Biomonitoring of traffic air pollution in Rome using magnetic properties of tree leaves

ARTICLE IN PRESS AE International – Europe Atmospheric Environment 37 (2003) 2967–2977 Biomonitoring of traffic air pollution in Rome using magnetic properties of tree leaves Eva Moreno*, Leonardo Sagnotti, Jaume Dinare" s-Turell, Aldo Winkler, Antonio Cascella Istituto Nazionale di Geofisica e Vulcanologia, Via Vigna Murata 605, Roma 00143, Italy Received 25 November 2002; received in revised form 14 March 2003; accepted 19 March 2003 Abstract We report a biomonitoring study of air pollution in Rome based on the magnetic properties of tree leaves. In a first step, magnetic properties of leaves from different tree species from the same location were compared. It was observed that leaves of evergreen species, like Quercus ilex, present much higher magnetic intensities than those of deciduous species, like Platanus sp., suggesting that leaves accumulate magnetic pollutants during their whole lifespan. In a second step, leaves from Q. ilex and Platanus sp. trees, both very common in Rome, have been used to monitor traffic emission pollution in two different periods. A Platanus sp. sampling campaign was undertaken in October 2001, at the end of the seasonal vegetational cycle, and 5 Q. ilex monthly sampling campaigns from April to August 2002. The strong difference observed in the magnetic susceptibility from leaves collected in green areas and roads allowed the realization of detailed pollution distribution maps from the south of Rome. Magnetic properties indicate that high concentrations and relatively larger grain-sizes of magnetic particles are observed in trees located along roads with high vehicle traffic and in the vicinity of railways. The decrease in concentration and grain size of magnetic particles with distance from the roadside confirms that magnetic properties of leaves are related to air pollution from vehicle emissions. The results indicate that a magnetic survey of tree leaves, which is relatively rapid and inexpensive, may be used in addition to the classical air quality monitoring systems to identify and delineate high-polluted areas in urban environments. r 2003 Elsevier Science Ltd. All rights reserved. Keywords: Magnetic properties; Biomonitoring; Leaves; Traffic emission; Rome 1. Introduction Magnetic properties of soils (Hay et al., 1997; Hoffmann et al., 1999; Hanesch et al., 2001; Leocoanet et al., 2001; Shu et al., 2001), filters (Muxworthy et al., 2001; Xie et al., 2001) and leaves (Georgeaud et al., 1997; Matzka and Maher, 1999) have been used for identifying spreading of pollution derived from vehicular or industrial emissions. *Corresponding author. Tel.: +39-0651-860386; fax: +390651-860397. E-mail address: moreno@ingv.it (E. Moreno). In aerosols, magnetic minerals are derived from combustion processes, such as industrial, domestic or vehicle emissions (Hunt et al., 1984; Flanders, 1994) or from abrasion products from asphalt and from vehicles brake systems (Hoffmann et al., 1999). Depending on the fuel type and the temperature of combustion, the magnetic fine particles mostly consist of spherules and grains of irregular shapes that contains variable amounts and grain size of magnetite and hematite (Matzka and Maher, 1999). Air pollution in Rome is most likely due to emissions by vehicular motors and, during the winter, by domestic heating systems, as industrial activity is low. 1352-2310/03/$ - see front matter r 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S1352-2310(03)00244-9 ARTICLE IN PRESS 2968 E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 Usually, air quality is measured by specific monitoring stations for different gases and suspended particles, especially PM10. However, at the moment, there are only 4 monitoring stations for the PM10 and 12 stations for the measurement of benzene, CO2, ozone and other gases in the city of Rome. The limited number of these stations in the city does not allow the production of high-resolution spatial distribution maps of air pollution in the urban area. In the last few years, biomonitoring, based on the analysis of trace elements in plants like lichens, mosses, ryegrass or tree leaves have been proposed as a solution to the air pollution monitoring problem. Mainly, these studies are based on the estimation of trace elements concentration originated from traffic or industrial . et al., 1998; Alfani et al., 2000; Monaci emissions (Bohm et al., 