Selected ions and major- and trace elements as contaminants in coal-waste dump water from the Lower- and Upper Silesian Coal Basins (Poland)


 Many temporary- and permanent reservoirs of water occur on or in the immediate vicinity of coal-waste dumps in the Lower- and Upper Silesian Coal Basins (Poland). Little or nothing is known of the degree to which their water chemistry might reflect (a) reservoir type, i.e., whether permanent or temporary, (b) level of coal-waste thermal activity, i.e., whether inactive or self-heating or burnt-out or (c) region, i.e., whether the dumps lay in the Upper- or Lower Silesian basins. To provide some answers, concentrations of selected ions (NH4+, HCO3−, F−, Cl−, Br−, NO2−, NO3−, PO43−,SO42−) were determined by ion chromatography and of nineteen elements (Al, B, Ba, Ca, Cd, Cr, Cu, Fe, K, Li, Mg, Mn, Na, P, Pb, S, Si, Sr, Zn) by inductively coupled plasma mass spectrometry (ICP-MS). The data allow a number of observations. In temporary reservoirs, concentrations of ions and of major- and trace elements are relatively much lower and any correlations between components less significant; reservoirs exist for too short a time to allow any balance between coal waste- and water components to be established. A clear relationship does exist between concentrations of ions and of major- and trace elements and dump thermal activity. The highest concentrations occur where thermal activity is high and inorganic components are mobilized. Finally, a regional pattern of elemental- and ion concentrations in the dump waters reflects important regional differences in thermal activity.


Coal waste and coal-waste dump waters
The coal industry all over the world generates large amounts of waste rocks predominantly stored in dumps and land lls. Only small percentage of them can be used for road building or other purposes since mostly they do not full ll requirements for inert mining waste (Younger 2004;Sokol et al. 2005; Klojzy-Karczmarczyk et al. 2016, Klatka et al. 2017). In recent years, there has been a growing concern about the environmental impact of long-term storage of these rocks (e.g. Basins (USCB and LSCB, Poland), two large coal basins where bituminous coal has been mined since the 1800s. Coal waste minerals mainly include clay minerals (50-70%), quartz (20-30%) and other minerals (10-20%), e.g., chlorite, pyrite, siderite, ankerite, gypsum and jarosite (Skarżyńska 1995;Kruszewski 2019a, and b). In these rocks, elements may occur in the inorganic matrix as main-or trace components of minerals or as adsorbed elements in the inorganic fraction while, in the organic fraction, they occur as organometallic-, chelate-or ion-exchange compounds (Baruah et al., 2003). Due to the sorption capacity of clay minerals, waste rock can host a variety of trace elements, including rare earth elements (Finkelman 2004 . During coal-waste weathering, these can be transferred to soil and water. In addition, oxidative weathering of iron sulphides, e.g., pyrite and marcasite, Generally, coal-waste self-heating hot spots can be detected on satellite images where the re is intense. These spots have surface temperatures between 4-14 °C higher than background surface temperatures (Fig. 1). As the resolution of the thermal bands of the applied Landsat series varies from 60-120 meters, hots spots related to res falling below these resolutions, or with low surface temperatures, cannot be detected.
Thermal activity in LSCB dumps Self-heating of variable intensity occurring in the Lower Silesia dumps has been investigated by remote sensing. In the Wałbrzych dump, a hot spot associated with intensive re within the dump in 1987 is waning now. The Przygórze dump showed no intense heating. Thermal activity is ongoing in the Słupiecand Nowa Ruda dumps. Since 1987, a new hot spot has appeared in the northern part of the Słupiec dump and a similar hot spot has been detected in Nowa Ruda (Fig. 1a), despite sub-zero background temperatures.
However, it is possible that the dump was in a state of low thermal activity prior to its opening. It is di cult to distinguish thermal processes on satellite images of the Wełnowiec dump. The area of intense self-heating located on the inclined slope is relatively small. However, in 2003, a small self-heating spot appeared on the eastern side of the dump. On the northern side, hot spots fall below the satellite sensitivity limits. In the Czerwionka-Leszczyny dump, the hot spot at the top of the highest cone has been waning since the early 1985s (Nádudvari 2014;Nádudvari and Ciesielczuk 2018). The eastern side has also undergone self-heating ( Fig. 1b). At present, the burned-out waste there is being excavated and used for highway construction, railroad ballast, etc. . Today, activity is waning though the size of the visible spot is stable (Fig. 1b). Several hot spots were detected on the Dębieńsko coal-waste dump in 2015 and 2018 (Fig. 1b). Temperature ranges indicate relatively mild self-heating with smallish hot spots. However, residential houses are situated very close to these spots (http://www.nowiny.rybnik.pl/artykul,40981,halda-plonie-od-miesiecy.html). The Waleska dump shows no evidence of self-heating.

