T E M P O R A L VARIABILITY OF T H E E L E M E N T A L C O M P O S I T I O N
IN
URBAN STREET DUST S T E P H E N J. V E R M E T T E
Illinois State Water Survey, Atmospheric Science Division, Air Quality Unit, 2204 Griffith Dr., Champaign, IL 61820, U.S.A. K I M N. I R V I N E
Department of Geography & Planning, State University College at Buffalo, 1300 Elmwood, Ave., Buffalo, NY 14222, U.S.A. and J O H N J. D R A K E
Department of Geography, McMaster University, Hamilton, Ontario, Canada LSS 4Ml
(Received November 1989)
Abstract. Urban Street dust is recognized as a source of urban air and runoff degradation. This paper takes a preliminary step toward a better understanding of temporal variability in street dust chemistry and of the controlling mechanisms. Street dust samples, collected over four seasons in the city of Hamilton, Canada, show a variability dependent on element and source - anthropogenic sources exhibiting the greatest temporal variability. In addition, elements attributed to c o m m o n sources exhibit similar temporal patterns. The use of 'generic' or even one-time samples may seriously misrepresent the elemental make-up of urban street dust. Based on the samples collected in this study, a number of questions/insights are posed to further the study of street dust temporal variability.
Introduction
It is now well-established that urban street dust, and fugitive dust sources in general, can contribute significantly to the degradation in quality of the urban atmosphere and urban stormwater runoff (e.g. Novotny and Chesters, 1981; Lucas and Casuccio, 1987; Watson et al., 1989; Vermette et al., 1989). The efforts of various researchers have been directed at characterizing the elemental composition of various urban dusts to better ascertain their sources a n d / o r the degree to which they may contribute to pollution in a number of cities (eg. Harrison, 1979; Hopke et aL, 1980; MOE, 1982; Malmqvist, 1983; Dong et al., 1984; Hamilton et aL, 1984; Nriagu and Davidson, 1986; Ramlan and Badri, 1989). Accurate elemental source profiles are essential to source apportionment techniques, and urban street dust is commonly characterized to represent an urban dust source. Therefore, it is essential to evaluate the variability of the street dust elemental profde. Sampling has demonstrated compositional variability between and within cities such that a 'canned' or 'generalized' street dust source profile, as would be provided in the literature (e.g.U.S. EPA, 1984; Sartor and Gadboury, 1984; James, 1985), should be used only after careful consideration of the unique aspects of a particular study area, its local industry and parent materials. For example, pollutants in stormwater runoff often are modeled as a function of the street dust and the pollutant concentration is assumed a Environmental Monitoring and Assessment lg: 69-77, 199t. 9 1991 Kluwer Academic Publishers. Printed in the Netherlands.
70
STEPHEN
J. VERMETTE
ET AL.
representative constant (eg. Novotny and Chesters, 1981; James, 1985; Irvine, 1989). However, modelers should be aware that there may be spatial and temporal variability inherent in this 'representative constant'. The spatial variability of the chemical composition of street dust has been examined to some extent (e.g. Vermette et al., 1987), but surprising little research has dealt with temporal variability. Duggan (1984) examined the temporal variability of Pb from weekly samples over a period of two years in London, England. While a seasonal trend was not apparent, concentrations of Pb were shown to vary considerably from week to week. Hamilton et al., 1984(b) indicated that there was limited temporal variability of Cd, Cu, Pb, and Zn concentrations in the dust taken from three streets in London, England. However, only four samples were taken at each site between August 1982 and March 1983 and the interpretive emphasis was placed on the final three samples. In Canada, the longer, colder winters with more permanent snowpacks, intuitively could have some temporal effect on elemental concentrations in street dust. There undoubtedly is a need to assess the extent of the temporal variability of elemental concentrations in street dust. To this end, road dust samples were collected over four seasons, from two road sites in the industrial city of Hamilton, Ontario (population 300 000), While representative of a specific city, these data are presented as a preliminary step toward the identification of temporal variability in road dust chemistry and the controlling processes.
Sample Collection Street dust samples were collected from the curbside near two intersections (one block apart) in the Central Business District of Hamilton, Ontario (King St. at Bay St., and Main St. at Bay St.). Samples were taken within 0.25 m of the curb and over a parallel distance of approximately 2 m. The samples were collected between 20 February, 1987 and 5 October, 1987, for a total of six samples at each site (the length of time between collection ranged from 14 days to four months). At each site, surface material was removed from the gutter with a polyethylene scoop and dry-sieved on-site through a plastic 2 mm diameter filter into a polyethylene bag. Samples were air-dried prior to chemical analysis. Concentrations of AI, Ca, C1, Mn, Na, and V were determined by short-lived instrumental thermal neutron activation analysis at the McMaster University (Hamilton, Ontario) Reactor Facility. Bulk samples (particles < 2 mm in diameter) of approximately 0.1 g were analyzed with no pre- or post-irradiation chemistry. Three replicates taken from a sample bag were analyzed and elemental determinations were found not to vary more than the analytical errors.
