subsequent rainfall or from inappropriate disposal methods (WHO, 2006). The properties of the involved substances, their behaviour in the soil environment and the aquifer characteristics and vulnerability are also important factors (Vighi and Funari, 1995). If the pesticide and metabolite degradation rates exceed their percolation rates through the soil, contamination of groundwater is less probable (Guzzella et al., 2006), but the occurrence of preferential flow increases the pesticide contamination risk (Hallberg, 1989). In the USA, pesticide detection in groundwater has a lower frequency than in streams but with similar correspondence to land use (Gilliom, 2007). From all the classes of pesticides, herbicides are the most frequently detected pesticides in groundwater, which is related to their uses and properties (Funari et al., 1995). Groundwater contamination with pesticides is generally not dependent of the season and has a strong inertia (Funari, 1995). The parametric value for each pesticide (including their relevant metabolites, degradation and reaction products) in drinking water in the EU is 0.10 μg/l with a parametric value for total pesticides quantified as 0.50 μg/l (EC, 1998). Guidelines for drinking water from the World Health Organization (WHO, 2006) are generally less restrictive. The greatest risks of pesticides in groundwater supplies occur where: the pesticide is relatively stable, the aquifers are overlain by permeable soils with a short travel time to the water table and the aquifer dilution potential is small (Chilton et al., 1998).
The purpose of the present study is to employ a multi-method approach to the evaluation of nitrate and pesticide contamination levels and risk in shallow groundwater, influenced by lithology, land use (crop type) and irrigation water supply source. In addition to common uni- and bivariate analysis, more advanced tools such as multivariate factorial correspondence analysis (FCA) and indicator kriging are applied as interpretation and visualization tools. Multivariate analysis allows a rapid assessment of multiple correlations among quantitative and qualitative parameters in large data sets, which otherwise would be rather an arduous task. Indicator-kriging tools allow the incorporation of uncertainty in the spatial assessment of a contamination problem.
During FCA, the original variables are reduced to a small number of orthogonal factors that by definition are independent (Pereira and Sousa, 2000). As symmetry is conferred to the data matrix, correlations within and between variables and samples can be studied simultaneously (Benzécri, 1973, Pereira and Sousa, 2000). Moreover, both qualitative and quantitative variables may be used, by dividing them into classes (modalities). Since Benzécri (1973) proposed the FCA methodology, several applications to water quality studies have been published (e.g. Lachance et al., 1979, Usunoff and Guzman-Guzman, 1989, Oleson and Carr, 1990, Johannesson et al., 1996, Pacheco, 1998, Farnham et al., 2003, Pacheco and Landim, 2005, Stigter et al., 2008).
Indicator-geostatistical techniques result in maps that indicate the probability of concentrations in groundwater exceeding predetermined threshold values. According to Ribeiro (1998), such techniques were developed to allow the detection and modelling of so-called spatial dependency patterns of the variables and to evaluate the associated risk. In this non-parametric approach, by using a binary transformation at various cut-off levels and applying the kriging algorithm to these so-called indicator variables, the uncertainty around the unsampled value is modelled in a conditional distribution function (Deutsch and Journel, 1998, Stigter et al., 2005). For (groundwater) management purposes, the existence of a robust tool that specifically incorporates uncertainty into contamination assessment can be extremely useful, as was also concluded by Hu et al. (2005), who used a similar methodology to assess the risk of nitrate contamination in shallow groundwater in Quzhou County in the North China Plain. When the cut-off level corresponds to specific standards such as those defined for drinking water, the resulting maps illustrate the probability of exceeding such standards and are therefore easily interpretable for people outside the scientific domain, such as policy-makers.
For the pesticide compounds, there leaching potential is determined by calculating the Groundwater Ubiquity Score (GUS), created by Gustafson (1989), which is based on the partition coefficient between soil organic carbon and water—sorption coefficient (Koc) and the soil half-life for each pesticide compound. As adsorption is an indicator of the adherence of the pesticide to the soil components while is moving with water and persistence gives a measure of the time the pesticide stays in its original form in soil (Guzzella et al., 2006), the GUS index allows the evaluation of the pesticide leaching potential.