Page 6/54
The subject of the research was water from reservoirs occurring at or in the immediate vicinity of the coalwaste dumps from the LSCB and two areas of the USCB, namely, the Rybnik (USCB-R) and the Katowice (USCB-K) Regions. The LSCB self-heating dumps are in the towns of Słupiec and Nowa Ruda. All of the USCB dumps investigated were self-heating at the time of sampling, except the Waleska dump.
In total, 31 samples of water (ca 1500 mL) were collected from puddles, streams, and settlers (ponds) (Fig. 2). The sampling points are described in Table 1. The samples were collected and stored in accordance with PN -ISO5667-5: 2017-10. Prior to analyses, the water samples were ltrated on Millipore mixed cellulose ester lters φ 47 mm, 0.45µ m pore size. The total water-sample volume prepared for analysis was ca 1000 cm 3 . After pH and conductivity measurements but prior other analyses, all samples were acidi ed to prevent precipitation. Each sample was divided into two parts, one for analysis of ions by ion chromatography, the other for elements by ICP-MS.

pH and conductivity
Water pH was measured on an Elmetron CP-315 pH-meter with temperature compensation equipped with a glass-combined electrode. Water conductivity was measured with a CC-401 conductometer with temperature compensation.

ICP-MS
Nineteen elements, i.e., Al, B, Ba, Ca, Cd, Cr, Cu, Fe, K, Li, Mg, Mn, Na, P, Pb, S, Si, Sr, and Zn were quantitatively determined by inductively coupled plasma -mass spectrometry (ICP-MS) using a ICP JI 50p spectrometer. The Sequential Spectrometer Spectro ame M with an excitation in the ICP plasma made by Spectro Analytical Instruments (Germany) was used with the following parameters: frequency - All elements were determined with the monochromator except for Li for which the polychromator was used. In the preparation of the calibration curves, ve standards were used. The raw results are shown in Table 2S in the Supplement and their statistics in Tables 3S-7S.

Statistical analysis
Statistical analysis was used to nd the value and variability of the concentrations of elements and ions. All calculations were performed using the StatSoft STATISTICA and PQStat software (test version).
In the initial stage of the analysis, the basic statistical characteristics were calculated (measures of location and dispersion) which enabled the data to be summarized and basic conclusions and generalizations to be drawn. To determine the average level and distribution of element-and ion concentrations, the arithmetic means (X ar ) and quartiles Q 1 = 25%, Q 2 = median = 50%, Q 3 = 75% were calculated. To determine the variation area of element-and ion concentrations, the range (R = max-min), the standard deviation (S x ) and the coe cient of variation (V x ) were calculated. To illustrate the basic numerical characteristics (concentrations of elements and ions) box-whisker plots were made showing the minimum (X min )-and maximum (X max ) concentrations of a given element or ion and the differentiation between typical units (interquartile range, IQR = Q 3 -Q 1 ). The differentiation of the variable value (value of the element-and ion concentrations) was evaluated by comparing the length of four sections < X min -Q 1 >, <Q 1 -Me>, <Me -Q 3 >, <Q 3 -X max >.
Next, the normality of the distribution was tested using the Kolmogorov-Smirnov-, Lillefors-and Shapiro-Wilk tests and the Spearman's rank-order correlation coe cient was calculated to investigate the relationship between the concentrations of elements and ions for the permanent-and temporary reservoirs and the regions of the LSCB and USCB. The following scale was adopted (Stanisz 2006 In order to group reservoirs showing similar element-and ion concentrations, cluster analysis was performed. Initially, each object was treated as a separate cluster. The nearest reservoirs were then gradually merged into new clusters until one cluster was acquired (the agglomeration method). The dendrograms were created using the Euclidean distance and Ward's method. The Euclidean distance, a geometric distance in a multidimensional space, is the most frequently chosen measure. The Ward's method is very effective although it initially creates small-sized clusters. It uses the analysis of variance approach to estimate the distance between clusters. At each stage of all possible combinations of clusters pairs, those that give clusters with minimal differentiation were selected.

