Wind dispersal of metals from smelter waste tips and their contribution

Wind dispersal of metals from smelter waste tips and their contribution to environmental contamination. M. Harper, K. R. Sullivan, and M. J. Quinn. En...
0 downloads 0 Views 509KB Size
Environ. Sci. Technol. 1907, 21, 481-484

(7) Singh, B. K. Glueclzauf 1.978, 114, 513 (in German). Available from National Research Council, Ottawa, Ontario,

1021, 106946-84-3; Percol 352, 106946-88-7; H20, 7732-18-5.

Literature Cited (1) Hamza, H. A. CIMBull. 1979, 72, 116. (2) Vreugde, M. J. A.; Poling, G. W. CIMBull. 1975, 68, 54. (3) Chow, C. D.; Jewett, G. L. J. Liq. Chromatogr. 1980,3,419. (4) Beazley, P. M. Anal. Chem. 1985, 57, 2098. (5) Hasimoto, I.; Sasaki,H.; Airua, M.; Kato, Y. J. Polym. Sci., Polym. Phys. Ed. 1978, 16, 1789. (6) Skelly, N. E.; Husser, E. R. Anal. Chem. 1978, 50, 1959.

Canada. Received for review April 18,1986. Accepted January 15,1987. This work was supported by a contract (Contract Serial OS0 83-00146, DSS File 15SQ.23440-3-9189, Oct. 31,1983) from the Department of Energy, Mines and Resources, Government of Canada.

Wind Dispersal of Metals from Smelter Waste Tips and Their Contribution to Environmental Contamination M. Harper* and K. R. Sullivan TUC Centenary Institute of Occupational Health, London School of Hygiene and Tropical Medicine, London WCIE 7HT, U.K.

M. J. Qulnn EPSE Division, Central Directorate of Environmental Protection, Department of the Environment, London SW 1, U.K.

rn Results from a detailed investigation of metal concentrations in soil and house dust in the Lower Swansea Valley, based on work done in Swansea as part of a nationwide survey, showed that contamination of garden soils by wind erosion from waste tips had taken place. However, house dusts showed considerably higher levels than the corresponding soils, and there was no evidence for a similar route of contamination. Multiple-regression models indicated that only a small number of factors explained a large proportion of the variability in metal concentrations in soil; similar models could not explain the variability in house dust.

Introduction In recent years, much attention has been focused on the possible subclinical health effects of heavy metals, particularly lead (I),and especially in relation to areas of soft water (2) or to children with pica (3). It has been suggested that the dust-to-hand-to-mouth route is of primary importance in the exposure of children ( 4 ) . A number of prospective studies are currently investigating this route, and preliminary results from a major study of hand-tomouth activity in 2-year olds in an inner city area of the U.K. have recently been published (5). For cadmium, it has been suggested that it would be prudent to contain intakes within currently acceptable tolerance limits, as the metal serves no known useful biological function, accumulates in the body throughout life, and is acutely toxic in very low doses (6). Only quite recently has information become available for the U.K. on the concentrations of lead and other metals in soils and house dusts across the country (7). As a part of this survey, paired samples of house dust and garden soil were collected from the town of Swansea in Wales. This paper reports the results of an investigation into the metal concentrations of 80 such samples. At the beginning of the industrial revolution Swansea was ideally situated, as a port at the head of a valley leading deep into the Welsh coalfield (Figure I), to refine metal ores and export the finished product. The basis of much of the manufacturing wealth was copper production, which continued from 1717 until the turn of the century. Zinc, produced by a sublimation process, was also refined from 1738 until World War 11. These processes led to 0013-936X/87/0921-0481$01.50/0

prodigious quantities of waste slags and furnace ashes being discarded to form the bulk of the non-ferrous metal waste tips on the floor of the valley (8). Since eventual restoration of the land on termination of working was not proscribed by legislation, the Swansea valley contains a number of non-ferrous metal tips, as large mountains of unsorted and often uncemented materials scattered over a marshy, waterlogged valley plain and overlooked by around 70% of the local housing situated in terraces along the valley sides. The metal content is apparent from the strong colors predominating in both the fresh and weathered materials and the general toxicity of the substrate, which has prevented the development of soils. Only a few sparse areas of metal-tolerant grasses (9) were found to be binding the surface, and the greater extent was obviously subject to erosion and transport by wind.