Results and Discussion Concentrations, plotted over time, are presented in Figures 1, 3, and 4 for each of the studied elements. A range of concentrations showing spatial variability across the city from a previous one-time sampling carried out during July, 1986 (Vermette et al., 1987)
TEMPORAL
VARIABILITY
OF THE
ELEMENTAL
COMPOSITION
71
Road Salt 25-
Chloride (King St.) 20-
Chloride (Main St.) Sodium (King St.)
-8 15-
EI
Sodium (Main St.)
o
10-
~
5-
Mean & Range
Feb
March April April 1987
June
1986 SpatialVariabili~
Sept
Fig. 1. Chloride and sodium concentrations in street dust plotted over time. Spatial variability from a previous one-time sampling (Vermette et aL, 1987) taken July 1986.
Airborne Chloride 3500 30002500 2000(.9 z 1500
I,I
1000-
II Oc,o er II,
500 Winter
Spring Summer
.ll Ill
Autumn
1986 Fig. 2.
Seasonal variability of airborne chloride in Southeast Chicago (Vermette et aL, 1988).
72
STEPHEN
J.
VERMETTE
ET
AL.
Industrial Sources 2400-
Manganese (King St)
-50
S
2200 2000 m..
1800 o~
M a n g a n e s e (Main St)
-45
V a n a d i u m (King St) -40 ~" tX.
-35 E
.5 1600-
30 ~
tt~
1400
25
12001000
t
4
i
i
~
i
Feb MarchApril April June Sept 1987
V a n a d i u m (Main St)
20
Manganese Mean: 2600 ppm Range: 750-7200 ppm Vanadium Mean: 43 ppm Range: 22-72 ppm 11986 SpatialVariability I
Fig. 3. Manganese and vanadium concentrations in street dust plotted over time. Spatial variability from a previous one-time sampling (Vermette et aL, 1987) taken July 1986.
Indigenous Sources 16~I Calcium (King St)
140- r 120-
Calcium (Main St)
100
Aluminum (King St)
80
Aluminum (Main St)
"O r-
r
60
•
40 84
2o 0
~__-~--~~~-~
Mean & Range
~~
Feb Ma'rch Al~ril Al~ril Ju'ne 1987
Sept
1986 Spatial Variability
Fig_ 4. Calcium and aluminum concentrations in street dust plotted over time. Spatial variability from a previous one-time sampling (Vermette et al., 1987) taken July 1986.
TEMPORAL VARIABILITY OF T H E ELEMENTAL COMPOSITION
73
also are shown. The correlation matrices for the elements are presented in Table I. The coefficients of variation (CV's), representing the temporal variability for each element and the Vermette et al., (1987) study, are presented in Table II. C H L O R I D E AND SODIUM
The trend that is most striking and easiest to explain is observed for C1 and Na (Figure 1). The C1 and Na concentrations at both sample sites are high in the winter and decrease through the early autumn. The C1/Na ratio (King St. = 1.1; Main St. = 1.3) of the February sample is similar to that of road salt (1.4). The C1/Na ratios for the remaining samples range from 0.81 to 0.07, decreasing with time from the last application of road salt. The low ratio is similar to an average street dust ratio of 0.09 reported for Hamilton by Vermette et al., 1987 and approaches the average ratio for local soil (0.02). The lower CI and Na concentrations in the summer and autumn, combined with the lower CI/Na ratios, indicate that the seasonal source (i.e. road salt) is depleted by washoff and wind TABLE I
Correlation coefficients (r 2) for elements Ca
AI
Mn
V
Na
CI
1.00
-0,96 1.00
0.59 0.72 1.00
0.66 0.79 0.93 1.00
0.05 0.02 0.02 0.05 1.00
0.02 0.05 0.18 0.29 0.87 1.00
1.00
0.00 1.00
0.09 0.00 1.00
0.12 0.01 0.93 1.00
0.26 0.14 0.03 0.10 1.00
0.22 0.18 0.06 0.16 0.99 1.00
King Street Ca AI Mn V Na CI
Main Street Ca AI Mn V Na CI
TABLE II
Coefficients of variation representing the temporal and spatial variability of elemental concentrations
Temporal King Street Main Street
Ca
A1
Mn
V
Na
CI
0.142 0.127
0.157 0.152
0,265 0.190
0.268 0.155
0,480 0.590
1.700 1.380
0.201
0,382
0.676
0.364
0.237
0,308
Spatial
City Wide"
a After Vermette et al, (1987),
74
STEPHEN J. VERMETT E ET AL.
