Research Articles
Atmospheric Lead and Bromine in Germany
Research Articles
Atmospheric Lead and Bromine in Germany Post-abatement Levels, Variabilities and Trends Gerhard Lammel 1*, Andreas Röhrl 1,2 and Hermann Schreiber 3 1 Max
Planck Institute for Meteorology, Bundesstraße 55, D-20146 Hamburg, Germany University of Hamburg, Institute for Inorganic and Applied Chemistry, Germany 3 University of Hohenheim, Institute for Physics and Meteorology, Stuttgart, Germany 2
*Corresponding author (
[email protected])
DOI: http://dx.doi.org/10.1065/espr2002.10.137 Abstract
Intention, Goal, Scope, Background. Abatement measures since the 1970s have depleted lead and bromine levels in the atmosphere over large parts of Europe. Our knowledge of the atmospheric cycling of these elements while a several decade-long period of intensive mobilization reaches its end is incomplete. Objective. We have characterized the trends in the atmospheric levels of Pb and Br and present-day temporal and spatial variabilities. Methods. This was achieved by short-term (weeks) and long-term (years) measurements of particulate Pb and Br at various sites in Germany. Samples of atmospheric particulate matter were collected on filter membranes and analyzed by x-ray fluorescence. Results and Discussion. Average Pb levels at rural and urban inland sites did not exceed concentrations in background aerosols, sampled at a Baltic Sea coastal site, by more than a factor of 3. Due to sea salt, bromide inland particulate Br levels are below those at a coastal site. There, non-sea salt Br, however, is significant as well. Urban Pb and Br levels are not necessarily higher than rural levels. The concentration levels have decreased in such a way that the previously common source, local vehicular traffic emissions, is no longer predominant. Regional rather than local sources have been increasingly determining the concentrations since the 1990s. This is more pronounced for Br than for Pb. We found indications for coal burning and longrange transport as significant Pb sources. For particulate Pb species, a range of ages (elapsed time since Pb emission) has been found. This range shows two maxima corresponding to characteristic times of 72 and 24 h. Conclusions. The (mean) atmospheric residence time of particulate Pb is longer than the residence time of Br, in particular in the wintertime. The chemical species contributing to atmospheric Pb should be addressed in future studies. Recommendation and Outlook. Clearly, despite effective abatement measures, atmospheric Pb will continue to be dominated by anthropogenic mobilization. The influence from long-range transport can be expected to decrease with the effectiveness of abatement programmes in neighbouring countries of the region. Keywords: Air quality; atmospheric aerosols; bromine; central
Europe; lead; seasonality; vehicular traffic emissions
Introduction
Br in the environment has both strong natural (sea salt, marine biosphere) and anthropogenic sources (Sturges and Harrison 1986, Dowdell et al. 1994, Gribble 2000). Significant anthropogenic sources are automotive emissions, pesticide application (CH3Br), Br chemical manufacturing, coal burning, and PVC usage and disposal. The cycling of Pb has clearly been anthropogenically dominated since many centuries, while natural emissions (volcanic eruptions, mineral dust) are seen to play a minor role (Hutchinson and Meema 1987, Nriagu and Pacyna 1988, Renberg et al. 2000). Global Pb emissions reached ca. 0.43 Tg year–1 in the 1970s. The deposition flux of Pb from the atmosphere to the terrestrial ecosystems in the 1980s was estimated as being 5–6 mg m–2 year–1 for populated regions of the globe and 5–45 mg m–2 year–1 in Germany (Nriagu and Pacyna 1988, Höfken et al. 1983, Grosch 1986, von Storch et al. 2002). Br and Pb species cycling in the atmosphere each span a whole range of properties (phase state, water solubility, mass median diameter). Br compounds have been identified as crucial for stratospheric and significant for tropospheric ozone chemistry (WMO/UNEP 1999, Finlayson-Pitts et al. 1990, Vogt et al. 1996). Clear ecotoxic and human toxic evidence gives rise to concern about observed Pb levels in environmental media and humans (Hutchinson and Meema 1987, Stigliani et al. 1993, Gawel et al. 1996, Bard 1999, Herman et al. 2001). The WHO guideline value and the EU limit value for lead are 0.5–1 µg m–3 and 2 µg m–3, respectively (24-h means). Critical Pb limits for terrestrial ecosystems are currently derived under the auspices of the UN-ECE Convention on Long-range Transboundary Air Pollution. Atmospheric total Pb and total Br in polluted areas had been dominated by vehicular traffic emissions as these elements were bound in fuel additives, tetraethyl and tetramethyl lead and 1,2-dibromoethane. Other significant Pb sources are lead smelters and waste incineration. Abatement measures since the 1960s in North America and since the 1970s in Europe have effectively depleted Pb levels in the atmospheric environment globally (Nriagu 1990, Huang et al. 1996), and Pb and Br levels in Europe (Harrison and Sturges 1983, EEA 1996, Schreiber 1998, LUA 2000, Hagner 2002). In Germany, emission reductions were 97% of 1975 levels by the year 1995.
