The changes in the atmospheric concentrations of selected trace elements at rural locations, for the period 1972-1996, have already been reported (Baker, 1997). The long-term trends in changes in air quality, with respect to the anthropogenically-derived metals As, Cr, Ni, V, Zn and Pb together with Br, were investigated. Cl and Na (of mainly marine origin) and Ce and Sc (mainly soil-derived) were also included for comparison. Data for 1997 and 1998 have now been included.

The annual mean concentrations of As, Cr, Ni, Se, V and Zn in air at Chilton for the period 1972-1998 are shown in Figure 6a, with annual mean air concentrations for Pb, Br, Cl, Na, Ce and Sc at Chilton shown in Figure 6b. Corresponding data for Styrrup and Wraymires are plotted in Figures 6c to 6f, respectively. The trends in element concentrations in air have previously been examined by linear regression against time (Cawse, 1987), (Cawse et al., 1995). In each of these time-series plots (Figures 6a to 6f) a regression against time is also shown to indicate the direction of any trend. The data were found to fit better to an exponential rather than linear trendline. Also, a linear regression trendline implies that air concentrations of an element will reach zero at some future date, which is very unlikely to be the case.

Regression analysis is often used to determine whether measurements of concentration or deposition have a significant upward or downward trend with time. There are limitations to this technique, as discussed by Lee et al. (1994), due to autocorrelation. Essentially, although the rate of decline can be assessed by regression analysis, the significance attributed to the decline may be overestimated. However, this is unlikely to be a severe limitation with annual data, for which the autocorrelation in error terms due to persistence is less important than for shorter period data. Nevertheless, the statistical significance of the trends in air concentrations of elements has not been assessed in the current work.

The downward trends in annual mean air concentrations of the mainly anthropogenically-derived metals As, Cr, Ni (at Chilton and Styrrup only), V and Zn are quite apparent (although less so for Se) at all three locations. These are also evident for Ce and Sc, which are thought to be predominantly associated with material of direct terrestrial origin (Cawse, 1987). For Na and Cl (mainly marine-derived) a decline in average air concentrations was less apparent at all three locations, except for Cl at Styrrup.
 

Figure 6a - Changes in annual mean concentrations of As, Cr, Ni, Se, V and Zn in air at Chilton (1972-1998)

Figure 6b Changes in annual mean concentrations of Pb, Br, Sc, Ce, Na and Cl in airat Chilton (1972-1998)

Figure 6c Changes in annual mean concentrations of As, Cr, Ni, Se, V and Zn in air at Styrrup (1972-1998)

Figure 6d Changes in annual mean concentrations of Pb, Br, Sc, Ce, Na and Cl in air at Styrrup (1972-1998)

Figure 6e Changes in annual mean concentrations of As, Cr, Se, V and Zn in air at Wraymires (1972-1998)

Figure 6f Changes in annual mean concentrations of Pb, Br, Sc, Ce, Na and Cl in air at Wraymires (1972-1998)

Figure 7 Percentage reductions in annual mean air concentrations of elements relative to the period 1972-1979 at Chilton, Styrrup and Wraymires

The long-term changes in average atmospheric concentrations of these elements at Chilton, Styrrup and Wraymires have been quantified by comparison of their annual mean air concentrations for the periods 1972-1979, 1980-1989 and 1990-1998. The percentage reduction in the annual mean, relative to the period 1972-1979 was calculated for each element at each sampling location. These reductions, for the 12 elements shown in Figures 6a to 6f and for 10 other elements that are listed in Table A4.1 (Appendix 2), are illustrated in Figure 7. Those elements whose quarterly concentrations (and therefore annual mean concentrations) were frequently below analytical detection limits were excluded from this summary. In Figure 7, Ni has been excluded at Wraymires since many annual mean air concentrations of Ni (particularly in more recent years) were below analytical LODs at this site. Also, for the elements Ca, Cu K and Mg, the percentage reductions in annual mean air concentrations are relative to the period 1975-1979; also data for the period 1980-1989 are excluded as annual mean concentrations were frequently below LODs during the 1980s.

Where it was possible to compare both the periods 1980-1989 and 1990-1998 with 1972-1979, the majority of the elements showed an increase in the percentage reduction in annual mean air concentration with time (Figure 7). The greatest reductions (>70%) were recorded at two or more sites for As, Br, Cr, Cu, Mn, Pb and Zn when the 1990s were compared to the 1970s. Corresponding reductions of between 60-70% for Al, Ca, Co, Cs, Fe, K, Mg, Ni, Sc and V were seen at least at two of the sites. The smallest reductions (<60%) tended to occur for Ce, Cl, Na, Sb and Se.

