**Wildfire E**ff**ects on Groundwater Quality from Springs Connected to Small Public Supply Systems in a Peri-Urban Forest Area (Braga Region, NW Portugal)**

**Catarina Mansilha 1,2,\*, Armindo Melo 1,2, Zita E. Martins 3, Isabel M. P. L. V. O. Ferreira 3, Ana Maria Pereira <sup>1</sup> and Jorge Espinha Marques <sup>4</sup>**


Received: 11 March 2020; Accepted: 15 April 2020; Published: 17 April 2020

**Abstract:** Peri-urban areas are territories that combine urban and rural features, being particularly vulnerable to wildfire due to the contact between human infrastructures and dense vegetation. Wildfires may cause considerable direct and indirect effects on the local water cycle, but the influence on groundwater quality is still poorly understood. The aim of this study was to characterize the chemistry of several springs connected to small public supply systems in a peri-urban area, following a large wildfire that took place in October 2017. Groundwater samples were collected in four springs that emerged within burned forests, while control samples were from one spring located in an unburned area. Sampling took place from October 2017 until September 2018, starting 15 days after the wildfire occurrence, to evaluate the influence of the time after fire and the effect of precipitation events on groundwater composition. Groundwater samples collected in burned areas presented increased content of sulfate, fluoride and nitrogen and variability in pH values. Iron, manganese and chromium contents also increased during the sampling period. Post-fire concentrations of polycyclic aromatic hydrocarbons (PAHs), mainly the carcinogenic ones, increased especially after intense winter and spring rain events, but the levels did not exceed the guideline values for drinking water.

**Keywords:** wildfire; peri-urban area; groundwater quality; polycyclic aromatic hydrocarbons; major ions; metals

#### **1. Introduction**

Wildfires are one of the main obstacles to the sustainability of forests and related ecosystems. The resulting devastation may cause severe economic and social costs, with the loss of lives and infrastructures, as well as the disturbance of the provision of goods and services, including ecosystem services, together with environmental damages such as the loss of carbon sequestration [1–3].

The urbanization of forest areas constitutes a new fire risk scenario. Over the years, throughout the world, the contact areas between human infrastructures and wild vegetation increased (the so-called peri-urban or wildland-urban interfaces) [4–6]. In Portugal, these areas result mainly from the abandonment of croplands since the seventies, with fewer people to manage the land, creating great quantities of highly flammable combustible material, mostly from non-native species. This spatial pattern of human presence in the territory leads to particularly vulnerable areas to wildfire impacts, with disastrous consequences for populations and the environment [7,8]. Climate-related droughts and extreme weather are also propitious conditions for wildfires.

In 2017, Portugal suffered the greatest devastation caused by wildfires in a single year. Although in the last decades there have been a high number of fires, when compared to other countries and similar conditions, the wildfires that took place in 2017 largely exceeded the suppression capacity of the emergency services and were disastrous. Consequently, great damages occurred to population, with the loss of 115 lives, as well as to forests, rangelands, rural, industrial and urban areas. These catastrophic wildfires were worsened by climate change (it was the driest summer in nearly 90 years) and adverse meteorological conditions, and by the vegetation cover of fire-prone species, where the eucalyptus *(Eucalyptus globulus)* and maritime pine (*Pinus pinaster)* are dominant [2,9,10]. The burnt area was 539,921 ha, representing 498% of the average of the previous decennium, which was 90,269 ha, and nearly 60% of the total area burnt in the entire European Union, in which Portugal only represents about 2.1% of total landmass. The most critical month was October, with 3234 rural wildfires (15.4% of total annual rural wildfires) and 289,124 ha burnt (53.5% of total area). Wildfire occurrence prevailed mostly in the urban districts, often in peri-urban areas, which registered 55.6% of the total number of fires [10].

Regarding surface water and groundwater resources, the burning of forest catchments may result in a long-lasting legacy of water quality deterioration, whose magnitude and persistence can be observed from a few months to several years after the wildfire. Post-fire water quality concerns are complex and vary significantly from place to place depending on the severity, intensity, and duration of the fire, the soil and vegetation cover characteristics, the geological and geomorphological nature of the terrain, and the amount and intensity of precipitation during post-fire rain events [11,12]. Water quality impacts may also result from indirect effects associated with smoke and aerial deposition of ash [13,14].

According to the literature, wildfires may cause changes in several water quality parameters of interest or concern to water systems [11,15,16]. Contamination of streams and water reservoirs by post-fire inputs of suspended sediments and various trace elements present in ash may be problematic for both health and aesthetic reasons. High concentrations of iron (Fe), manganese (Mn), zinc (Zn), sodium (Na+) and cloride (Cl−) cause organoleptic problems (taste, color, staining of pipes and fittings). Sulfates (SO4 <sup>2</sup>−) have purgative effects for concentrations over 500 mg/L. Poisoning may occur from continued consumption of water containing high concentrations of copper (Cu), with gastrointestinal symptoms. Arsenic (As) and cromium (Cr) (specially hexavalent Cr) may be carcinogenic, while aluminium (Al), lead (Pb) and mercury (Hg) are toxic when consumed in sufficient quantities for prolonged periods. Following wildfire, increased exports of nitrogen (N) and phosphorous (P), in various forms, can also be problematic for managers of water supply catchments. High concentrations of nitrates (NO3 <sup>−</sup>) and nitrites (NO2 −) also present a potential risk to human health, primarily through reduction of NO3 <sup>−</sup> to NO2 −, which may affect oxygen transport in red blood cells, while high concentrations of ammonium NH3 +/NH4 <sup>+</sup> may corrode copper pipes and fittings. N and P are limiting nutrients for growth of aquatic plants, algae and cyanobacteria in water bodies. Eutrophication increases the risk of potentially toxic blooms, with implications for human health, aesthetic problems (taste, odor and color), and aquatic ecosystem function disturbance [14,17].

Polycyclic aromatic hydrocarbons (PAHs), such as benzo[*a*]pyrene, are known for their potential teratogenicity, carcinogenic and mutagenic properties, explained by the formation of adducts between the DNA bases and epoxides derived from hydrocarbons after an oxidizing process in the liver [18]. In addition to cancer, long-term exposition can also cause chronic bronchitis, skin problems and allergies. Some compounds are also classified as potential endocrine disruptors because they have estrogenic activity, generally coupled with a high potential for bioaccumulation [15,16,19,20]. The United States Environmental Protection Agency (USEPA) designated several PAHs as priority pollutants that should be regulated due to their high toxicity and adverse effects [21]. PAHs were also designated as priority

hazardous substances by the European Commission, in Directive on Environmental Quality Standards (Directive 2008/105/EC) [22].

The effects of fire on surface water are more evident that those on groundwater. Recharge rates, net infiltration and water balance can change after a wildfire, leading to difficulties in maintaining the supply of potable groundwater to populations [23]. Nowadays, despite the widespread access to tap water provided by the Portuguese municipalities, many people still prefer to drink groundwater from public fountains supplied by nearby springs, as it is associated with high purity and pleasant organoleptic characteristics. However, the impact of wildfires on the composition of groundwater from nearby springs is unknown.

The goal of this study was to identify the impact of the wildfire that occurred on October 2017, on the chemical characteristics of water from springs connected to small public supply systems used for human consumption in the peri-urban area of the city of Braga, in NW Portugal. The study included the analyses of major ions and trace elements, namely PAHs, which were carried out during one year after the wildfire.

#### **2. Materials and Methods**

#### *2.1. Hydrogeological Framework*

The city of Braga (41◦32 N; 8◦25 W) is located in NW Portugal (Figure 1), in the Minho Province. The municipality has a resident population of 181,494 inhabitants (in 2011) [24], representing the seventh largest municipality in Portugal (by population) whit an area of 183.40 km2. The city is surrounded by peri-urban areas consisting of agroforestry systems, especially in the higher and steeper terrain (Figures 1 and 2). To the southeast, the urban area is bordered by a mountainous ridge (Figure 1) with altitude above 500 m at the summits of Santa Marta (562 m), Monte Frio (548 m) and Sameiro (572 m). To the north, this region falls into the Cávado river catchment, while to the south it falls into the Ave river catchment.

**Figure 1.** Location of Braga region in Iberia and in NW Portugal; hypsometric features of the study area; location of the studied springs.

**Figure 2.** Some aspects of the study region: (**a**) burnt area of Eucalyptus globulus (first plan) and the Braga urban area (background); (**b**) Santa Maria Madalena sampling point (S1); (**c**) Monte de Dadim spring (S5); (**d**) Quercus suber burnt area at the vicinity of Santa Marta de Leão spring (S2); (**e**) Eucalyptus globulus burnt plantation at the recharge area of the Mina Spring; ash covered burnt soil immediately after the fire (**f**) and soil showing intense erosion one year after the fire (**g**); (**h**) deciduous forest at Monte de Dadim (S5, control point).

The Braga climate has Atlantic features, with a mean value of annual precipitation around 1449 mm at the local climatic station (located at 190 m of altitude). The temporal distribution pattern of precipitation is similar to the one usually observed in NW Portugal: December is the rainiest month (about 220 mm) and July is the driest month (about 22 mm). The mean annual air temperature is around 15 ◦C, with maximum in July and August (21 ◦C) and minimum in January (9 ◦C). The Braga Köppen–Geiger climate classification is Csb, which is dominant in Northwestern Iberia and corresponds to warm temperate (C), with dry and warm summer (sb) [25,26].

Braga region is located in the following geomorphological units [27]: Central System (1st level unit) and Entre Douro e Minho Open Valleys and Hills and Atlantic Front of NW Iberia Mountains (2nd level units). The region is also located in the Central-Iberian Zone of the Iberian Massif [28]. The dominant regional geological units are granitic rocks and metasedimentary rocks; sedimentary cover areas are residual. Therefore, the major processes in the local groundwater cycle take place mostly in fractured circulation media and, to a minor extent, in porous media.

The unsaturated zone depth often reaches more than 20 m at hilltops and less than 10 m at valley bottoms. In addition, the structure of the unsaturated zone encompasses a soil cover (mainly Leptosols, Regosols and Cambisols, with an umbric A horizon) overlying granite or metasedimentary rock.