ESPR – Environ Sci & Pollut Res 9 (6) 397 – 404 (2002) © ecomed publishers, D-86899 Landsberg, Germany and Ft. Worth/TX, USA • Tokyo, Japan • Mumbai, India • Seoul, Korea
397
Research Articles
Atmospheric Lead and Bromine in Germany We analysed particulate matter samples from a long-term and five short-term campaigns, 1998–2000, in Germany. These data are discussed together with the results of a longterm data set, 1972–1997 (Schreiber 1998). The aim of this study is to characterize present levels and the temporal and spatial variabilities of Pb and Br in the atmospheric environment in central Europe in the context of the question 'What are determinants of the present-day, i.e. post-abatement, environmental cycling, of Pb and Br?'. 1
Methods
The sampling sites and periods are listed in Table 1. The long-term campaign addressed weekly PM10 samples (modified Sierra Andersen sampler) at an urban site, Leipzig 1. In fact, these were quasi-weekly samples, i.e., in a one-week cycle, the sampler was turned on each seventh hour, leading to a 24 h sampling time per week. The short-term campaigns addressed mostly daily (ca. 20% of the samples were 12-h samples, or shorter) TSP samples (Digitel sampler) at several rural and urban sites in Germany, 1998–2000 (Table 1). One of these campaigns was conducted in Leipzig at a second site ('Leipzig 2'), June–July 1999. During this period, not weekly, but daily PM10 samples were synchronously collected at the first site, Leipzig 1. At the urban sites and one rural site, Merseburg, sampling took place in the height of the urban canopy, on rooftops (14–21 m above ground). At another rural site, Falkenberg, air was sampled in flat, agricultural land near the ground (h = 2.5 m) and at the Kap Arkona site air was sampled on a platform (h = 6 m) distanced some 20 m from the cliff. These samples were collected on quartz fibre filters (Munktell MK360, 150 mm) and analysed using x-ray fluorescence analysis without extraction or other pre-treatment (total element contents; Steinhoff et al. 2000). The analytical detection limits were 45 ng Pb and 3.8 ng Br (mean blank value + 3 σ). These values correspond to detectable levels of ca. 13 and 1.1 ng m–3 (weekly samples) and 10 and 0.9 ng m–3 (daily samples) of Pb and Br, respectively, depending on the sampled volume of air. In the context of the near-ground level aerosol monitoring activity of the University of Hohenheim, 24–48 h samples, 3 48-h and 1 24-h sample per week, were collected in a residential area of the city of Stuttgart. Agriculture is also significant at the site. Sampling was from 2.5 m above ground using cellulose nitrate filter membranes (Sartorius SM11301, 47 mm). The effective upper cut-off of the sampling method is 12 µm. These have been analysed by energy-dispersive x-ray fluorescence analysis (Cercasov et al. 1998, Schreiber 1998).