Methods The soil and house dust samples analyzed in this study were obtained as part of a survey into the nationwide concentrations of heavy metals in garden soils and house dusts, carried out by the Applied Geochemistry Research Group of Imperial College (University of London) on behalf of the Department of the Environment between 1980 and 1983 (7). The sample localities were selected to give as even a coverage as possible over each chosen urban area. Each garden soil sample (0-5-cm depth) was dried, sieved (2-mm mesh), and ground to -80 mesh. Samples (0.25 g) were digested in concentrated nitric acid for 1h at 105 f 5 "C before dilution and analysis. The contents of the vacuum cleaner bag were sieved twice (l-mm mesh) to remove carpet fluff and large objects. Samples (0.25 g) were carefully heated to 185 f 5 "C with a mixture of concentrated perchloric and nitric acids until dryness. The samples were then refluxed for 1h at 60 f 5 "C with 6 M hydrochloric acid followed by dilution and analysis, which was carried out on a Perkin-Elmer SP 5000 atomic absorption spectrophotometer. In addition to soil and house dust samples taken at each house in the national survey, a questionnaire was completed by the householder. A simplified version of that questionnaire was used in this study and covered the occupation of the adult occupants, number of preschool children, smoking habits, foam-backed carpets, age of house, type of heating, distance of house from road, road

0 1987 American Chemical Society

Environ. Sci. Technol., Vol. 21, No. 5, 1987

481

Table I. Metal Concentrations (mg/kg) in Soils and Housedusts Swansea metal copper cadmium zinc lead

Londonn

median

range

geo mean

range

152 232 2.5 8.3 573 1607 385 634

31-668 57-2040 0.3-15 2.0-32 92-3840 415-6400 56-2517 164-3984

73 208

13-2320 9-5300 1-40 1-336 58-13 120 81-114 800 1-13 680 5-36 900

soil dust soil dust soil dust soil dust

1.3

7.7 424 1324 647 1007

reconnaissance survevn geo mean range 53 204

5-16 800 3-48 800 1-17 1-804 13-14 568 128-70 100 13-14 125 13-34 530

1.2

6.8 260 1055 230 507

Table 11. Interelement Correlations for Swansea Soils and House Dusts

b-

cadmium BIRMINGHAM

SWANSEA

copper cadmium zinc

0.87

soil zinc 0.69 0.79

lead

cadmium

0.61 0.66 0.82

0.56

dust zinc 0.42 0.53

lead 0.28 0.35 0.52

10xIn CCdl

I.

BRISTOL

. . . .. . . 14- . . .. . . .. .,.. . . *

24..

Figure 1. Swansea, South Wales: geographical setting.

classification, and make of vacuum cleaner. Two further pieces of information were collected subsequently: the type of soil parent material and the distance from the house to the approximate center of the valley waste tips. The frequency distributions of the metal concentrations in the soil and house dust samples were found to be skewed; they were approximately log normal and so were transformed to their natural logarithms prior to statistical analysis. The discrete variables from the questionnaire were entered into multiple regression models as dummy variables. The data were analyzed by the SPSS system on the University of London’s Amdahl computer.

