Applied Geochemistry
Applied Geochemistry 21 (2006) 377–403 www.elsevier.com/locate/apgeochem
Rate controls on the chemical weathering of natural polymineralic material. II. Rate-controlling mechanisms and mineral sources and sinks for element release from four UK mine sites, and implications for comparison of laboratory and field scale weathering studies
K.A. Evans a,*, D.C. Watkins a, S.A. Banwart b a Groundwater Protection and Restoration Group, Department of Civil Engineering, University of Sheffield, Mappin St, Sheffield S1 3JD, UK b Camborne School of Mines, Tremough Campus, Treliever Road, Penryn, Cornwall TR10 9EZ, UK
Received 9 May 2005; accepted 24 October 2005 Editorial handling by R. Fuge
Abstract
Predictions of mine-related water pollution are often based on laboratory assays of mine-site material. However, many of the factors that control the rate of element release from a site, such as pH, water–rock ratio, the presence of secondary minerals, particle size, and the relative roles of surface-kinetic and mineral equilibria processes can exhibit considerable variation between small-scale laboratory experiments and large-scale field sites. Monthly monitoring of mine effluent and analysis of natural geological material from four very different mine sites have been used to determine the factors that control the rate of element release and mineral sources and sinks for major elements and for the contaminant metals Zn, Pb, and Cu. The sites are: a coal spoil tip; a limestone-hosted Pb mine, abandoned for the last 200 a; a coal mine; and a slate-hosted Cu mine that was abandoned 150 a ago. Hydrogeological analysis of these sites has been performed to allow field fluxes of elements suitable for comparison with laboratory results to be calculated. Hydrogeological and mineral equilibrium control of element fluxes are common at the field sites, far more so than in lab- oratory studies. This is attributed to long residence times and low water–rock ratios at the field sites. The high water stor- ativity at many mine sites, and the formation of soluble secondary minerals that can efficiently adsorb metals onto their surfaces provides a large potential source of pollution. This can be released rapidly if conditions change significantly, as in, for example, the case of flooding or disturbance. 2006 Elsevier Ltd. All rights reserved.
1. Introduction
* Corresponding author. Present address: Australian National University, RSES, Building 61, ANU, Mills Road, Canberra Mining activities expose large volumes of fresh ACT 0200, Australia. Fax: +61 2 61250738. rock to atmospheric conditions. Subsequent weath- E-mail address: [email protected] (K.A. Evans). ering and release of rock constituents, especially via
0883-2927/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.apgeochem.2005.10.002 378 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 sulphide oxidation, can lead to acidic solutions and oratory results to field situations. Laboratory- elevated concentrations of environmentally undesir- derived mineral dissolution rates are often 2–4 able elements in surrounding waterways (Younger, orders of magnitude faster than those measured 2000; Banks et al., 1997). This global problem is for minerals in the field (e.g. Schnoor, 1990; White exacerbated by the lack of an effective framework et al., 1996). for determination of legal liability, which means A number of factors have been proposed to that resources for remediation of contaminated account for this discrepancy. These include differ- abandoned sites are limited (Younger, 1997). Fortu- ing pH, grain size, temperature, hydrology, cation nately, pre-mining environmental assessments are exchange characteristics, availability of reactants now required in the majority of countries, and for such as O2, degree of physical and chemical heter- this, as well as for cost-effective remediation, it is ogeneity, secondary mineral behaviour, mineral necessary to develop the ability to predict the sever- surface characteristics and proximity to chemical ity and longevity of mining-related contamination saturation with respect to the dissolving minerals, on a site-specific basis. An understanding of the pro- between weathering environments (e.g. Sverdrup cesses affecting weathering is also relevant to cli- and Warfvinge, 1995; Malmstrom et al., 2000; matic modelling, where a better knowledge of White and Brantley, 2003). Algorithms that quan- field-scale mineral dissolution rates is necessary to titatively account for some of these factors have facilitate assessment of relationships between tem- been devised (e.g. Sverdrup and Warfvinge, 1995; perature, precipitation, elevation and weathering Malmstrom et al., 2000). These are based on the rate (e.g. White and Blum, 1995). premise that each of the different parameters that Laboratory assessments are the most conve- affect dissolution rates can be considered sepa- nient method of measuring the contamination rately, and a scaling factor calculated for each. potential of spoil. Detailed studies of the dissolu- The combination of the scaling factors then allows tion of individual mineral phases have determined extrapolation of laboratory-based weathering rates dissolution mechanisms and relationships between to the field. Calculations are based on fundamental dissolution rates and pH, temperature, surface relationships between reaction rate and the physi- and crystallographic characteristics, and rate- cal parameter of interest, and so site-specific cali- determining concentrations of reactants that influ- bration is not required for most parameters. ence kinetic mass action (e.g. .Wieland et al., Thus, the algorithms should prove to be robust 1988; Xie and Walther, 1992; Martello et al., and generally applicable. The algorithm of Malm- 1994; Peiffer and Stubert, 1999; Holmes and Crun- strom et al. (2000) has been demonstrated to be dwell, 2000), but it is difficult to use such work to successful in a study of the weathering of granitoid account for interactions between the components waste rock from the Aitik Mine, Sweden (Malm- of the phases in a polymineralic assemblage. Rudi- strom et al., 2000). mentary acid–base measurements on polymineralic However, application of this type of algorithm materials are popular (e.g. Adam et al., 1997), involves the implicit assumption that the same because of the relative ease and rapidity with mechanism determines the rate of element release which they can be undertaken and interpreted. in the laboratory and in the field, and that mineral Such methods measure the total potential of a sources and sinks play the same role in the two sample for contamination, but do not provide environments. This is not necessarily the case. The information on the rate at which contamination present study investigates rate-determining