2 2.1
Results and Discussion Concentrations, correlations, seasonality
Concentrations of the weekly samples (Leipzig 1) were 3.2 (0.2–12.7) ng m–3 Br and 19.2 (5.3–90) ng m–3 Pb (mean (min-max)). In the daily samples from short-term campaigns at rural and urban sites, 4.4 (1.6–12.2) ng m–3 Br and 11.2 (1.3–44) ng m–3 Pb were found (Table 2). Urban Pb and Br levels are not necessarily higher than rural levels. Fig. 1 shows the Leipzig 1, 1998–2000, together with the Stuttgart, 1972–1997, time series. The latter shows trends of –(12 ± 1)% year–1 for Pb and Br (exponential regressions). The Leipzig 1 summer data follow the previous trend.
Fig. 1: Combination of Stuttgart (1972–1997) and Leipzig 1 (1998–2000) data: Observations of (a.) the yearly, wintertime and summertime means of Pb and Br particulate matter concentrations (ng m–3) and (b.) the annual mean, wintertime and summertime Br/Pb ratios. Wintertime = months December–February (DJF), summertime = months June–August (JJA).
Table 1: Sampling sites and periods
Site Leipzig 1
Type
Location
Urban
Halle-Leipzig-Bitterfeld conurbation, 51°19‘N/12°25‘E
Measurement period April 1998–January 2000
Leipzig 2 Eichstädt/Berlin
June–July 1999 Urban
at the rim of the Berlin conurbation, located some 20 km NW of the centre, 52°42‘N/13°08‘E
Stuttgart
Urban, residential
48°43'N/9°13'E
Kap Arkona
Rural, coast
Rügen Island, 54°41‘N/13°26‘E
July–August 1998 1972–1997 February–March 1998
Merseburg
Rural
at the rim of the Halle-Leipzig- Bitterfeld conurbation, 51°21‘N/11°58‘E
Falkenberg
Rural
80 km SE of Berlin, 52°13‘N/14°08‘E
398
November–December 1999 July–August 1998
ESPR – Environ Sci & Pollut Res 9 (6) 2002
Research Articles
Atmospheric Lead and Bromine in Germany
For the period 1970–1995, the Pb concentration changes over time are in agreement with the emission trends and climate in Europe (recent model study; von Storch et al. 2002). However, the wintertime levels at Stuttgart and Leipzig 1 seem to indicate a stabilization since ca. 1994. Pb and Br in the daily samples from 1998–2000 were correlated, however not strongly (r(Pb, Br) = 0.56 (all samples, cf. Table 2 for individual campaigns)). According to a onesided F-test, correlations r >0.36 are statistically significant on the 0.1%-level. Some other anthropogenically dominated heavy metal particulate concentrations were also significantly correlated in the same data set, e.g. r(Pb, Zn) = 0.49, r(Pb, As) = 0.93. In most European urban environments zinc, Zn, is dominated by vehicular traffic emissions and arsenic, As, is a tracer for coal and lignite burning (Steiger 1991, de Miguel et al. 1997, Borbély-Kiss et al. 1999, del Carmine et al. 1999). Lignite burning is significantly used in the Leipzig area. Pb and Br were higher correlated in the weekly samples, r(Pb, Br) = 0.74, and even more when based on monthly means (r(Pb, Br) = 0.93) (Fig. 2). For weekly samples, we found r(Pb, Zn) = 0.79, r(Pb, As) = 0.88. Pb and Br show a winter maximum in phase with Zn, while elements associated with the crustal composition including iron, Fe, are highest in late summer (mineral dust source) (Fig. 2). A seasonal variation of the correlation coefficient is observed showing Pb and Br not correlated during the summer campaigns (Table 2). The data analysis shows that it is clearly linked to the variabilities of the concentration levels and, hence, dominated by the local variability of atmospheric mixing and surface exchange. The Stuttgart long-term data set shows that present-day atmospheric Pb and Br are less coupled than previously: r(Pb, Br) was 0.64 in 1992–1994, but 0.93 in 1979–1981. Lower
