3.5 Deposition of elements, nitrate and sulphate

From the concentration of an element and the rainfall recorded at each location, the deposition of that element (in a soluble form) can be derived. Annual deposition estimates for some elements between 1994 and 1998 are listed in Table 9 (section 3.4)
 
 

Seasonal variability in the deposition of elements

The seasonal variability of the deposition of some elements has been investigated for the period 1992-1996 (Baker, 1997). The quarterly deposition data for 1997 and 1998 has now been included in this examination. For a given element, the ratio of quarterly mean to annual mean deposition was determined and the average ratio for each quarter during 1992-1998 calculated. This is shown for Cu, Pb and Zn, together with As, Cd and Co (1994-1998 only) at Chilton, Styrrup and Wraymires in Figure 12a. The seasonal variation in rainfall is also shown. The corresponding ratios for Al, Ca, Fe, K, Mg, Mn and Na are shown in Figure 12b.

There is a larger seasonal variation in rainfall in the west of the UK. This is reflected in Figure 12a where a larger increase in ‘winter’ rainfall relative to ‘summer’ was seen at Wraymires, than at Chilton and Styrrup. Consequently, more elements have shown a stronger seasonal variation in bulk deposition at Wraymires than at the other two sites. All 13 elements that were included in this investigation showed ‘winter’ increases in deposition at this location (Figures 12a and 12b).

Air concentrations of many heavy metals from anthropogenic sources, e.g. As, Cr, Cu, Ni and V, have exhibited seasonal variations with ‘winter’ increases at all three sampling locations. The ‘winter’ increases are attributed to the increased combustion of fossil fuels during these periods, together with several meteorological factors (section 3.1.1).

However, when the seasonal variations in the deposition of these metals were investigated at Chilton and Styrrup (i.e. the sites where the ‘winter’ to ‘summer’ differences in rainfall are small), less consistent patterns in seasonal differences were found.

Of the heavy metals, the deposition of Cd, Co, Cu and Zn was elevated in the winter quarters at Chilton, with the lowest average quarterly mean to annual mean ratios for As and Pb recorded in the third quarter July to September. For these metals such seasonal variations were not as apparent at Styrrup (Figure 12a).

Generally, the measured concentrations of many elements in rainwater are more variable than in corresponding air particulate samples. The dry deposition of elements (which is deposited onto the collection funnel during periods of low rainfall) will influence the measured concentrations. It is assumed that any dry deposits will be washed into the collecting bottle by subsequent rainfall. However, it is possible that the ‘subsequent’ rainfall could occur in a different sampling period. This sampling artefact could mask or distort any seasonal variations in the deposition of some elements, e.g. heavy metals from anthropogenic sources. Also, no consistent seasonal pattern in precipitation is apparent at Chilton and Styrrup (Figure 12a). These factors are likely to contribute to the reason why some elements do not exhibit a seasonal variability in their deposition when they do in their measured air concentrations.

The strongest seasonal variations in total deposition were seen for Mg and Na at all locations (Figure 12b). This is indicative of increased inputs of marine-derived aerosol during the ‘winter’ quarters.

For the mainly soil-derived elements (although these will have anthropogenic sources), the deposition of Ca, Fe, K and Mn appeared to be elevated during the ‘summer’ quarters at Styrrup. However, this was only seen for Ca at Chilton (Figure 12b). This feature could be attributed to the increased likelihood of the resuspension of soil dust during the ‘summer’ quarters. It was not evident that rainfall tended to be lower during the ‘summer’ at these sites (Figure 12a). However, warmer temperatures, increases in the length of drier periods and increased agricultural activities are more likely during April to September. These factors will contribute to the increased likelihood of soil dust being resuspended.

At Chilton, seasonal differences were not apparent for Al (also at Styrrup), K and Mn, although Fe showed a tendency for increased deposition in the ‘winter’ quarters (Figure 12b). At Wraymires, the deposition of Al, Ca, Fe, K and Mn were clearly influenced by the more strongly seasonal rainfall patterns (Figure 12b).
 