distribution processes and that other sources (i.e. soil) dominate. The seasonality of road dust C1, as measured in this study, is reflected in airborne monitoring at another city, Chicago, IL (Figure 2) - a similar seasonal variability in Hamilton's airborne CI could be expected. The higher Na summer/autumn concentrations for King St., as compared to Main St., are attributed to soil erosion from a nearby (across the sidewalk) garden. The seasonal influence of the garden further illustrates the temporal and spatial variability inherent in street dust sampling. Other trends are more difficult to explain and can be site-specific. However, the correlation matrices and CV's indicate two trends: (1) elements having anthropogenic sources tend to have greater temporal variability than non-anthropogenic derived elements; and (2) similar temporal patterns may occur for elements having predominately anthropogenic sources. For example, the large CV's (Table II) and correlation coefficients for C1 and Na (King St. r 2 = 0.87; Main St. r 2 = 0.99) may be attributed to the effects of road salt input and removal. M A N G A N E S E AND VANADIUM
For Mn and V (Figure 3), similar temporal patterns (King St.: r 2 = 0.93; Main St.: r 2 = 0.93) showing a quick buildup of concentrations in the spring and wide fluctuations in the summer/autumn suggests a common source. Manganese has been attributed to iron & steel emissions (Winchester and Nifong, 1981), while V has been attributed to the burning of residual oil and to various industrial processes (Usero and Gracia, 1986; Dzubay et al., 1988; Kalogeropoulos et al., 1989)- the city of Hamilton boasts the largest assemblage of steel mills in Canada. The buildup and fluctuation in concentrations appears a characteristic of anthropogenic (industrial) sources. Variability appears not to be affected by a single factor but it is thought by several interacting factors including: the total number of dry days in the sample period; rainfall washoff characteristics of storm events prior to sampling; and prevailing wind direction during the sampling period (affecting input rates from sources in different locations (i.e. Irvine et al., 1989)). A visual inspection indicated that street cleaning immediately prior (e.g. one-half to one and a half days) to sampling decreased the mass of road dust in the gutter, however, street cleaning did not appear to have a significant effect on Mn and V concentrations. The between street variability shown for Mn and V concentrations ((Figures 3) for the June and September samples offers a perplexing problem in defining the controlling mechanisms, but offers yet another example of the spatial and temporal variability inherent in road dust sampling. C A L C I U M AND ALUMINUM
The Ca and AI concentrations probably exhibit smaller temporal variability (Table 4) because of limited anthropogenic inputs. Calcium concentrations are enriched as compared to rural soil levels (Vermette et al., 1987), but Ca is an integral part of the urban landscape - indigenous source (i.e. cement, concrete, bricks and aggregates) and exhibits little spatial or temporal variability. Aluminum typically is associated with crustal or soil
TEMPORAL
VARIABILITY
OF THE
ELEMENTAL
COMPOSITION
75
sources. Soil inputs from the adjacent garden explains the higher correlation of A1 with Mn and V at the King St. site, and also the high negative correlation with Ca (varying dominance of Ca-enriched and Al-enriched sources). Finally, a comparison of the temporal and spatial variability in elemental concentrations can be revailing. This comparison can be done using the CV's for the King St. and Main St. samples (Table II) and the CV's for the street samples from throughout Hamilton that were obtained on a one-time basis (Vermette et al., 1987). The temporal variability is less than the spatial variability for AI, Mn, V and Ca. This reflects the variability of sources throughout the city. The temporal variability is greater than the spatial variability for CI and Na. Again, the influence of road salt is evident (the samples from Vermette et al., 1987 were taken in mid-summer, 1986). Together, the spatial and temporal variability of elemental concentrations can have a significant impact on a snap-shot sample that could be used to represent urban street dust. In comparing the CV's and elemental temporal patterns one is struck by the overall similarities between the two street samples. While spatial (Vermette et al., 1987) and temporal variability (this study) have been demonstrated individually it remains to determine to what degree temporal patterns are linked spatially. In this study, similarities between the two sampling sites may reflect a general linkage or may be more a product of their proximity. Conclusion
The small sampling period and limited number of elements studied restricts our ability to make any firm conclusions, however, a temporal variability in the elemental composition of urban street dust is apparent from the available data. This variability appears to be a function of element, source and different hydrometeorological variables affecting dispersion, although more information is needed to understand controlling mechanisms. In general, elements dominated by anthropogenic sources exhibit the greatest temporal variability with some temporal patterns being similar. The strong correlations between CI and Na and between Mn and V, for example, underline a potential for temporal characteristics as a means to identifying elemental sources. In addition, the chemical data hints of a linkage between spatial and temporal patterns. More information is required to understand other sources of variability in urban street dust and includes: (1) variabilities of scale (i.e. multiple samples within +__20 m; and (2) variabilities in sample collection techniques (e.g. does a brush and scoop sampling method bias against smaller particles as compared to the use of a vacuum cleaner). The spatial and temporal characteristics of street dust elemental concentrations and an understanding of their controlling mechanisms should be considered when developing and applying models to describe atmospheric and urban runoff quality. Recent efforts have begun to characterize spatial variability, but more research needs to be directed to temporal variability. Certainly, the use of 'generic' or even one-time samples may seriously misrepresent the elemental make-up of the street dust, and, perhaps as important, the determination of realistic uncertainties.
76
STEPHEN J. VERMETTE ET AL.