mecha- is produced, and are thus of limited use (e.g. Jam- nisms and mineral sources and sinks at four UK bor, 2000). Batch and column experiments have mine sites via a year-long mine-water monitoring also been used (van Grinsven and van Riemsdijk, program and bulk and mineralogical analysis of 1992; Stromberg and Banwart, 1999; Banwart samples from the sites. Results are related to those et al., 2002), to measure rates and to distinguish produced by laboratory experiments using results of rate-controlling mechanisms of mineral dissolu- hydrogeological analyses of the sites. Results for tion. Column experiments can be more useful than one site are then compared to column and batch batch experiments because they involve water:rock experiments on material from the site (Evans and ratios and hydrological solute transport processes Banwart, 2006). The implications of results for similar to those found in the field. Regardless of prediction and treatment of mine-water related method, however, it is difficult to extrapolate lab- pollution are discussed. K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 379
2. Field sites (Gandy, 2002) with associated publications (Gandy and Evans, 2002; Gandy and Younger, 2003), which Locations of all sites are shown in Fig. 1. The have developed groundwater flow and chemical criteria for field sites were that they should: be models for the site. hydrologically well defined with point dis- charges;have a well-documented mining history; 2.2. Grattendale exhibit commonly observed contaminant charac- teristics; be representative of the selected type of Limestone hosted Pb/Zn deposits in the valley of mining environment; have available data for rain- Grattendale in the Peak District, Derbyshire, Eng- fall, discharge, topology, and geology, and have a land (Fig. 1b), were mined up to the end of the reasonably simple geological structure. 18th Century. Grattendale is in the White Peak area, which comprises an inlier of Carboniferous 2.1. Quaking Houses limestone surrounded, except to the extreme SW, by younger fluvio-deltaic sandstones (Stevenson The spoil tip at Quaking Houses, near Newcastle- et al., 1982). Mississipi Valley Type (MVT) Pb upon-Tyne, County Durham, England (Fig. 1a), and Zn ores formed when cooling metal-rich fluids comprises spoil excavated during working of the passed sub-horizontally through the limestones Morrison Busty pit between 1922 and 1973. The (Ewbank et al., 1995). Ore minerals include galena, tip overlies sand and clay drift deposits, which them- sphalerite, baryte, and fluorite. Mineralisation was selves overlie Carboniferous interbedded mudstones, controlled by lithology and by pre-existing struc- sandstones and coal seams (Pritchard, 1997). The ture, and resulted in a geometry of near vertical or area of the tip is approximately 35 ha and the height horizontal sheets, known as rakes and flats respec- varies from 4.25 to 11 m. The spoil material is a het- tively, and occasional linear pipes (Ford and Rieuw- erogeneous mix of shale, ash, coal, and coal dust, erts, 1970). Mineralised zones rarely penetrate the with scattered cobbles, sandstone boulders, timber, volcanic horizons, known locally as toadstones, and traces of red burnt shale. The vast majority of and are thus found mainly within higher levels of particles are smaller than 5 cm diameter. Recently the formation (Ford and Rieuwerts, 1970). Gratten- formed secondary minerals are common in samples dale is a steep sided valley with walls of clean, from the upper 5 m of the tip. Hydrology of the tip fossiliferous (rugose, fragmented corals, and bra- is simplified by the underlying clay-rich drift deposits chiopods) limestone. The Matlock Lower Lava, which act as an aquitard, isolating the tip from known locally as the Grattendale Lava, outcrops groundwater flow (Pritchard, 1997). The discharge partway up the valley; where it is around 9 m thick at Stanley Burn (output G, OS grid ref: 41770 and dips to the NE (Oakham, 1979). Surface expres- 55095) is thought to collect water from approxi- sion of mineralisation is rare because most outcrop- mately 10% of the tip, plus relatively uncontami- ping deposits were extensively worked to a depth of nated water that flows from a surface drain 10–15 m and then filled with earth. (Pritchard, 1997). Mining in the White Peak since Roman times has Significant tip-related pollution was not recorded produced a poorly mapped network of largely inter- between closure of the mine in 1974, and 1986, connected mines and drainage channels, or soughs when construction of the A693 Annfield Plain (Rieuwerts, 1987). The resultant hydrology is com- bypass cut through almost the full depth and plex; however, Grattendale is effectively hydrologi- breadth of the tip. During construction, drains that cally bounded by topography, geology and soughs serviced the tip were incorporated into the road and so presents a relatively simple system. The dis- drainage system, the bulk of which discharges into charge (OS grid reference: 42083 36078) is continu- the nearby Stanley Burn. A subsequent decrease in ous year round, with a flow rate between 0.1 and 2 1 pH and increases in Fe, Al, and SO4 contents in 2Ls . Mining exposed large areas of fresh sul- the Stanley Burn have been attributed to drainage phide surfaces to relatively oxidising conditions. from the tip. There have been a number of previous Regional S oxidation rates higher than those that studies of water quality at the site. Local authority pertained before mining began are indicated by S reports (e.g, Markey-Amey, 1995; Newbegin, isotopes (Bottrell et al., 1999), and analysis of 1997) have been complemented by two M.Sc. theses sediments in cave systems shows fine grained sulp- (Pritchard, 1997; Srour, 1998) and a Ph.D. thesis hides with a relatively high potential for dissolution 380 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403
(a) (b)
(c)
(d)