 



Figure 12a

Seasonal variability in the depositions of heavy metals in rainwater

at Chilton, Styrrup and Wraymires (1992-1998)
 

Figure 12b

Seasonal variability in concentrations of Al, Ca, Fe, K, Mg, Mn and Na in rainwater

at Chilton, Styrrup and Wraymires (1992-1998)


 












Long-term changes in the deposition of elements

The long-term trends in the annual bulk depositions of Pb, V, Zn and Na at the three rural locations were previously investigated for the period 1972-1996 (Baker, 1997). The time-series plots produced then have now been updated to include the years 1997 and 1998 (Figures 13a to 13c). Plots for Cu and Ni (for the Chilton and Styrrup sites only) have also been included. In recent years, many quarterly mean concentrations of Ni in rainwater at Wraymires were below analytical limits of detection (section 3.4). Other anthropogenically-derived metals, e.g. As and Cr could not be included since many of the historical measurements were below analytical LODs.

In the time-series plots a regression analysis (as for the air concentrations, section 3.3) has been used to indicate the direction of the trend in annual deposition since the 1970s. Downward trends in the deposition of Cu, Ni, Pb, V and Zn were apparent over the 27-year period. However, for Na the downward trend was less marked with even an upward trend shown at Wraymires (Figures 13a to 13c).

The long-term changes in average annual bulk depositions of these five metals at Chilton, Styrrup and Wraymires have been quantified by comparison of their average annual means 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 data, together with those for other elements, nitrate and sulphate are listed in Table A4.2, Appendix 2 and are illustrated in Figure 14 (annotations (a), (b) and (c) are explained in footnotes to Table A4.2). To complete the dataset, the corresponding changes in the precipitation-weighted annual mean concentrations are also listed in Table A4.3, Appendix 2. Those elements whose quarterly concentrations (and therefore annual mean concentrations) were frequently below analytical detection limits were excluded from this summary.
 
 


Figure 13a

Changes in annual mean depositions of Cu, Ni, Pb, V, Zn and Na

at Chilton (1972-1998)
 

Figure 13b

Changes in annual mean depositions of Cu, Ni, Pb, V, Zn and Na

at Styrrup (1972-1998)
 

Figure 13c

Changes in annual mean depositions of Cu, Ni, Pb, V, Zn and Na

at Wraymires (1972-1998)
 
 

Figure 14

Percentage reductions in annual mean deposition of Cu, Ni, Pb, V, Zn and Na

relative to the period 1972-1979 at Chilton, Styrrup and Wraymires


 










All the elements, except Na (and Mg at Wraymires), showed an increase in the percentage reduction in annual mean deposition with time (Figure 14). As was noted for air concentrations of Na (section 3.3), the changes in the deposition of Na were not as clearly seen as those for the metals of anthropogenic origin. This reflects the relatively unchanging contribution of marine-derived aerosols at all three locations. Further, the deposition of Mg at Wraymires has hardly changed since the 1970s (Figure 14). Inputs of Mg at this site are primarily marine-derived; this is reflected in the strongest seasonal variations in deposition shown by Na and Mg at Wraymires (section 3.5.1).

It is apparent that for each element investigated, the long-term reductions in their deposition were less at Wraymires than at the other two sites (Figure 14). The annual rainfall at Wraymires is about 3 to 4 times greater than that at Chilton and Styrrup and also experiences larger variations in rainfall (Table A4.2, Appendix 2). These influences appear to have reduced the magnitude of the long-term changes in the deposition of elements at this site, relative to the other two.

For the predominantly industrially-derived metals and Pb, large reductions (>60%) in their deposition have occurred relative to the 1970s at Chilton, Styrrup and Wraymires (Figure 14). The long-term reductions in the deposition of Cu, Ni, Pb, V and Zn at these rural locations are discussed, in relation to corresponding changes in atmospheric emissions in the UK in section 3.5.3.