#### *2.2. Wildfire Description*

Between 13 and 18 October of 2017, more than 7900 wildfires affected Northwestern Iberia. During this month, 15,440 wildfires were active, 33 of which were of important size, which spread quickly due to the drought and to particular meteorological conditions: strong winds caused by the Ophelia hurricane that swept the western coast of the Iberian Peninsula and unusual air temperatures, above 30 ◦C. September 2017 was the driest month in 87 years [29]. Around 81% of the territory was under severe drought and 7.4% in extreme drought. In the district of Braga, 23 wildfires were registered on 14 October and 43 wildfires on 15 October. The wildfire of greatest impact on the study region started in Leitões, Guimarães, and quickly reached Braga (particularly the mountain area between Santa Marta and Sameiro, see Figure 2), consuming over 1200 ha of forest [29].

The wildfire started at 13:00h on 15 October and was controlled within in a few hours. However, a strong reactivation projected the fire in several directions, with a very significant speed of propagation, and by 17:17h the authorities requested additional means to protect the houses, since the fire had two very long and intense fronts. Only by 10:09h, on 16 October, the wildfire was finally controlled [30].

#### *2.3. Water Sampling*

Groundwater samples were collected in springs that supply non-treated water used for human consumption, located in near Braga city (Figure 1): (i) Santa Maria Madalena Fountain (S1); (ii) Santa Marta de Leão Fountain (S2); (iii) Tanque de Dadim Fountain (S3); (iv) Depósitos Spring (S4); (v) Monte de Dadim Spring (S5). The recharge areas of S1 to S4 springs were affected by the wildfire, while S5 corresponds to a spring located in an unburned area, which was used as control.

All springs are characterized by relatively shallow water circulation. In fact, water is abstracted by means of hand dug galleries which, in this region, are typically associated with springs connected to the upper saturated zone and, sometimes, also to interflow. Moreover, springs S1 and S2 are located at hilltops and, therefore, have very short circulation paths. Finally, S3, S4, and S5 springs are all located at slopes and abstract water circulating in the granite weathering mantle.

The site selection criteria were the existence of permanent flow throughout the year, the location of the recharge areas regarding the burnt region, and the sampling feasibility. These sampling points were also chosen to be preserved as much as possible from other anthropogenic impacts, reducing the risk of water contamination by sources of pollution other than wildfires (Figures 1 and 2 and Table 1). The control spring analytical results were used to provide a framework of reference for contaminant concentrations on burned areas. The discharge flow in all springs was under 0.5 L/s.


**Table 1.** Features of the sampling points and the corresponding spring recharge areas.

Samples collected at S1, S2, S3 and S4 were designated as Burnt samples (BR), and the ones collected at S5 as unburnt samples (NB). The sampling plan included five campaigns, carried out during one year, from October 2017 until September 2018, starting 15 days after the wildfire occurrence (before the first post-fire rain event), and included 25 samples.

Samples were collected according to ISO 5667-3:2003(E) Water quality—Sampling—Part 3: guidance on the preservation and handling of water samples, during non-storm conditions, before and after the first overland flow event and aquifer recharge following the fire.

The influence of wildfire on water quality constituents was investigated. The physicochemical analyses of key water quality constituents included the major ions, organic matter and heavy metals, which may be mobilized according to fire intensity, post-fire precipitation, geological and geomorphological conditions and vegetation cover of each site. Compounds that can serve as specific markers, indicative of wildfire, such as the high molecular weight PAHs, were also monitored. Water pH, electrical conductivity (EC) and temperature were measured *in situ* during sampling.

The temporal distribution of precipitation and climate data (measured at Braga meteorological station) during the study period, provided the knowledge of the influence of rainfall on the input of pollutants into groundwater [31].

#### *2.4. Laboratory Analyses*

Analyses were performed according to procedures outlined in *Standard Methods for the Examination of Water and Wastewater 23rd edition* and in *Le Rodier*—*L'analyse de l'eau 10e édition*. The laboratory has been accredited under ISO/IEC 17025 since 2007. Precision and accuracy were calculated for all analytical methods with values <10%. Uncertainties were also calculated with results varying from 2% to 10%.

Water turbidity was measured in a Hach 2100N Laboratory Turbidity Meter. Electrical conductivity (EC) and pH were determined in a Crison MultiMeter MM 41. Total alkalinity and total hardness were analyzed by titration and reported as milligrams per liter of calcium carbonate (mgL−<sup>1</sup> CaCO3). Color, PO4 <sup>2</sup><sup>−</sup> and total phosphorus, expressed as P (TP) were analyzed in a Shimadzu UV-1601 Spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Total nitrogen, expressed as N (TN) and chemical oxygen demand (COD), were evaluated in a Hach DR 2800 Spectrophotometer (Hach Company, Loveland, CO, USA). Major inorganic ions (Na+, K+, Mg2+, Ca2+, Li+, Cl−, NO3 −, F− and SO4 <sup>2</sup>−) were analyzed by ion chromatography (DionexTM system DX-120/ICS-1000, Dionex Corporation, Sunnyvale, CA, USA). Total organic carbon (TOC) was analyzed in a Shimadzu TOC-V (TOC-ASI-V, Shimadzu Corporation, Kyoto, Japan), heavy metals (Cr, Mn, Ni, Cu, Zn, As, Cd and Pb) and other components, such as Al, Fe, NO2 <sup>−</sup>, NH4 <sup>+</sup> and SiO2, were, analyzed in a Varian AA240 Atomic Absorption Spectrometer (Varian Inc., Palo Alto, CA, USA) and in a Continuous Segmented Flow Instrument (San-Plus Skalar, Skalar Analytical, Breda, The Netherlands), respectively. PAHs were analyzed by dispersive liquid-liquid microextraction coupled to gas chromatography/mass spectrometry (DLLME–GC/MS) methodology in a Shimadzu GCMS-QP2010 gas chromatograph mass spectrometer equipped with an auto injector AOC5000 (Shimadzu Corporation, Kyoto, Japan), according the procedure described in Borges et al. [32].

Analytical standards were supplied by Sigma–Aldrich (Steinheim, Germany) and Merck (Darmstadt, Germany). The reference standard mixture containing the 15 EPA PAHs (acenaphthylene, Acy; acenaphthene, Ace; fluorene, Flu; phenanthrene, Phe; anthracene, Ant; fluoranthene, Flt; pyrene, Pyr; benz[*a*]anthracene, BaA; chrysene, Chr; benzo[*b*]fluoranthene, BbF; benzo[*k*]fluoranthene BkF; benzo[*a*]pyrene, BaP; dibenz[*a,h*]anthracene, DahA; benzo[*ghi*] perylene, BghiP; and indeno[*1,2,3-cd*]pyrene, Ind) was purchased from Sigma-Aldrich (Steinheim, Germany).

Methanol, dichloromethane and acetonitrile were organic trace analysis grade SupraSolv and were supplied by Merck (Darmstadt, Germany). Ultrapure water was highly purified by a Milli-Q gradient system (18.2 mΩ/cm) from Millipore (Milford, MA, USA).

#### *2.5. Statistical Studies*

Data on chemical concentration was analyzed throughout time, location, and throughout time in burned areas. All dependent variables from every analyzed parameter were tested for distribution of the residuals with the Shapiro–Wilk's test. Chemical concentrations were studied using a one-way analysis of variance (ANOVA), if normal distribution of the residuals was confirmed. Welch correction was applied when the homogeneity of variances was not verified. Whenever statistical significances were found, Tukey's test or the Tamhane's test post-hoc tests were applied for mean comparison, depending on variances assumption or not.

If normal distribution of the residuals was not found, parameters analyzed were studied using a Kruskal–Wallis test. Whenever statistical significances were found, Dunn's post-hoc test was applied for median comparison.

All analyses were performed at 5% significance level, using XLSTAT for Windows version 2014.5 (Addinsoft, Paris, France).

#### **3. Results and Discussion**

The effects of wildfires on the catchment's hydrologic responses were noted worldwide, although the specific impacts are unpredictable in terms of both the magnitude of potential effects and the persistence of the influence. So, there is no clear pattern for wildfire effects on water bodies, which may not be affected or, on the other hand, experience fire-related changes that can range from aesthetic concerns (taste or appearance) to potential toxicity or carcinogenicity with prolonged exposure, as well as environmental damages [14,33].

The concentration of major and trace elements monitored in groundwater collected at the peri-urban area of Braga, after the wildfire of October 2017, shown significant differences between samples for several compounds. However, for others, a clear tendency that could be attributed to the wildfire was not observed, or the variation over time was similar to that observed in control samples.

Post-fire analytical results are shown in Table 2 and corresponding statistical analysis on Supplementary Table S1.

The analyses of the results show a slight decrease followed by an increase in pH values in S1 and S2, and an increase in pH values in S3 and S4 during the sampling period, wherein 18 January and April 2018 were grouped, whereas the remaining sampling times were independent of each other (p = 0.006). This is in accordance with studies that described wood ash as being alkaline and rich in carbonates and metal oxides [14]. However, ash chemical composition is highly variable as it reflects the type of vegetation and the part of the plant burned, as well as the soil type and combustion conditions, with different implications on water quality parameters [14,34]. No statistical differences (p > 0.050) were found for turbidity, color and silica. The electrical conductivity, which is a water quality indicator for estimating the amount of mineralization and total dissolved solids, remained also constant during the study period, with no significant variations in BR samples (p > 0.050).

Wildfires often induce quantitative and qualitative changes in soil organic matter, sometimes with significant losses due to the partial or total removal of the litter layer and, possibly, some organics from the upper few centimeters of mineral soil [35]. However, in this study, the total organic carbon concentrations appeared to be unaffected by fire, with mean values of 0.37 <sup>±</sup> 0.15 mgL−<sup>1</sup> and 0.33 <sup>±</sup> 0.17 mgL−<sup>1</sup> in BR and NB samples, respectively, and the chemical oxygen demand increased immediately after the fire, reaching a peak during spring, but then slowly declined as vegetation re-established. Nevertheless, these variations could not be attributed to the wildfire, as they occurred in the NB samples too (p > 0.050).



