ارزیابی حضور، منشا و تغییرات فصلی آلاینده های آلی دیر پا در آب زیر زمینی از طریق نمونه گیری غیر فعال و تحلیل آماری چند متغیره Assessment of presence, origin and seasonal variations of persistent organic pollutants in groundwater by means of passive sampling and multivariate statistical analysis

۱٫ Introduction During recent years intensive research on urban areas and on the impact of agriculture on groundwater has been conducted worldwide. This increased interest can be traced to the fact that large cities and other urban agglomerations are, as a rule, supplied by drinking water from the aquifers above which they are located, or from aquifers that have urban or agricultural zones within the vicinity of their recharge areas. Trends pointing to increased pollution in large aquifer systems continue to increase year by year. Studies have shifted from identifying pollutants to determining the origin of contamination. Researchers’ attention today is focused on determining the presence of anthropogenic organic pollutants in the groundwater (Jurado et al., 2012; Stuart et al., 2012; Lapworth et al., 2015; Postigo and Barceló, ۲۰۱۵). These pollutants have been recognised as one of the important factors in environmental pollution (Wille et al., 2011). Anthropogenic organic pollutants have been found as contaminants in sewage, surface waters and groundwater, as well as in drinking water. This group of pollutants appears as pesticides and pharmaceuticals with metabolites, hormones, steroids, industrial additives, personal care products, water treatment by-products, fire retardants, surfactants and food additives (Stuart et al., 2012). Some of these contaminants can have a serious impact on human and environmental health, which underlines the need to better understand their role in the environment. Groundwater source protection starts with improving drinking water quality and effective implementation of environmental protection policies. In order to realize these goals, the possible contaminants and their sources have to be determined. With the aim of effective water management, the European Union in 2000 adopted the Water Framework Directive (Directive 2000/60/EC) of the European Parliament and of the Council, establishing a framework for community action in the field of water policy (Ambrožič et al., 2008). In order to satisfy the requirements of legislative frameworks and directives it is necessary to monitor pollutants in the aquatic environment, as many of these compounds can pose a threat to both human health and ecosystems (Vrana et al., 2005). Commonly used monitoring techniques usually record only a limited number of chemical parameters and samplings of pollutant concentrations at specific points in time, which, however, substantially increases the cost of monitoring. Contrary to such techniques, passive sampling is less sensitive to accidental extreme variations in the concentration of organic pollutants in natural waters (Kot et al., 2000) and has been proved a useful monitoring tool in detecting a large range of contaminants in the aquatic environment (Seethapathy et al., 2008; Vermeirssen et al., 2009; Nyoni et al., 2011; Wille et al., 2011). A passive sampler can cover a long sampling period by integrating the pollutant concentration over time, and the analytical costs over the monitoring period can be reduced substantially. Several types of passive samplers are in use all over the world. Semipermeable membrane devices (SPMEs) have been used to monitor various persistent contaminants in sea and surface water (Bennett and Metcalfe, 2000; O’Toole et al., 2006; Metcalfe et al., 2000, 2008). Silicone rubber passive samplers were deployed to estimate concentrations of polycyclic aromatic hydrocarbons (PAHs) resulting from wildfires, in streams in Victoria, Australia (Schäfer et al., 2010). It is believed that siliconebased passive samplers represent a promising tool for determining organic toxicants. One of the most common substances in passive samplers for adsorbing organic contaminants from both air and water is activated carbon, which has been recognised and employed for decades (Rivera et al., 1987; Kadokami et al., 1990; Hale et al., 2009; Yu et al., 2009). This type of samplers was chosen precisely for general identification purposes in our research. Strategies for sampler design, calibration, in situ sampling and quality-control issues, and advantages and challenges associated with passive sampling in aqueous environments, are considered in various studies (Kot et al., 2000; Stuer-Lauridsen, 2005; Vrana et al., 2005; Seethapathy et al., 2008; Metcalfe et al., 2011; Nyoni et al., 2011; Wille et al., 2011). Research efforts are directed towards developing a passive sampling technology for monitoring organic and inorganic pollutants in water. New monitoring techniques and data processing procedures are being developed, enabling fast and costeffective determination of the presence of pollutants and their origin. Multivariate statistical analysis is a quantitative and independent approach to groundwater classification that allows for objective grouping of groundwater samples and establishing correlations between chemical parameters and groundwater samples for normally distributed data (Prasanna et al., 2010) on continuous scale. It has been successfully applied in a number of hydrogeochemical studies designed to identify contaminated aquifer zones using nonparametric statistical tests (Robinson and Ayuso, 2004) and multivariate analysis (Suk and Lee, 1999; Love et al., 2004; Suvedha et al., 2009; Yidana and Yidana, 2009; Prasanna et al., 2010). It has been shown that multivariate statistical analysis significantly helps to classify groundwater data and identify major mechanisms impacting the groundwater chemistry (Kim et al., 2005; Singh et al., 2005; Cloutier et al., 2008). The common factor analysis method has also proved to be an appropriate approach to analysing data on the ordinal scale (Basto and Pereira, 2012), and is used in our study. Many researchers demonstrated, based on theoretical models, that polychoric correlation should be used when performing factor analysis for ordinal data instead of Pearson’s correlation matrix (Gilley and Uhlig, 1993; Holgado–Tello et al., 2010; Basto and Pereira, 2012). But there is a shortage of studies that use ordinal data and factor analysis in hydrogeochemistry and in other studies related to the natural sciences. Only a few studies were found to use polychoric correlation of ordinal data in the natural sciences (Souter et al., 2010; Norris and Michalski, 2010; Cavallero et al., 2012). Variables characterized by an ordinal scale are more common in many empirical investigations within the social and behavioural sciences (Onyx and Bullen, 2000; Cutter et al., 2003; Basto and Pereira, 2012). Using multivariate statistical analysis with ordinal data from a passive sampling technique to solve problems related to groundwater management presents a challenge. Therefore our study is focused on verifying the suitability of a methodology in which the results of groundwater passive sampling on the ordinal scale are supported by various statistical methods that could provide additional information on the studied subject. The goal of the presented study was to assess the presence of anthropogenic organic substances in the Vrbanski plateau (Maribor, Slovenia) aquifer, to study the relationships between them, and to determine their origin and seasonal variations depending on the type of anthropogenic activity using effective monitoring (passive samplers) and multivariate statistical methods to interpret the results. The factors indicating contamination according to different types of pollution should be defined using multivariate statistical methods. With this methodology we tried to explain the detected substances in groundwater in terms of hydrogeological conditions, in the framework of the dynamics of water and land use. The obtained results should provide a quick and transparent overview of potential pollution loads linked to the type of pollution (agricultural, urban or industrial). The presented methodology should be used in groundwater management for an effective assessment of the state of groundwater quality as a basis for determining the measures necessary to improve the status of groundwater.