Considerable reductions in air concentrations of Al, Fe and Mn have occurred between the 1970s and the 1990s, with >60% reductions in annual means being measured at all three sites during this time (section 3.3). At Chilton and Styrrup, similar reductions in their annual mean depositions have occurred (1990-1998 relative to 1972-1979) (Figure 14). Again, the long-term decreases in emissions of fly ash is the likely explanation (see section 3.3).
 

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

The long-term changes in the estimated annual emissions of Pb, Ni (Chilton and Styrrup only), V, Cu and Zn to the atmosphere (Salway, 1999) are compared to the changes in their annual depositions at Chilton, Styrrup and Wraymires between 1972 and 1997 in Figures 15a to 15e, respectively. Also, the annual mean depositions (m g m-2) of these metals have been plotted against annual estimates of total emissions (t) at each location (Figures 15a to 15e).

The percentage reduction in the annual mean depositions of these metals, for the period 1993-1997 relative to 1972-1979 together with the corresponding reductions in their estimated annual emissions are presented in Table 12. For each metal at each location, these two parameters have also been correlated. The significances of the respective correlation coefficients are listed in Table 13.
 
 

Table 12
Temporal Changes in Annual Mean Annual Depositions and

Estimated UK Atmospheric Emissions of Heavy Metals

at Rural Locations, (1972-1997)
Annual Mean Deposition
(m g m-2)
Cu
Ni
Pb
V
Zn
 
CHILTON
1972-1979
9.47E+03
5.17E+03
1.86E+04
3.44E+03
4.67E+04
1993-1997
4.32E+03
8.26E+02
1.59E+03
6.01E+02
1.51E+04
% reduction
54%
84%
91%
83%
68%
 
STYRRUP
1972-1979
1.34E+04
3.34E+03
2.60E+04
3.27E+03
8.19E+04
1993-1997
3.96E+03
9.92E+02
2.49E+03
6.84E+02
2.67E+04
% reduction
70%
70%
90%
79%
67%
 
WRAYMIRES
1972-1979
1.43E+04
7.17E+03
1.49E+04
3.70E+03
4.12E+04
1993-1997
1.08E+04
(a)
4.36E+03
1.42E+03
1.58E+04
% reduction
25%
-
71%
61%
62%
           
Annual Mean Estimated Total UK Atmospheric Emissions (t) (b)
 
Cu
Ni
Pb
V
Zn
1972-1979
147
1023
8573
3373
1647
1993-1997
76
360
1701
1149
1255
% reduction
48%
65%
80%
66%
24%

Notes

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

(b) Derived from data from Salway (1999)
 
 


Figure 15a

Comparison of annual mean depositions of Pb at Chilton, Styrrup and Wraymires

with the estimated annual total UK emissions to the atmosphere (1972-1997)
 
 

Figure 15b

Comparison of annual mean depositions of Ni at Chilton and Styrrup

with the estimated annual total UK emissions to the atmosphere (1972-1997)
 
 
 
 
 

Figure 15c

Comparison of annual mean depositions of V at Chilton, Styrrup and Wraymires

with the estimated annual total UK emissions to the atmosphere (1972-1997)
 
 
 
 

Figure 15d

Comparison of annual mean depositions of Cu at Chilton, Styrrup and Wraymires

with the estimated annual total UK emissions to the atmosphere (1972-1997)
 
 
 
 

Figure 15e

Comparison of annual mean depositions of Zn at Chilton, Styrrup and Wraymires

with the estimated annual total UK emissions to the atmosphere (1972-1997)
 
 
 

Table 13

Correlations between Annual Mean Depositions at Rural Locations and Estimated Annual Total Atmospheric Emissions of Heavy Metals in the UK

(1972-1997)