References Dong, A., Chesters, G., and Simsiman, G. V.: 1984, 'Metal Composition of Soil, Sediments and Urban Dust and Dirt Samples From The Menomonee River Watershed, Wisconsin, U.S.A', Water, Air andSoilPollution 22, 257-275. Duggan: 1984, 'Temporal and Spatial Variations of Lead in Air and in Surface Dust - Implications for Monitoring', The Science of the Total Environment 33, 37-48. Dzubay, T.G., Stevens, R.K., Gordon, G.E., Olmez, I., Sheffield, A.E., and Courtney, W.J.: 1988, 'A Composite Receptor Method Applied to Philadelphia Aerosol'. Environmental Science Technology 22, 46--52. Hamilton, R. S., Ellis, J. B., and Revitt, D.M.: 1984, 'Highway Pollutions: Proceedings of the International Symposium, Middlesex Polytechnic, United Kingdom, 6-9 September, 1983', The Science of the Total Environment 33, 1-200. Hamilton, R. S., Revitt, D. M., and Warren, R. S.: 1984(b), 'Levels and Physico-Chemical Associations of Cd, Cu, Pb, and Zn in Road Sediments', The Science of the TotalEnvironment 33, 59-74. Harrison, R. M.: 1979, 'Toxic Metals in Street and Household Dusts', The Science of the TotalEnvironment 11, 89-97. Hopke, P.K., Lamb, R. E., Natusch, D. F.S.: 1980, 'Multielemental Characterizations of Urban Roadway Dust', Environmental Science Technology 14, 164-172. Irvine, K.N.: 1989, 'The Effect of Pervious Urban Land on Pollutant Movement Through an Urban Environment', Unpub. Ph.D. Thesis, Department of Geography, McMaster University, Hamilton, Canada, 2lip. Irvine, K. N., Murray, S. D., Drake, J. J., and Vermette, S.J.: 1989, 'Spatial and Temporal Variability of Dry Dustfall and Associated Trace Elements: Hamilton, Canada', Environmental Technology Letters IO, 527-540. James, W.: 1985, User Manual PCSWMM3.2 Runoff Module, Computational Hydraulics Inc., Guelph, Ontario, 199p. Kalogeropoulos, N., Scoullos, M., Vassilaki-Grimani, M., and Grimanis, A. P.: 1989, "Vanadium in Particles and Sediments of the Northern Saronikos Gulf, Greece', The Science of the TotalEnvironment 79, 241-252. Lucas, J. H., and Casuccio, G. S.: 1987, 'The Identification of Sources of Total Suspended Particulate Matter by Computer Controlled Scanning Electron Microscopy and Receptor Modeling Techniques', Presented at the 80th Annual Meeting of the Air Pollution Control Association, New York, NY, June 21-26. Malmqvist, P.: 1983, Urban Stormwater Pollutant Sources, Chalmers University of Technology, Department of Sanitary Engineering, Goteborg, Sweden, 371 p. Nriagu, J. O. and Davidson, C. I., (eds.): 1986, Toxic Metals in the Atmosphere, Volume 17 in the Wiley Series in Advances in Environmental Science and Technology, John Wiley & Sons, New York, 63 p. Novotny, V. and Chesters, G.: 1981, Handbook of Nonpoint Pollution, Van Nostrand Reinhold Company, New York, 555p. Ontario Ministry of the Environment (MOE): 1982, An Assessment of Street Dust and Other Sources of Airborne Particulate Matter in Hamilton, Ontario (ARB-28-82-ARSP), 325 p. Ramlan, M. N. and Badri, M. A.: 1989, 'Heavy Metals in Tropical City Street Dust and Roadside Soils: A Case of Kuala Lumpur, Malaysia', Environmental Technology Letters 10, 435 ~4'!.. Sartor, J. D. and Caboury, D.R.: 1984, 'Street Sweeping as a Water Pollution Control Measure: Lessons Learned over the Past Ten Years', The Science of the Total Environment 33, 171-183. United States Environmental Protection Agency: 1984, Receptor Model Source Composition Library EPA450/4-85-002, Research Triangle Park, NC. Usero, J. and Gracia, I.: 1986, 'Trace and Major Elements in Atmospheric Deposition in the "Campo de Gibraltar" Region', Atmospheric Environment 20, 1639-1646. Vermette, S. J., Sweet, C. W., Gatz, D. F., Landsberger, S., and Hopke, P. K.: 1989, 'Receptor Modeling for PM-10 Source Apportionment in Southeast Chicago', Presented at the 82nd Annual Meeting & Exhibition of the Air & Waste Management Association, Anaheim, CA, June 25-30, 1989 (Paper ~ 89-33.7). Vermette, S.J., Sweet, C.W., and Landsberger, S.: 1988, 'Airborne Fine Particulate Matter (PM-10) in Southeast Chicago: Preliminary Report II', Illinois State Water Contract Report 48h, 2204 Griffith Dr., Champaign, IL 61820. Vermette, S. J., Irvine, K. N., and Drake, J. J.: 1987, 'Elemental and Size Distribution Characteristics of Urban
TEMPORAL VARIABILITY OF THE ELEMENTAL COMPOSITION
77
Sediments: Hamilton, Canada', Environmental Technology Letters 8, 619-634. Watson, Frazier, J. C., Chow, J., Farber, R. J. and Countess, R.J.: 1989, 'Survey of Fugitive Dust Control Methods for PM-10', Presented at the 82nd Annual Meeting ' Exhibition of the Air & Waste Management Association, Anaheim, CA, June 25-30, 1989 (Paper ~: 8%32.2). Winchester, J. W., and Nifong, D.: 1971, 'Water Pollution in Lake Michigan by Trace Elements from Pollution Aerosol Fallout', Water, Air and Sail Pollution 1, 50-64.