Few compositional changes regarding major ions were observed. The Piper diagram presented in Figure 3 highlights the main changes: all sampling points reveal a minor shift from the first campaign (October 2017, at the end of the dry season) to the third campaign (April 2018, after the intense March precipitation events). The cations content shift results mainly from a decrease in calcium and an increase in sodium, while the anions content shift results from an increase in chloride and sulfate. However, changes in calcium and chloride ions observed in the burnt area are similar to those observed in the unburnt area and, thus, may not be connected to the wildfire. A decrease in bicarbonate is also observed in S1 to S4 points, but not in S5 point. Bicarbonate is originated by water–rock interaction and depends on the nature of the groundwater flowpaths that, in this case, are poorly known.

**Figure 3.** Piper diagram comparing the October 2017 campaign to the April 2018 campaign.

Regarding the sampling date, ANOVA analyses of major ions also reveal that chloride, sodium, potassium, calcium and magnesium concentrations did not vary significantly (p > 0.050). In contrast, the increase of sulfate in groundwater from the burnt areas was statistically significant (p = 0.034) and is probably due to the oxidation of sulfur in soil organic matter after the wildfire. Similar results were already observed in an earlier study carried out in Caramulo region, in Central Portugal [36] and reported in the literature by other authors [14].

The levels of fluoride have also risen (p = 0.003) in May by approximately two orders of magnitude regarding the control point (S5), decreasing to levels similar to the initial ones in the last campaign.

Nutrient export from burnt soils usually increases after wildfires, and this process may affect groundwater composition. However, wildfire effects on stream exports of total phosphorus (TP) and total nitrogen (TN) vary significantly [14,37,38]. In the present study, it was observed a small decline (despite this, with no statistical significance) of phosphate, but an increase of TP with multiple change of 1.2 to 4.2 times the initial values. The wildfire effect on nitrogen should be examined with caution, because small differences regarding pre-fire values or control samples could be very important [39,40]. During the first six months after fire, the concentration of nitrogen species increased significantly (p < 0.001), especially after precipitation, compared with concentrations in the reference samples. Combined concentrations of post-fire nitrite, nitrate and ammonium, which are dissolved forms, varied from 0.52 mgL−<sup>1</sup> in control to 6.28 mgL−<sup>1</sup> (mean values) in S4 sample. The land cover at the S4 recharge area consists only of forest, without other pollution sources besides the wildfire. The higher nitrate content observed in S4 should be a result of the wildfire and is in agreement with the thicker layer of ash observed in the recharge area of this spring. Several factors could explain the increase of

nitrogen exports in post-fire situations. On one hand, there is a lower plant demand and the nitrogen mineralization is stimulated (due to changes in pH and electrolytes). On the other hand, the nitrate form is mobile in soil-water systems and leaches through soil into catchments and drainages after heavy overland flow events that follow the wildfire [39,41]. In addition, because the wildfire removes forest cover and litter, rain interception decreases and nutrient transport via infiltration may increase. As well as nitrate, ammonium loading may increase as it is volatilized during fire and can dissolve into water. This compound may be retained in soil in its exchangeable form and subsequently be leached.

Hazardous chemicals, such as metals, were also monitored in Braga, as wildfires may influence the concentration of trace metals differently, with potentially harmful effects to human health and the environment. In this study no significant differences between samples collected in NB and BR locations were observed for cadmium (Cd), arsenic (As), lead (Pb), nickel (Ni), copper (Cu) and zinc (Zn). In contrast, the mean levels of iron (Fe), manganese (Mn) and chromium (Cr) were about 1.3, 7.0 and 2.8 times above the NB sample levels, respectively. Substantial post-fire increases in total iron and total manganese have been reported in the literature, indicating an added influx of these metals as part of an increase in particulates [42]. Even at values below the criteria established for aquatic systems, results must be considered as iron and manganese are related to aesthetic issues of the water (taste and color), and chromium, namely the hexavalent form, is carcinogenic. Concentrations of metals in groundwater samples are displayed in Table 3.


**Table 3.** Descriptive statistics of selected trace elements in groundwater samples after the wildfire.

Min, Max, Avg, for minimum, maximum and average. Values of trace elements are in μg/L. Limits of detection (LD): 0.1 μg/L for Cd, As, Pb, Ni, Cu, Zn and Cr; 5 μg/L for Fe.

#### *PAH Analyses*

This study also included the analyses of the 15 PAHs designated as priority hazardous pollutants. Thirteen hydrocarbons were found during the sampling period. The most abundant compounds in samples collected in burnt areas were Ant (26%), Acy (17%), BaA (15%) and BaP (14%), while in control samples Acy was the dominant compound (33%), followed by BaA (21%) and Ant (18%) (Figure 4). The sum of total concentrations, as well as the variety of PAHs, has increased throughout the year in all sampling points, with maximum values in May, after the wet season precipitation events. Mean values ranged from negligible amounts in S5 sample in January to 0.029 μgL−<sup>1</sup> in S3 in May. Comparing the

initial to the last campaign, total values increased by factors of 1.3 to 2.2 times the initial concentrations. Furthermore, in control samples, only light PAHs, with three to four rings (Acy, Ace, Flu, Phe, Ant, Flt, Pyr, BaA, and Chr), were detected, in contrast to the samples of burnt areas, where a different profile was observed throughout the study period. Heavy PAHs, with five (BbF, BkF, BaP, and DahA) and six rings (InP, BghiP), have increased until May followed by a decrease in September, which probably indicates a natural remediation concerning these compounds. The compositional pattern of PAHs by ring size for the water samples is shown in Figure 5.

**Figure 4.** Average concentrations of polycyclic aromatic hydrocarbons (PAHs) (individual fractions) in S5 unburnt samples (NB) and in burnt areas (BR) (n = 4) and monthly precipitation.

**Figure 5.** PAHs profiles of the samples collected in burned and control areas in the five sampling campaigns, according to their structural composition (number of benzene rings).

Significant differences for some PAHs were observed if the date of sampling was considered. Results from Shapiro–Wilk's test shown that PAHs do not have a normal distribution of the residuals. Kruskal–Wallis analysis for the sum of PAHs revealed that October 2017 and May 2018 sample results were significantly different. (Table S1).

Regarding the carcinogenic PAHs (BaA, Chr, BbF, BkF, BaP, Ind and DahA), the concentrations were also significantly higher in May 2018 (p = 0.022), at levels around 0.010 μgL−<sup>1</sup> (0.0098 to 0.013 μgL<sup>−</sup>1) in BR samples, representing 60.4%, 52.9%, 36.2% and 62.6% of total PAHs in S1, S2, S3 and S4, respectively. Approximately half of these values correspond to BaP, with concentrations ranging from 0.0046 μgL−<sup>1</sup> in S4 to 0.0077 μgL−<sup>1</sup> in S1. In S3, concentrations of BaP increased in April 2018 and remained approximately constant up to September 2018, with a mean value of 0.0052 <sup>±</sup> 0.0006 <sup>μ</sup>gL−1. In S5 samples (NB), BaP was not detected during the study period (Figure 6).

**Figure 6.** Total levels of carcinogenic PAHs (benz[*a*]anthracene (BaA), chrysene (Chr), benzo[*b*]fluoranthene (BbF), benzo[*k*]fluoranthene (BkF), benzo[*a*]pyrene (BaP), indeno[*1,2,3-cd*]pyrene (Ind) and dibenz[*a,h*]anthracene (DahA)) in NB (S5-control) and BR (S1 to S4) samples in the five campaigns.

Groundwater quality standards for PAHs (referred to as "threshold values") were established by several European Member States taking into account identified risks [43]. Ranges of threshold values throughout Europe varied from 0.005 to 0.03 μgL−<sup>1</sup> for BaP, 0.01 to 0.1 for BghiP and BbF, and 0.05 to 0.1 μgL−<sup>1</sup> for BkF. Regarding these values, our results reveal several nonconformities for BaP, namely in samples collected in April and May in S1, S2 and S3. BaP, is the most extensively studied carcinogenic PAH, classified by IARC as a *Group 1* or a known human carcinogen [44], and is in the top ten priority pollutants designated by the Agency for Toxic Substances and Disease Registry (ATSDR) in 2015 [45]. BaP is also considered a powerful endocrine disruptor compound [46]. Despite the parametric value for drinking water established by the European Union Council Directive 98/83/EC for BaP has not been exceeded (0.010 μgL−1) [47], attention is needed especially regarding mixture cancer potency, as individual PAHs occur as part of environmental mixtures and cumulative risk assessment should be considered [48]. The established limit of 0.100 μgL−<sup>1</sup> for the sum of concentrations of BbF, BkF, BghiP and Ind was also not exceeded.

Regarding PAHs sources, they may have a pyrogenic origin, linked to processes of incomplete combustion of organic matter (in particular, vegetation or fossil fuels) or a petrogenic origin, resulting from the transformation of organic matter that occurs in geological materials [49]. Another source of PAH is the activity of plants, algae/phytoplankton, and microorganisms (biogenic PAHs) [50,51]. Diagnostic ratios based on PAHs physical and chemical properties and stability against photolysis can be used to distinguish between petrogenic and pyrogenic origins. Congener ratios of (Ant/Ant + Phe)

> 0.1 and (Phe/Ant) < 10 indicate pyrogenic sources. The (BaA/BaA + Chr) and (Flt/Flt + Pyr) ratios are used to further distinguish combustion sources: high ratios (>0.35 and >0.50, respectively) indicate grass, wood or coal combustion, intermediate ratios (0.20–0.35 and 0.40–0.50, respectively) indicate liquid fossil fuel combustion or mixed petrogenic and pyrogenic origin and low ratios (<0.20 and <0.40) usually imply petrogenic sources [52,53].

The ratios of (Ant/Ant + Phe) and (Phe/Ant) were calculated with results between 0.68–0.82 and between 0.23–0.50, respectively, for samples in which PAHs were detected. The application of (BaA/BaA + Chr) and (Flt/Flt+Pyr) ratios was also performed, with values > 0.35 and >0.50, also suggesting a pyrogenic source of PAHs (Figure 7).