   
Chilton
Styrrup
Wraymires
         
Cu r
0.407
0.670
0.320
  p
<0.05
<0.001
not significant
Ni r
0.580
0.262
(a)
  p
<0.01
not significant
-
Pb r
0.812
0.740
0.519
  p
<0.001
<0.001
<0.01
V r
0.926
0.907
0.766
  p
<0.001
<0.001
<0.001
Zn r
0.665
0.333
0.413
  p
<0.001
<0.1
<0.05

Notes

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

Deposition of lead

The reductions in the annual mean depositions of Pb between 1972-1979 and 1993-1997 ranged from 71% at Styrrup to ~90% at both Chilton and Styrrup, which was comparable to the corresponding reduction in estimated annual emissions (80%) (Table 12). Further, the temporal changes in the annual deposition of Pb at Chilton and Styrrup were highly significantly correlated (at the 99.9% level) with the changes in emissions, although the significance was less (at the 99% level) at Wraymires (Table 13).

A sharp reduction in annual mean air concentrations of Pb, which corresponded with the reduction in the Pb content of petrol between 1985 and 1986 was clearly seen at all three sites (Figure 8) (section 3.3.1.1). No such discernible step change in the annual deposition of Pb was apparent at any of the sites (Figure 15a). However from the end of the 1980s to 1997, Pb deposition has clearly fallen below the levels that were recorded pre-1986 at all three sites. This is attributed to the increase in the use of unleaded petrol since 1988. Sales of unleaded petrol accounted for 47% of the annual total for 1992 (Stevenson, 1994). These have progressively increased to the current level of over 70% of the market (DETR, 1998). Further, in the plots of Pb deposition against atmospheric emissions the points clearly fall into two groups, pre- and post-1986. At all three locations, the majority of the lower annual mean deposition points fall within the post-1986 group (Figure 15a). The influence of vehicle emissions on Pb in bulk deposition at rural locations is shown by these data.

Over the period 1972-1998, the annual deposition of Pb was highly significantly correlated with annual mean air concentrations at all three sites. The correlations were highly significant at Chilton (r = 0.841, p < 0.001) and Styrrup (r = 0.766, p < 0.001), but less so at Wraymires (r = 0.412, p < 0.05).
 
 

Deposition of nickel

Reductions in the depositions of Ni at both Chilton and Styrrup exceeded the corresponding reductions in emissions. Compared to that for the 1970s, the annual mean for Ni during the period 1993-1997 was reduced by 84% at Chilton and 70% at Styrrup. The reduction in emissions was estimated to be 65% (Table 12). Annual mean air concentrations of Ni were highly significantly correlated (at the 99.9% level) with estimated annual emissions between 1972 and 1997 (section 3.3.1.3). However, the correlations between annual deposits and annual emissions appears to be less significant (at the 99% level) at Chilton and not significant at Styrrup (Table 13). From the plots of annual deposition against annual emissions, it is clear that one data point at Chilton and two data points at Styrrup appear anomalous and but for these both correlations would be highly significant (Figure 15b). During 1987, unusually high quarterly concentrations of Ni in rainwater were recorded at both sites and these have given rise to the anomalously high depositions seen in Figure 15b. Exclusion of these data points increases the significance of the correlations at both sites. At Chilton, the correlation coefficient, r, becomes 0.625 (p < 0.001) while that for Styrrup increases to 0.404 (p < 0.05), compared to the values listed in Table 13.

The other apparently anomalous high annual deposition value for Ni that was observed at Styrrup corresponds to the year 1984 (Figure 15b). However, during this year the estimated emissions for Ni also increased (Figure 15b). This increase coincided with a period of industrial action that was experienced in the coal mining industry. During this time, more heavy fuel oils (the major source of Ni emissions in the UK, section 3.3.1.3) were consumed for public power generation and by industry. The relative proximity of the Styrrup site to the major power stations in Yorkshire and the East Midlands perhaps explains the peak in Ni deposition recorded at this location in 1984.
 