The association of Pb and Br in vehicle emissions is discussed in section 3.1 and the long-term reductions in Pb concentrations in air are addressed in section 3.3.1.1. Similarly, the long-term changes in air concentrations of the industrially-derived As, Cr, Ni, Se, V and Zn are discussed in relation to corresponding changes in their atmospheric emissions in sections 3.3.1.2 and 3.3.1.3.

Average annual mean air concentrations of Cu have reduced by ~70% at all three locations between the periods 1975-1979 and 1990-1998 (Figure 7). The main source of UK atmospheric emissions is coal combustion, with lesser contributions from waste incineration, iron and steel manufacture, non-ferrous metals production and the combustion of heavy fuel oil (Salway et al., 1996 and Salway, 1999). The corresponding reduction in average annual atmospheric emissions in the UK was estimated to be ~40%.

The considerable reductions in air concentrations of Sc (and to a lesser extent Ce), together with Al, Fe and Mn, that were recorded between the 1970s and the 1990s, may seem unexpected at first since these elements are thought to be of mainly crustal origin. However, they may also be emitted as fly ash from the combustion of coal and as dust from mining and other mineral-related activities (Lee et al., 1994). Consequently, the observed reductions can be explained by reductions in fly-ash emissions caused by reductions in coal and fuel oil combustion by industry and changes in domestic energy sources for space heating, e.g. from coal to gas.

The percentage reductions in average air concentrations of Na and Cl were among lowest observed (Figure 7). Also, the trends in their air concentrations (particularly for Na) were not as apparent as for the industrially-derived heavy metals at Chilton, Styrrup and Wraymires (Figures 6a to 6f). This indicates that little change in concentrations of marine-derived aerosols has occurred at any of the rural locations since the 1970s.

The decline in air concentrations of Ca, K and Mg between the 1970s and the end of 1993 at Chilton, Styrrup and Wraymires was discussed by Lee et al. (1995). The major source of these elements has commonly and long been assumed to be soil dust. However, sources of Ca have been identified to include fly-ash from coal and fuel oil combustion, with other industrial emissions from cement manufacture and quarrying of limestone (Lee et al., 1995). The large long-term reductions in air concentrations of Ca, K and Mg are similar to those for Sc (Figure 7), another element thought to be of predominantly crustal origin, but present in fly-ash. Therefore, Lee et al. (1995) concluded that the decline in air concentrations of Ca, K and Mg at these three rural locations could be, at least in part, explained by reductions in fly-ash emissions.
 

Comparison of long-term changes in air concentrations of heavy metals with changes in estimates of industrial emissions to the atmosphere

The long-term changes in the annual mean air concentrations of As, Cr, Ni, Pb, V and Zn at the three rural locations, have been compared to corresponding changes in the estimated annual atmospheric emissions of these metals by industry (Salway et al., 1996) during the period 1972-1995 (Baker, 1997). Cd and Cu were excluded since many historical measurements of Cd and Cu (particularly during the 1980s) were frequently below the limits of detection achievable by the analytical technique used at the time (section 2.3.1).

The National Atmospheric Emissions Inventory (NAEI) has revised the UK emissions inventories for these heavy metals since the report by Salway et al. (1996). For the metals of concern in the current work, estimates of annual emissions have tended to increase. The main changes include the following. Emissions from power stations have been revised. The estimates have been improved by use of emission factors provided by the Electricity Supply Industry rather than those previously used by the NAEI. The emission factor for V for fuel oil has been revised downwards. Emissions estimates have included contributions from the combustion of coke and solid smokeless fuel. This has caused increases in estimates for the industrial, commercial and domestic sectors. The estimates for waste incineration now take account of the new regulations from 1996 onwards (Salway, 1999). Emissions have been included from aircraft, coastal shipping, fishing, military aircraft and naval vessels, however these are comparatively rather small. For Pb, a large emission from the manufacture of alkyl lead is included under industrial processes.