**Figure 7.** Ratios of (anthracene (Ant)/Ant + phenanthrene (Phe)) and (fluoranthene (Flt)/Flt + pyrene (Pyr)) for samples in which PAHs were detected.

#### **4. Conclusions**

This study identified the main impacts of a large forest wildfire on groundwater quality of springs connected to small public supply systems in a peri-urban area. Results pointed out that the best practice for assessing wildfire hydrochemical effects is to start monitoring programs immediately after the wildfire event, and proceed with sampling campaigns for at least 12 months, since several parameters show considerable variation through time. It was also found that extreme events, like intense precipitation, were much more important for groundwater contamination than long-term average changes.

An increase in several parameters, such as sulfate, fluoride, phosphorous, nitrogen compounds, iron, manganese and chromium was observed in Braga peri-urban aquifers. Concerning PAHs, the results reflected the fire impact mainly through the profile of the compounds that appear in the BR springs, which differ significantly from the control spring. Six months after the wildfire, and after the first intense rain event, carcinogenic PAHs, including BaP, began to be detected in considerable concentrations, corroborating the idea of it being difficult to predict the long-term impacts of wildfires on groundwater quality.

Although the connection between groundwater depletion and destructive wildfires might seem tenuous at first glance, and the parametric values for drinking water established by international guidelines have not been exceeded, the results clearly demonstrate the vulnerability of aquifers to wildfires, especially for PAHs, which constitutes an issue yet poorly understood in terms of both the magnitude and persistence.

More information is needed on appropriate monitoring strategies in order to identify a standard set of trace and major compounds to be analyzed and establish protocols to effectively assess water quality, which is essential for developing sustainable water resources management practices.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4441/12/4/1146/s1, Table S1: Values for statically analyses trends in the different sampling time for burned areas.

**Author Contributions:** Conceptualization, C.M. and J.E.M; Data curation, C.M., A.M., Z.E.M., J.E.M.; Formal analysis, A.M., A.M.P.; Writing—original draft, C.M., A.M., J.E.M.; Writing—review and editing, C.M., I.M.P.L.V.O.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work received financial support from the European Union (FEDER funds POCI/01/0145/FEDER/007265) and National Funds (FCT/MEC, Fundação para a Ciência e Tecnologia and Ministério da Educação e Ciência) under the Partnership Agreement PT2020 UID/QUI/50006/2013. The author J. Espinha Marques acknowledges the funding provided by the Institute of Earth Sciences (ICT), under contracts UIDB/04683/2020 with FCT (the Portuguese Science and Technology Foundation), and COMPETE POCI-01-0145-FEDER-007690.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## *Article* **Irrigation with Coal Mining Effluents: Sustainability and Water Quality Considerations (São Pedro da Cova, North Portugal)**

**Catarina Mansilha 1,2,\*, Armindo Melo 1,2, Deolinda Flores 3,4, Joana Ribeiro 3,5, João Ramalheira Rocha 4, Vítor Martins 4, Patrícia Santos 3,4 and Jorge Espinha Marques 3,4**


**Abstract:** Two water effluents that drain from the abandoned coal mine of São Pedro da Cova (NW Portugal) were characterized in terms of their physic-chemical properties and suitability for irrigation purposes. Samples were also collected in a local surface stream, upstream and downstream from the mine drainage points, also used for irrigation by local farmers. Water samples were analyzed for major and minor ions and for trace element concentrations. Sampling campaigns started in 2017 and ended in 2019 and there were 46 water quality parameters tested. There were also proposed allinclusive indices (the Water Quality Index and the Contamination Index, and also the Trace Element Toxicity Index) based on specific groups of 18 and 17 physic-chemical parameters, respectively, to achieve adequate monitoring requirements for mine effluents and surface water from coalfield. From the physical and chemical aspects of mine water it is inferred that the mine is not producing acid mine drainage. The coal mine water is of medium to high salinity, having almost neutral pH and a high thermal stability during the year, which is a distinguishing feature of the effluents. When compared to international irrigation water quality standards, as Food and Agriculture Organization of the United Nations admissible concentrations, the impacted waters are unsuitable for irrigation. The major outliers to the guidelines were iron, manganese, potassium, magnesium and bicarbonates, being also detected carcinogenic polycyclic aromatic hydrocarbons. Cost-effective ways of monitoring water quality parameters are needed to help control and manage the impact of coal mine effluents that should be treated before releasing into a ditch system that could be then used by local farmers to irrigate their crops.

**Keywords:** coal mine wastewater quality; irrigation; heavy metals; water quality index; environmental impact

#### **1. Introduction**

Irrigation is fundamental for agriculture but policies that push towards a restrained use of water are not popular among farmers, who are also not prepared to respond to drastic increases in water costs, which could decrease the economic profitability of their activities. Therefore, the use of unconventional free water sources, such as mine wastewaters, is regarded as a possible choice for irrigation. Effluents from coal mines are often considered severe and persistent forms of pollution, with environmental impact not only throughout the mine's life cycle, but also long after the end of mining activities. Some deleterious impacts on the environment include the disruption of hydrological pathways,

**Citation:** Mansilha, C.; Melo, A.; Flores, D.; Ribeiro, J.; Rocha, J.R.; Martins, V.; Santos, P.; Espinha Marques, J. Irrigation with Coal Mining Effluents: Sustainability and Water Quality Considerations (São Pedro da Cova, North Portugal). *Water* **2021**, *13*, 2157. https:// doi.org/10.3390/w13162157

Academic Editor: William Frederick Ritter

Received: 21 June 2021 Accepted: 31 July 2021 Published: 5 August 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

contamination of surface and groundwater, depression of the water table, soil contamination and loss of biodiversity [1,2]. The concentration of contaminants in coal mine drainage waters vary greatly and depends on a series of geological, hydrological and mining conditions, which are different from mine to mine. Therefore, the effluents can be alkaline, acidic, ferruginous, highly saline, or even clean [3]. Frequently, coal mine drainage is acid metal-rich waters with high concentrations of iron (Fe), copper (Cu), manganese (Mn) and nickel (Ni), formed during water–rock interaction involving sulfur-bearing minerals, such as pyrite (FeS2). These processes may cause red, orange, or yellow sediments with negative impacts on ecosystems and water resources. Additionally, mine effluents often contain high levels of total dissolved and suspended solids. The dissolved cations include mainly calcium (Ca), magnesium (Mg), sodium (Na) and potassium (K); the major anions are sulfate (SO4), chloride (Cl), fluoride (F), nitrate (NO4), bicarbonate (HCO3) or carbonate (CO3) [4].

The Douro Coalfield (NW Portugal) represents the most important coal-bearing deposit in Portugal [5–8] with 53 km length and width between 30 and 250 m (Figure 1).

São Pedro da Cova mining area was one of the principal centers of mining activity in Portugal (Figure 2a,b), with great economic and technological impacts, and a cultural significance from the end of the 17th century (1795), until the 20th century (1970) [9].

**Figure 1.** Some aspects of the study area: (**a**) São Pedro da Cova mine in the mid 20th century; (**b**) São Pedro da Cova mine in 2021; (**c**) Silveirinhos stream upstream from the mining effluents discharge; (**d**) mining effluents discharge; (**e**) water monitoring in Silveirinhos stream downstream from the mining effluents discharge; (**f**) irrigation with polluted water from Silveirinhos stream; (**g**) agricultural area irrigated with polluted water from Silveirinhos stream.

The São Pedro da Cova coal mine is an abandoned mine located in a peri-urban area, with a landscape consisting of a mosaic of urban, industrial, agricultural and forest areas (Figure 2a,b). The mine is located very close to a densely populated zone with several basic facilities including schools, a healthcare center and a leisure center.

The coal mining effluents from two mine drainage galleries are discharged around 1 km to SE of the mine, producing an ocher-colored sediment that is continuously accumulated in local watercourses (Silveirinhos stream and Ferreira river). Water from Silveirinhos stream is collected downstream from the mine drainage discharges and is locally used for agriculture irrigation. Local farmers have been using this polluted water for decades to

produce a wide variety of products, including green vegetables, corn, barley and fruits (Figure 2c–g).

**Figure 2.** São Pedro da Cova location and geological setting (modified from [10]).

This study aimed at assessing the impact of coal mining effluents on the quality of irrigation water, by means of water quality individual parameters and indices.

This approach intended to contribute to the development of a specific evaluation methodology for coal mining regions in terms of i) the suitability of this type of water for irrigation purposes and ii) the environmental impacts resulting from the uncontrolled applications of these unconventional water resources in irrigated areas.

#### **2. Assessment of Irrigation Water Quality: A Short State of the Art**

Mine effluents are used for irrigation purposes worldwide, as they constitute an easily accessible and inexpensive source of water.

The quality of irrigation water is considered a key factor for safe food production [11]. However, when good quality water is scarce, water of marginal value is often considered for use in agriculture [12]. The prolonged and uncontrolled use of polluted water could result in reduced crop yields, deterioration of soil properties and severe environmental and health damages, requiring more complex management practices and more stringent monitoring procedures [13]. Therefore, the agricultural use of water resources in areas affected by mining activities requires not only a baseline water quality data, but also continuous monitoring.

There are several quality standards and guidelines for irrigation water proposed by various countries and organizations, combining conservative and liberal approaches. However, despite being useful, they are not always satisfactory due to the wide variability of hydrological and hydrogeological settings [14–17].

In 1985, the Food and Agriculture Organization of the United Nations (FAO) published a document concerning water quality for agriculture, which was reprinted in 1994, presenting a set of guidelines modified to give more practical procedures for evaluating and managing water quality-related problems, emphasizing long-term effects [18]. These general water quality classification guidelines help identify potential crop production problems associated with the use of conventional water sources, and are equally applicable to evaluate wastewaters for irrigation purposes in terms of their chemical constituents. As wastewater effluents may contain a number of toxic compounds, FAO also presented threshold levels for some selected trace elements.