Results and Discussion The range of concentrations and the median values for each of the four metals, copper, cadmium, zinc, and lead, in the soils and house dusts are presented in Table I, along with comparison data for a reconnaissance survey of towns and villages over the whole country (excluding London boroughs and geochemical “hot spots”) and for London boroughs alone. The comparison results are given as geometric means, although these should be sufficiently similar to the median values to allow general comparisons to be made. As is to be expected, Swansea lead values were comparable with those of other towns and villages but about half those found in a large city. The soil values of copper and cadmium were enriched by a factor of 2 over even London Levels, indicating a powerful source of industrial contamination. That the house dusts are not similarly enriched may be a reflection of the comparatively minor role of soil contamination of houses compared with internal sources. The correlations between the concentrations of the various metals in soil and those in house dust are given in Table 11; the weakest association was between copper and lead. Ranking these correlations in order of strength of their associations indicates an ap482

Environ. Sci. Technol., Vol. 21, No. 5, 1987

*6-16 -

2000

4000

6000

8000

Metres

Figure 2. Soil cadmium levels with distance from waste tips, r = -0.80. 651. lox In [CUI

I_.

.

. .

I . .

1

..

*

I

. . . . ,.. .

* .

e

I

351 d

2000

.

.

4000 o

0

Figure 3. Soil copper levels with distance from waste tips, r = -0.73. 1 10 x In C znl

:. . .

..

.

50

1

.

. 2000

4000

6000

Metres

8000

Flgure 4. Soil zinc levels with distance from waste tips, r = -0.54.

Table 111. Proportions of the Variability in Soil Metal Concentrations Explained by Multiple Regression Models metal

distance from tip

age of house

distance from road

type of soil

copper cadmium zinc lead

0.53 0.64 0.30 0.07

0.13 0.06 0.10 0.29

0.013 0.013 0.078 0.028

0.031 0.024 0.028 0.059

Table IV. Proportions of the Variability in House Dust Metal Concentrations Explained by Multiple-Regression Models variable

R2

total R2

metalworkers distance from road distance from road preschool children make of vacuum cleaner age of house type of soil distance from road

0.077

0.13

metal copper cadmium zinc lead

type of heating

total R2 0.70 0.73 0.52 0.45

0.020

7n

0.055 0.069 0.066

0.14

0.076

0.08

0.130 0.055 0.052

0.24

Flgure 6. Histogram of soil metals with age of house.

.. .

60-

.

40

I

. . ... . .. .

2000

*.

.

4000

6000

Metres

Figure 5. Soil lead levels with distance from waste tips, r

8000

= -0.37.

parently two-phase sample consisting of Cu-Cd-(Zn) and Pb-Zn-( Cd). Scatterplots of the metal concentrations in soil against distance from tip are shown in Figures 2-5. The linear correlation coefficients were all significant but are used here as a means of ranking the importance of the associations. Copper and cadmium were very strongly correlated (0.73 and 0.80, respectively) and zinc and lead rather less so (0.54 and 0.37, respectively). The results for copper and cadmium strongly suggest a source for these metals in the va11ey. Results from the multiple-regression models with dummy variables created from the nominal-scale questionnaire data are presented in Tables I11 and IV. Only those variables giving rise to an increase of at least 0.01 (1%) in R2 have been presented, and attention is directed primarily to those variables that give rise to a greater than 0.05 (5%) increase in R2. It is noteworthy that quite high proportions of the variability in soil metal concentrations were explained (from 45 to 73%) and that only four variables (distance from tip, age of house, type of soil, and distance from road) explained the bulk of the variance within each model. Following distance from the tip, the next most significant variable affecting the garden soil concentrations of copper, cadmium, and zinc was the age of the house. For lead, this was the major variable in the model; this is presumably due to the reduction of metalliferous building and decorating materials (particularly lead-based exterior paints) used over the years. Histograms showing the relationship between the age of the