2000; Caggiano et al., 2001). Leaves with large surface areas per unit of weight and/ or a long lifespan, like conifer needles or evergreen tree leaves, are considered to be good accumulators (Alfani et al., 2000). Moreover, the obtained data represent a time-averaged result, which is more useful than the direct determination of the pollutant concentration on the air over a short period (Lau and Luk, 2001) to estimate the long-term effects of pollutions, that are probably the most influential on the health of the citizens. The main advantage is that plants are wide-spread, providing a high density of sampling points and the possibility of building high-resolution maps of air pollution in urban areas. The drawbacks are mainly related to the measurement quality, in terms of reproducibility and sensitivity, because of the high heterogeneity of the living conditions (Caggiano et al., submitted). Environmental magnetism is a useful tool as a potential biomonitoring method. Magnetite spherules have been observed on dust deposited on leaves near a motorway (Freer-Smith et al., 1997). On the other hand, Matzka and Maher (1999) shown that vehicular derived urban particulate matter includes a magnetite-like magnetic phase in the grain size range of 0.3–3 mm, whereas a specific study of atmospheric particulate matter collected in Munich (Germany) pointed out that the primary magnetic minerals derived from vehicular combustion and street-trams were maghemite and metallic iron, respectively, in the grain size of 0.1– 0.7 mm (Muxworthy et al., 2002). This grain size is particularly dangerous to humans because of its facility to be inhaled into the lungs. Moreover, in aerosols, magnetite is associated to other heavy metals like zinc, cadmium and chrome (Georgeaud et al., 1997) and to mutagenic organic compounds (Morris et al., 1995), also dangerous to human health. In this work, a new biomonitoring study of air pollution in Rome has been performed based on the magnetic properties of tree leaves. The aim of the work was to test the validity of the method in selected urban areas of Rome and its suburbs and to delineate effective sampling strategies and experimental protocols for conducting a magnetic biomonitoring study using tree leaves as natural dust collectors. 2. Location of the study areas and sampling Different sampling strategies at different times and areas of Rome (Fig. 1) were followed. A sampling test was conducted in a very restricted area, along Via Ostiense, a high-traffic road, collecting samples from various different tree species widely diffused in Rome, to test their suitability for our purposes. A first sampling campaign was carried out on Platanus sp. that is probably the most common tree species along roads in the town of Rome. Such sampling was carried out in a single day, in October 2001 and included leaves collections from 77 trees distributed in the southern half of Rome, from the Tiber River on the west to the Via Tuscolana on the east. A systematic study of Quercus ilex leaves was also undertaken in a selected area in the southeast of Rome, with collection of samples and measurements repeated each month from April to August 2002. The studied areas in Rome are characterized by very different traffic conditions. They include large suburban parks (Appia Antica Natural Park, including the Parco della Caffarella and the Parco degli Acquedotti) with low to null car circulation and major traffic axes like Via Tuscolana, Via Appia Nuova and Via Casilina, running from the outskirts toward the centre, Via di Porta Furba/ Via di Tor Pignattara connecting tangentially Via Casilina to Via Tuscolana, and Piazza Re di Roma, a roundabout square on the northwest sector of the study area. 3. Magnetic measurements At each site, 5–10 leaves were detached from the tree on the proximal side of the road about 1.5–2 m above the ground at the lower section of the crown. The leaves were place in 8 cm3 cubic plastic boxes, specifically designed for sampling of paleomagnetic specimens. Magnetic measurements were carried out on all the samples in the paleomagnetic laboratory of the Istituto Nazionale di Geofisica e Vulcanologia, within a day after sampling. The low-frequency (0.92 kHz) magnetic susceptibility (w) was measured at low-field (0.38 mT) using an AGICO Kappabridge KLY-2 instrument. We also measured the susceptibility of 10 empty plastic boxes that gave an average value of
2.1570.3 (10