FAO's guidelines have been widely incorporated into national regulations all over the world and the following criteria have been considered the most relevant in defining quality [13,15,18]:

(i) Salinity hazard: the concentration of soluble salts in irrigation water, estimated in terms of electrical conductivity. Salinity has been deemed as the most important factor of irrigation water quality because high salinity in soil can create a hostile environment for the crop to absorb nutrients and cause specific ion toxicity;

(ii) Infiltration and permeability problems: the two most common water quality factors which influence the normal infiltration rate are water salinity (total quantity of salts in the water) and the sodium content relative to the calcium and magnesium content;

(iii) Specific toxicity hazard: certain ions and metals can accumulate in sensitive crops in concentrations high enough to cause damage and reduce yields. The ions of primary concern are boron, sodium and chloride;

(iv) Miscellaneous effects: these include high nitrogen concentrations in water, which supplies nitrogen to the crop and may cause excessive vegetative growth, lodging and delay crop maturity; unsightly deposits on fruit or leaves due to overhead sprinkler irrigation using water with high bicarbonate, gypsum, or iron contents; various abnormalities often associated with an unusual water pH.

In order to better classify water based on its specific characteristics and possible uses, a number of models of water quality, named Water Quality Indices (WQI), have also been developed since the 60s [19]. WQI are simplified representations of complex realities, used to assess the suitability of water for certain purposes based on specific characteristics. The use of WQI allows the representation of a large number of parameters in a single numerical value, facilitating the operational management of water resources and their allocation for different uses [20]. However, for this number to accurately represent the reality of a water body, the correct selection of environmental quality parameters is essential [21].

The first modern WQI was proposed by Horton (1965) [22] and was followed by numerous studies in this field. In recent years, many modifications have been considered in the WQI concept and several indices have been proposed and used by governmental agencies and researchers [19,20,23].

Misaghi et al. (2017) [24] introduced the first systematic WQI for irrigation purposes. However, this index considers a limited set of parameters for estimating water quality and does not take into account all potential impacting properties that could be critical, especially regarding wastewaters or other "marginal" quality waters.

The selection of variables is of major importance in calculating WQI as they should be independent and the most relevant ones, in order to define water quality and detect water quality deterioration. Several indices use parameters selected according to the opinion of specific expert panels, and the consequence is that the final evaluation can be highly subjective and variable. Other authors propose that the selection of parameters should be made according to the water management objectives, the location of the studied waters and the sampling periodicity [20].

To assess the impact of mining activities on irrigation water, some indices seemed to be more appropriate than others to be used as additional tools, as the Weight Arithmetic Water Quality Index method (WQIA), proposed by Cude (2001) [25], which is based on Horton's principles but has been modified by introducing the normative values of the major factors of the water [20,25], the Contamination Index (CI), developed by Backman et al. (1998) [26–28] and the Trace Element Toxicity Index (TETI), by Ali et al. (2017) which is based on contaminant hazard intensity [28].

Furthermore, despite the usefulness and importance of normative guidelines, the effect of unusual or special water constituents is not always considered. An example is the contamination by polycyclic aromatic hydrocarbons (PAH), which are persistent, semi-volatile organic pollutants that can result from the oxidation and self-combustion

of mine wastes and should be analyzed in mining surroundings due to their genotoxic, mutagenic and carcinogenic properties [29]. Although there are many PAHs, scientists and regulators have focused on 16 compounds that have been identified as priority control pollutants by the U.S. Environmental Protection Agency (USEPA). PAHs in underground mining environments may dissolve in mine water and eventually pollute the groundwater system. In addition, mine effluents can bring PAHs to the surface environment, polluting surface water and soils as well [30].

#### **3. Materials and Methods**

#### *3.1. Study Area*

The region of São Pedro da Cova (N 41◦09 ;W8◦30 ) is situated in NW Portugal (Figure 1), in the city of Porto peri-urban region and has a resident population of around 16,500 inhabitants [31] and an area of 16 km2, where the territory comprises residential, industrial, and agroforestry areas (Figure 2a,b,g).

The region is located in the Central-Iberian Zone of the Iberian Massif [32]. The regional geological units consist of metasedimentary rocks (Figure 1) with minor sedimentary cover areas. For that reason, the prevailing groundwater circulation media are fractured.

The study region is located along the western limb of the Valongo Anticline, a regional megastructure [33], which created mountainous landforms reaching 367 and 385 m of altitude at the Santa Justa and Pias summits, respectively, on its western and eastern flanks. The regional morphology is dominated by two hill alignments that originated from differential erosion and are crosscut by Douro River. This structure controlled significantly the regional drainage network, that is part of the Douro river. In São Pedro da Cova it is possible to identify different metasedimentary formations with ages between the Precambrian and/or Cambrian, Ordovician, Silurian, Devonian and Carboniferous.

The old mine of São Pedro da Cova, where exploitation of anthracite A occurred for nearly 200 years, is located in one of the multiple coal deposits hosted in Douro Coalfield (from Upper Pennsylvanian), that represents the most significant Portuguese coal-bearing deposit. This deposit is elongated along NW–SE, presenting an approximate length of 53 km and variable width, ranging from 30 to 250 m. The sedimentary sequence comprises a basal breccia, followed by fossiliferous shales, siltstones and sandstones, along with interlayered conglomerates and coal seams [10].

The regional climate is Atlantic, the normal annual precipitation is around 1254 mm (with 195 mm in December and 18 mm in July) and the normal annual air temperature is around 15◦C (with 20◦C in July and August and 9◦C in January). The Köppen-Geiger climate classification is Csb: warm temperate, with dry and warm summers [34,35].

#### *3.2. Water Sampling*

Groundwater samples were collected from two mine drainage galleries (G1 and G2). Additionally, surface water from Silveirinhos stream was sampled in two points, one located upstream (SS-U) and the other downstream (SS-D) from the mine galleries discharge (Figure 2d). Nine sampling campaigns were carried out in two periods, from April 2017 until December 2017 (Apr, Jun, Sep, Dec 2017), and from November 2018 until December 2019 (Nov 2018; Feb, May, Sep, Dec 2019), in a total of 34 samples. In the campaigns of October 2017 and September 2019 it was not possible to collect water from Silveirinhos stream upstream from the drainage galleries because, during the dry season, this part of the stream does not flow.

The sampling points were chosen in order to be preserved from other anthropogenic impacts besides coal mining, reducing the risk of water contamination by other pollution sources. Sampling was carried out according to standard methods: ISO 5667-3:2018 (E) Water quality—Sampling—Part 3 [36]. Samples were collected in glass or polyethylene bottles, were stored at low temperature (<5 ◦C) in the dark, and delivered to the laboratory within 5 h. The samples were collected with as little agitation or disturbance as possible. Special preservatives were required for some parameters. In this case, care was taken not to

flush any preservative out of the bottle during sampling. Conditions such as strong winds or heavy rain were avoided during sampling.

Water samples were analyzed for a range of physical and chemical constituents in the laboratory, while temperature, pH and EC were measured in situ at the moment of sampling (Figure 2e), using a multiparametric meter from Hanna Instruments, model HI-991300, Woonsocket, RI, USA.

This set of sampling points provided a monitoring network for investigating the impact of mine drainage on environment, namely on irrigation water chemistry, including its seasonal variation.

#### *3.3. Laboratory Analysis*

Analyses were performed according to procedures outlined in Standard Methods for the Examination of Water and Wastewater 23rd edition [37] and in Le Rodier—L'analyse de l'eau 10e édition [38]. The laboratory has been accredited under ISO/IEC 17025 since 2007. Precision and accuracy were calculated for all analytical methods with values <10%. Uncertainties were also calculated with results varying from 2% to 10%.

Water turbidity was measured in a Hach 2100N Laboratory Turbidity Meter (Hach Lange, Düsseldorf, Germany). Electrical conductivity (EC) and pH were determined in a Crison MultiMeter MM 41 (Hach Lange Spain, S.L.U., Barcelona, Spain). Total alkalinity, carbonates (CO3 <sup>2</sup>−) and bicarbonates (HCO3 −), were analyzed by titration. Phosphate (PO4 <sup>2</sup>−) was analyzed in a Shimadzu UV-1601 Spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Chemical oxygen demand (COD) was evaluated in a Hach DR 2800 Spectrophotometer (Hach Company, Loveland, CO, USA). Major inorganic ions (Na+, K+, Mg2+, Ca2+, Li+, Cl−, NO3 −, F− and SO4 <sup>2</sup>−) were analyzed by ion chromatography (DionexTM system DX-120/ICS-1000, Dionex Corporation, Sunnyvale, CA, USA). Total organic carbon (TOC) was analyzed in a Shimadzu TOC-V (TOC-ASI-V, Shimadzu Corporation, Kyoto, Japan), heavy metals (Cr, Mn, Ni, Cu, Zn, As, Cd and Pb) and other components, such as Al, Fe, NO2 −, NH4 <sup>+</sup> and SiO2 were analyzed in a Varian AA240 Atomic Absorption Spectrometer (Varian Inc., Palo Alto, CA, USA) and in a Continuous Segmented Flow Instrument (San-Plus Skalar, Skalar Analytical, Breda, The Netherlands), respectively. PAHs were analyzed by dispersive liquid–liquid microextraction coupled to gas chromatography/mass spectrometry (DLLME–GC/MS) methodology in a Shimadzu GCMS-QP2010 gas chromatograph mass spectrometer equipped with an auto injector AOC5000 (Shimadzu Corporation, Kyoto, Japan), according the procedure described in Borges et al. 2018 [39].

Analytical standards were supplied by Sigma–Aldrich (Steinheim, Germany) and Merck (Darmstadt, Germany). The reference standard mixture containing the 16 EPA PAHs (naphthalene, Nap; acenaphthylene, Acy; acenaphthene, Ace; fluorene, Flu; phenanthrene, Phe; anthracene, Ant; fluoranthene, Flt; pyrene, Pyr; benz[*a*]anthracene, BaA; chrysene, Chr; benzo[*b*]fluoranthene, BbF; benzo[*k*]fluoranthene BkF; benzo[*a*]pyrene, BaP; dibenz[*a,h*]anthracene, DahA; benzo[*ghi*]perylene, BghiP; and indeno[1,2,3-*cd*]pyrene, Ind) was purchased from Sigma-Aldrich (Steinheim, Germany).