Fig. 1. Locations and details of the four field sites. Grid references are to the Ordnance Survey grid. (Bottrell et al., 2000). However, solubility of mineral low under high pH conditions such as those found phases bearing Pb, Zn, and Ba, the principal poten- in limestone-rich areas because formation of spar- tially toxic contaminants in the area, are relatively ingly soluble Pb and Zn carbonates and hydroxides K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 381 effectively removes most of the metals from solution River Dulais, the bed of which passes over the Bla- (Benevenuti et al., 2000). enant workings. However, the coal barriers between the Blaenant and Ynysarwed mines (Fig. 1c) proved 2.3. Ynysarwed ineffective and a substantial discharge began at the Ynysarwed adit in 1993 (Younger and Adams, The Ynysarwed minewater discharge flows from 1997). Water discharges from the adit portal at an old adit connected to the lower Ynysarwed Ynysarwed are up to 36 L s 1. Discharges were ini- workings in the Neath Valley, South Wales tially acidic with a pH as low as 3, but are currently (Fig. 1c). The regional geology consists of faulted circumneutral. The discharge is reducing, and rich 2 1 sandstones, mudstones, siltstones and coals of the in dissolved Fe and SO4 (200 and 1800 mg L Upper and Middle Coal Measures (Westphalian), respectively), although a gradual decline in Fe con- which dip at approximately 6 SW (Barclay centrations since 1993 has been observed (Ranson et al., 1988). The topography comprises steep sided and Edwards, 1997). Precipitation of ochre from narrow valleys, separated by larger areas of rela- discharging water has adversely affected fish and tively flat upland. Upper Coal Measures fluvial invertebrate populations in the Neath Canal, which Pennant sandstones, interspersed with a number receives the discharge, and the discharge was given of small, largely unworked coal seams, overlie the 11th priority in a 1993 water quality survey of Rhondda number two seam, which is the principal British minewater discharges. This led to a signifi- worked seam in the area. Below the coal lie Lynfi cant amount of work on both the hydrochemistry Middle Coal Measures mudstone beds, which show and hydrology (e.g. Younger, 1996; Younger and a marked marine influence (Barclay et al., 1988). It Adams, 1997). is this marine influence which gives the anthracitic Rhondda number 2 coal its high S content (2–4% 2.4. Church Coombe pyritic S). The hydrogeology of the area is dominated by The Church Coombe minewater discharge, Corn- the highly conductive mine workings. The Ynysar- wall, England flows into a small stream (Fig. 1d). wed discharge (OS grid reference: 28094 20186) is The surrounding area is a geometrically complex fed by recharge through the relatively permeable mix of Upper Devonian Mylor slates (killas), and Pennant Sandstone (approximately 20% of precipi- Late Carboniferous/Early Permian granites which tation) and groundwater flows from the Blaenant form part of the Cornubian Batholith. These are system to the west (Younger and Adams, 1997). intruded by NNE-WSW striking coarse granite por- Underlying Lynfi mudstones are relatively imperme- phyries (elvans) and dolerites (greenstones) (Dines, able and are assumed to prevent downward escape 1956). Mineralisation was associated with circula- of water from the system. Flows from the neigh- tion of basinal and meteoric fluids driven by heat bouring unsaturated Upper Ynysarwed, Crynant, from the intrusion of the Cornubian batholith and Lwynon workings, which bound the area under (Gleeson et al., 2001). Mines near Church Coombe consideration to the north, are thought to be minor are located in a trough of slate between the Carn- as all drain to adits well above the groundwater mellis granite to the south, and the Carn Brea gran- table (C. Rees, per. comm, 1999; Younger and ite to the north. The main ore-bearing body dips at Adams, 1997). Piezometer measurements show that around 30 S in this region and follows the granite- groundwater flow is from NW to SE, that is, from slate contact in the area shown in Fig. 1d. A number Blaenant (water table 75 m AOD) to Ynysarwed of ENE-trending sub-vertical off-shoots from the (water table 20 m AOD). Prior to the cessation of main body were also worked in the area (Dines, mining, the Blaenant works were pumped, while 1956). Principal ore minerals are chalcopyrite and the lower Ynysarwed workings drained through cassiterite, with a tourmaline/chlorite gangue the discharging adit. Lower Ynysarwed was closed (peach). The sub-vertical lodes were worked mainly in 1938, and although an occasional ferruginous dis- for Cu, although Sn and As were also produced charge was noted at the adit this was not of any vol- (Dines, 1956). Mining was well established in the umetric significance. The Blaenant mines closed in area by 1699, and by 1743 the adit was being used 1991, the pumps were turned off, and groundwater to produce water power (Hamilton-Jenkin, 1965). levels within the workings began to rise. Minewater Mining flourished until the 19th Century, but discharges were expected to begin at or near the underground activity had effectively ceased by 382 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403
1919. The dumps were reworked for secondary and field, to quantify rarer elements such as Cd, U, accessory minerals in the 1930s (Dines, 1956). and Th. Significant concentrations of these elements Exploration and mapping of the adit (portal at were not found. Iron-rich samples from Ynysarwed OS grid reference: 16916 04076) suggest that it were not analysed by IC because Fe forms insoluble drains the Wheal Bassett system, which includes compounds inside the ion exchange columns, and, 2 North Bassett and Wheal Bassett (Fig. 1d). Flow in any case, it was known that SO4 is the dominant rates are up to 90 L s 1. The region is classed as a anion (Younger, 1997). pH, redox potential, con- minor aquifer, with permeability controlled largely ductivity and temperature were measured with a by the depth of weathering, rather than by the Myron 6P Ultrameter. Alkalinity was measured by hydrological properties of the slate or the granite. titration with 0.035 N H2SO4 to pH 4.5 using bro- The water is circumneutral and contains 1 mg L 1 mocresol green/methyl orange indicator. Flow rate of Cu all year round, and thus potentially poses was also measured at each sampling site. At Quak- some threat to receiving water courses. However, ing Houses and Grattendale this was accomplished this is mitigated by subsequent dilution. using a bucket and stopwatch. At Ynysarwed, flow rates were measured electronically with stream 3. Methods gauging by the Environment Agency. These mea- surements were affected by the construction of a 3.1. Element release mechanisms treatment plant for the minewater effluent and thus results from this site are subject to additional 3.1.1. Water sampling and analysis sources of uncertainty, discussed below. At Church Sites were monitored monthly (Table 1) for a per- Coombe, flow rates were calculated from the veloc- iod of at least 12 months between June 1999 and ity of flow in a channel of known cross-sectional June 2001. Quaking Houses samples were taken area. from outfall G (Fig. 1a), Grattendale samples from Wraithe Sough (Fig. 1b), Ynysarwed samples from 3.1.2. Interpretation the marked adit portal (Fig. 1c), and Church Coo- Relationships between element concentrations mbe samples from the adit portal marked in Fig. 1d. (moles L 1), discharge flow rates (L s 1) and ele- At Quaking Houses, unfiltered acidified (1% ment fluxes (mole s 1; obtained by multiplying con- HNO3) and non-acidified samples were analysed centration and