Section snippets
Geographic location and climate
The study area is located west of Coimbra city and it is part of the Mondego River drainage basin—Central Portugal (Fig. 1). It covers an area approximately 16 km long and 3–4 km wide, totalling nearly 51 km2. It is a plane region with a maximum elevation of nearly 13 m above mean sea level on the eastern side and a minimum elevation of 4 m on the western side, close to the upper limit of the river estuary.
It is important to emphasize that the study area is strongly affected by anthropogenic
Sampling and chemical analysis
For this study 29 observation wells, ranging in depth from 2.3 to 5.1 m were implanted in the area. Groundwater samples for nitrate analyses were collected from these observation wells, in 20 campaigns from October 2000 to September 2002, using a bailer. River water samples were also collected at two points located at both extremities of the study area, except for the first two campaigns. The location of the sampling points is shown in Fig. 2. In January and February 2001 it was not possible to
Descriptive statistics
The descriptive statistics for NO3−, NO2− and PO43− concentrations and in situ measured parameters in groundwater samples of the study area are presented in Table 4; for river samples, statistics of NO3− are also presented. Box-and-whisker plots of NO3− concentrations in groundwater are presented in Fig. 4, and their distribution over several intervals (≤5.0 mg/l, >5.0–25 mg/l, >25–50 mg/l and >50 mg/l) is represented for each sampling location in Fig. 5. It can be noted that nitrate concentrations
Herbicide presence in groundwater—results and discussion
In EEA (1999) atrazine and metolachlor are on the list of most important pesticides; atrazine is frequently considered one of the three most important. Data for molinate and propanil are not available in this report, whereas pesticides in Portugal are not considered a problem, although data are not available.
In all the sample campaigns at least one of the analysed herbicide compounds was detected with a concentration above 0.1 μg/l in one of the groundwater samples, corresponding to a total of
Conclusions
The range of nitrate values found in the shallow groundwater of the Mondego River alluvial body indicates the existence of anthropogenic contamination. FCA allows understanding that the contamination in the area is controlled by redox state and dissolved oxygen levels, as well as land use and lithology. Nitrate contamination mainly occurs under oxidizing conditions as the aerobic environment allows rapid nitrification of other forms of N. Areas where rice culture is prevalent have a
Acknowledgments
The authors wish to express their thanks to the Departamento de Ciências da Terra, Instituto de Investigação da Água and Centro de Geociências da Universidade de Coimbra for the support during the progress of this work, to the farmers who allowed the installation of the observation wells on their lands, to the Escola Superior Agrária de Coimbra for the availability of the equipment for well installation, to Eng. Filipe Melo and Sr. José Borralho for the help during the installation of the wells
References (95)
- et al.
Assessment of groundwater contamination by nitrate leaching from intensive vegetable cultivation using geographical information system
Environ. Int.
(2004) - et al.
Fertilizer-N use efficiency and nitrate pollution of groundwater in developing countries
J. Contam. Hydrol.
(1995) - et al.
Nitrate and pesticide contamination of groundwater under rice-based cropping systems: past and current evidence from the Philippines
Agric. Ecosyst. Environ.
(2002) - et al.
Monitoring water and NO3-N in irrigated maize fields in the Sorraia Watershed, Portugal
Agric. Water Manage.
(2003) - et al.
Assessment of irrigation and environmental quality at the hydrological basin level II. Salt and nitrate loads in irrigation return flows
Agric. Water Manage.
(2004) - et al.
Pesticides in Portuguese surface and ground waters
Water Res.
(2003) - et al.
Factor analytical approaches for evaluating groundwater trace element chemistry data
Anal. Chim. Acta
(2003) EU Policies for the reduction of nitrogen in water: the example of the Nitrates Directive
Environ. Pollut.
(1998)- et al.
Herbicide contamination of surficial groundwater in Northern Italy
Environ. Pollut.
(2006) Pesticide pollution of groundwater in the humid United States
Agric. Ecosyst. Environ.
(1989)