Methanol, dichloromethane and acetonitrile were organic trace analysis grade Supra-Solv and were supplied by Merck (Darmstadt, Germany). Ultrapure water was highly purified by a Milli-Q gradient system (18.2 mΩ/cm) from Millipore (Milford, MA, USA).

#### *3.4. Irrigation Water Quality Parameters*

Water quality evaluation is necessary to assess the suitability of water to serve a specific purpose, and to determine appropriate treatments or precautions, if necessary. However, monitoring all parameters involved in a water source could be time-consuming and expensive. Therefore, reducing the subjectivity and the effective cost for assessing water quality is a great challenge.

This study focuses on the parameters adopted by FAO guidelines 29 [18] and on a set of quantitative assessment ratios which included the widely applied Sodium Adsorption Ratio (SAR), the Total Hardness (TH), the Residual Sodium Carbonate (RSC) and the Permeability Index (PI) (Table 1).


**Table 1.** Water quality classification as per different water quality ratios/parameters.

Moreover, in order to meet the hydrogeological and hydrogeochemical specificity of coal mining effluents, the water quality indices WQIA, CI and TETI were considered of significant importance and, therefore, were also calculated.

The water quality index based on the weighted arithmetic method (WQIA) was amended to be specific for irrigation, being adjusted taking into consideration the FAO recommendations, that is, the weights were defined as functions of the standards proposed in this guideline. For computation of the WQIA index, 18 water quality parameters were used, namely, the EC to estimate the salinity hazard; 3 elements with specific ion toxicity (Na+, Cl<sup>−</sup> and B); 3 elements with miscellaneous effects (NO3 −, HCO3 − and pH) and 11 trace elements with Recommended Maximum Concentration Values (Al3+, As, Cd2+, Pb2+, Cu2+, Cr3+, Fe2+, Mn2+, Ni2+, Zn2+ and F−).

WQIA was calculated by using the following equation [20,43,44]:

$$\text{WQI}\_{\text{A}} = \sum\_{i=1}^{n} \text{Q}\_{\text{i}} \text{W}\_{\text{i}} / \sum\_{i=1}^{n} \text{W}\_{\text{i}} \tag{1}$$

The quality rating scale (Qi) for each parameter for a total of *n* water quality parameters is calculated by using this expression:

$$\text{Qi} = 100[(\text{V}\_{\text{i}} - \text{V}\_{0}/\text{S}\_{\text{i}} - \text{V}\_{0})] \tag{2}$$

where, Vi is the actual value of the ith water quality parameter obtained from laboratory analysis, V0 is the ideal value of that water quality parameter obtained from standard Tables (V0 = 0, except for pH = 7.0) and *Si* is the recommended standard value of ith parameter.

The relative unit weight (Wi) for each water quality parameter is calculated by adopting the following formula:

$$\mathbf{W}\_{\mathrm{i}} = \mathbf{K} / \mathbf{S}\_{\mathrm{i}} \tag{3}$$

where, K is the proportionality constant and can also be calculated by using the following equation:

$$\mathbf{K} = \frac{1}{\sum\_{i=1}^{n} (1/\mathbf{S}\_i)} \tag{4}$$

The proposed index ranges from 0 to 100 and plain descriptions for index data were developed in order to provide a qualitative description of the index outcome [44,45]. The calculation of WQIA following the 'weighted arithmetic index method' involves the estimation of 'unit weight', assigned to each physic-chemical parameter considered for the calculation. By assigning unit-weights, all the concerned parameters of different units and dimensions are transformed to a common scale.

Weightage of each parameter means its relative importance in the overall water quality, and it depends on the permissible limits. Those parameters which have low permissible limits and can influence water quality to a large extent allocate high weighting, while parameters having high permissible limits are less harmful to the water quality and allocate low weighting.

Table 2 shows the irrigation water quality standards and the unit weights assigned to each parameter used for calculating the WQIA index. Maximum weights were assigned to cadmium (0.7135), arsenic and chromium (0.07135) and to copper, manganese and nickel (0.03568), thus suggesting the key significance of these trace elements in water quality assessment and their considerable impact on the index.


**Table 2.** Standards for irrigation water and relative weight of parameters.

<sup>1</sup> All values are in mg/L, except pH and EC (mS·cm−1); <sup>2</sup> FAO [18]; source: own elaboration.

The CI was also calculated as a sum of the contamination factors of individual components (the 18 water quality parameters chosen for WQIA calculation, analyzed in the nine sampling campaigns), some of these exceeding the trigger values recommended by FAO [18]. The CI is determined by the following formula [26]:

$$\text{CI} = \sum\_{i=1}^{n} \left[ \left( \frac{\text{C}\_{\text{Ai}}}{\text{C}\_{\text{Ni}}} \right) - 1 \right] \tag{5}$$

where CAi and CNi represent the analytical value and upper permissible concentration of the ith component, respectively. Note that CNi is taken as maximum allowable concentration.

Based on the CI index, values less than 1, 1–3, and more than 3 indicate low, medium and high levels of contamination, respectively [26].

TETI [28] was calculated based on the contaminants hazard intensity. The hazard intensity, or total score, of each parameter was determined according to the Toxicological Profiles of the Priority List of Hazardous Substances prepared by the Agency for Toxic Substances and Disease Registry (ATSDR), the Division of Toxicology and Environmental Medicine, Atlanta, GA, USA [46]. The ATSDR prioritization of substances is based on a combination of their frequency, toxicity, and potential for human exposure.

The concentration of each trace element detected was multiplied by its total score, and products were added to calculate TETI. The proposed TETI only considers toxic elements in water and is calculated by using the expression:

$$\text{TETI} = \sum\_{i=1}^{n} \text{C}\_{i} \times \text{TS}\_{i} \tag{6}$$

where Ci is the concentration of each individual trace element and TSi is its Total Score (ATSDR). This index clearly represents the impact of mine activities on the aquatic environment, where the lower TETI value represents better water quality.

In addition to the conventional parameters of water quality, studies on organic pollutants as PAHs are also very important, as several mining activities such as coal mining, processing or storage of coal provide the basic conditions for the generation and release of these compounds.

As PAHs can also become a source of pollution after the abandonment of coal mines, the 16 priority hydrocarbons were analyzed in São Pedro da Cova samples in order to investigate contamination levels and distribution [47].

#### *3.5. Statistical Analysis*

Data obtained for different parameters were tested for distribution of the residuals with the Shapiro–Wilk's test. Chemical concentrations were studied using a one-way analysis of variance (ANOVA), if normal distribution of the residuals was confirmed. Welch correction was applied when the homogeneity of variances was not verified. Whenever statistical significances were found, Tukey's test or the Tamhane's test post-hoc tests were applied for mean comparison, depending on variances assumption or not. If normal distribution of the residuals was not found, the analyzed parameters were studied using a Kruskal–Wallis test. Whenever statistical significances were found, Dunn's post-hoc test was applied for median comparison. All statistical analyses were performed at 5% significance level using R version 4.0.2 (R Project for Statistical Computing).

#### **4. Results and Discussion**

#### *4.1. Water chemistry*

The values of the physical and chemical parameters used to evaluate water quality from the two mine drainage galleries (G1 and G2, groundwater) and from Silveirinhos stream, collected upstream (SS-U) and downstream (SS-D) from the mine effluents discharge points are reported in Table 3. Results were compared to FAO guidelines in order to assess the water suitability for irrigation. The mining effluents G1 and G2 correspond to groundwater which circulates in the exploited rock massif as well as in the mine galleries. The SS-U water corresponds to surface water without mining influence and the SS-D water originates from the mixture of SS-U, G1 and G2 waters.


**Table 3.**

Summary statistics of physical and chemical parameters

 analyzed in mining effluents and in Silveirinhos

 stream.

Comparison

 with FAO Guidelines

 for



**Table 3.** *Cont.*

The water samples from Silveirinhos stream collected upstream from the galleries (SS-U) showed results within FAO's permissible limits for irrigation purposes, with the exception of pH values, which were slightly below 6.5 due to the geological features of the catchment. Water samples from the mine drainage galleries G1 and G2 reflect the geochemical system of the coal seams and overlying strata. These mining effluents are not as acidic as one might expect, being neutral to slightly acidic, with pH values close to FAO's inferior permissible limit.

Acidity in coal mine waters results mainly from the dissolution of oxidized pyritic materials associated with coal, during mining operations, which explains the existence of iron and sulphate in the water. The pH values of samples G1 and G2, concomitantly with their high levels of iron and sulphate, suggest the existence of an underground neutralization process. The scarcity of pyrites in particular layers and the predominance of carbonate minerals constitute the most common explanation for neutral or alkaline mine drainages [48]. However, the calcareous materials in this region are very scarce and do not constitute a reasonable justification. In São Pedro da Cova, a plausible origin of the neutralization process could be the mixture of acid groundwater, circulating in the shallow rock massif along the mine galleries and wells, with alkaline thermomineral water following deeper circulation paths, possibly through major faults. This hypothesis is corroborated by the relatively high fluoride content in G1 and G2 samples (a hydrogeochemical signature of thermomineral circulation), when compared to SS-U samples, and by the water temperature measured *in situ* during the nine sampling campaigns. Data analysis shows that the temperatures measured in mine effluents were higher than the average annual air temperature for the study area (15 ◦C), ranging from 9 ◦C in January to 20 ◦C in July. G1 and G2 temperatures remained constant during the study, with mean values of 18.9 ± 0.5 and 19.1 ± 0.4 ◦C, respectively, which are significantly higher than the mean water temperature in SS-U (15.4 ± 4.7 ◦C). SS-U water temperature is seasonal, and it closely follows changes in air temperature as can be seen in Figure 3. In SS-D the influence of the mining drainages is notorious, with a mean value of temperature of 17.3 ± 1.8 ◦C.

**Figure 3.** Average monthly air temperature measured between 1981 and 2010 in Serra do Pilar weather station (IPMA). Water temperature recorded at the sampling points SS-U, G1, G2 and SS-D on April, June, October and December 2017, January and November 2018, and February, May, September and December 2019.