discharge flow rate) were used to by atomic absorption spectroscopy (AAS) and ion determine the rate-controlling mechanism for each chromotography (IC) for metals and common element. Rate-controlling mechanisms were split 2 anions (F ,Cl,NO2 ,Br,NO3 ,SO4 ) at the into those that involved principally (a) surface Department of Civil Engineering, University of kinetic-, (b) mineral equilibrium-, and (c) trans- Newcastle. pH, alkalinity, Eh and conductivity were port-related factors. determined in the field using portable meters. At Grattendale, Ynysarwed and Church Coombe, 3.1.2.1. Surface kinetic (S). If surface kinetic pro- water samples were filtered through 0.2 lm filters cesses control the rate of mineral dissolution then (Schleicher and Schuell) into new Nalgene bottles the flux of elements sourced from the dissolving which were rinsed in the discharge before use. Acid- mineral is constant, so long as other rate-determin- ified (1% HNO3) and unacidified samples were col- ing factors such as the quantity of source mineral, lected at each sampling visit. Acidified samples pH, temperature and the concentration of reactants were analysed by inductively coupled plasma- remain constant (e.g. Wieland et al., 1988). If pro- atomic emission spectroscopy (ICP-AES) for major cesses such as secondary mineral precipitation do and abundant trace elements at the Assay Office, not interfere then the flux of the element from the Sheffield. Unacidified samples were analysed, in mine should also be constant. Clearly, temperature, most cases, by Ion Chromatography (IC: Dionex) pH, and other rate-controlling factors are likely to for light alkali metals and alkaline earths (Na+, change in a mine over the course of a year and so K+,Ca2+,Mg2+), and common anions (F ,Cl , some variation in surface kinetic-controlled element 3 2 NO3 ,NO2 ,Br ,PO4 ,SO4 ). One or two samples fluxes would be expected. An approximately con- from each site were screened using inductively cou- stant element flux was thus used to infer that the pled plasma-mass spectroscopy (ICP/MS) at the principal mineral source or sources for that element Centre for Analytical Services, University of Shef- were dissolving at a rate controlled by surface-kinetic Table 1a Quaking Houses discharge data Conc. (mg L 1 ) 15/10/98 19/11/98 17/12/98 26/1/98 23/2/99 23/3/99 23/4/99 14/5/99 3/6/99 2/7/99 23/7/99 30/7/99 16/9/99 8/10/99 Al 3.23 5.36 5.48 5.13 8.49 7.734 3.854 2.85 4.091 7.334 8.5 10.18 5.79 2.24 Ca 193.2 171.4 362.2 381.9 182 263.7 168.1 144.8 196.5 257 314 Fe 6.61 3.73 5.24 4.92 11.46 2.486 2.851 2.706 4.39 7.5 9.41 9.89 16.95 7.14 K 100 140.6 294 297.2 54 106.6 164.2 130.2 188.3 327.3 36.5 Mg 52.1 54.4 126.1 120.9 46.4 82.3 51.6 38.8 65 100.6 98 Mn 3.16 3.56 3.72 3.18 2.37 3.786 0.948 1.301 2.552 4.127 5.6 6.18 4.95 4.15 Na 840 1074 2493 2572 566 1148 1263 792 1149 1922.1 2163
Zn 0.84 1.6 1.14 1.53 1.09 2.405 0.948 0.575 1.01 1.391 0.91 377–403 (2006) 21 Geochemistry Applied / al. et Evans K.A. Cl 2236 3132 2049 1691 2062 440 1851 3047 2621 2446 1581 3939 2 SO4 565 552 544 921 706 404 242 512 873 1227 1257 881 910 pH 6.46 6.37 6.58 6.48 5.7 6.09 6.14 7.1 6.45 5.45 5.22 4.2 6.2 5.73 Conductivity (lS) 5765 7672 7555 7130 7320 7138 8010 4787 6880 11,090 10,140 8760 5970 12,190 1 Alk. (mg L as CaCO3) 77 67 54 27 58 107 70 30 <5 0 31 <5 Flow (L min 1 ) 55.2 107.7 71.2 76.3 39.7 89.3 110.5 168 76.36 131.2 20.5 23.8 28.6
Table 1b Wraithe Sough discharge data Conc. (mg L 1 ) 14/10/99 19/11/99 02/12/99 19/01/00 02/02/00 16/02/00 01/03/00 30/03/00 04/05/01 07/06/00 18/08/00 19/09/00 28/10/00 Ca 92.40 92.60 80.20 78.89 98.23 95.23 91.23 70.25 67.74 76.18 91.38 96.65 82.29 K 0.80 1.00 0.74 0.47 1.25 1.08 1.19 0.67 0.65 0.26 0.56 0.63 0.99 Mg 34.10 31.60 26.04 24.64 43.67 41.53 39.39 27.32 26.03 28.84 29.24 31.89 32.09 Na 5.85 6.66 5.00 5.03 7.54 7.22 7.46 5.51 5.63 4.70 7.14 7.24 5.37 Si 2.58 2.32 2.41 2.24 3.02 4.24 3.73 2.37 2.29 2.00 2.51 2.73 2.52 Pb 0.01 0.02 0.06 0.01 0.04 0.02 0.01 0.00 0.00 0.00 0.08 0.00 0.00 Sr 0.03 0.03 0.04 0.04 0.04 0.03 0.03 0.03 0.03 0.04 0.03 Zn 0.20 0.16 0.06 0.18 0.15 0.11 0.11 0.15 0.15 0.23 0.15 0.18 0.20 Cl 14.42 13.38 14.05 15.19 15.71 14.87 15.65 11.04 11.16 11.04 11.21 11.14 11.64 F 1.87 1.55 1.65 1.77 1.81 1.65 1.91 1.67 1.68 1.58 1.60 1.58 1.65 NO3 23.83 25.58 26.90 35.41 28.66 27.20 29.23 30.14 25.73 25.26 27.37 27.22 2 SO4 24.41 22.79 25.23 24.96 24.80 24.78 25.04 19.46 19.35 19.87 22.01 23.65 22.45 pH 8.12 7.30 7.57 7.31 8.19 7.37 6.97 7.89 7.87 7.69 7.84 8.08 7.60 Conductivity (lS) 623.30 612.00 597.20 604.00 606.00 606.00 574.00 605.00 588.00 598.00 621.50 629.00 597.00 Alkalinity (mg L 1 ) 270.00 168.00 192.00 192.00 161.00 161.00 176.00 184.00 200.00 200.00 330.00 250.00 Temp. ( C) 9.30 7.80 8.90 8.80 8.80 8.40 9.40 8.90 9.20 9.40 10.00 9.30 9.50
ORP (mV) 134 205 178 205 16 246 153 91 109 n/a 97 521 101 383 Flow rate (ml min 1 ) 142 205 142 563 889 800 1720 875 950 845 290 200 548 384 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403
Table 1c Ynysarwed discharge data Conc. 17/08/99 20/09/99 13/10/99 15/11/99 10/12/99 26/01/00 22/02/00 28/03/00 26/04/00 16/05/00 28/06/00 (mg L 1) Al <0.05 0.0 <0.01 0.2 0.2 0.2 0.8 0.7 0.8 0.6 0.6 Ca 285.0 314.0 313.0 300.0 239.8 241.4 289.4 238.7 226.9 288.8 241.0 Fe 173.0 170.0 169.0 163.0 150.6 147.6 154.8 138.1 131.8 159.5 139.8 K 23.1 24.9 25.2 25.5 23.3 24.3 29.1 26.8 25.8 30.2 26.9 Mg 156.0 170.0 171.0 154.0 119.6 120.5 176.5 121.8 115.0 144.6 119.2 Mn 3.9 3.9 0.9 3.7 3.6 3.5 4.0 3.7 3.5 4.1 3.4 Na 108.0 115.0 159.0 152.0 120.0 123.2 171.3 155.6 146.9 141.3 118.6 Si 7.5 7.1 7.2 7.8 7.1 7.3 8.9 8.3 28.9 7.1 6.2 As <0.01 0.1 0.1 0.1 0.1 0.1 0.2 <0.01 <0.01 <0.01 0.1 Ni 0.2 0.2 0.2 3.7 0.2 0.2 0.3 0.4 0.7 0.2 0.2 Pb 0.1 <0.1 <0.01 0.1 0.1 <0.1 0.2 0.1 0.1 <0.1 <0.1 Sr <0.01 2.2 <0.01 1.9 1.2 1.2 1.3 1.3 1.2 1.0 0.9 Zn <0.1 <0.1 0.1 <0.1 <0.1 <0.1 <0.1 <0.1 0.6 0.1 <0.1 2 SO4 1839.0 1950.0 1830.0 1980.0 1589.2 1595.9 2055.9 1911.6 1811.7 1520.3 1346.1 pH 6.44 6.47 6.38 6.20 6.08 6.37 6.46 6.35 5.98 6.28 6.36 Conductivity 2859 2763 2750 2800 2573 2709 2619 2642 2539 2583 2517 (lS) Alkalinity 100 160 180 98 112 112 112 96 112 128 140 (mg L 1) Flow rate 26.31 24.07 25.97 26.74 29.09 28.43 29.05 28.25 26.52 28.37 27.21 (L min 1) Temp. ( C) 13 15 15 11 14 11 12 12 13 15 15 ORP (mV) 56 31 31 5 44 28 44 116 0 n.a. 54