Fig. 2: Time series (monthly means) of (a.) Pb and Br as well as (b.) 2 other heavy metal particulate matter concentrations, Fe and Zn (ng m–3)
values in recent years might, however, also be influenced by the analytical error which is more significant at low concentration levels. For the same periods in Stuttgart, r(Pb, Zn) was 0.70 in 1992–1994 and 0.77 in 1979–1981. In agreement with
Table 2: Mean atmospheric concentration, ci, particulate phase mass fractions, xi = ci / cparticulate mass, and crustal enrichment factors for i = Pb, Br in the various data sets. n = number of samples, crustal enrichment factor EF = (E/Al)aerosol/(E/Al)crustal, E = Pb, Br (Rahn 1976), r(E1,E2) = correlation coefficient between the concentrations of elements E1, E2 in particulate matter
Pb Site, time
Type of sample
n
Leipzig 1
PM10, weekly
87
Leipzig 2
Eichstädt/ Berlin
Kap Arkona
TSP, mostly daily
TSP, 21 or 12 h
Merseburg
Falkenberg
TSP, mostly daily
T (°C) (a)
Br r (Pb,Br) (d)
c (ng m–3)
x (mg/g)
Crustal EF
C (ng m–3)
x (mg/g)
Crustal EF
19.2
0.66
550
3.2
0.11
490
0.74
23
18.0 (11–30)
10.3
0.35
240
3.4
0.12
410
0.14
15
17.1 (14–20)
7.0
0.23
160
4.4
0.14
500
0.24
26 (b)
3 (0 – 8)
10.7
0.44
560 (c)
8.3
0.45
2300 (c)
0.69
22
0.3 (–1–3)
16.3
0.72
1180
6.0
0.30
2560
0.63
15
19.7 (14–31)
9.5
0.39
640
3.7
0.17
1460
0.35
(a)
12 hourly means; (b) n = 23 for Pb; Only for measurements with Al above detection limit (90 ng m–3 for most samples), i.e. marine air masses mostly excluded (d) According to a one-sided F-test, correlations r > 0.18 and r > 0.33–0.44 are statistically significant on the 5%-level for n = 87 and n = 15–26, respectively (c)
ESPR – Environ Sci & Pollut Res 9 (6) 2002
399
Research Articles
Atmospheric Lead and Bromine in Germany
Table 3: Mean summer to winter ratios of the atmospheric concentrations, c, crustal enrichment factors for Pb and Br and the Br/Pb ratio in the various data sets from 1998–2000 (data sets from intensive field campaigns combined, cf. Table 1)
Type of sample
n
csummer/cwinter (1) Pb
Br
EF(summer)/EF(winter) (a,b) Pb
Br
Br/Pb Summer
Winter (a)
(a)
Leipzig 1
PM10 weekly
Intensive field campaigns
TSP, mostly daily
87
0.38
0.36
0.43
0.40
0.15
0.15
102 (c)
0.68
0.52
0.38
0.30
0.41
0.54
(a)
defined as: summer = June–August, winter = December–February; (b) crustal enrichment factor EF = (E/Al)aerosol/(E/Al)crustal (Rahn 1976) (c) n = 99 for Pb and Br/Pb values
the long-term Pb trend, the Zn/Pb ratio is seen to increase significantly: Zn/Pb = 0.53, 1.0 and 1.9 for 1978–80, 1988– 90 (Stuttgart) and 1998–2000 (Leipzig 1), respectively (same period each, i.e. 22 months, April through January). This indicates that Pb is also increasingly decoupled from Zn. We conclude from this, and the relatively weak correlation between Br and Pb, that the previously common source, vehicular traffic emissions, is no longer predominant and another Pb source, common with As, is now significant.
From the enrichment factors (cf. Table 2) we conclude that, although Pb levels are much lower than in the 1970s and 1980s (when up to 1 µg m–3 had been measured; e.g. Harrison and Sturges 1983), this element is still anthropogenically dominated. The air mass analysis shows that Br enrichment is much higher in marine influenced air masses (EFcrustal (marine)/ EFcrustal (continental) = 1.9 and 1.3 for Br and Pb, respectively). Even at the coastal site, however, Br had significant non-sea salt sources, EFsea salt = (Br/Na)aerosol/(Br/Na)sea salt = 1.6 i.e., 38% non-sea salt Br (based on Br/Na = 0.62% in sea salt). We use non-sea salt Br as a tracer for recent continental influence and