8 m3/kg). ARTICLE IN PRESS E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 2969 Fig. 1. Location of the areas sampled in Rome. The asterisk (*) indicates the location of the test site shown in Fig. 2. The large area A refers to the sampling of October 2001 (Platanus sp. leaves). The smaller area B refers to the sampling carried out from April to August 2002 (Quercus ilex leaves). The area B is also shown in the detail on the right of the figure, with the main locations discussed in the text. In some samples, the isothermal remanent magnetization (IRM) produced on a pulse magnetizer in fields of 900 mT (IRM900) and of –300 mT (IRM
300) was measured with a 2G Enterprises 755R cryogenic magnetometer. Since for all the samples the IRM was saturated in fields of ca. 300–400 mT, the IRM900 can also be defined as the saturation IRM, or SIRM. All the values were normalized by the leaf wet mass. Magnetic susceptibility depends on the whole composition of the dust deposited on the leaves and is however dominated by ferrimagnetic minerals that have much higher susceptibility values than other common paramagnetic and diamagnetic minerals like clay or quartz. SIRM is mainly influenced by the concentration of lowcoercivity, magnetite-type, minerals and high-coercivity, hematite-type, minerals (e.g. Thompson and Oldfield, 1986). On the contrary, IRM
300 mT is mainly influenced by the low-coercivity fractions only. Therefore, the
IRM
300 mT/SIRM ratio or S
300 (Bloemendal et al., 1988) is used for estimating the relative contribution of high-coercivity and low-coercivity minerals. When the S
300 ¼ 1; the magnetic mineralogy is composed by magnetite-type minerals only. The lower is the S
300, the highest is the content of high-coercivity minerals. The SIRM/w ratio depends on the composition and the grain-size of the magnetic particles. When the magnetic mineralogy is homogeneous, the SIRM/w ratio indicates changes in the grain size of the magnetic minerals assemblage or in the contribution of paramagnetic minerals (e.g. clays). Magnetic susceptibility was measured in all campaigns whereas IRM was only measured in the April and August 2002 collections. IRM values show a good linear correlation coefficient of R ¼ 0:88 with susceptibility. The two parameters can therefore be assumed as representative of the amount of ferrimagnetic particles on the leaves surface. For systematic measurements, magnetic susceptibility was selected, since the measurements take only a few seconds per sample and a whole batch of samples can be promptly measured soon after the collection. 4. Results and discussion 4.1. Comparison between different tree species Different species of tree leaves were collected within a distance of 2 m from the road in order to find out which species is the most suitable for air pollution monitoring by means of magnetic measurements. Samples were collected on a limited stretch (ca. 500 m) of Via Ostiense, a tree-lined road where traffic is very intensive and with usual traffic jams (Fig. 2). The vegetation in the selected stretch of Via Ostiense is composed by different tree ARTICLE IN PRESS 2970 E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 Fig. 2. Magnetic susceptibility values shown as bar length (10
8 m3/kg) of different tree leaves collected in Via Ostiense, a heavy traffic road. species very abundant in Rome: two evergreen species (Q. ilex tree and Nerium Oleander shrub) and three deciduous tree species (Platanus sp., Cercis siliquastrum, and Robinia Pseudoacacia). As all samples have been collected in a limited area under the same living conditions (distance to the street, car circulation and meteorological factors), the differences due to problems of reproducibility are minimized. However, intra-specific differences appear, that are due to relative position of the tree with respect to the traffic source. In the case of Q. ilex, the highest value corresponds to the cross-light between Via Ostiense and the LungoTevere San Paolo that is the point with the maximum traffic and the longer halt of circulating vehicles. There are different factors that can influence the ability of the leaves to retain atmospheric fine particles. These include the duration of the exposure, the surface and the texture of the leaf and the capacity of the stomata to adsorb pollutants. Broad leaves with large and rugose leaves seems more efficient (Bussoti et al., 1995). The duration of the exposure seems to be one of the main factors controlling in the intensity of the magnetic