Inorganic geochemistry of coal-waste dump water
The ICP-MS results are considered below in terms of: Average concentrations of leached elements and ions and variability of their occurrence in the waters in ltrating the coal-waste dumps.
Differences in the concentrations of the elements investigated in waters originating from currently inactive self-heated coal-waste dumps, thermally-active dumps and dumps that never ignited.
Differences in element-and ions concentrations in permanent-and temporary water reservoirs.
Regional trends and correlations between the LSCB and the Rybnik-and Katowice regions of the USCB.
Correlations between concentrations of ions and major-and trace elements.
Statistics characterizing the concentrations (arithmetic mean X ar ) and variability of element occurrence (variation coe cient V x ) in the dump waters are given in Tables 2 and 3.
Of the sample population, 52% of the sample population came from small but permanent water reservoirs (ponds) located at the bases of dumps and 48% from stagnant waters in puddles on the dumps and in their vicinity. In the USCB-R set, 78% of samples come from lakes and, in the USCB-K set 20%. In the LSCB set, 52% of samples came from puddles and the remainder from ponds.  (Table 2 and Table 3S in Supplement).
The highest average variation of occurrence (> 100%) was found for the following elements in the following places:   3.2 Regional variability of trace element-and ion occurrence in coal-waste dump water The regional variability of concentrations of ions and major-and trace elements is shown on box and whiskers plots (Fig. 3). Cd (≤ 0.00049 ppm), Cr (≤ 0.0028) and Pb (≤ 0.0903) showing the smallest degree of variability are not included on the plots. Other elements were divided into three groups, i.e., those with the lowest-, intermediate-and highest concentrations (groups 1, 2 and 3, respectively). The ions were divided into two groups, i.e., those with lower-and higher concentrations (groups 1 and 2, respectively).

LSCB
Group 1, elements with low concentrations (< 540 ppb) include Cu, Fe, Al, Zn, P, Ba, B, Mn, Li and Sr. Among these, Al, P, B and Sr show a large range (X max -X min , R > 312 ppb) and the most diversi ed interquartile range (Q 3 -Q 1 ). For P, B and Sr smaller diversity in element concentration occurs between X min -Q 1 , and larger between Q 3 -X max . For Ba and Al the trend is reversed; smaller diversity in element concentration occurs between Q 3 -X max , and larger between X min -Q 1 (Fig. 3 and Tables  regions. Sr in USCB water has the smallest diversity in element concentration between X min -Q 1 , and the largest between Q 3 -X max . This trend was not found in the LSCB. Group 3 elements with the highest concentrations (< 13199 ppm) include K, Mg, Ca, Na, and S. Sulphur has a large range (X max -X min , R = 13193 ppm), interquartile range (Q 1 -Q 3 , IQR = 2885 ppm) and shows the highest concentrations. For S, the smallest diversity in element concentrations occurs between X min -Q 1 , and the greatest between Q 3 -X max , as in other regions ( Fig. 3 and Tables 2S, 3S in Supplement).   in USCB-R waters are much higher than in the LSCB or USCB-K waters (Tables 2 and 3 The best coal-re ngerprint minerals are salammoniac, elemental S, rostite and mikasaite (Kruszewski et al. 2019b). Apart from elemental S, these minerals easily dissolve in water. Thus, high sulphur contents in water from coal-waste dumps also re ect combustion.
In the studied waters, the highest variability of occurrence for most elements is characteristic for waters from the Rybnik dumps and the lowest, waters from the LSCB dumps. Only Na, P, S, and Zn show greater variability in LSCB waters (Tables 2 and 3). The variability of element elution in the coal-waste dumps can be summarized in series from the highest-to the lowest value of V x as follows: The variability in element occurrence testi es to the diversity of the material stored in the dumps and of the sources from which the elements originate. It also testi es to, e.g., the degree of their binding with organic-and inorganic substances in coal, or the mineral substances present. Sulfur has a lower variability of occurrence in waters from USCB dumps, but a much higher concentration, compared to waters from LSCB dumps. This may re ect a less diverse S source in the USCB dumps, but one that was undoubtedly richer in this element than the material put into the LSCB dumps.
Signi cant differences in element concentrations characterize waters emanating from burned-out dumps (LSCB) and those where self-heating is on-going. Waters from the former have much higher concentrations of the elements and ions analyzed, particularly, HCO 3 − , Cl − and SO 4 2− (Table 4 and   Table 5S in Supplement).