The revised emissions data for As, Cr, Ni, Pb, Se, V and Zn (Salway, 1999), which now include annual estimates for 1997, have now been compared with the changes in their annual average air concentrations at Chilton, Styrrup and Wraymires for the period 1972 to 1997 (Figures 8, 9a, 9b and 9c). Data for As, Cr, Se and Zn from Styrrup (Figure 9b) have been plotted separately to the data from Chilton and Wraymires (Figure 9a) for clarity.

The percentage reduction in the annual mean air concentrations of these metals, for the period 1993-1997 relative to 1972-1979 (Baker, 1997), together with the corresponding reductions in their estimated annual emissions are presented in Table 6.
   

Table 6 Temporal Changes in Annual Mean Air Concentrations and Estimated UK Atmospheric Emissions of Heavy Metals at Rural Locations, (1972-1997)

Notes
(a) Not derived as many annual mean concentrations were below analytical limits of detection.
(b) Derived from data from Salway (1999)
 

In addition, the annual mean air concentrations (ng m^-3) of these metals have been plotted against annual estimates of total emissions (t) at each location (Figures 8, 10a to 10c). The significances of the respective correlation coefficients are listed in Table 6.
 

Notes

(a) Not derived as many annual mean concentrations were below analytical limits of detection.
   

Lead in air

The temporal changes in annual mean air concentrations of Pb at these three locations are plotted, along with estimated annual total atmospheric emissions for the UK, during the period 1972-1997 in Figure 8. It is apparent that the decreases in air concentrations of Pb at Chilton, Styrrup and Wraymires were consistent with corresponding decreases in estimated emissions during this period.

The percentage reductions in the annual mean air concentrations at these three sites, between the periods 1972-1979 and 1993-1997, were in the range 82% - 84% compared with a 80% reduction in estimated annual emissions (Table 6). Also, annual average air concentrations of Pb at all three locations were highly significantly correlated with estimated emissions (Table 7). UK atmospheric emissions of Pb are dominated by road transport and are better quantified than other pollutants (Salway et al., 1996 and Salway, 1999). The step change in estimated emissions (when the last reduction in the limit on the Pb content of petrol was introduced) is clearly seen between 1985 and 1986 (Figure 8). This is reflected in the plots of annual mean air concentrations, both against time and estimated annual emissions, for each of the three sites.

Similar reductions have been reported for other rural locations (i.e. East Ruston, High Muffles and Banchory) in the UK (Playford and Baker, 1998). The percentage reductions in average air concentrations of Pb were in the range 49% - 66% between 1987-1990 and 1994-1997, compared to a 55% reduction in estimated UK emissions. Again at all three locations, the changes in annual mean air concentrations were significantly correlated with changes in estimated emissions of Pb between 1987 and 1997.
 

Figure 8 Comparison of annual mean air concentrations of Pb at rural locations with the estimated annual total UK emissions to the atmosphere (1972-1997)

Arsenic, chromium, selenium and zinc in air

At Chilton, Styrrup and Wraymires, the percentage reductions in annual mean air concentrations of As (75% - 83%) and Cr (61% - 72%) since the 1970s were in excess of corresponding reductions in atmospheric emissions estimates, i.e. 57% and 51%, respectively (Table 6). However, the average air concentrations of both metals were significantly correlated with their estimated emissions between 1972 and 1997 (Table 7, Figure 10a).

At Chilton and Wraymires (Figure 9a) and at Styrrup (Figure 9b), it is apparent that the pattern in annual mean air concentrations of As follows that of the estimated emissions, particularly since the mid-1980s. For Cr, this feature is not so apparent (Figures 9a and 9b). However, reductions in annual mean air concentrations of Cr at East Ruston and High Muffles (1987-1990 to 1994-1997) were 26% and 38%, respectively(Playford and Baker, 1998) and were comparable to corresponding reductions in Cr emissions (44%). As was not measured at these locations over the whole 10 year period.

It is possible that analytical uncertainties in the historical data were greater than in the more recent data. Although the major contribution to the estimated emissions of As and Cr is coal combustion for public power generation and industrial energy requirements (Salway et al., 1996 and Salway, 1999), other sources, e.g. waste incineration, iron and steel production processes, also contribute. Consequently, estimates of emissions of these metals are less well quantified and are more uncertain than for Pb (predominantly of motor vehicular origin). These uncertainties in the data may be responsible for the discrepancies between the percentage reductions, calculated for air concentrations and for estimated emissions, of As and Cr shown in Table 6.