IN
URBAN STREET DUST S T E P H E N J. V E R M E T T E
Illinois State Water Survey, Atmospheric Science Division, Air Quality Unit, 2204 Griffith Dr., Champaign, IL 61820, U.S.A. K I M N. I R V I N E
Department of Geography & Planning, State University College at Buffalo, 1300 Elmwood, Ave., Buffalo, NY 14222, U.S.A. and J O H N J. D R A K E
Department of Geography, McMaster University, Hamilton, Ontario, Canada LSS 4Ml
(Received November 1989)
Abstract. Urban Street dust is recognized as a source of urban air and runoff degradation. This paper takes a preliminary step toward a better understanding of temporal variability in street dust chemistry and of the controlling mechanisms. Street dust samples, collected over four seasons in the city of Hamilton, Canada, show a variability dependent on element and source - anthropogenic sources exhibiting the greatest temporal variability. In addition, elements attributed to c o m m o n sources exhibit similar temporal patterns. The use of 'generic' or even one-time samples may seriously misrepresent the elemental make-up of urban street dust. Based on the samples collected in this study, a number of questions/insights are posed to further the study of street dust temporal variability.
Introduction
It is now well-established that urban street dust, and fugitive dust sources in general, can contribute significantly to the degradation in quality of the urban atmosphere and urban stormwater runoff (e.g. Novotny and Chesters, 1981; Lucas and Casuccio, 1987; Watson et al., 1989; Vermette et al., 1989). The efforts of various researchers have been directed at characterizing the elemental composition of various urban dusts to better ascertain their sources a n d / o r the degree to which they may contribute to pollution in a number of cities (eg. Harrison, 1979; Hopke et aL, 1980; MOE, 1982; Malmqvist, 1983; Dong et al., 1984; Hamilton et aL, 1984; Nriagu and Davidson, 1986; Ramlan and Badri, 1989). Accurate elemental source profiles are essential to source apportionment techniques, and urban street dust is commonly characterized to represent an urban dust source. Therefore, it is essential to evaluate the variability of the street dust elemental profde. Sampling has demonstrated compositional variability between and within cities such that a 'canned' or 'generalized' street dust source profile, as would be provided in the literature (e.g.U.S. EPA, 1984; Sartor and Gadboury, 1984; James, 1985), should be used only after careful consideration of the unique aspects of a particular study area, its local industry and parent materials. For example, pollutants in stormwater runoff often are modeled as a function of the street dust and the pollutant concentration is assumed a Environmental Monitoring and Assessment lg: 69-77, 199t. 9 1991 Kluwer Academic Publishers. Printed in the Netherlands.
70
STEPHEN
J. VERMETTE
ET AL.
representative constant (eg. Novotny and Chesters, 1981; James, 1985; Irvine, 1989). However, modelers should be aware that there may be spatial and temporal variability inherent in this 'representative constant'. The spatial variability of the chemical composition of street dust has been examined to some extent (e.g. Vermette et al., 1987), but surprising little research has dealt with temporal variability. Duggan (1984) examined the temporal variability of Pb from weekly samples over a period of two years in London, England. While a seasonal trend was not apparent, concentrations of Pb were shown to vary considerably from week to week. Hamilton et al., 1984(b) indicated that there was limited temporal variability of Cd, Cu, Pb, and Zn concentrations in the dust taken from three streets in London, England. However, only four samples were taken at each site between August 1982 and March 1983 and the interpretive emphasis was placed on the final three samples. In Canada, the longer, colder winters with more permanent snowpacks, intuitively could have some temporal effect on elemental concentrations in street dust. There undoubtedly is a need to assess the extent of the temporal variability of elemental concentrations in street dust. To this end, road dust samples were collected over four seasons, from two road sites in the industrial city of Hamilton, Ontario (population 300 000), While representative of a specific city, these data are presented as a preliminary step toward the identification of temporal variability in road dust chemistry and the controlling processes.
Sample Collection Street dust samples were collected from the curbside near two intersections (one block apart) in the Central Business District of Hamilton, Ontario (King St. at Bay St., and Main St. at Bay St.). Samples were taken within 0.25 m of the curb and over a parallel distance of approximately 2 m. The samples were collected between 20 February, 1987 and 5 October, 1987, for a total of six samples at each site (the length of time between collection ranged from 14 days to four months). At each site, surface material was removed from the gutter with a polyethylene scoop and dry-sieved on-site through a plastic 2 mm diameter filter into a polyethylene bag. Samples were air-dried prior to chemical analysis. Concentrations of AI, Ca, C1, Mn, Na, and V were determined by short-lived instrumental thermal neutron activation analysis at the McMaster University (Hamilton, Ontario) Reactor Facility. Bulk samples (particles < 2 mm in diameter) of approximately 0.1 g were analyzed with no pre- or post-irradiation chemistry. Three replicates taken from a sample bag were analyzed and elemental determinations were found not to vary more than the analytical errors.