 

Deposition of vanadium

The reductions in estimated atmospheric emissions of V over the last 26 years are reflected in the corresponding decreases in V deposition at all three rural sites (Figure 15c). Annual mean emissions fell by 66% between 1972-1979 and 1993-1997, while the reductions in annual mean depositions were in the range 61% to 83% (Table 12). In addition, annual emissions estimates were highly significantly correlated (at the 99.9% level) with annual depositions at Chilton, Styrrup and Wraymires between 1972 and 1997 (Table 13). The combustion of heavy fuel oils is the major contributor to atmospheric V emissions. Further, the annual deposition of V was highly significantly correlated with annual mean concentrations between 1972 and 1998 at Chilton (r = 0.844, p < 0.001) and Styrrup (r = 0.854, p < 0.001); at Wraymires the significance was less (r = 0.419, P < 0.05).
 
 

Deposition of copper

The reduction in the mean estimated annual atmospheric emissions of Cu between the periods 1972-1979 and 1993-1997 was 48%. This reduction, unlike some of the other heavy metals, was not as well reflected by the corresponding changes in annual depositions at the three rural locations (range 25% to 70%); although the reduction at Chilton was comparable (54%) (Table 12). Large variations in the estimated annual deposits are apparent, particularly at Chilton and Wraymires (Figure 15d). However, annual Cu emissions were well correlated (at the 99.9% level) with the annual Cu deposits recorded at Styrrup between 1972 and 1997 (Table 13, Figure 15d). At Chilton the corresponding correlation was much less significant with no significance apparent at Wraymires (Table 13).

The major source of atmospheric emissions of Cu is coal combustion, with other notable contributions from waste incineration and metals (ferrous and non-ferrous) production (Salway et al., 1996 and Salway, 1999). Therefore it is possible that the better correlation between emissions and deposition observed at Styrrup is attributable to its relative nearness to the major power stations (section 3.5.3.2) and the metals industries, e.g. iron and steel in South Yorkshire.
 
 

Deposition of zinc

Like the Zn concentrations in air (section 3.3.1.2), the reductions in the annual mean depositions between 1972-1979 and 1993-1997 were consistent at all three locations. These were in the range 62% to 68% (Table 12). Estimated annual emissions of Zn have fallen by only 24% over this period of time. Further, annual mean air concentrations and depositions of Zn were significantly correlated at Chilton and Wraymires (p < 0.01) and especially at Styrrup (p < 0.001). With the exception of Chilton, the annual depositions of Zn were less significantly correlated with emissions between 1972 and 1997, than for Pb, Ni and V (Table 13). The proportion of emissions from coal combustion to the total Zn emissions is estimated to be much less for Zn, with larger contributions from other sources (section 3.3.1.2).
 
 

Long-term changes in the deposition of nitrate and sulphate

Although the operational protocols are not as appropriate for nitrate and sulphate measurement as those used in the Acid Deposition Monitoring network (see section 3.4.2), it is nevertheless of use to examine the long-term changes in nitrate and sulphate deposition at the three rural locations that comprise the Rural Trace Elements network. However, it must be noted that some changes in the analytical methods used in their determination have taken place over the years. An ultra-violet spectrometric method was used to determine nitrate concentrations until the late 1980s, while a colorimetric method was used to measure sulphate concentrations up until this time (Cawse et al.,1995). Ion-chromatography has been used to determine both since (section 2.3.2).

Time-series plots of the annual bulk depositions of nitrate and non-marine sulphate at Chilton, Styrrup and Wraymires between 1973 and 1998 are shown in Figure 16. A regression analysis has been used to indicate the direction in the trend in annual deposition during this period.

Peaks in the annual deposition of nitrate were recorded at Chilton and Styrrup in 1983 and at Wraymires in 1982. These were due to anomalously high concentrations that were measured in some quarterly samples, which biased the annual mean. Since the early 1980s a gradual reduction in the average deposition of nitrate has occurred at Styrrup, but was not so apparent at Chilton and Wraymires (Figure 16).