susceptibility. The two evergreen species sampled show much higher values of susceptibility than the deciduous ones (Fig. 2). This is true for values normalized by the wet and the dry mass so there is not due to the different content in water within the leaf (Table 1). Q. ilex shows a susceptibility value between 100 and 1000 times higher than in Robinia Pseudoacacia and 65 times higher than in Platanus sp. However this is not the only controlling factor because differences between groups also appear within the sampled evergreen and deciduous species, that are probably due the characteristics of the leaf: texture, specific surface, etc (Table 1). We selected Platanus sp. and Q. ilex from all the tested species for doing further experiments because they present the highest susceptibility values within the deciduous and the evergreen sampled species and they are very abundant in the urban environment (roadsides and green areas). We compared samples from Platanus sp. and Q. ilex collected in areas of different traffic intensity (Fig. 3). The leaves susceptibility from both trees increase with the intensity of traffic, indicating that both species are affected by vehicular emissions. However, leaves from Q. ilex always show higher magnetic intensities confirming the finding of the test site and suggesting that the duration of the exposure to pollutants play a main role in the accumulation/ adsorption of magnetic dust. Q. ilex has long and narrow leaves with a lifespan can reach 3 yr. However, it has been shown that when the leaves are exposed to high traffic level conditions, their lifespan is reduced to 1 yr and rarely to 2 yr (Gratani et al., 2000). Numerous studies have shown that Q. ilex is a suitable biomonitor (Alfani et al., 2000, 2001) and the analysis of trace elements content in their leaves have pointed out that they accumulate lead and other elements as function of the exposure time (Monaci et al., 2000). On the other side, Platanus sp. is a deciduous species with broad and large leaves. It has a leaf lifespan of only a few months, accumulating dust only during the vegetational period. Q. ilex has been used for 5 sampling campaigns from April to August 2002 and Platanus sp. for one sampling campaign in October 2001. 4.2. Comparison between different traffic levels Different sampling campaigns of Q. ilex leaves were undertaken from April to August 2002. Table 2 shows the descriptive statistics of magnetic parameters measured. As magnetic properties are log-normally distributed, the mean and the standard deviation have been calculated from the log-transformed values. The samples have been classed in three groups: green areas that ARTICLE IN PRESS E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 2971 Table 1 Susceptibility values of the different tree and shrub species sampled in Via Ostiense normalized by the wet and the dry mass Type Specie Type of leaf Wet w (10
8 m3/kg) Dried w (10
8 m3/kg) Evergreen Quercus ilex Leathery green leaves, dark and glossy above, downy beneath, sometimes toothed, but very variable in form. 17.01 49.57 54.35 7.68 Long thin, leathery grey– 4.05 green leaves. 4.61 79.90 31.20 18.55 Nerium oleander Deciduous Cercis siliquastrum Large heart-shaped leaves. Platanus sp. Very large alternate leaves, 3–5 lobed, almost hairless. Robinia pseudoacacia Small pinnate leaves almost hairless and bluish-green beneath. 18.14 0.23 0.71 1.11 0.59 1.01 3.04 2.50 3.97 0.83 0.05 3.40 0.22 Fig. 3. Magnetic susceptibility (in 10
8 m3/kg) of selected Platanus sp. and Quercus ilex leaves collected in areas of different traffic intensities. correspond to suburban parks and small urban gardens, low/medium traffic roads and high traffic roads. Comparison between groups (Table 3) shows that the difference between the w and SIRM means estimated in Table 2 are statistically very significant for w and SIRM/ w (po0:01) and significant for SIRM (po0:05). The difference between medium/low and high traffic, are also statistically different except for the SIRM from April and SIRM/w in August. Finally, the difference between green areas and medium/low traffic roads are only statistically significant in April suggesting that there is a gradual trend on the magnetic properties of leaves with not definite boundary in which local effects may be important. Samples at high distance from high traffic road can have relatively low magnetic intensities, medium/low group can contains samples collected in streets with low traffic level that it is little affected by vehicular emissions and samples collected in green areas but close to roads can have relative high magnetic intensities (Fig. 4). ARTICLE IN PRESS E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 2972 Table 2 Descriptive statistics Month w (10
