SEASONAL CHANGES OF METAL FORMS IN A MINING REGION RECIPIENT MEASURED BY FILTER AND DEAE FRACTIONATION

Åsa Sjöblom (Department of Water and Environmental Studies, Linköping University, SE-581 83 Linköping, Sweden – asasj@tema.liu.se), Karsten Håkansson (Swedish Geotechnical Institute, SE-581 93 Linköping, Sweden)

ABSTRACT

The fractionation and total concentration of Zn, Cd, Cu, Fe, Al, Ca, and organic carbon (OC) in surface water samples from seven stations along a small river receiving treated mine drainage were investigated in the spring, summer, and autumn of 1999. Zn, Cd, and Cu in the river water were redistributed to particles (>0.2 mm) when mixing with the limed effluent water from the mine resulted in a pronounced increase in pH. A larger part of the Cu was also found in particles later in the year, together with higher concentrations of particulate Fe, Al, and OC. Based on comparisons with Ca, no or little reduction in heavy metal concentration occurred due to immobilization in wetlands along the river.

INTRODUCTION

The understanding of key metal immobilization processes in natural wetlands is important for the design and optimization of constructed low maintenance treatment wetlands. Metal forms in the transporting water phase give indirect information on what wetland features that will have the most impact on the initial immobilization of metals. The aim of this study is to identify if any changes in metal forms occurred during the investigated period, or upon passing the different wetlands along a small river receiving treated mine drainage. The importance of different sources of metals within the catchment area, as well as the overall metal retention capacity of the investigated wetlands are also briefly discussed.

The River Vormbäcken is situated in the county of Västerbotten, northern Sweden. Within the 370 km² catchment area, mining of complex sulfidic ores has taken place since the 1940s. Today, only the Kristineberg mine is still in operation. The drainage from the mine is treated by slaked lime, resulting in a pH well above 7, before it is released into the River Vormbäcken between sampling stations S and A (MiMi 2000; see Table 1 for position of the sampling stations). Between A and R, the river is quite channelized and at least in parts lined by aquatic macrophytes. After passing R, it widens and meanders through a vast mire. Between B and V, the river flows through a 8 km² lake, Lake Vormträsket. After V it joins the River Vindelälven, thereby leaving the investigated area.

METHODS

During 1999, surface water samples were collected from seven stations along the River Vormbäcken (Table 1) at three occasions representing snowmelt (middle of May), summer (late July), and low flow before snowfall (early October).

All bottles used in this study were brand new wide neck HDPE bottles (Nalgene^®). Samples for metal analyses were collected and stored in acid washed bottles, while samples for analyses of pH and OC were taken and kept in ordinarily washed bottles. For acid washing, a procedure similar to that described by Ledin et al. (1996) was used. Filters for metal analyses were mounted in the filter holders (Micro Filtration Systems^®) in the laboratory, washed with 10 ml warm 0.02 M suprapur nitric acid (Merck^®), and rinsed with 30 ml MilliQ^®-water.

The initial treatment, and acidification, of the samples was always performed within 5 h of sampling and generally in the field immediately after sampling. Three subsamples of approximately 30 ml were taken from each sampling bottle: a) no treatment, b) filtered through a 0.2 mm filter, c) filtered through a 0.4 mm filter (Millipore^® isopore polycarbonate membrane filters). In order to investigate the fraction of metals bound to dissolved natural organic matter (NOM), approximately 30 ml of samples filtered through a 0.2 mm filter were collected in acid washed bottles containing approximately 0.7 g of a preconditioned weak anion exchanger, diethylaminoethyl-cellulose (DEAE), according to a procedure adopted by Pettersson et al. (1993). The flasks were thoroughly shaken and the DEAE was allowed to settle prior to the analysis of the water phase.

The pH was measured on site. The concentration of OC in the samples was measured using a Shimadzu^® TOC-5000. Metal analyses were performed by a Perkin Elmer^® Elan 6000 ICP-MS.