Results and Discussion Concentrations, plotted over time, are presented in Figures 1, 3, and 4 for each of the studied elements. A range of concentrations showing spatial variability across the city from a previous one-time sampling carried out during July, 1986 (Vermette et al., 1987)
TEMPORAL
VARIABILITY
OF THE
ELEMENTAL
COMPOSITION
71
Road Salt 25-
Chloride (King St.) 20-
Chloride (Main St.) Sodium (King St.)
-8 15-
EI
Sodium (Main St.)
o
10-
~
5-
Mean & Range
Feb
March April April 1987
June
1986 SpatialVariabili~
Sept
Fig. 1. Chloride and sodium concentrations in street dust plotted over time. Spatial variability from a previous one-time sampling (Vermette et aL, 1987) taken July 1986.
Airborne Chloride 3500 30002500 2000(.9 z 1500
I,I
1000-
II Oc,o er II,
500 Winter
Spring Summer
.ll Ill
Autumn
1986 Fig. 2.
Seasonal variability of airborne chloride in Southeast Chicago (Vermette et aL, 1988).
72
STEPHEN
J.
VERMETTE
ET
AL.
Industrial Sources 2400-
Manganese (King St)
-50
S
2200 2000 m..
1800 o~
M a n g a n e s e (Main St)
-45
V a n a d i u m (King St) -40 ~" tX.
-35 E
.5 1600-
30 ~
tt~
1400
25
12001000
t
4
i
i
~
i
Feb MarchApril April June Sept 1987
V a n a d i u m (Main St)
20
Manganese Mean: 2600 ppm Range: 750-7200 ppm Vanadium Mean: 43 ppm Range: 22-72 ppm 11986 SpatialVariability I
Fig. 3. Manganese and vanadium concentrations in street dust plotted over time. Spatial variability from a previous one-time sampling (Vermette et aL, 1987) taken July 1986.
Indigenous Sources 16~I Calcium (King St)
140- r 120-
Calcium (Main St)
100
Aluminum (King St)
80
Aluminum (Main St)
"O r-
r
60
•
40 84
2o 0
~__-~--~~~-~
Mean & Range
~~
Feb Ma'rch Al~ril Al~ril Ju'ne 1987
Sept
1986 Spatial Variability
Fig_ 4. Calcium and aluminum concentrations in street dust plotted over time. Spatial variability from a previous one-time sampling (Vermette et al., 1987) taken July 1986.
TEMPORAL VARIABILITY OF T H E ELEMENTAL COMPOSITION
73
also are shown. The correlation matrices for the elements are presented in Table I. The coefficients of variation (CV's), representing the temporal variability for each element and the Vermette et al., (1987) study, are presented in Table II. C H L O R I D E AND SODIUM
The trend that is most striking and easiest to explain is observed for C1 and Na (Figure 1). The C1 and Na concentrations at both sample sites are high in the winter and decrease through the early autumn. The C1/Na ratio (King St. = 1.1; Main St. = 1.3) of the February sample is similar to that of road salt (1.4). The C1/Na ratios for the remaining samples range from 0.81 to 0.07, decreasing with time from the last application of road salt. The low ratio is similar to an average street dust ratio of 0.09 reported for Hamilton by Vermette et al., 1987 and approaches the average ratio for local soil (0.02). The lower CI and Na concentrations in the summer and autumn, combined with the lower CI/Na ratios, indicate that the seasonal source (i.e. road salt) is depleted by washoff and wind TABLE I
Correlation coefficients (r 2) for elements Ca
AI
Mn
V
Na
CI
1.00
-0,96 1.00
0.59 0.72 1.00
0.66 0.79 0.93 1.00
0.05 0.02 0.02 0.05 1.00
0.02 0.05 0.18 0.29 0.87 1.00
1.00
0.00 1.00
0.09 0.00 1.00
0.12 0.01 0.93 1.00
0.26 0.14 0.03 0.10 1.00
0.22 0.18 0.06 0.16 0.99 1.00
King Street Ca AI Mn V Na CI
Main Street Ca AI Mn V Na CI
TABLE II
Coefficients of variation representing the temporal and spatial variability of elemental concentrations
Temporal King Street Main Street
Ca
A1
Mn
V
Na
CI
0.142 0.127
0.157 0.152
0,265 0.190
0.268 0.155
0,480 0.590
1.700 1.380
0.201
0,382
0.676
0.364
0.237
0,308
Spatial
City Wide"
a After Vermette et al, (1987),
74
STEPHEN J. VERMETT E ET AL.