processes. The definition of constant flux used here an element are likely to remain constant over the was that of a standard deviation of the monthly flux course of a year at any of the field sites studied. measurements less than 20% of the average value. Thus, an approximately constant concentration of The value of 20% is arbitrary, but choices of values an element was taken to infer equilibrium control. for the constant flux threshold between 15% and Constant concentration was diagnosed when the 40% give very similar results. Thresholds for the standard deviation of the monthly measurements other categories described below are also arbitrary, was less than 20% of the average value, and where and results show a similar weak sensitivity to the a suitable source/sink mineral was observed, and choice of value. calculated to be in equilibrium with the solution (see below). The second criterion is included because 3.1.2.2. Equilibrium (E). If the concentration of an a constant concentration in a mine discharge may element is controlled by chemical saturation with also result from constant concentration in some respect to a mineral then the concentration of that external source. element is fixed by the presence of the mineral, so long as factors such as pH, temperature, the chem- 3.1.2.3. Transport (T). Transport control of element ical activity of other aqueous species, and the activ- release occurs when changes in hydrogeological ity of the mineral phase remain constant. A solution parameters that affect flow rate, such as the rate of may be chemically saturated with respect to a source recharge, the percentage of water saturation in the mineral present in rock prior to weathering or to unsaturated zone, or water table height, determine some secondary mineral that incorporates the ele- the rate of mineral dissolution and element trans- ment of interest. Cation exchange surfaces on clay port. The transport control described here is similar minerals also exert some control on the concentra- to the transport-limited erosion described by Stal- tions of elements in solution, although the systemat- lard and Eddmond (1987). However, it is slightly ics of this are more complicated. It is unlikely that different to the process described by these authors factors that affect the equilibrium concentration of because it does not include situations in which Table 1d Church Coombe discharge data Conc. 15/08/00 13/09/00 26/10/00 29/11/00 18/12/00 30/01/01 27/02/01 29/03/01 25/04/01 31/05/01 26/06/01 25/07/01 377–403 (2006) 21 Geochemistry Applied / al. et Evans K.A. (mg L 1 ) Al 0.41 0.73 0.45 0.41 0.38 0.39 0.44 0.39 0.39 0.38 0.38 0.38 Ca 8.13 11.94 11.47 9.97 10.96 10.39 10.05 11.01 10.43 8.69 9.38 9.49 Fe 0.32 0.04 0.01 0.01 0.02 0.01 0.02 0.03 0.70 3.78 1.06 0.72 K 2.76 4.34 3.96 0.56 3.31 3.28 3.14 3.59 3.26 13.88 11.76 5.96 Mg 4.92 5.84 5.57 4.05 4.93 5.02 4.76 4.39 4.17 4.65 5.03 5.16 Mn 0.05 0.08 0.06 0.05 0.05 0.05 0.06 0.05 0.06 0.04 0.04 0.05 Na 13.24 24.26 18.28 18.96 19.80 17.89 18.78 19.29 17.19 19.41 21.33 20.63 Si 3.23 4.83 4.06 3.39 3.60 3.66 4.28 3.64 3.63 4.19 4.42 4.46 Cu 1.47 2.09 1.28 1.11 1.08 1.20 1.36 1.14 1.27 1.57 1.67 1.74 Zn 0.23 0.30 0.22 0.18 0.17 0.19 0.33 0.19 0.22 0.23 0.26 0.26 Cl 27.57 28.13 28.58 27.00 27.27 26.52 24.61 24.30 23.69 25.79 24.47 25.18 F 0.86 0.86 0.74 0.59 0.64 0.59 0.61 0.60 0.69 0.79 0.83 0.81 NO3 12.06 11.76 13.28 13.55 13.01 13.91 14.89 13.94 12.44 13.08 n.a. 12.44 2 SO4 26.74 28.41 28.60 26.65 26.73 28.15 25.04 27.17 23.77 26.08 29.25 32.99 pH 5.6 6 5.7 5.4 6 5.6 5.6 6 6.1 5.3 6.2 5.44 Conductivity (lS) 200 190 210 210 240 230 250 240 270 220 210 220 Alkalinity (mg L 1 ) 11.8 15 15 10 10 10 15 10 10 10 10 10 Flow rate (L s 1 ) 14 2 40 60 45 90 60 50 40 14 16 12 Temp. ( C) 12 11.7 12.3 12 12 12 12 12.1 12.2 12.2 12.5 12.6 ORP (mV) 186 220 215 235 253 261 244 247 235 215 198 200 385 386 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 mineral equilibrium is approached, although 3.2. Mineralogy transport-related processes are implicated in the attainment of mineral equilibrium. Mechanisms 3.2.1. Sample collection for transport-controlled element release include: Twenty kilograms of each of the main lithologies flushing, where an increase in the height of the water was collected from each site. Representative sam- table dissolves soluble secondary minerals precipi- pling of 20 kg of material from heterogeneous sys- tated in previously hydrologically-unsaturated parts tems containing many tons of material is not of a hydrogeological system; dilution, where remotely achievable. However, the samples can be increased water flow decreases element concentra- used to identify source and sink minerals present tions and may enhance the rate of mineral dissolu- at the sites. tion; and changes in the ratio of mobile to At Quaking Houses samples were taken from immobile water in the unsaturated zone (Evans material retrieved during borehole construction; and Banwart, 2006; Banwart et al., 2004), which and consisted of two types, black coal waste affect the efficiency of removal of dissolved mineral (QB), and orange clay-rich material from boulder constituents. Physical processes that accompany clay underlying the tip (QO). QB was made by transport control are complex (Clow and Drever, mixing approximately equal proportions of spoil 1996) and so the flux–flow characteristics of trans- taken at 1 m intervals down the tip. Weathering port control are variable; however, all are likely to of the top portion of the tip has caused significant result in a correlation between flow rate and element mineralogical variation with depth (Evans et al., release that could be positive (flushing) or negative 2003). Ore-bearing limestone from Grattendale (dilution). Transport control of element release was collected at a spoil tip 5 km from the adit was inferred when the absolute value of the correla- portal, and non-ore bearing limestone was col- tion coefficient between flow rate and element con- lected from Grattendale itself. All samples were centration was greater than 0.7. large (1–5 kg), partly weathered limestone boul- ders. Material from Ynysarwed was collected from 3.1.2.4. Depletion (D). It may be that the supply of a spoil tip that is known to contain material from a mineral may be depleted by dissolution, and that both Blaenant and Ynysarwed mines (C. Rees, this process is the main control on the flux of con- pers. comm.). Samples were almost unweathered stituent elements of that mineral from the system because recent landscaping of the area had of interest. If this is the case then the flux of this ele- exposed relatively fresh spoil at the surface. ment from the system will decrease with time. Twenty kilograms of each of pyrite-bearing mud- Depletion control of an element was inferred when stone (YM), clean, mature sandstone (YS) and the absolute value of the correlation coefficient pyrite-bearing coal (YC), in 1–5 kg lumps, were between element flux and time was less than 0.7. collected, to allow differences in mineral dissolu- tion rates and composition between lithologies to 3.1.2.5. General transport (I). The categories be determined. At Church Coombe only one set defined above are end-members in a continuum of material (CC) was collected from a spoil tip of system behaviour that is created by the