Here, the heavy metals Zn, Cd, and Cu have been selected for a more thorough discussion since they represent metals that were found to have the highest concentrations related to reference values for Sweden (Naturvårdsverket 1999). These metals may interact with particulate phases; here Fe, Al, and NOM were studied, or with dissolved NOM (e.g. Mantoura et al. 1978). Ca has been used to estimate dilution from surface and ground waters in the catchment area.

RESULTS AND DISCUSSION

The results obtained in the study are summarized in Table 1.

Table 1. Total metal and organic carbon (TOC) concentrations, and pH values measured in surface water samples from the River Vormbäcken, collected during 1999. S-V are the investigated stations in downstream order. M=May, J=July, O=October. Stations T and M were not sampled in May.

Station Distance		S 0 km			A 4 km			T 11 km		R 15 km			M 23 km		B 29 km			V 50 km
Month		M	J	O	M	J	O	J	O	M	J	O	J	O	M	J	O	M	J	O
Zn	mg/l	543	456	438	419	413	431	291	189	157	289	190	147	122	86	122	119	59	58	51
Cd	mg/l	1.7	1.3	1.2	1.3	1.2	1.1	0.9	0.5	0.4	0.8	0.5	0.4	0.3	0.2	0.3	0.3	0.1	0.1	0.1
Cu	mg/l	142	104	83	114	86	67	58	26	42	54	23	25	13	19	21	12	9	10	8
Ca	mg/l	2.7	2.8	3.0	23	64	111	44	39	11	41	39	21	17	6.4	18	18	8.0	8.6	10
Al	mg/l	210	163	125	410	264	450	214	230	200	213	186	122	124	142	116	117	134	83	66
Fe	mg/l	0.3	0.2	0.3	0.6	0.8	2.2	1.3	1.6	0.7	1.6	2.1	1.2	1.1	0.6	1.2	1.1	0.7	0.5	0.5
TOC	mg/l	7.4	7.3	6.2	7.3	6.6	5.6	8.6	8.2	8.7	8.6	7.1	10	7.5	10	11	7.5	11	11	7.7
pH	field	5.6	6.3	6.7	9.7	6.1	6.8	6.3	6.9	7.6	6.4	5.6	6.0	5.7	5.5	6.3	5.5	7.4	6.5	6.0

At all sampling occasions, the concentrations of Zn, Cd, and Cu were highest (sic) at the sampling station S, upstream from the effluent from the Kristineberg mine, and decreased along the stretch of the river. This indicates that the contributions of these metals from the active mine are low compared to contributions from the natural background and older abandoned minor mines in the area (c.f. Naturvårdsverket 1998). With the exception of the samples taken at A in May, more than 80% of the Zn and Cd was found in the dissolved (<0.2 mm) fraction (Figure 1). The increase of particulate Zn and Cd at A in May is probably attributed to the pronounced increase in pH between S (5.6) and A (9.7) on this particular occasion, which might be connected to a low buffering capactity of the melt water, resulting in an increased sorption to particles (e.g. Håkansson et al. 1989). Between A and R, the concentrations of Zn and Cd are then reduced more than could be expected from comparisons to Ca, indicating that such particles may act as metal scavengers when they settle. However, some of this effect seems to be counteracted as the pH starts to approach background values again at R (7.6). Also, at this time of the year, the flow of water is high and only remnants of the macrophytes from the previous year remain, which do not favor the settling of particles.

Figure 1. Total concentration of Zn and Cu (mg/l) fractionated into particles >0.4 mm and 0.2-0.4 mm, dissolved (<0.2 mm) and bound to dissolved NOM. S-V are the investigated stations in downstream order.

Similar to Zn and Cd, dissolved (<0.2 mm) Cu is redistributed to particles (>0.2 mm) at A. For Cu, this effect is pronounced also in July and October. Also, a greater part of the available Cu is present in particles (>0.2 mm) at higher concentrations of particulate Fe, Al, and OC respectively (Figure 2), that is, later during the investigated period (see below). Along the river, total concentrations of Cu are reduced more or less as could be expected from comparisons to Ca. However, between sampling stations R and M, and B and V, there is a tendency towards a larger reduction in the concentration of Cu in the particulate phase than in the dissolved one, which suggests that settling of above particles in the wetland at R and Lake Vormträsket may immobilize at least parts of the Cu in the system.