The number of particles in water can be expressed by turbidity. With the exception of the SS-U samples that recorded a turbidity median value of 1.1 NTU (Table 3), all the impacted waters have values that were far above the limits proposed by the US EPA of 2 NTU for directly consumed crops and unrestricted irrigation [49], and by Spain that recommends levels lower than 10 NTU for vegetable irrigation water [15].

High levels of turbidity can affect the performance of irrigation facilities, causing the clogging of the equipment, can lower the hydraulic conductivity of the soil, pollute the soil surface, and lead to aesthetic impairment of the water and of the vegetables produced. In addition, irrigating vegetables with turbid water could affect the quality of the products since microorganisms, such as parasites, bacteria and viruses can be attached to the solid particles and contaminate the crops [18].

The EC is also a significant parameter in determining the suitability of water for irrigation use, as it affects water salinity, which subsequently affects the productivity and yield of crops. The EC levels obtained in this study were all below the 3.0 dS/m permissible limit set by FAO for irrigation water [18]. However, the levels recorded in G1 and G2 were significantly higher than those recorded in Silveirinhos stream upstream from mine drainage galleries, these waters having a slight to moderate restriction on use.

Regarding major ions, the levels in SS-U were within FAO permissible limits, and significantly lower than in the impacted waters. Coal mining pollution significantly increases mineralization as a result of a greater water–rock interaction. Mg2+ and K<sup>+</sup> were above the usual range for irrigation water, being the abundance of the ions as follows: SO4 <sup>2</sup><sup>−</sup> > Mg2+ > Ca2+ > HCO3 <sup>−</sup> > Na+ > Cl<sup>−</sup> > K+ > CO3 <sup>2</sup>−.

The hydrogeochemical effect of coal mining, in terms of hydrogeochemical facies and major ion content, is illustrated by means of a Piper diagram (Figure 4) and a Stiff diagram (Figure 5). Surface water without mining influence (SS-U) has an intermediate SO4/Cl-Na/Mg classification while mine drainage waters (G1 and G2) as well as the water collected downstream from the mining effluents discharge (SS-D) have a hydrogeochemical SO4-Mg facies.

**Figure 4.** Piper diagram of the studied waters (average values from April 2017 to December 2019, *n* = 9).

**Figure 5.** Stiff diagram of the studied waters (average values from April 2017 to December 2019, *n* = 9).

Sulfate is relatively common in water and has no major impact on soil other than contributing to the total salt content. Despite being within the usual range for irrigation (0–20 meq/L), concentrations of sulfate in G1 and G2 mine effluents, and in SS-D, were 26 and 17 times higher, respectively, than in SS-U, highlighting the mining influence. Chloride also contributes to the salinity of soils. It is necessary for plant growth in small amounts, but in high concentrations can inhibit plant growth or be toxic to some plants. In the studied waters, chloride levels were low, although G1 and G2 recorded two times the concentration of the upstream water samples.

Regarding the bicarbonate levels, G1 and G2 samples exhibited more than 25 times, and the SS-D samples more than 14 times, the concentration found in the SS-U samples, confirming the mining impact. High levels of bicarbonates can be directly toxic to some plant species. Levels greater than 1.5 meq/L are sufficient to cause concern. Concentrations of bicarbonates greater than 3.3 meq/L may pose a severe potential hazard. Bicarbonate reacts with calcium forming deposits of calcium carbonate and render calcium unavailable. Bicarbonate is also toxic to roots and reduces shoot growth, the uptake of phosphorus and of several micronutrients [13,15,18].

Calcium and magnesium are essential plant nutrients that occur naturally in water through the weathering of geological materials that contain these elements. However, in high concentrations, they are associated to soil aggregation and friability, being important qualitative criteria in the assessment of irrigation water quality. The average concentrations of magnesium in G1 and G2 mine waters (5.43 and 5.10 meq/L, respectively) were above FAO guidelines (0–5 meq/L). Calcium concentrations varied from 0.19 in SS-U to 4.12 meq/L in G2, far below the limit of 20 meq/L of FAO.

These results correlate with the TH values that were also calculated to categorize the water samples considering their calcium and magnesium contents, using the formula shown in Table 1. TH calculated mean values for G1, G2 and SS-D samples were 476, 461 and 308 mg/L, respectively, indicating very hard waters that can be considered harmful and unsuitable for irrigation use. In contrast, the SS-U samples are classified as soft, with a mean value of TH of 22 mg/L.

The RSC index was also calculated as it is an important parameter for irrigation used to indicate the alkalinity hazard for soil. RSC compares the relative concentrations of bicarbonate and carbonate ions with the concentrations of calcium and magnesium. The average RSC values were negative for all samples (<1.25 meq/L), indicating that according to this index waters are safe for irrigation.

Sodium and potassium also occur naturally in groundwater and surface water due to normal water–rock interaction. Among the soluble constituents of irrigation water, sodium is considered the most hazardous. High concentrations of sodium are undesirable because it adsorbs on to the soil cation exchange particles, causing deflocculation and pore sealing, decreasing soil permeability. In the studied water samples, sodium concentrations were low, with mean values varying from 0.32 in SS-U to 1.03 meq/L in G1 and G2. Otherwise, the mean concentrations of potassium in G1 and G2 were 5.2 and 6.7 mg/L, respectively, about three times higher than FAO's limit of 2 mg/L [18]. In SS-D the mean concentration was 3.1 mg/L, also above the recommended limit.

Decades of research on the effect of irrigation water quality on soil permeability have established that the decreasing order of negative impacts of the four major cations follows the sequence: Na > K > Mg > Ca, although the current guidelines are still based on SAR and assume that potassium and magnesium pose no hazard. However, recent studies demonstrated that the negative effects of high K and Mg concentrations on soil permeability are substantial and that they should be taken into account through a new irrigation water quality parameter, the Cation Ratio of Structural Stability (CROSS) that can be directly incorporated into existing irrigation water quality guidelines by replacing SAR [50]. CROSS quantifies both the differing effects of Na and K as dispersing cations diminishing soil permeability and the differing effects of Mg and Ca as flocculating cations enhancing soil permeability. The interpretative guidelines for irrigation water quality involving SAR and CROSS are similar [51].

As mine waters are non-conventional irrigation waters and results revealed high levels of K and Mg, the CROSS ratio was also calculated by the following formula:

$$\text{CROSS} = \frac{\text{Na}^+ + 0.56 \text{ K}^+}{\sqrt{\left(\text{Ca}^{2+} + 0.6 \text{Mg}^{2+}\right)/2}} \tag{7}$$

where the concentrations of ions (Na, K, Ca, and Mg) are expressed in mmol/L. CROSS results were similar to SAR values, varying from 1.1 in SS-D to 0.8 in G1, G2 and SS-U. According to Richards (1954) [40] and FAO guidelines [18], all samples fall in the excellent section for irrigation, which is in accordance with RSC values. Concerning the permeability index (PI) results, only the SS-U samples were classified as of good quality, with more than 75% of maximum permeability. The impacted waters, although classified in Class II, presented values close to 25%, which can be considered unsuitable for irrigation (Table 3).

Thus, considering the characteristics discussed above, it can be concluded that the SS-U water is excellent for irrigation, but the mutual balancing of cations and anions leads to contradictory classifications of the aptness of G1, G2 and SS-D water samples for irrigation purposes. According to SAR, CROSS, and RSC values waters can be classified as good/excellent for irrigation, whilst according to TH and PI they are considered unsuitable.

Plant nutrient concentrations (nitrate, nitrite, ammonium and phosphate) and organic matter content (TOC and COD) were low in all water samples during the entire study period. Nitrate and phosphate are essential plant nutrients, but when in excessive amounts can cause water quality problems and accelerate eutrophication, altering the density and types of aquatic plants found in affected water bodies, promoting their degradation.

Boron is essential for the normal plant growth, but its occurrence in toxic concentrations makes it necessary to consider this element in assessing the water quality. Boron mean values in water samples ranged from 0.07 mg/L in SS-U to 0.26 mg/L in G2. Values are within acceptable threshold, not included in the restriction categories of the FAO classification.

Finally, several trace elements, mainly metals, were also analyzed as they are necessary for crop growth but when in high doses can cause serious environmental and health hazards. Their quantitative determination has shown that the waters affected by coal mining activities have higher metal content, especially iron, manganese, aluminum, nickel and arsenic, which may cause various health hazards such as cancer and environmental pollution. Only iron and manganese exceeded the FAO standards, with values far above the recommended concentrations, but the levels of aluminum almost duplicated in the impacted waters, and nickel and arsenic increased significantly in G1, G2 and SS-D, in a proportion of 7, 3, 3 times and 21, 48, 11 times, respectively, in comparison with the values in water samples collected upstream from the discharges.

Iron was the most abundant metal detected in the mine wastewater samples and in the samples collected downstream from the mine galleries. The median concentration of iron ranged from 0.12 mg/L in SS-U to 52.28 and 84.66 mg/L in G1 and G2, respectively, and 18.24 mg/L in SS-D. Iron can be a complex water quality problem, which not only affects plant growth, as it can compete with other needed micro-nutrients, but also can clog irrigation equipment. For micro-irrigation systems, iron levels need to be below 0.3 mg/L to prevent clogging. Above 1.0 mg/L, iron may lead to rust stains and discoloration on foliage plants in overhead irrigation systems, and above 5 mg/L iron is toxic to plant tissues.

Manganese presents many of the same issues as iron in irrigation water. It can clog irrigation equipment and cause foliar staining. The recommended drinking water standard for manganese is 0.05 mg/L, which is also the level where black staining and irrigation clogging may occur. Concentrations above 2.0 mg/L can be directly toxic to some plant species. In this study the mean concentration of manganese ranged from 0.06 mg/L in SS-U to 4.96 mg/L in G1.

#### *4.2. Irrigation Water Quality Indices*

As the results suggest that no unique parameter can sufficiently describe water quality, thus, the chemical status of the water samples was also assessed by using Water Pollution Indices. Indices were calculated covering a wide range of variables that were gathered in a single numerical value, allowing a simplified representation of a complex reality and also the evaluation of historical trends. The most significant parameters for the water quality evaluation were selected according to FAO's guidelines, and for TETI according to the Toxicological Profiles of the Priority List of Hazardous Substances of ATSDR, in order to proceed with the calculation of the indices in a robust but simple way, using the smallest number of analytes. Indices calculations were also performed including other parameters, but no differences were found in the results and in the outcomes of the evaluation.