8 m3/kg) Traffic SIRM (10
5 m3/kg) SIRM/w (kA/m) S
300 Mean SD Mean N Mean SD Mean SD SD April Total H M/L G 21 3 8 10 8.15 19.65 10.11 5.27 1.90 1.34 1.38 1.75 116.46 178.76 132.28 92.49 1.53 1.28 1.43 1.48 14.29 9.10 13.08 17.56 1.35 1.18 1.20 1.23 0.97 0.98 0.97 0.97 0.01 0.01 0.01 0.01 August Total H M/L G 58 24 14 20 5.85 9.74 5.26 3.41 2.29 2.30 1.83 1.77 84.44 121.36 76.78 58.40 2.03 2.22 1.74 1.57 14.44 12.46 14.60 17.11 1.37 1.36 1.18 1.40 0.98 0.99 0.98 0.98 0.01 0.01 0.01 0.01 Month Traffic w (10
8 m3/kg) May Total H M/L G N 21 6 7 8 Mean 14.59 33.40 14.05 8.10 w (10
8 m3/kg) Month SD 2.25 1.95 1.49 1.93 N 29 11 7 11 June Month Mean 5.01 11.78 3.30 2.78 SD 4.03 3.84 3.48 3.25 July w (10
8 m3/kg) N 59 25 17 17 Mean 4.66 8.04 3.62 2.70 SD 2.75 2.33 2.32 2.82 Mean; N: number of samples; SD: standard deviation of the log-transformed distribution of w; SIRM, SIRM/w and S-300 measured in Quercus Ilex leaves sampled from April to August 2002. Samples have been classed as function of the traffic intensity in three categories: High traffic (H), medium/low traffic (M/L) and green areas (G). Table 3 One-way analysis of variance between means from Table 2 among the different levels of traffic using the Student–Newman–Keuls test Comparison p-values among groups April H vs. G H vs. M/L M/L vs. G May June July August w SIRM SIRM/w w w w w SIRM SIRM/w 0.001 0.044 0.007 0.033 >0.05 >0.05 o0.001 0.005 0.014 o0.001 0.016 >0.05 0.032 0.047 >0.05 o0.001 0.007 >0.05 o0.001 0.012 >0.05 0.001 0.038 >0.05 0.002 >0.05 >0.05 Values in bold indicate statistically significant differences. High traffic (H), medium/low traffic (M/L) and green areas (G). The S
300 is higher than 0.97 and there is not a statistically significant difference for S
300 suggesting that the magnetic mineralogy is independent on the location and dominated by low-coercivity minerals. These results indicate that magnetic properties are well affected by vehicular emissions and therefore good indicators of air pollution. The mean SIRM/w values found in April and August are around 14 kA/m close to those reported in English polluted topsoils (Hay et al., 1997) and in leaves collected in Leoben, Austria (Hanesch et al., in preparation). These values were interpreted as due to magnetite multidomain grains, very common in industrial fly ash. An increase in the SIRM/w ratio was observed from high traffic roads to green areas (Table 2) and with the distance to the roadside (Fig. 4). As the magnetic mineralogy indicated by the S
300 is homogeneous, it can be assumed that these increments indicate a decrease in the magnetic particles grain size with an increasing distance from the pollution sources. 4.3. Influence of the distance to the roadside The values of susceptibility, IRM and IRMXw in various locations of Rome with different traffic intensities, or from suburban parks, have been plotted in Fig. 4 as a function of the distance from the roadside. ARTICLE IN PRESS E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 2973 ibility and IRM values was also measured with the distance to the roadside. These results are in agreement with the findings by Hoffmann et al. (1999) and Matzka and Maher (1999) and confirm that the main source of magnetic pollution is traffic emissions and that the load of fine magnetic particles significantly reduces in the first meters distance from the roadside, where the coarser grains are deposited, while only the finest magnetic particles reach farther distances. 4.4. Spatial distribution of the Platanus sp. survey The rather uniform spatial distribution of Platanus sp. leaf sites collected in October 2001 in the southern part of Rome allows a visualization of magnetic measurements in the form of a contouring map (Fig. 5). The wet weight normalized low-field magnetic susceptibility (w) varies from 0.1 to 10.4 (10