Figure 2. The relationship between the fraction of Cu found in the particulate (>0.2 mm) fraction and the concentration (mg/l) of particulate Fe, Al, and organic carbon (OC).

As a result of liming of the drainage water, the concentration of Ca found at the sampling station A is at least eight times higher than the background values found at S, and in all samples more than 95% of the Ca was present in dissolved (<0.2 mm) forms, which motivates its use as a measure of dilution, at least close to the mining area. The concentration of Al also increase between S and A. The reduction in Al concentration then follows that of Ca reasonably well. There is a tendency towards a greater part of the Al being found in the particulate fractions later in the investigated period.

The concentration of Fe also increases between S and A. However, it increases between a number of downstream sampling stations as well, which indicates significant contributions from the catchment area. Increasing concentrations of Fe over the investigated period, when the diluting effect of the melt water disappears, give further support for this assumption. The fraction of dissolved (<0.2 mm) Fe is fairly constant, approximately 0.2 mg/l.

NOM measured as OC is also supplied along the river, presumably from the wetlands nearby. In May, >90% of the NOM was found in the dissolved (<0.2 mm) fraction, while in July and October, particles could account for up to 30%. These particles may be biological (plankton), the result of Ca induced flocculation between S and A (e.g. Sholkovitz and Copland 1981), or the OC may be associated with particulate Fe and Al (e.g. Tipping 1981), which is supported by correlations between these parameters. Such particles seem to be more efficient sorbers of Cu than “pure” OC particles (white diamond in Figure 2). Up to 40% of the Fe, 35% of the Cu, and 20% of the Al could be found to be bound to dissolved NOM. Generally, this fraction was most important at V, downstream from all the investigated wetlands. In July, Cu was redistributed to NOM between S and A, despite the addition of potentially competing Ca from the effluent water, which may reduce Cu sorption to particles, see Figure 1.

Without water flow measurements or a hydrological model coupled to the concentration data, only a first estimate of any reduction in heavy metal (Zn, Cd, and Cu) transport along the river could be made. It seems likely that dilution from surface and ground water is the major mechanism explaining the decreasing heavy metal concentrations along the river, and that no dramatic decrease in metal transport occurs upon passing the investigated wetlands. However, if the surface and ground waters in this highly mineralized area contribute metals other than Fe to the river, our results may be more encouraging for treatment purposes.

Since there is limited contact between the bulk of the water and the river sediments, adsorption of dissolved metals to the river bed is probably of minor importance for metal immobilization in the investigated system compared to the settling of particles. Thus factors favoring formation of and adsorption to particles, as well as settling of such particles, should be optimized in constructed treatment wetlands. Based on our results, a supply of a naturally iron rich water, and the building of a wetland or pond where the particles formed can settle, seems to be a good option if the aim is to create a low maintenance wetland.

ACKNOWLEDGEMENTS

This study is part of the Swedish program Mitigation of the Environmental Impact from Mining Waste, financed by the foundation for Strategic Environmental Research.

REFERENCES

Håkansson K, Karlsson S, Allard B (1989), Sci. Tot. Environ. 87/88: 43-57.

Ledin A, Karlsson S, Håkansson K, Sandén P, Düker A (1996), Sci. Tot. Environ. 188: 87-99.

Mantoura RFC, Dickson A, Reily JP (1978), Est. Coastal Mar. Sci. 6: 387-408.

MiMi (2000), Årsrapport 1999 för MISTRA-programmet MiMi. Stockholm, MiMi-print. In Swedish.

Naturvårdsverket (1998), Rapport 4948. Stockholm, Naturvårdsverket förlag. In Swedish.

Naturvårdsverket (1999), Rapport 4913. Stockholm, Naturvårdsverket förlag. In Swedish.

Pettersson C, Håkansson K, Karlsson S, Allard B (1993), Wat. Res. 27: 863 - 871.

Sholkovitz ER, Copland D (1981), Geochim. Cosmochim. Acta 45: 181-189.

Tipping E (1981), Geochim. Cosmochim. Acta 45: 191-199.