Pergamon Atmospheric Environment Vol. 31, No. 17, pp. 2733-2740, 1997 © 1997 Elsevier Science Ltd All rights reserved. Printed in Great Britain PII: S1352-2310(97)00101-5 1352-2310/97 $17.00 + 0.00 ORIGIN AND PATTERNS OF DISTRIBUTION OF TRACE ELEMENTS IN STREET DUST: UNLEADED PETROL AND URBAN LEAD EDUARDO de MIGUEL,* JUAN F. LLAMAS,* ENRIQUE CHACON,* TORUNN BERG,t STEINAR LARSSEN,t ODDVAR ROYSET t and MARIT VADSET t * E. T. S. Minas Madrid (Madrid School of Mines), Rios Rosas 21, E-28003 Madrid, Spain; and t Norwegian Institute for Air Research, P.O. Box 100, N-2007 Kjeller, Norway (First received 23 February 1996 and in final form 29 January 1997) Abstract--The elemental composition, patterns of distribution and possible sources of street dust are not common to all urban environments, but vary according to the peculiarities of each city. The common features and dissimilarities in the origin and nature of street dust were investigated through a series of studies in two widely different cities, Madrid (Spain) and Oslo (Norway), between 1990 and 1994. The most comprehensive sampling campaign was carried out in the Norwegian capital during the summer of 1994. An area of 1,1 km 2, covering most of downtown Oslo and some residential districts to the north of the city, was divided into 1 km 2 mapping units, and 16 sampling increments of approximately 150 g were collected from streets and roads in each of them. The fraction below 100 I~m was acid-digested and analysed by ICP-MS. Statistical analyses of the results suggest that chemical elements in street dust can be classified into three groups: "urban" elements (Ba, Cd, Co, Cu, Mg, Pb, Sb, Ti, Zn), "natural" elements (AI, Ga, La, Mn, Na, Sr, Th, '() and elements of a mixed origin or which have undergone geochemical changes from their original sources (Ca, Cs, Fe, Mo, Ni, Rb, Sr, U). Soil resuspension and/or mobilisation appears to be the most important source of "natural" elements, while "urban" elements originate primarily from traffic and from the weathering and corrosion of building materials. The data for Pb seem to prove that the gradual shift from leaded to unleaded petrol as fuel for automobiles has resulted in an almost proportional reduction in the concentration of Pb in dust particles under 100 lim. This fact and the spatial distribution of Pb in the city strongly suggest that lead sources other than traffic (i.e. lead accumulated in urban soil over the years) may contribute as much lead, if not more, to urban street dust. © 1997 Elsevier Science Ltd. Key word index: Street dust, trace elements, urban, lead, sampling, ICP-MS. 1. INTRODUCTION One of the numerous geochemical features of urban environments is that they can be regarded as spatial domains where a steady flow of incoming trace ele- ments is subject to changes in their physicochemical state before being disposed of and accumulated within their limits. Among other possibilities, trace elements concentrate on urban surfaces as dust particles of a wide range of sizes (often referred to as "street dust"). The health implications of inhalation and/or inges- tion of dust particle, s with high concentrations of trace elements is the subject of intense debate and the re- suits of research to date are contradictory (Harvey et al., 1985a; McBride et al., 1985). It seems clear, nevertheless, that children are the sector of the popu- lation at highest risk (Rundle et al., 19851 Thomas et al., 1985; Harvey et al., 1985b; Biggins and Har- rison, 1980). There is a credible need, therefore, for assessing the concentration levels of trace elements in street dust. Maps can be used as a convenient tool to present the information gained from analysis of individual sam- ples and to facilitate interpretation of the results. The choice of mapping unit size and the number and mass of the sampling increments taken in each one (samp- ling strategy) determines the accuracy and precision of the final graphic display and the validity of the conclusions drawn from the information conveyed therein. 2. MAPS OF TRACE ELEMENTS AND SAMPLING STRATEGIES The most common approach to the production of geoehemical maps uses a regular grid to divide the area under investigation in a number of mapping units, all equal in shape and size. Each mapping unit is represented by a single concentration value. This value results either from analysis of one composite sample made up of a number of sampling increments 2733