Annual average air concentrations of Se have decreased at Chilton (41%), Styrrup (22%) and Wraymires (45%), 1993-1997 relative to 1972-1979 and are comparable to the corresponding reduction (35%) in estimated UK atmospheric emissions (Table 6). However at all three sites, the average air concentrations of Se were less significantly correlated with estimated annual emissions between 1972 and 1997, than As and Cr (Table 7, Figure 10b). Like As and Cr, the major UK emission source of Se is the combustion of coal (Salway et al., 1996 and Salway, 1999).

Correlations between Zn concentrations in air at Chilton, Styrrup and Wraymires and its estimated emissions to the atmosphere in the UK are also less significantly correlated than is the case for the other metals (Table 7, Figure 10b). Further, the reduction in estimated emissions was only 24% between the 1970s and 1993-1997, while corresponding reductions in annual mean air concentrations of Zn were consistent (~66%) at the three sampling sites (Table 6). The contribution of coal combustion to the total emissions is much less for Zn, compared to the other metals, with Zn contributions from other sources, e.g. iron and steel, non-ferrous metals production, waste incineration and road transport (i.e. tyre wear) being more predominant (Salway et al., 1996 and Salway, 1999). It may be that emissions from such sources, while significant to the UK as a whole, were not as important as local sources at these sites during the 1970s.

Figure 9a Changes in annual mean air concentrations of As, Cr, Se and Zn at Chilton and Wraymires compared with changes in their estimated annual total UK emissions to the atmosphere (1972-1997)

Figure 9b Changes in annual mean air concentrations of As, Cr, Se and Zn at Styrrup compared with changes in their estimated annual total UK emissions to the atmosphere (1972-1997)

Nickel and vanadium in air

The temporal changes in annual mean air concentrations of Ni at Chilton and Styrrup are plotted along with the estimated annual total atmospheric emissions of Ni for the UK, during the period 1972-1997 in Figure 9c. Many of the more recent annual mean concentrations of Ni at Wraymires have been below analytical limits of detection. Corresponding data for V at all three locations is also plotted in Figure 9c.

The percentage reductions in the annual mean air concentrations of Ni at Chilton and Styrrup, between the periods 1972-1979 and 1993-1997, were 66% and 59%, respectively (Table 6). These reductions were very consistent with the 65% reduction in annual emissions that has been estimated to have occurred since the 1970s (Table 6). Also, annual average air concentrations of Ni at these locations were highly significantly correlated with estimated emissions (Table 7, Figure 10c).

Similar observations have been made at other rural locations in the UK. Between the periods 1987-1990 and 1994-1997, the reductions in average air concentrations of Ni at East Ruston (22%) and High Muffles (33%) (Playford and Baker, 1998) were comparable to the corresponding estimated decrease in UK emissions (27%).

For V, reductions in annual average air concentrations of V at all three rural locations have followed the changes in estimated atmospheric emissions in the UK since the 1970s (Figure 9c). The percentage reductions in the annual mean air concentrations of V were in the range 69% - 78%, compared with a 66% reduction in estimated annual emissions, 1993-1997 relative to 1972-1979 (Table 6). At all three sites, annual mean air concentrations of V were significantly correlated with estimated annual emissions between 1972 and 1997 (Table 7, Figure 10c).

The combustion of heavy fuel oils is the major source of atmospheric emissions of Ni and V, with large contributions also from coal combustion (Salway et al., 1996 and Salway, 1998).

The decline in atmospheric emissions of As, Cr, Ni, Se and V in the UK since the 1970s has been reflected in the decreases observed in their average air concentrations at rural locations. Changes in methods of energy production and in the nature of the manufacturing sector, together with tighter limits on emissions from prescribed processes are the major reasons. A decline in the use of coal for power generation, in favour of other fuels, e.g. increased use of natural gas since 1992, has occurred. For Ni and V, combustion of heavy fuel oils is the major source of atmospheric emissions. Heavy fuel oil was replaced by orimulsion in some power stations in 1995 (Salway et al., 1996); whilst the Ni and V content of this fuel is greater, the impact of larger emissions would be unlikely to be significant at these monitoring sites.

Figure 10a Comparison of annual mean air concentrations of As and Cr at rural locations with the estimated annual total UK emissions to the atmosphere (1972-1997)

Figure 10b Comparison of annual mean air concentrations of Se and Zn at rural locations with the estimated annual total UK emissions to the atmosphere (1972-1997)