distribution processes and that other sources (i.e. soil) dominate. The seasonality of road dust C1, as measured in this study, is reflected in airborne monitoring at another city, Chicago, IL (Figure 2) - a similar seasonal variability in Hamilton's airborne CI could be expected. The higher Na summer/autumn concentrations for King St., as compared to Main St., are attributed to soil erosion from a nearby (across the sidewalk) garden. The seasonal influence of the garden further illustrates the temporal and spatial variability inherent in street dust sampling. Other trends are more difficult to explain and can be site-specific. However, the correlation matrices and CV's indicate two trends: (1) elements having anthropogenic sources tend to have greater temporal variability than non-anthropogenic derived elements; and (2) similar temporal patterns may occur for elements having predominately anthropogenic sources. For example, the large CV's (Table II) and correlation coefficients for C1 and Na (King St. r 2 = 0.87; Main St. r 2 = 0.99) may be attributed to the effects of road salt input and removal. M A N G A N E S E AND VANADIUM
For Mn and V (Figure 3), similar temporal patterns (King St.: r 2 = 0.93; Main St.: r 2 = 0.93) showing a quick buildup of concentrations in the spring and wide fluctuations in the summer/autumn suggests a common source. Manganese has been attributed to iron & steel emissions (Winchester and Nifong, 1981), while V has been attributed to the burning of residual oil and to various industrial processes (Usero and Gracia, 1986; Dzubay et al., 1988; Kalogeropoulos et al., 1989)- the city of Hamilton boasts the largest assemblage of steel mills in Canada. The buildup and fluctuation in concentrations appears a characteristic of anthropogenic (industrial) sources. Variability appears not to be affected by a single factor but it is thought by several interacting factors including: the total number of dry days in the sample period; rainfall washoff characteristics of storm events prior to sampling; and prevailing wind direction during the sampling period (affecting input rates from sources in different locations (i.e. Irvine et al., 1989)). A visual inspection indicated that street cleaning immediately prior (e.g. one-half to one and a half days) to sampling decreased the mass of road dust in the gutter, however, street cleaning did not appear to have a significant effect on Mn and V concentrations. The between street variability shown for Mn and V concentrations ((Figures 3) for the June and September samples offers a perplexing problem in defining the controlling mechanisms, but offers yet another example of the spatial and temporal variability inherent in road dust sampling. C A L C I U M AND ALUMINUM
The Ca and AI concentrations probably exhibit smaller temporal variability (Table 4) because of limited anthropogenic inputs. Calcium concentrations are enriched as compared to rural soil levels (Vermette et al., 1987), but Ca is an integral part of the urban landscape - indigenous source (i.e. cement, concrete, bricks and aggregates) and exhibits little spatial or temporal variability. Aluminum typically is associated with crustal or soil
TEMPORAL
VARIABILITY
OF THE
ELEMENTAL
COMPOSITION
75
sources. Soil inputs from the adjacent garden explains the higher correlation of A1 with Mn and V at the King St. site, and also the high negative correlation with Ca (varying dominance of Ca-enriched and Al-enriched sources). Finally, a comparison of the temporal and spatial variability in elemental concentrations can be revailing. This comparison can be done using the CV's for the King St. and Main St. samples (Table II) and the CV's for the street samples from throughout Hamilton that were obtained on a one-time basis (Vermette et al., 1987). The temporal variability is less than the spatial variability for AI, Mn, V and Ca. This reflects the variability of sources throughout the city. The temporal variability is greater than the spatial variability for CI and Na. Again, the influence of road salt is evident (the samples from Vermette et al., 1987 were taken in mid-summer, 1986). Together, the spatial and temporal variability of elemental concentrations can have a significant impact on a snap-shot sample that could be used to represent urban street dust. In comparing the CV's and elemental temporal patterns one is struck by the overall similarities between the two street samples. While spatial (Vermette et al., 1987) and temporal variability (this study) have been demonstrated individually it remains to determine to what degree temporal patterns are linked spatially. In this study, similarities between the two sampling sites may reflect a general linkage or may be more a product of their proximity. Conclusion
The small sampling period and limited number of elements studied restricts our ability to make any firm conclusions, however, a temporal variability in the elemental composition of urban street dust is apparent from the available data. This variability appears to be a function of element, source and different hydrometeorological variables affecting dispersion, although more information is needed to understand controlling mechanisms. In general, elements dominated by anthropogenic sources exhibit the greatest temporal variability with some temporal patterns being similar. The strong correlations between CI and Na and between Mn and V, for example, underline a potential for temporal characteristics as a means to identifying elemental sources. In addition, the chemical data hints of a linkage between spatial and temporal patterns. More information is required to understand other sources of variability in urban street dust and includes: (1) variabilities of scale (i.e. multiple samples within +__20 m; and (2) variabilities in sample collection techniques (e.g. does a brush and scoop sampling method bias against smaller particles as compared to the use of a vacuum cleaner). The spatial and temporal characteristics of street dust elemental concentrations and an understanding of their controlling mechanisms should be considered when developing and applying models to describe atmospheric and urban runoff quality. Recent efforts have begun to characterize spatial variability, but more research needs to be directed to temporal variability. Certainly, the use of 'generic' or even one-time samples may seriously misrepresent the elemental make-up of the street dust, and, perhaps as important, the determination of realistic uncertainties.
76
STEPHEN J. VERMETTE ET AL.