8 m3/kg) and it can be observed that high values define highs in the contouring map, which are located in zones of high traffic intensity. The highest value we measured was found in the Via Ostiense toward the city centre. The tree was located in the proximity of the General Markets, the centre for urban food distribution, daily crowded by trucks and commercial vehicles. However there are also contrasting relative ‘‘lows’’ and ‘‘highs’’ close to each other like to the south of the Via Ardeatina (Fig. 5) which can be explained by local effects (high susceptibility values from a site very close to a bus stop). The results suggest that although data contouring is a useful way to a snapshot appreciation of pollution in urban areas, care has to be taken in the selection and distribution of the sampling sites. 4.5. Spatial distribution of the Quercus ilex survey Fig. 4. Susceptibility, SIRM and SIRM/w versus the location and the distance to the road. For each area, the average values for samples located within the same distance from the road were taken in account. In general, both susceptibility and IRM are observed to decrease and the SIRM/w ratio increase as the distance to the roadside increases, suggesting a decrease in the concentration and grain size of magnetic minerals with the distance to the road. This is particularly clear in the more polluted points. In Via di Porta Furba, the susceptibility decreases from around 45 to 9 (10
8 m3/ kg) moving from 2 to 25 m of distance to the roadside and a value of 2 (10
8 m3/kg) was measured in an adjacent park. In Piazza Re di Roma, a significant decrease in susceptibility and IRM was measured between the trees along the roundabout and those located in the central garden at a 25 m distance from the roadside. In Via Lemonia, the samples were collected at the edge of a suburban garden: a decrease in suscept- The spatial distribution maps of the magnetic susceptibility values from the 5 sampling campaigns undertaken from April to August 2002 are shown in Fig. 6. The sampling sites distribution is random and slightly clustered. Due to the scattered presence of Q. ilex trees in the study area, the data were plotted in points rather that in a contour map. For each month, the results were classed in 6 groups of equal intervals of magnetic susceptibility. A point, whose diameter increases with the magnetic susceptibility, represented each group. To visualize better the spatial distribution of the data, the range of magnetic susceptibility values in each group are based on the absolute values obtained in each specific sampling campaign. In each survey, the highest susceptibility values were found in Via di Porta Furba/Via di Tor Pignattara, Via Casilina and Piazza Re di Roma. All these places are characterized by a high traffic density with several traffic jams during the day. Therefore magnetic measurements of Q. ilex leaves confirm that vehicles constitute the ARTICLE IN PRESS 2974 E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 Fig. 5. (A) Location of the sampling campaign of the Platanus sp. trees sampled in October 2001, (B) magnetic susceptibility contour map and isolines. The contour map has been built using the kriging gridding method and a grid size of 30  30 m. main source of atmospheric fine particles in Rome. The maps also suggest a possible influence of city railways given that the highest susceptibility value measured in Via di Tor Pignattara, is situated close to a city railway bridge and in Via Casilina the railway runs parallel to the road. In Munich, fragment of maghemite and metallic iron have been found in atmospheric particulate matter in the vicinity of street-trams (Muxworthy et al., 2002). It also seems that in Rome, railways debris may have a certain influence in magnetic properties of leaves. On the other side, the lowest susceptibility values are found in urban gardens and suburban parks situated mostly on the southwest sector of the selected area. The minimum absolute values varies for each month but usually they are lower than 2 (10
8 m3/kg) except in May where the minimum value is 4.5 (10
8 m3/kg). They are found in trees situated in the Parco degli Acquedotti and Parco della Caffarella far from the roadside. These values can be considered as the natural background values for the Roman area. 4.6. Temporal distribution Temporal variations of the mean susceptibility and IRM values (Table 2) are partially due to the difference in the number of trees sampled in each campaign. In Fig. 7, the leaves susceptibility values from different streets from Rome have been compared with the daily PM10 concentrations recorded in Magna Grecia, the nearest automatic monitoring station. The temporal variations of susceptibility are not always correlated to the PM10 concentration fluctuations. Various possible explanations can be