play- adjacent to the North Basset Mine. Here, 1–3 kg off between mineral reactivity, fluid flow rates, samples of a reddish-brown hornfelsed mudstone, and other hydrological parameters. There are a showing evidence of mineralisation, were taken number of situations for which none of the above from a freshly turned spoil tip. criteria would be met, which implies that the rate- controlling factor for element release varies over the course of a year. The rate-controlling element 3.2.2. Analysis release mechanism in these cases cannot be deter- Samples were jaw-crushed to a median grain size mined by a simple inspection of element flow and of 1 mm and homogenised in an industrial mixer flux characteristics and is classed as general trans- at the Department of Civil Engineering, Sheffield. port (I). This preparation was necessary for the laboratory Figs. 2–5 illustrate examples of the relationships experiments (Evans and Banwart, 2006). Major between element concentrations, fluxes, time, and and trace element compositions for each of the eight discharge flow rate. Inferred rate-controlling element lithologies were determined at the British Geologi- release mechanisms are summarised in Table 2. cal Survey, Keyworth, using X-ray fluorescence K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 387
(a) 9 (b) 1400 T 8 I 1200
) 7 ) 1000 -1 -1
6 800 5 600 4 400 3 Concentration (mg L Concentration Concentration (mg L Concentration 200 2 2- QuakingHouses: Al Quaking House: SO4 1 0 Oct Feb Jun Oct Oct Feb Jun Oct Date Date
(c) 0.25 (d) 26 E I 25
) 0.20 ) -1 -1 24
23 0.15 22
21 0.10 Concentration (mg L Concentration Concentration (mg L Concentration 20 Grattendale: Zn Grattendale: SO42- 0.05 19 Oct Feb Jun Oct Oct Feb Jun Oct Date Date
(e) 4.5 (f) 180 E D 4.0 170 ) ) 3.5 -1 -1 160 3.0
2.5 150
2.0 140 1.5 Concentration (mg L Concentration
Concentration (mg L Concentration 130 1.0 Ynysarwedd: Mn Ynysarwedd: Fe 0.5 120 Aug Nov Feb May Aug Nov Feb May Date Date
(g) 2.2 (h) 15.5 Church Coombes: Cu T T 15.0 2.0
) ) 14.5 -1 -1 1.8 14.0
1.6 13.5
13.0 1.4 12.5 Concentration (mg L Concentration Concentration (mg L Concentration 1.2 12.0 - Church Coombes: NO3 1.0 11.5 Aug Nov Feb May Aug Aug Nov Feb May Aug Date Date
2 Fig. 2. Concentration versus time for representative elements from the four field sites. (a) Quaking Houses: Al; (b) Quaking Houses: SO4 ; 2 (c) Grattendale: Zn; (d) Grattendale: SO4 ; (e) Ynysarwedd: Mn; (f) Ynysarwedd: Fe; (g) Church Coombes: Cu; (h) Church Coombe: NO3 . Shaded bands indicate the region of ±20% around the average value. 388 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403
(a) 9 (b) 1400 IT 8 1200 ) ) -1 -1 7 1000 6 800 5 600 4
Concentration (mg L Concentration 400
Concentration (mg L Concentration 3
2 200 Quaking Houses: Al Quaking Houses: SO42- 1 0 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 Flow Rate (L min-1) Flow Rate (L min-1)
(c) 0.25 (d) 2- I 26 E Grattendale: SO4 ) )
-1 0.20 -1 24
0.15 22 Concentration (mg L Concentration 0.10 (mg L Concentration 20
Grattendale: Zn 0.05 18 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 Flow Rate (ml min-1) Flow Rate (ml min-1)
(e) 4.5 (f) 180 E D 4.0 170 ) ) -1
3.5 -1 160 3.0
2.5 150
2.0 140
Concentration (mg L Concentration 1.5 Concentration (mg L Concentration 130 1.0 Ynysarwedd: Mn Ynysarwedd: Fe 0.5 120 23 24 25 26 27 28 29 30 23 24 25 26 27 28 29 30 Flow Rate (L min-1) Flow Rate (L min-1)
(g) 2.2 16 (h) T T Church Coombes: Cu 2.0 ) ) 15 -1 -1 1.8 14 1.6 13 1.4 Concentration (mg L Concentration Concentration (mg L Concentration 12 1.2 - Church Coombes: NO3 1.0 11 0 20406080100 0 20406080100 Flow Rate (L s-1) Flow Rate (L s-1)
Fig. 3. Concentration versus flow rate for representative elements from the four field sites. (a) Quaking Houses: Al; (b) Quaking Houses: 2 2 SO4 ; (c) Grattendale: Zn; (d) Grattendale: SO4 ; (e) Ynysarwedd: Mn; (f) Ynysarwedd: Fe; (g) Church Coombes: Cu; (h) Church Coombe: NO3 . K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 389
(a) 1200 (b) 1.2e+5 Quaking Houses: Al I Quaking Houses: SO42- T 1000 1.0e+5 r = 0.20 r = -0.11 ) )
800 -1
-1 8.0e+4
600 6.0e+4 Flux (mg min
Flux (mg min 400 4.0e+4
200 2.0e+4
0 0.0 Oct Feb Jun Oct Oct Feb Jun Oct Date Date
(c) 14 (d) 3000 Grattendale: SO42- 12 I E 2500 10 r = 0.00 ) ) 2000 -1 -1 8
6 1500
4 1000 Flux (mg min Flux (mg min 2 r = 0.21 500 0 Grattendale: Zn 0 Oct Feb Jun Oct Oct Feb Jun Oct Date Date
(e) 8000 (f) 2.8e+5 E I 7000 2.7e+5 2.6e+5 6000 ) ) -1
-1 2.5e+5 5000 2.4e+5 4000 2.3e+5 Flux (mg min Flux (mg min 3000 r = 0.11 2.2e+5 r = --0.51 2000 2.1e+5 Ynysarwedd: Mn Ynysarwedd: Fe 1000 2.0e+5 Aug Nov Feb May Aug Nov Feb May Date Date
(g) 7000 (h) 80000
6000 T T 60000 5000 ) ) -1 4000 -1 40000 3000 20000
Flux (mg min 2000 Flux (mg min 1000 r = 0.05 r = 0.00 0 0 - Church Coombes: Cu Church Coombes: NO3
Aug Nov Feb May Aug Aug Nov Feb May Aug Date Date
2 Fig. 4. Flux versus time for representative elements from the four field sites. (a) Quaking Houses: Al; (b) Quaking Houses: SO4 ; 2 (c) Grattendale: Zn; (d) Grattendale: SO4 ; (e) Ynysarwedd: Mn; (f) Ynysarwedd: Fe; (g) Church Coombes: Cu; (h) Church Coombe: NO3 . Shaded bands indicate the region of ±20% around the average value. 390 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403
(a) 1200 (b) 1.2e+5 I T 1000 1.0e+5 ) )
-1 800 8.0e+4 -1
600 6.0e+4 Y Data
Flux (mg min 400 4.0e+4
200 2.0e+4 Quaking Houses: Al Quaking Houses: SO42- 0 0.0 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 Flow Rate (L min-1) Flow Rate (L min-1) (c) 14 (d) 3000
12 I E 2500 10 ) ) Flux (mg min -1 -1 2000 8
6 1500
4 Flux (mg min 1000 2 500 0 Grattendale: Zn Grattendale: SO42- 0 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 Flow Rate (ml min-1) Flow Rate (ml min-1) (e) 8000 (f) 2.8e+5 E D 7000 2.7e+5 2.6e+5 6000 ) Flux (mg min )
-1 -1 2.5e+5 5000 2.4e+5 4000 2.3e+5 Flux (mg min 3000 Flux (mg min 2.2e+5
2000 2.1e+5 Ynysarwedd: Mn Ynysarwedd: Fe 1000 2.0e+5 23 24 25 26 27 28 29 30 23 24 25 26 27 28 29 30 Flow Rate (L min-1) Flow Rate (L min-1)
(g) 7000 (h) 80000 T 6000 T 60000
) 5000 ) -1 -1 4000 40000 3000 Flux (mg min Flux (mg min 2000 20000 1000 - Church Coombes: Cu Church Coombes: NO3 0 0 0 20 40 60 80 100 0 20406080100 Flow Rate (L s-1) Flow Rate (L s-1)
2 Fig. 5. Flux versus flow rate for representative elements from the four field sites. (a) Quaking Houses: Al; (b) Quaking Houses: SO4 ; 2 (c) Grattendale: Zn; (d) Grattendale: SO4 ; (e) Ynysarwedd: Mn; (f) Ynysarwedd: Fe; (g) Church Coombe: Cu; (h) Church Coombe: NO3 . K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 391
Table 2 Element release mechanisms Field Si Al Fe Mg Ca Mn Na K Zn Cu Pb As S N Cl F QH I T I I T I I I T I DI EE EII I EIIE Y IIIEEEEEI IIE CC E I E I E I E I I T E D I T I: general transport; T: transport control; E: equilibrium control; D: depletion control.