Page 23/54
During combustion, the mineral substance of coal is enriched with elements, e.g., B, Ca, K, Mg, Mn, Na, P, S, Si, and Sr associated with the organic matter of the coals, and are genetically bound to biogenic-and sorptive ash; they are released on combustion of the organic-carbon substance (Graf, 1960;Turekian and Wedephol 1961;Yudovich and Ketris, 2002). Thus, the increased concentrations of these elements in the water collected in the dumps where the coal waste has burnt may be explained, and their easy leaching from the coal waste by rainwater. The diversity in the element concentrations and the variability of occurrence of elements present in the coal-dump waters re ects the diversity of the waste stored in dumps in both basins, and some inter-basin differences.

Impact of thermal activity on the composition of coalwaste dump water
There is an evident relationship between thermal activity and concentrations of trace elements and ions in coal-waste dump water. Regional variations in ion and trace-element concentrations can be explained by regional differences in thermal activity.
The intensity of self-heating and the frequency of its occurrence varies greatly within the USCB. This regional pattern can be explained by the differences in organic matter coali cation. The westernmost part of the basin contains coals of a much higher rank (R o <1.0%) than the easternmost part (R o ca 0.50%) (Jureczka and Kotas 1995). Average temperatures within in the coal-bearing Mudstone Series ranges from 24.9 o C in the eastern part to 58.9 o C in the southwestern part (Kędzior 2015). As these differences re ect the geological history of the basin (Kędzior et al. 2007, Kędzior 2015, it is no surprise that coal-waste organic matter shows the same trend in thermal maturity. Higher coali cation also means better combustion, important in the later stages of burning. A signi cant factor is the presence of methane in the USCB rocks. To date, the role of methane in the spontaneous ignition of coal waste has not been fully appreciated. Methane can promote the process in coal mines (e.g., Xia et al., 2015). There are regional differences in methane occurrence in USCB coal seams. During the Mesozoic and Paleogene, coal seams in the northern-and central (Katowice area) parts of the basin were degassed. Two zones of high methane content occur in the south-western part of the USCB, i.e., an upper (secondary) zone beneath the Miocene impermeable cover and a lower (primary) zone at greater depth. In the south-eastern part of the basin, methane contents are variable (Kędzior et al. 2013). The fact that the general pattern of methane contents accords with the regional pattern of thermal activity within the dumps suggests that methane content is a signi cant factor in self-heating occurrence.
Possibly, ignition in the dumps is caused by oxygen reacting with methane desorbed from pores as happens with self-heating in coal mines (Xia et al. 2014). Thus, combined methane-coali cation factors, with methane affecting ignition and coali cation promoting continued burning, are re ected in the regional pattern of self-heating.
LSCB coal-waste dumps show a much lower level of thermal activity than those in the USCB. Many dumps there are burnt out and now show temperatures close to ambient (Fig. 1a).