suggested to explain the lack of correlation. First, the 4 PM10 monitoring stations are situated out of our sampling area (Fig. 1). We also think that the magnetic properties of leaves represent a time-averaged dust accumulation and that magnetic biomonitoring can be used to indicate a long-time exposure to urban atmospheric particulate matter. However, the temporal fluctuations also indicate that magnetic properties can be controlled by other atmospheric and meteorological processes. In agreement with previous works, we did not observe a decrease in susceptibility after rainfalls (Matzka and Maher, 1999; Muxworthy et al., 2001) suggesting that part of the magnetic properties are due to pollutants adsorbed by the leaf and not only by the dust deposited on its surface. However, other meteorological factors such a sun hours, wind speed, air pressure and relative humidity can induce to the capacity of adsorbing pollutants (Muxworthy et al., 2001). ARTICLE IN PRESS E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 2975 Fig. 6. Location of the Quercus ilex trees and magnetic susceptibility spatial distribution for the sampling campaigns carried out from April to August 2002. These results indicate that magnetic properties of tree leaves cannot be used for an instantaneous reading of air pollution like PM10 station but they are a useful tool for the identification of areas exposed to long-term air pollution. 5. Conclusions Magnetic properties of leaves in Rome reveal that the magnetic fraction of the urban dust is dominated by magnetite-type minerals. The enhancement in magnetite concentration in areas with heavy traffic circulation and the decrease in magnetite concentration and grainsize with the distance to the roadside indicate that the main source of pollution is derived from vehicles emissions. The high susceptibility values obtained in sites close to railways suggest that railways debris may also constitute also an important source of fine magnetic particles. Both Q. ilex and Platanus sp. leaves are affected by traffic emissions but we relate the higher values observed in Q. ilex leaves to the longer leaf-lifespan and therefore to a longer duration of exposure to atmospheric pollutants. Because Q. ilex is evergreen, the air pollution biomonitoring can be done all year around. ARTICLE IN PRESS 2976 E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 Fig. 7. Comparison between susceptibility values in Quercus ilex and PM10 concentration of the collecting day. The lack of correlation with the hourly and daily PM10 records, as measured in the automatic monitoring stations, and the small effect of the rainfalls on the magnetic parameters suggest that magnetic properties of leaves cannot be used as an instantaneous record of dust load. However, the results strikingly demonstrate that magnetic parameters can provide an excellent measure of the long-term pollution loadings, which are probably more meaningful in terms of human health hazards than instantaneous readings from air filter stations. We suggest that this new method could be used as a complement to the traditional monitoring stations in the cities air pollution networks as a rapid a inexpensive proxy for mapping traffic emission pollution in urban areas. However, much more work is necessary to identify the links between magnetic parameters and trace elements concentration in the plant as well as the influence of meteorological factors in the magnetic properties of the leaf. Acknowledgements This research was funded by EU TMR Network contract ERBFMRXCT98-0247 (Mag-Net). We thank the Roma Environmental service for the contribution of the PM10 data. References Alfani, A., Baldantoni, D., Maisto, G., Bartoli, G., Virzo de Santo, A., 2000. Temporal and spatial variation in C, N, S and trace element contents in the leaves of Quercus ilex within the urban area of Naples. Environmental Pollution 109, 119–129. Alfani, A., Maisto, G., Pratti, M.V., Baldantoni, D., 2001. Leaves of Quercus ilex as biomonitors of PAHs in the air of Naples (Italy). Atmospheric Environment 35, 3553–3559. Bloemendal, J., Lamb, J.B, King, J., 1988. Paleoenvironmental implications of rock-magnetic properties of late quaternary sediment cores from the eastern equatorial Atlantic. Paleoceanography 3 (1), 61–87. . Bohm, P., Wolterbeek, H., Verburg, T., Mulisek, L., 1998. The use of tree bark for environmental pollution monitoring in the Czech Republic. Environmental Pollution 102, 