The results of the WQIA and the CI indices are shown in Table 4. From WQIA values waters were classified into five categories: excellent, good, poor, very poor and unsuitable, and CI values indicate a low, medium or high level of contamination.


**Table 4.** Calculated values of WQIA (Water Quality Status) and CI (Level of Contamination).

Water Quality Status (WQIA): Excellent (0–25 •); Good (26–50 •); Poor (51–75 •); Very Poor (76–100 •); Unsuitable (>100 •). Level of contamination (CI): Low (<1 •); medium (1–3 •); high (>3 •).

The relation of precipitation events with the WQIA values is represented in Figure 6, with data highlighted in different colors according to the classification for irrigation purposes (Legend Table 4).

**Figure 6.** Relation between precipitation and water quality assessed by means of WQIA. Precipitation measured at Porto meteorological station. Data from the Portuguese Institute for Sea and Atmosphere, I. P. (IPMA, IP) [52].

Results comparing upstream and downstream sampling sites point out the impact of mining effluents on surface water, being SS-U classified as excellent (WQIA values ranging from 1.4 to 3.2, with a mean value of 2.0) and SS-D as poor (WQIA values ranging from 16.2 to 105.4, with a mean value of 56.2). Analysis also revealed that G1 and G2 were, as expected, the two most polluted waters, reported as very poor, with values ranging from 66.2–129.8 and 69.4–140.1, respectively. Out of the 18 parameters considered for this study, iron and manganese were the two deciding parameters, followed by arsenic and EC, which exhibit the maximum influence (Qi x Wi) in the WQIA calculations.

Figure 6 shows fluctuations in WQIA values in both study periods: from April 2017 to December 2017 and from November 2018 to December 2019. In the first period, the effect of draught in water quality is clear: the highest WQIA values correspond to the driest months due to the lack of mixture of mine drainage with recently infiltrated precipitation. In the second period, a similar trend is observed in September 2019, with G1, G2 and SS-D presenting similar values. It was not possible to collect samples on SS-U because there was no streamflow due the drought conditions. The lower WQIA values from this period were observed in February and in December 2019 as a result of a dilution effect due the infiltration of precipitation in the previous months.

The results of the computed CI index are, in general, comparable with the WQIA values. The CI results for impacted waters exceeded the value of 3, ranging from a mean value of 10.1 in SS-D to 29.4 in G1 and 35.8 in G2, which indicates a high degree of pollution due mainly to iron, manganese and bicarbonate content, which exceed the limits of FAO guidelines. The SS-U samples have their computed CI values below 1, reflecting the absence of coal mining influence.

TETI results, based on the elemental toxicological impact, are shown in Table 5.


**Table 5.** Trace element toxicity index values.

*Ci*—mean concentration values.

TETI indicates that manganese and potassium had the highest impact on the toxicological profiles of the polluted waters (G1, G2 and SS-D) with a total score above 2000, according to the ATSDR assessment, followed by ammonium, nitrate and aluminum [43]. Regarding SS-U water samples, nitrate was the most important constituent in terms of water quality concerns, followed by potassium, aluminum and zinc. G1 had the highest index value followed by G2 and SS-D, which clearly demonstrates the impact of mine activities on the water environment, where higher TETI values represent lower quality.

The evaluation of these three selected indices (WQIA, CI and TETI) highlights the coal mine inputs of metals and other pollutants in the study area. Although the three indices specify similar levels of contamination, their outcomes regarding the most important constituents are not uniform.

For example, for TETI calculation, potassium had a high impact with a total score 5–11 times higher than in SS-U. However, it is not considered in FAO or any international guideline for restriction on use purposes, hence it is not accounted for either the WQIA or CI indices. The same occurs regarding ammonia.

These findings clearly highlight the limitations of each index and of the international water quality guidelines that are, firstly, non-standardized between different countries and, secondly, do not provide guidelines for a number of pollutants of importance for specific matrices as is the case of coal mine effluents.

#### *4.3. PAH Analyses*

Regarding PAHs, analyses were performed in six campaigns (April 2017, September 2017, December 2017, November 2018, February 2019 and May 2019). According to the ring numbers, PAHs can be classified into three classes: 2–3 rings, 4 rings, and 5–6 rings composition, which represent low, medium, and high molecular weight hydrocarbons, respectively.

PAHs were detected in water samples at very low concentrations, with prevalence of low molecular weight compounds, with average concentration percentages that varied from 62% to 80%, which indicate a petrogenic origin consistent with the water circulation through the coal bearing rocks in the mine [53] (Figure 7).

**Figure 7.** Composite model and average concentrations of PAHs in the studied samples.

The highest average concentration was detected in G1 (42.5 ng/L). SS-U and SS-D showed similar values of 35.8 and 34.1 ng/L, respectively.

The carcinogenic PAHs (BaA, Chr, BbF, BkF, BaP, InP, and DahA) were detected mostly in G1 and G2, accounting for 28.8% and 27.6% of the total average concentration of 16- PAHs, in contrast to 5.6% and 6.7% found in SS-U and SS-D, being the main components in mine effluents the BaA and the BkF. The carcinogenic PAHs are high-ring PAHs, i.e., 4–6 ring, which is related to the degree of coal maturation [47].

PAHs are of high environmental and human health concern, as they are toxic and persistent in the environment, susceptible to long-range atmospheric transport, and able to bioaccumulate. The uptake of PAHs by plants is important when considering their transfer from soils and water into the food chain. Recent studies also demonstrate that PAHs-metal co-contamination also alters PAHs uptake, attributing to the metal–soil or metal–root interactions [54].

As food chain is the most important pathway for pollutants' entry into the human body, the uptake of carcinogenic PAHs and heavy metals through soil-to-root system and their translocation/accumulation in plant tissues is very important, particularly for food crops cultivated on non-treated wastewater-irrigated soils, as is the case.

#### **5. Conclusions**

This study points out the suitability of the coal mining effluents and the polluted surface water for irrigation purposes in São Pedro da Cova abandoned coalfield. The results allow proposing a cost-effective assessment methodology adjusted to specific problems of the water, minimize pollution of natural watercourses and soils and increase the potential of use of these effluents.

The evaluation of the use of mining effluents for irrigation can be an issue as specific water quality guidelines or legislation does not exist, and several dangerous chemicals are not included in routine water quality assays. In an attempt to standardize decision-making regarding irrigation with mining effluents, the criteria and data have been combined in user-friendly indices, which could assist in the practical implementation of mining effluents irrigation control plans as part of optimal mine-water management and reuse strategy. Eighteen parameters including selected anions, cations and trace elements were chosen after an exhaustive analysis and were used for water quality indices calculation. Source specific water quality indices are a very helpful tool to represent water quality in a simple and understandable manner, minimizing the data volume to a great extent and simplifying the expression of water quality status, giving efficiently the overall water quality of a specific area and for a specific use.

Results revealed that samples associated with mining activities have unacceptable index values due mainly to the high concentration of iron, manganese, bicarbonates, magnesium and potassium, and are not suitable for irrigation. WQIA indicated an overall Poor/Very Poor quality status, which is in accordance with the level of contamination (CI) of these waters. The TETI index, which only reflects the elemental toxicological impact of the waters, not considering other fundamental water quality parameters, indicated that the impacted water samples had higher index values, and manganese had the highest impact on the toxicological profiles, also showing the influence of coal activities on surface water quality. Risks to human health also arise from water pollution by organic substances such as PAHs with several carcinogenic compounds detected in G1 and G2. Levels in the water collected upstream from mine drainage points are within acceptable range for irrigation.

Scientific community and local authorities of mine areas are committed to mitigating the effects of past actions through the development of better management strategies for reducing environmental and health impacts in the mining area. The understanding of the mine water chemistry is fundamental for the design of an effective treatment system. Simple passive treatments could be an option for the contaminated drainage at this abandoned mine site according to Robert Hedin el al. 2013 (Effective Passive Treatment of Coal Mine Drainage. Robert Hedin, Ted Weaver, Neil Wolf, George Watzalaf. Paper presented at the 35th Annual National Association of Abandoned Mine Land Programs Conference, 2013). If properly designed, constructed and maintained, passive systems provide highly reliable treatment at a fraction of the cost of active alternatives.

As the overall quality of groundwater from mine drainage galleries revealed contamination, it is very important to raise awareness for rapid intervention in the area, since the mine is located near a population center and social infrastructures, as well as to mitigate the pollution on adjacent agricultural lands. Despite the evident deleterious impact for local communities, the health effects of mine drainage remain neglected in research and policy arenas, mainly because of the lack of documented evidence. Such communities suffer the effects of mine drainage principally through a perpetual risk posed by water pollution. This project intended to study the impact of mine drainage contamination towards an investment in temporary to long-term solutions, in order to reduce the risks caused by mining externalities.

**Author Contributions:** Conceptualization, C.M.; methodology, C.M., A.M., J.E.M.; validation, C.M., J.E.M.; formal analysis, C.M., A.M., J.E.M.; investigation, C.M., J.E.M., V.M., P.S., J.R., J.R.R., A.M. and D.F.; resources, J.E.M., C.M., D.F.; writing—original draft preparation, C.M., J.E.M., A.M., P.S., J.R.; writing—review and editing, C.M., J.E.M., A.M.; supervision, C.M., J.E.M., D.F.; project administration, D.F.; funding acquisition, D.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the project CoalMine—Coal mining wastes: assessment, monitoring and reclamation of environmental impacts through remote sensing and geostatistical analysis—financed by the Portuguese Science and Technology Foundation, FCT, call AAC nº 02/SAICT/2017 (POCI-01-0145-FEDER-030138) and framed within the activities of the ICT (Ref. UIDB/04683/2020).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data supporting the findings of this study are available within the article.

**Acknowledgments:** The authors acknowledge the Mining Museum of São Pedro da Cova for the permission to use the photograph from Figure 1a. This work received support and help from the UID/QUI/50006/2020 with funding from FCT/MCTES through national funds.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


*Article*