The magnitude of the changes in the deposition of nitrate was similar at all three locations when the period 1990-1998 was compared with 1973-1979. The percentage reduction in annual mean depositions over this period were in the range 35-39% and were of similar magnitude to those for Na (Table A4.2, Appendix 2). Relatively little long-term change, compared to those recorded for the heavy metals, in the annual deposition of nitrate has been observed at these rural locations. Between the mid-1980s and 1995 for the UK as a whole, the Acid Deposition Monitoring network has detected no significant overall change in wet nitrate deposition (Vincent et al., 1996).
 
 




Figure 16

Changes in annual mean depositions of nitrate and non-marine sulphate

at Chilton, Styrrup and Wraymires (1973-1998)


 














The non-marine derived component of the sulphate deposition has been calculated by subtracting the estimated marine-derived sulphate from the total deposited sulphate. The marine-derived component was estimated from the Na/SO42- ratio in bulk seawater of 0.23 (Suess and Urey, 1956) and the assumption that all deposited Na is of marine origin.

Although the annual deposition data were quite variable, relatively little overall changes in the annual wet deposits of non-marine sulphate were apparent at Chilton, Styrrup and Wraymires between the 1970s and 1980s (Figure 16). The percentage reductions in annual mean deposits (1980-1989 relative to 1973-1979) were only 5% and 15% at Wraymires and Chilton, respectively. A larger reduction (52%) occurred at Styrrup over this period but this was mainly due to the high annual wet sulphate deposits that were measured during 1973 and 1974 (Figure 16).

However, a step-change in the deposition of non-marine sulphate occurred between 1990 and 1991 at all three locations (Figure 16). Therefore the percentage reductions in the annual mean depositions, when the 1990s were compared with the 1970s, increased and were 44%, 49% and 69% at Wraymires, Chilton and Styrrup, respectively.

The changes in the estimated UK total annual emissions of sulphur dioxide (SO2) for the period 1973 to 1997 (Salway, 1999) may be compared to corresponding changes in the wet deposition of non-marine sulphate at Chilton, Styrrup and Wraymires. The percentage reduction in estimated annual mean emissions between the periods 1973-1979 and 1990-1997 was 48%. The corresponding values for the annual deposition of non-marine sulphate were 47% and 48% at Wraymires and Chilton, respectively; at Styrrup this was 68%.

Further, the largest contribution to SO2 emissions in the UK is from power stations, which was estimated to account for 65% of the total in 1994 (Salway et al., 1996). Since 1990, the decline in estimated SO2 emissions has accelerated because of the increase in the proportion of electricity generated by combined cycle gas turbine (CCGT) power stations, which have negligible SO2 emissions. Also, increases in the use of abatement technology in the 1990s, e.g. flue gas desulphurisation plant at coal-fired power stations, have also had a significant effect on SO2 emissions (Salway et al., 1996).

When the annual means for the deposition of non-marine sulphate for the periods 1980-1989 and 1990-1997 are compared, the percentage reductions at the three rural locations were in the range 33-44%, compare with a 29% reduction in estimated UK total annual mean emissions.
 
 

3.5 Washout factors

the efficiency of removal of an element by rain is assessed by the ‘washout factor’ (W) defined (Chamberlain, 1960) as follows:
 
 



W = Concentration of an element in rainwater (m g kg-1)

Concentration of an element in air (m g kg-1)


 










Cawse (1974 and 1987) concluded that it was preferable to derive washout factors using rainwater concentrations of elements in the soluble fraction only, i.e. Wsol; values of W derived from the total (wet + dry) concentrations of elements were found to be excessive, due to a relatively large contribution from dry deposition.

High values of Wsol (e.g. >1000) are associated with pollutants present at the rain-forming altitude (~3 km) that are readily scavenged by in-cloud processes. Low values (e.g.<500) are typical of emission sources close to ground-level.

The ‘washout factor’ is also known as the ‘scavenging ratio’. This ratio is most meaningful when applied to event data, however Campbell (1992) calculated ‘scavenging ratios’ based on quarterly measurements of Pb in air and rainwater.