Table 3 Composition of samples QO QB Da WYSYMYCb CC Major elements (wt.%)
SiO2 66.59 34.83 1.3 3.22 82.76 44.89 2.76 75.28 TiO2 0.88 0.72 0.01 <0.01 0.37 0.85 nd 0.62 Al2O3 16.04 20.3 0.14 0.13 7.65 22.17 1.83 11.85 Fe2O3 5.9 5.01 0.35 0.15 2.46 4.61 13.61 5.22 Mn3O4 0.05 <0.01 0.11 0.05 0.05 0.11 0.08 0.52 MgO 1.07 0.5 0.13 0.83 0.59 0.95 0.78 0.88 CaO 0.07 0.2 37.9 53.7 0.34 0.43 1.59 0.03
Na2O 0.89 0.19 0.13 <0.05 0.71 0.47 <0.05 0.08 K2O 2.72 2.04 0.03 <0.05 1.63 3.65 <0.05 2.50 P2O5 0.13 0.05 <0.01 0.02 0.08 0.29 nd 0.01 SO3 <0.1 <0.1 10 <0.1 <0.1 <0.1 nd nd Cr2O3 0.02 0.02 <0.01 <0.01 0.01 0.02 nd nd SrO 0.01 <0.01 0.21 0.04 <0.1 0.02 nd nd
ZrO2 0.03 <0.02 <0.02 <0.02 <0.02 0.02 nd nd BaO 0.07 0.07 14.35 0.09 0.03 0.08 nd nd NiO <0.01 <0.01 <0.01 <0.01 <0.01 0.01 <0.05 nd CuO <0.01 <0.01 0.01 <0.01 <0.01 0.01 <0.05 nd ZnO <0.01 <0.01 7.28 0.1 <0.01 <0.01 nd nd PbO <0.01 <0.01 0.18 <0.01 <0.01 <0.01 nd nd LOI 4.89 35.48 24.41 41.93 2.82 21.05 75.34 3.78 Total 99.36 99.41 96.54 100.26 99.5 99.63 95.99 101.97
Trace elements (ppm) Ni 38 42 31 4 24 52 nd 35 Cu 28 129 164 7 16 125 nd <150 Zn 108 114 >2000 819 73 62 nd 232 As 6 131 <1 2 10 13 nd 56 Sr nd nd >1500 348 nd nd nd 18 Ba nd nd >5000 367 nd nd nd 361 Pb 18 52 >1000 88 8 36 nd 70 WSS (% S) 0.04 1.28 0.00 0.00 0.00 0.04 0.21 0.00 ASS (% S) 0.03 0.24 0.22 0.00 0.04 0.21 0.01 Total S (wt.%) 1.90 2.86 5.36 0.06 0.31 0.18 8.4 0.03 BET (m2 g 1) 8.66 2.86 0.47 0.8 0.99 1.05 0.52 6.7 nd: not analysed. a Low totals for D result from incomplete oxidation of sulphides during ashing process and/or the presence of fluorine-bearing minerals. b YC was digested in aqua-regia before analysis of solution by ICP-AES. All other samples by XRF (bead and pellet).
(XRF) on glass beads and pellets respectively at the Department of Materials Engineering, Shef- (Table 3). Mineral assemblages and textural rela- field University, and optical microscope examina- tionships were identified using X-ray diffraction tion of thin sections. Total S was determined by (XRD) and scanning electron microscopy (SEM) LECO combustion analysis at the Assay Office, 392 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403
2 Sheffield, and water- and acid-soluble SO4 were poorly constrained for minerals that exhibit signif- obtained using standard methods (Czerewko, icant compositional variation. This is because the 1997). Disulphide-S was calculated from the differ- fine-grained nature of the samples made representa- ence between total S and the sum of acid plus water tive mineral compositions difficult to quantify using 2 soluble SO4 . Organic S was assumed to be negligi- techniques such as electron microprobe on single ble (Taylor, 1989). The specific surface area of the grains. For this reason, important mineral compo- crushed samples was measured by BET analysis at sition variables, such as Tschermak’s substitution the Department of Engineering Materials, Univer- in illite, were accounted for by specifying the sity of Sheffield. exchange component to be an unknown in the set of equations: for example, the anorthite component 3.2.3. Mineral modes in albite was specified as CaAlNa 1Si 1 (notation Mineral modes (Table 4) were calculated for the of Thompson, 1982). The use of exchange compo- observed mineral assemblage by solving the set of nents is discussed further by Thompson (1982). simultaneous equations that relates bulk composi- Water-soluble, acid-soluble and disulphide-S were tion to mineral modes and compositions (c.f. included as compositional constraints. This allowed Ferry, 1988). This set consists of i equations of a maximum of 10 variables, including mineral the form: modes and composition variables, to be deter- X mined. The plausibility of results was checked by N ¼ n p ; ð1Þ i i;j j a comparison of the mass of the calculated assem- j blage to measured mass. Ideally, the assemblage where Ni is the number of moles of element i per calculated from the analysis of 100 g of material unit mass, ni,j is the number of moles of element i should weigh 100 g, and incorrect choices of assem- in mineral j and pj is the number of moles of min- blage or mineral formulae will result in deviation eral j per unit mass. Mineral compositions are from this value.