house and soil metal levels are presented in Figure 6. Distance of the house from the road was also significant for both lead and zinc, the two metals associated with automotive emissions, but it is interesting that the road classification variable did not enter into any of these models. The greater base exchange capacity of clay-rich soils is indicated by the presence of a small influence from the type of soil. The most noticeable feature of the results of the analysis of metals in house dusts is that the concentrations were much higher than in the soils (Table I). The most likely reason for this is that the house itself was a greater source of metals to the indoor environment than was the garden. This is confirmed in the first instance by the almost complete absence of any correlation between house dust metal concentrations and distance from the tips and in the second instance by the general similarity to the levels found in London and the reconnaissance survey. The correlations between the soil and dust concentrations of each metal were very low (< 0.1) except in the case of lead (0.38), which may be a result of greater mobility of the finer, automotive emissions contaminatingboth house dust and soil. None of the regression models for house dust explained more than 24% of the variability in metal concentrations (Table IV). This could be a function of the wide variability in the samples owing to dilution to differing extents by extraneous materials. A second possibility is influence(s) by a factor or factors unknown and not covered by the questionnaire. The relationship between lead concentrations and the age of the house is likely to be related to the use of high-lead paints. The occurrence of distance from the nearest road in three of the models once again probably indicates the greater mobility of the finer, vehicle-derived particles. In the mid-19709, a collaborative study on certain elements in the Swansea-Neath-Port Talbot area had been carried out (IO). Aerosol measurements by filtration and cascade impactor at two sites in the area of our study, Mount Pleasant and Llansamlet, showed that concentrations of copper, lead, and zinc were all enhanced by a factor of 2-3 compared with a rural site. Lead-enriched particles tended to be small (average diameter of 0.7 pm with 75% less than 2 pm) and had low settling velocities (0.19 cm/s). Environ. Sci. Technol., Vol. 21, No. 5 , 1987 483

Copper-enriched particles tended to be large (average diameter 1.5 pm with 55% less than 2 pm) and had higher settling velocities (1.05 cm/s). This indicates a separate origin for the two metals. Zinc-enriched particles had an intermediate settling velocity suggesting an association with both lead and copper. Clear associations have been demonstrated between soil levels of lead, cadmium, and zinc and those of automotive emissions (11, 12). The particles formed during vehicle operation have a bimodal size distribution with small particles (C2 pm predominating). They consist of crystalline lead halides growing directly from the vapor phase and often contain ammonium halides and sulfates from condensation reactions with the exhaust gases (13, 14). Larger particles are lead halides condensed on the surface of iron- and silicate-rich particles. The lead-enriched particles found in the above survey were almost certainly the product of such vapor-phase condensation from automobile emissions. The copper-enriched particles, because of their larger size, were more likely to have been due to mechanical generation (15). As it has been many years since the cessation of large-scale copper refining, it would seem most likely that these particles were derived from waste tips by wind entrainment. That such tips are a source of localized dust emissions is well-known (16). This bimodal distribution of particle types appears to correspond well with the two different chemical associations found in the soil metal analysis. Once within the garden soils, metals appear to be strongly associated with the surface horizon of the soil, and there is rapid depletion with depth (17). This biochemical concentration is due to chelation by humic and fulvic acids from plant decay (18). In the strongly acid brown earth soils developed on the fluvioglacial sands and gravels or coal measure rocks of the Swansea area, carbonate and iron/manganese oxide coprecipitation will be minimal (19), and the base exchange capacity of the clay minerals will be lowered by competition with hydrogen ions (20). Large areas of these soils are subject to gleying, which causes a large increase in the solubility of metals (21). Under such conditions, it is unlikely that soil residence times would be long, and it is felt that the present contamination is of recent origin and not an artifact of smelting in the previous century, Constant replenishment of soil metals by wind deposition is the most likely source.