References Dong, A., Chesters, G., and Simsiman, G. V.: 1984, 'Metal Composition of Soil, Sediments and Urban Dust and Dirt Samples From The Menomonee River Watershed, Wisconsin, U.S.A', Water, Air andSoilPollution 22, 257-275. Duggan: 1984, 'Temporal and Spatial Variations of Lead in Air and in Surface Dust - Implications for Monitoring', The Science of the Total Environment 33, 37-48. Dzubay, T.G., Stevens, R.K., Gordon, G.E., Olmez, I., Sheffield, A.E., and Courtney, W.J.: 1988, 'A Composite Receptor Method Applied to Philadelphia Aerosol'. Environmental Science Technology 22, 46--52. Hamilton, R. S., Ellis, J. B., and Revitt, D.M.: 1984, 'Highway Pollutions: Proceedings of the International Symposium, Middlesex Polytechnic, United Kingdom, 6-9 September, 1983', The Science of the Total Environment 33, 1-200. Hamilton, R. S., Revitt, D. M., and Warren, R. S.: 1984(b), 'Levels and Physico-Chemical Associations of Cd, Cu, Pb, and Zn in Road Sediments', The Science of the TotalEnvironment 33, 59-74. Harrison, R. M.: 1979, 'Toxic Metals in Street and Household Dusts', The Science of the TotalEnvironment 11, 89-97. Hopke, P.K., Lamb, R. E., Natusch, D. F.S.: 1980, 'Multielemental Characterizations of Urban Roadway Dust', Environmental Science Technology 14, 164-172. Irvine, K.N.: 1989, 'The Effect of Pervious Urban Land on Pollutant Movement Through an Urban Environment', Unpub. Ph.D. Thesis, Department of Geography, McMaster University, Hamilton, Canada, 2lip. Irvine, K. N., Murray, S. D., Drake, J. J., and Vermette, S.J.: 1989, 'Spatial and Temporal Variability of Dry Dustfall and Associated Trace Elements: Hamilton, Canada', Environmental Technology Letters IO, 527-540. James, W.: 1985, User Manual PCSWMM3.2 Runoff Module, Computational Hydraulics Inc., Guelph, Ontario, 199p. Kalogeropoulos, N., Scoullos, M., Vassilaki-Grimani, M., and Grimanis, A. P.: 1989, "Vanadium in Particles and Sediments of the Northern Saronikos Gulf, Greece', The Science of the TotalEnvironment 79, 241-252. Lucas, J. H., and Casuccio, G. S.: 1987, 'The Identification of Sources of Total Suspended Particulate Matter by Computer Controlled Scanning Electron Microscopy and Receptor Modeling Techniques', Presented at the 80th Annual Meeting of the Air Pollution Control Association, New York, NY, June 21-26. Malmqvist, P.: 1983, Urban Stormwater Pollutant Sources, Chalmers University of Technology, Department of Sanitary Engineering, Goteborg, Sweden, 371 p. Nriagu, J. O. and Davidson, C. I., (eds.): 1986, Toxic Metals in the Atmosphere, Volume 17 in the Wiley Series in Advances in Environmental Science and Technology, John Wiley & Sons, New York, 63 p. Novotny, V. and Chesters, G.: 1981, Handbook of Nonpoint Pollution, Van Nostrand Reinhold Company, New York, 555p. Ontario Ministry of the Environment (MOE): 1982, An Assessment of Street Dust and Other Sources of Airborne Particulate Matter in Hamilton, Ontario (ARB-28-82-ARSP), 325 p. Ramlan, M. N. and Badri, M. A.: 1989, 'Heavy Metals in Tropical City Street Dust and Roadside Soils: A Case of Kuala Lumpur, Malaysia', Environmental Technology Letters 10, 435 ~4'!.. Sartor, J. D. and Caboury, D.R.: 1984, 'Street Sweeping as a Water Pollution Control Measure: Lessons Learned over the Past Ten Years', The Science of the Total Environment 33, 171-183. United States Environmental Protection Agency: 1984, Receptor Model Source Composition Library EPA450/4-85-002, Research Triangle Park, NC. Usero, J. and Gracia, I.: 1986, 'Trace and Major Elements in Atmospheric Deposition in the "Campo de Gibraltar" Region', Atmospheric Environment 20, 1639-1646. Vermette, S. J., Sweet, C. W., Gatz, D. F., Landsberger, S., and Hopke, P. K.: 1989, 'Receptor Modeling for PM-10 Source Apportionment in Southeast Chicago', Presented at the 82nd Annual Meeting & Exhibition of the Air & Waste Management Association, Anaheim, CA, June 25-30, 1989 (Paper ~ 89-33.7). Vermette, S.J., Sweet, C.W., and Landsberger, S.: 1988, 'Airborne Fine Particulate Matter (PM-10) in Southeast Chicago: Preliminary Report II', Illinois State Water Contract Report 48h, 2204 Griffith Dr., Champaign, IL 61820. Vermette, S. J., Irvine, K. N., and Drake, J. J.: 1987, 'Elemental and Size Distribution Characteristics of Urban
TEMPORAL VARIABILITY OF THE ELEMENTAL COMPOSITION
77
Sediments: Hamilton, Canada', Environmental Technology Letters 8, 619-634. Watson, Frazier, J. C., Chow, J., Farber, R. J. and Countess, R.J.: 1989, 'Survey of Fugitive Dust Control Methods for PM-10', Presented at the 82nd Annual Meeting ' Exhibition of the Air & Waste Management Association, Anaheim, CA, June 25-30, 1989 (Paper ~: 8%32.2). Winchester, J. W., and Nifong, D.: 1971, 'Water Pollution in Lake Michigan by Trace Elements from Pollution Aerosol Fallout', Water, Air and Sail Pollution 1, 50-64.