Correlations between trace metals and ions
The element concentrations in coal-waste dump waters from the LSCB and USCB were were examined for potential mutual relationships between elements and between elements and leached ions. The values of correlation coe cients between elements and between elements and ions in permanent-and temporary reservoirs, and for individual regions, were considered. In evaluating the results, correlation coe cients with values > 0.500, both negative and positive, were taken into account.
Of note is the difference between the structures of the correlations in the USCB-R (36 signi cant correlations), in the waters from the USCB-K (19 signi cant correlations) and in the LSCB waters (9 signi cant correlations). The structure is understood as a network of mutual correlation relationships between the concentrations of individual elements and ions (Table 6 and Table 7S in Supplement).    Table 6S). Most probably, a longer time is required to achieve a state of balance as re ected in these correlations than is provided by the temporary reservoirs on the dump surface. ions. This is re ected in the high values of correlation coe cients between these elements and ions.
The activity of hydrogen ions in natural waters has a very signi cant impact on geochemical processes in these environments and the possibility of the occurrence of various chemical elements in them. The importance of pH for geochemical processes re ects the fact that different metals precipitate as hydroxides at different pH values. It plays an important role during the migration of various elements including such a common element as iron, in surface waters and thus during hypergenic processes. The Fe solubility in water at pH = 6 is about 100,000 times greater than at pH = 8.5 (White 2013). Similarly, the separation of SiO 2 and Al 2 O 3 occurring in weathering processes is the result of the dissimilar solubility of these compounds at various pH in the aqueous environment. In strongly acidi ed environments with pH < 4, Al 2 O 3 is mobilized and SiO 2 is immobile, but in aquatic environments where pH lies in the range from 5-9, Al 2 O 3 is immobilized but SiO 2 is increasingly soluble with increasing pH (White 2013 will precipitate later and in another place (White, 2013). This can explain the negative correlations between Mn and Fe in waters from permanent reservoirs, r = -0.47 (Table 6Sa). Mn and Fe are also negatively correlated in the USCB-R waters. rocks. There must be another factor, namely, organic material that carries the given elements. Thus, the chemical composition of coal-waste dump water is related to the origin of elements in the material stored, the manner in which elements occur in this material, their a nity to organic-and mineral phases, and to the geochemical environment and the stage of hypergenesis.
Several factors related to self-heating cause an increase in element concentrations in dump waters. Selfheating results in the formation of pyrolytic water in rock particles. It, as does rainwater seeping through the dump, acts as a solvent forming solutions rich in elements (e.g., Chen et al. 2008). High temperatures promote cracking of coal-waste fragments aiding water penetration into their interiors, promoting further dissolution of rock elements. On a larger scale, ssures and vents formed in dump surfaces allow rainwater to enter dump interiors. Apart from the increase of coal-waste surface accessible to water, heating also induces changes in mineral composition (Ciesielczuk et al. 2014, Pierwoła et al. 2018).
The regional pattern of element concentrations con rms their origin from coal waste and not from, e.g., atmospheric particulate matter (APM) as is the case with organic contamination in these waters (Fabiańska et al., in print). USCB and LSCB are regions of high levels of APM emission from both road tra c and bituminous-coal combustion (Klejnowski et al. 2010). However, APM mineral phases are mostly composed of insoluble compounds such as metal oxides or metal alloys (Jabłońska and Janeczek 2019). Thus, their metals are not transferred to waters in amounts comparable to those leached from coal waste. The only exceptions are Ba sulphate and Fe, Pb, and Cd sulphides. As Ba, Li, and Cu, (in a few samples only) concentrations do not follow the regional pattern, it is tentatively assumed that they mostly come from atmospheric sources. An additional indicator of atmospheric input comes from the cluster analysis (Fig. 5). Both temporary-and permanent reservoirs show a tendency to form clusters re ecting their localization; the closest are the reservoirs from the same dump, e.g., all Dębieńsko waters (Deb) or from the same area, e.g., WA1 and P1. The pattern is preserved whether ions and trace metals are considered separately or together.
Other components that do not follow the regional pattern are Fe, P, Si, and among ions, HCO 3 − .
Bicarbonate ion concentration in water is mainly governed by pH to form the balance with CO 2 (gas) and CO 3 − 2 . Thus, it may be surmized that even if self-heating promoted its dissolution, subsequent CO 3 − 2 precipitation changed concentrations to such an extent that the regional pattern was obscured.

Impact of reservoir size on element concentrations
Water reservoirs on coal-waste dumps and in their vicinity can be permanent (ponds, streams) or temporary (puddles). Typically, concentrations of ions and of major-and trace elements are much lower in the latter as they dry up after a short period; there is less time to leach elements and achieve a balance in concentrations between rock and water.
In both types of reservoirs, S shows the highest and most diverse concentration values. Other highconcentration elements are Na, Ca, and Mg (Tables 3 and 4S The concentrations of the elements determined in the coal-waste waters are mostly much higher than legally allowable concentration levels in the wastewater discharged into surface waters and the ground. Acceptable concentrations were observed only in 12 samples out of 31, i.e., 4 from USCB-R, 3 from LSCB and 5 from USCB-K. Only 5 samples (3 from LSCB and 2 from USCB-R) had contents of analyzed elements demanded for drinking water. Of course, in order to qualify water as drinking water or noninvasive wastewater, a number of additional conditions must be met that were not considered in this research. 5. Concentrations of ions and trace elements are relatively much lower in the temporary reservoirs that exist on dump surfaces for short periods (weeks).

Conclusions
6. There is a clear relationship between concentrations of trace elements and ions in dump waters and thermal activity within dumps; a regional pattern of trace element-and ion concentrations is explained by regional differences in thermal activity. The highest concentrations occur in the Rybnik area where thermal activity is currently greatest and the lowest in LSCB reservoirs where activity is low and waning.
7. Concentrations of the elements determined in the coal-waste waters far exceed, in most cases, those allowed in non-invasive wastewater discharged into surface waters or the ground.   Correlations between element triads.

Figure 5
Pattern (dendrogram) of main ion-and trace-element concentrations in: a) Permanent reservoirs and b) Temporary reservoirs.