The values of Wsol derived from the annual mean concentrations of elements in air and rainwater for 1996, 1997 and 1998 are listed in Table 14. The range of elements is limited owing to many of the concentration data being below analytical limits of detection.
 
 



Table 14

Annual Mean Washout Factors for Elements in a Soluble Form

at Rural Locations (1996-1998)

Element
Annual Mean Washout Factor (Wsol)
 
Chilton
Styrrup
Wraymires
 
1996
1997
1998
1996
1997
1998
1996
1997
1998
                   
Al
-
60
-
330
240
-
-
170
-
As
420
140
370
230
80
230
360
110
640
Ca
4960
6070
10700
3290
1780
1950
1470
1260
1610
Cd
610
370
-
1310
1350
-
-
-
-
Co
880
290
790
1020
280
1000
640
240
1170
Cr
-
60
190
-
20
150
-
30
820
Cu
2240
790
3940
2110
320
2850
4390
1180
26340
Fe
20
-
30
40
40
50
60
40
140
K
1450
1120
580
1170
590
620
-
550
460
Mg
3080
3350
1440
3120
1670
1400
3880
3540
2270
Mn
1410
560
1380
1300
310
1030
800
420
1230
Na
5210
4380
2000
4310
1620
1880
5620
2690
2950
Ni
-
320
1760
-
170
1070
-
-
-
Pb
50
20
220
150
50
320
340
90
530
Sb
320
110
360
-
-
740
1100
970
1020
Se
550
170
240
120
50
160
-
110
530
V
-
110
270
-
130
450
-
160
480
Zn
1030
870
470
660
260
590
1120
440
 

 
 

A general classification of Wsol values from measurements of annual mean concentrations of elements in air and rainwater at Chilton, Styrrup and Wraymires during the period 1996-1998 shows:

Low Wsol (<500): Al, As, Cr, Fe, Pb, Se and V.

Intermediate Wsol (500-1000) Cd, Co, K, Ni, Sb and Zn.

High Wsol (>1000) Ca, Cu, Mg, Mn and Na.

Some variability in the annual mean washout factors, both between years at a given site and between the individual sites is apparent. For example, Wsol values for Ca at Chilton are higher than at Styrrup and at Wraymires (Table 14). These variations are expected to arise from meteorological differences, e.g. annual rainfall, and sources and chemical forms of emissions to the atmosphere, e.g. particle size distributions of elements, hygroscopicity of the particles, solubility of the elements in rainwater and of the dry deposited particulate material.

The annual mean Wsol values, for many of the elements listed in Table 14, for the periods 1972-1981 and 1982-1991 (Cawse et al., 1995) and 1992-1996 at Chilton, Styrrup and Wraymires were compared in Baker (1997). The Wsol values for 1997 and 1998 were of the same order as the mean value for 1992-1996.

Allowing for the variability in washout factors, as discussed above, Wsol values for the majority of each of the elements examined were of the same order in each of the three periods at a given location. However, increases or decreases in the Wsol values over the three periods sometimes altered classification of an element (Baker, 1997).

The uncertainties associated with the sampling and measurement of the particulate material should also be considered, in addition to those factors already mentioned, in examining the variability of the Wsol values.

The relative collection efficiencies of the air and rainwater samplers should be acknowledged (section 2.2). For example, the larger particle size fraction of >10 m m diameter will dominate in rainwater. Consequently, an increase in a source of airborne particulate material that is mainly associated with larger particles will increase the value of Wsol. Conversely, an increase in an emission source of a combustion-derived aerosol, with particles <2 m m diameter, will decrease the Wsol value.

Changes in the analytical techniques used to determine concentrations of the elements in air and rainwater have occurred since the 1970s (sections 2.3.1 and 2.3.2). Therefore, accompanying changes in analytical uncertainties will also contribute to the variations in Wsol values.