243–250. Bussoti, F., Grossoni, P., Batistoni, P., Ferreti, M., Cenni, E., 1995. Preliminary studies on the ability of plant barriers to capture lead and cadmium of vehicular origin. Aerobiologia 11, 11–18. Caggiano, R., D’Emilio, M., Macchiato, M., Ragosta, M., 2001. Ryegrass species as biomonitors of atmospheric heavy metals emissions. Fresenius Environmental Bulletin 10, 31–36. Caggiano, R., D’Emilio, M., Macchiato, M., Ragosta, M. Heavy metals in ryegrass species versus metal concentrations in atmospheric particulate. Atmospheric Environment, submitted for publication. Flanders, P.J., 1994. Collection, measurement, analysis of airborne magnetic particulates from pollution in the environment. Journal of Applied Physics 75, 5931–5936. Freer-Smith, P.H., Holloway, S., Goodman, A., 1997. The uptake of particulates by an urban woodland, site description and particulate composition. Environmental Pollution 95 (1), 27–35. Georgeaud, V.M., Rochette, P., Ambrosi, J.P., Vandamme, D., Williamson, D., 1997. Relationship between heavy metals and magnetic properties in a large polluted catchments, the Etang de Berre (South France). Physics and Chemistry of the Earth 22 (1–2), 211–214. Gratani, L., Crecente, M.F, Petruzzi, M., 2000. Relationship between leaf-life span and photosynthetic activity of Quercus ilex in polluted urban areas (Rome). Environmental Pollution 110, 19–28. Hanesch, M., Scholger, R., Dearing, J.A. Recording pollution in cities by measuring magnetic parameters of tree leaves. Atmospheric Environment, submitted for publication. Hanesch, M., Scholger, R., Dekkers, M.J., 2001. The application of fuzzy c-means cluster analysis and non-linear mapping to a soil data set for the detection of polluted sites. Physics and Chemistry of the Earth 26 (11–12), 885–891. Hay, K.L., Dearing, J.A., Baban, S.M.J., Loveland, P., 1997. A preliminary attempt to identify atmospherically derived pollution particles in English topsoils from magnetic susceptibility measurements. Physics and Chemistry of the Earth 22 (1–2), 207–210. Hoffmann, V., Knab, M., Appel, E., 1999. Magnetic susceptibility mapping of roadside pollution. Journal of Geochemical Exploration 66, 313–326. Hunt, A., Jones, J., Oldfield, F., 1984. Magnetic measurements and heavy metals in atmospheric particles of anthropogenic origin. The Science of the Total Environment 33, 129–139. ARTICLE IN PRESS E. Moreno et al. / Atmospheric Environment 37 (2003) 2967–2977 Lau, O.W., Luk, S.F., 2001. Leaves of Bauhinia blakeana as indicators of atmospheric pollution in Hong Kong. Atmospheric Environment 35, 3113–3120. Leocoanet, H., Leveque, F., Ambrosi, J.-P., 2001. Magnetic properties of salt-marsh soils contaminated by iron industry emissions (Southeast France). Journal of Applied Geophysics 48, 67–81. Matzka, J., Maher, B.A., 1999. Magnetic biomonitoring of roadside tree leaves, identification of spatial and temporal variations in vehicle-derived particulates. Atmospheric Environment 33, 4565–4569. Monaci, F., Moni, F., Lanciotti, E., Grechi, D., Bargagli, R., 2000. Biomonitoring of airborne metals in urban environments, new tracers of vehicle emission, in place of lead. Environmental Pollution 107, 321–327. Morris, W.A., Versteeg, J.K., Bryant, D.W., Legzdins, A.E., McCarry, B.E, Marvin, X.H., 1995. Preliminary comparisons between mutagenic and magnetic susceptibility of respirable airborne particle. Atmospheric Environment 29, 3441–3450. 2977 Muxworthy, A., Matzka, J., Petersen, N., 2001. Comparison of magnetic parameters of urban atmospheric particulate matter with pollution and meteorological data. Atmospheric Environment 35, 4379–4386. Muxworthy, A., Schmidbauer, E., Petersen, N., 2002. Magnetic . properties and Mossbauer spectra of urban atmospheric particulate matter, a case study from Munich, Germany. Geophysical Journal International 150, 558–570. Shu, J., Dearing, J.A., Morse, A.P., Yu, L., Yuan, N., 2001. Determining the sources of atmospheric particles in Shanghai, China, from magnetic and geochemical properties. Atmospheric Environment 35, 2615–2625. Thompson, R., Oldfield, F., 1986. Environmental Magnetism. Allen & Unwin, London. Xie, S., Dearing, J.A., Boyle, J.F., Bloemendal, J., Morse, A.P., 2001. Association between magnetic properties and element concentrations of Liverpool street dust and its implications. Journal of Applied Geophysics 48, 83–92. All in-text references underlined in blue are linked to publications on ResearchGate, letting you access and read them immediately.