Conclusions The results obtained in this survey are very clear. The waste tips of the lower Swansea Valley were undoubtedly the source of the elevated levels of copper, cadmium, and, to some extent, zinc found in the garden soils. In the case of lead, the small contributions from the tips was overshadowed by that from the house itself; the house was also a minor source for the other metals. In addition, for lead and zinc there was a small but significant input from roads. The results for house dusts from the regression models did not explain nearly as much of the variability as did those for soils. However, it would seem that the larger tip ma: terial particles did not form a substantial part of the household metal levels whether derived directly from the tip or indirectly through contaminated soil. There would seem t o have been a risk of enhanced human intake of metals, especially close t o the tips where the soil concen-

484

Environ. Sci. Technol., Vol. 21, No. 5, 1987

trations were many orders of magnitude higher than “background. However, as ingestion of metals in dusts, for instance through the dust-to-hand-to-mouthroute, is more likely within the house, the higher house dust metal concentrations should be of greater concern with regard to health.

Acknowledgments This work was originally carried out by M.H. for an M.Sc. Thesis at Kingston Polytechnic, and grateful thanks are extended to all the staff of the Applied Geochemistry Research Group of Imperial College whose cooperation made this possible. During preparation of this paper substantia1 reclamation of the derelict area of the Lower Swansea Valley has begun. Registry No. Cd, 7440-43-9;Cu, 7440-50-8 Zn, 7440-66-6; Pb, 7439-92-1.

Literature Cited (1) Barltrop, D. Postgrad. Med J . 1975, 51, 776. (2) Moore, M. R. Lead in Drinking Water and Its Significance to Health, Drinking Water Quality and Public Health; Water Research Centre: U.K., 1975. (3) Barltrop, D.; Burman,D.; Tucker, F. Arch. Dis. Child. 1976, 51, 809. (4) Duggan, M. J. Water, Air, Soil Pollut. 1980, 14, 309. (5) Thomas, J. F. A.; et al. Proceedings of the International Conference on Heavy Metals in the Environment, Athens, 1985; in press. (6) U.K. Department of the Environment Pollut. Pap.-Cent. Unit Enuiron. Pollut. (G.B.) 1980, No. 17. ( 7 ) Thornton, I.; Culbard, E.; Moorcroft, S.; Watt, J.;Wheatley, M.; Thompson, M.; Thomas, J. F. A. Environ. Technol. Lett. 1985, 6 , 137. (8) Hilton, K. J. in The Lower Swansea Valley Project; Longmans: London, 1967. (9) Ecology and the Industrial Society; Goodman, G. T.; Edwards, R. W.; Lambert, J. M., Eds.; B.E.S. Symposium 5; Blackwells: Oxford, 1965. (10) Welsh Office A Collaborative Study on Certain Elements in Air, Soil, Plants and Humans in the Swansea-NeathPort Talbot Area Together with a Moss Bag Study of Atmospheric Pollution across South Wales; HMSO: London, 1975. (11) Lead in the Environment;Boggess, R.; Wixson, B. G., Eds.; CRC Press: New York, 1977. (12) Cadmium in the Environment; Nriagu, J. O., Ed.; Wiley: New York, 1980; Part 1. (13) Biggins, P. D. E.; Harrison, R. M. Environ. Sei. Technol. 1980, 14, 336. (14) Olsen, K. W.; Skogerboe, R. K. Environ. Sci. Technol. 1975, 9, 227. (15) Whitby, K. T. Atmos. Environ. 1978, 12, 135. (16) Davies, B. E.; White, H. M. Sci. Total Enuiron. 1981,20, 57. (17) Lagenverff,J. V.; Specht, A. W. Environ. Sci. Technol. 1970, 4 , 583. (18) Zimdahl, R. L.; Skogerboe, R. K. Environ. Sei. Technol. 1977, 11, 1202. (19) Cranipton, C. B. Soils of the Vale of Glamorgan;Mem. Soil Surv. Eneland + Wales: HMSO: London, 1972. (20) Mitchell,-R. L. in Trace Elements in Soils and Crops; M.A.F.F. Tech. Bull. 21; HMSO: London, 1971. (21) Ward, N. I.; Brooks, R. R.; Roberts, E.; Boswell, C. R. Environ. Sci. Technol. 1977, 11, 917.

Received for review July 8, 1986. Revised manuscript received December 12, 1986. Accepted January 27, 1987.