Table 4 Mineral modes (Weight %) QO QB D W YS YM YC CC Albite 6.2 1 Illite 40 31 0.07 20 51 0.8 30 Kaolinite 34 8 20 3.3 4 Montmorillonite 7.7 6.9 Pyrite 0.2 3.9 0.6 0.2 15 0.04 Quartz 41 11 1.2 6.2 65 15 1 58 Galena 2.9 Barytes 22 0.1 Sphalerite 7.0 Calcite 66 89 0.6 Dolomite 0.6 7 16 Siderite 7.4 4.2 Alunite 0.25 0.04
Fe(OH)3 5.0 10 6 Gibbsite 0.6 Gypsum 0.13 0.01 0.01 0.01 0.96 0.9 0 Jarosite 2.6 0.38 1.7 0
Pb(OH)2/CO3 3.9 Zn(OH)2/CO3 1.8 Total X K,ill (set) 0.6 0.6 0.6 0.7 0.7 0.7 0.7 X Mg,ill 0.23 0.08 0.8 0.2 0.2 X Na,mont 0.99 0.58 X an 0.004 0.1 0.15 Total (g) 97 77 101 104 98 85 42 100 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403 393
3.2.4. Identification of source and sink minerals ion balance of ±10% were used to calculate mineral If a solution is saturated with respect to a mineral saturation indices. then the saturation index of that mineral in solution is 0. The saturation index is the base 10 logarithm of 3.3. Area-normalised element fluxes the observed Ion Activity Product of the stoichiom- etric mineral dissolution reaction of interest divided Comparison between laboratory and field ele- by the conditional thermodynamic constant for the ment release rates requires some methods of conver- corresponding stoichiometric reaction for mineral sion between the element fluxes from field sites solubility (e.g. Appelo and Postma, 1993). A posi- (M T 1), and the element fluxes per unit area of tive saturation index indicates that the solution is reacting mineral surface obtained from laboratory supersaturated with respect to the mineral phase, a experiment (M L 2 T 1). Element fluxes per unit negative value, that the solution is undersaturated, area are related to element fluxes via and a value near zero suggests that saturation is Qi reached, and that equilibrium exists between the F i ¼ ; ð2Þ mineral and the aqueous solution. Mineral satura- V qSSA tion indices for minerals observed by XRD, SEM where Fi is the flux of element i in units of 2 1 or thin section observation (Table 5) were calculated moles m s , Qi is the flux of element i calculated using PhreeqC for Windows, v1.0 (Parkhurst and from field observations (moles s 1), V is the volume Appelo, 1999) utilising the PhreeqC database (Ball of interacting rock in m3, q is the bulk density in and Nordstrom, 1991). This was used to test diag- kg m 3, and SSA is the reacting surface area in m2 noses of equilibrium-controlled element release kg 1. Uncertainties in the values of V and SSA rates (Table 2). Only solutions with a calculated are significant. This is because V is controlled partly by poorly known effects of hydrological focussing such as fracture flow, and because the SSA of samples in the field is less than that of the crushed Table 5 material for which BET analyses are available. Mineral saturation indices SSA, which is sensitive to both the freshness of QD Y CCthe mineral surface and to grain size, is difficult to Albite S U quantify, even for laboratory samples. Chlorite U U Here, three different approaches are considered. Illite S+ S A maximum value for the SSA, and thus a mini- Kaolinite S+ S+ mum value for the area-normalised flux, is given Montmorillonite S+ S Pyrite U U U U by the use of BET-derived SSA (SSABET). This flux Quartz S S S is described and tabulated (Tables 6 and 7)asFB. Chalcopyrite U Alternatively, SSABET can be scaled using assump- Galena U U tions of particle size variation between laboratory Sphalerite U U U and field. The SSA decreases by an order of magni- Alunite S S+ S Fe(OH)3 S S S tude for each order of magnitude increase in particle Gibbsite S S S radius. Flux estimates derived by this method are Gypsum S U S S referred to here as adjusted BET-derived fluxes Jarosite S U U (FABET: not tabulated). However, this method does Calcite U S S S not account for the increase in mineral surface reac- Dolomite U S S S Siderite U S S tivity that is caused by crushing of the samples Anglesite U S S S before BET measurement. A minimum value for Cerussite U S S the SSA, and thus a maximum value for the area- Cuprite S normalised flux, is given by a purely geometric Pb(OH) US S 2 approach to surface area calculation. The simplest Smithsonite U S S S method is to assume that the volume of material U: undersaturated, SI < 5. S: close to saturation 5 < SI < 1. that interacts with infiltrating fluids is divided into S: Saturated: 1 < SI < 1; S+: close to supersaturation, 1 < SI < 5. These relatively wide limits are used to compensate for cubic blocks with side length d m. This assumption errors in the calculated S.I. as a result of poorly known ther- gives a total reacting surface area of 6V/d, and a modynamic data for the phases of interest. surface area normalised flux of Qid/6V. The ratio 394 K.A. Evans et al. / Applied Geochemistry 21 (2006) 377–403
Table 6 Area-normalised field fluxes of elements 2 1 FB (moles m s ) QH flux QH Ln QH Ln D Flux D Ln DLn Y flux Y Ln YLn CC Flux CC Ln CC Ln flux r(flux) flux r(flux) flux r(flux) flux r(flux) Si 1.1E 14 40 2.2 1.3E 15 34 2.1 2.0E 18 41 2.1 Al 4.4E 16 35 2.1 6.7E 17 37 2.4 2.2E 19 43 2.1 Fe 2.1E 16 36 2.1 1.0E 14 32 2.0 7.0E 20 44 2.7 Mg 7.9E 15 32 2.1 1.5E 13 37 2.2 2.3E 14 31 2.0 2.9E 18 40 2.1 Mn 1.2E 16 37 2.1 2.4E 16 36 2.1 1.4E 20 46 2.1 Ca 1.5E 14 32 2.0 4.9E 13 36 2.1 2.6E 14 31 2.0 3.9E 18 40 2.1 Na 1.6E 13 29 2.1 2.9E 14 39 2.2 2.3E 14 31 2.0 1.2E 17 39 2.1 K 1.2E 14 32 2.1 2.3E 15 41 2.3 2.5E 15 34 2.0 1.4E 18 41 2.1 Zn 4.8E 17 38 2.1 2.4E 16 44 2.1 4.9E 18 40 5.3 5.0E 20 44 2.1 Pb 6.9E 18 47 2.4 1.4E 18 41 2.5 Cu 3.0E 19 43 2.1 S 1.4E 14 32 2.1 2.5E 14 39 2.1 7.0E 14 30 2.0 4.2E 18 40 2.1 Cl 1.2E 13 30 2.1 4.0E 14 39 2.2 1.1E 17 39 2.1 F 9.5E 15 40 2.1 5.1E 19 42 2.1 N 4.5E 14 38 2.2 3.4E 18 40 2.1
FG/FB:d = 0.01 8.6 E3 1.9 E3 3.0 E3 2.9 E4 FG/FB:d = 0.1 8.6 E4 1.9 E4 3.0 E4 2.9 E5 FG/FB:d = 1 8.6 E5 1.9 E5 3.0 E5 2.9 E6
Table 7 Comparison between laboratory and field results for Ynysarwedd mudstone Mechanism Si Al Fe Mg Mn Ca Na K S Field I I I E I E E E E Column E E ID X, T X, T X, T X, T X T Batch S ID E S S S S S