- Email: [email protected]
The influence of altitude on rainfall composition at great dun fell
Atmospheric Environment Vol. 22, No, 7, pp. 1355-1362, 1988. Printed in Great Britain.
0004-6981/88 $3.00+0.00 Pergamon Press plc
THE INFLUENCE OF ALTITUDE ON RAINFALL COMPOSITION AT GREAT DUN FELL
D. FOWLER,J. N. CAPE, I. D. LEITH Institute of Terrestrial Ecolog2~,Bush Estate, Penicuik, Midlothian EH26 0QB, U.K.
T. W. CHOULARTON,M. J. GAY and A. JONES Department of Pure and Applied Physics, University of Manchester Institute of Science and Technology, P.O. Box 88, Manchester M60 1QD, U.K. (First received 19 May 1987 and received for publication J 1 January 1988)
Abstract--The influence of altitude on rainfall composition and wet deposition has been investigated at Great Dun Fell in northern England. Measurements of rainfall at eight altitudes between 250 and 850 m on the western slopes of the hill show marked changes in both amount and composition when orographic cloud is present and a west or southwest wind is blowing.On average (20 precipitation events from autumn 1984 and spring 1985), the rainfall at the summit (847 m) exceeded that at 250 m by a factor of 2, and concentrations of SO2-, NO3, CI-, NH,~and H ÷ were larger at the summit by factors of between 2.2 and 3.1. Thus, wet deposition at the summit was larger than at 250 m by about a factor of 5. The concentrations of major ions in orographic cloud at 847 m exceeded concentrations in rain by a factor of between 2.0 and 3.9. A large change occurred in the concentrations of major ions in orographic cloud with altitude, decreasing with increasing altitude from cloud base. Such changes could generally be explained by the expected dilution as cloud liquid water content increased adiabatically. When the wind was from the east or with blocked flow no increase in concentration or rainfall amount was observed. Key word index: Wet-deposition, orographic enhancement, cloud chemistry.
INTRODUCTION
Contemporary interest in the chemical composition of rain was largely stimulated by Scandinavian studies in the 1950s and 1960s which linked the acidification of rainfall with the long-range transport of S compounds in the atmosphere over Europe (Sweden, 1972). Research during the last 15 yr into the meteorological and chemical processes involved in the atmospheric transport and chemical conversion of gaseous pollutants has identified the major processes controlling the long-range transport of S compounds in the atmosphere over Europe (Royal Society, 1984). Monitoring studies have improved the quality and the quantity of data to describe the spatial patterns of wet deposition (BAPMoN, 1986). Within northwest Europe and Scandinavia the areas receiving the largest amounts of wet-deposited acidity are the highrainfall regions of southwest Norway and western and northern Britain (Semb and Dovland, 1986; Barrett et al., 1983). Precipitation chemistry networks in Europe and North America have been designed to obtain regionally representative measurements of the average composition of precipitation. To obtain the wet deposition fields from these data, the much more extensive data on precipitation amount from meteorological and hydrological precipitation samplers have been combined with the data from precipitation chemistry networks. In general, annual rainfall
weighted average concentrations of SO ] - , H +, NO 3 and NH2 have been shown to vary much less than precipitation amounts. There are, however, areas of uncertainty, in particular the tops of hills, where rainfall amounts and wet deposition are larger and where for practical reasons few measurements are made. For estimates of wet deposition on the high ground in northwest Britain, Scandinavia and continental Europe, for which representative sites are not available, the composition of precipitation has been assumed to remain constant with altitude. As the areas experiencing the largest wet deposition are the upland districts which have been a focus of interest for freshwater acidification in Britain and Scandinavia, a measurement study of the variation in the composition of precipitation with altitude is of considerable importance. Recent measurements of the ionic composition of rainfall at different elevations in the Alps of France, in Austria and in western North America show generally smaller concentrations of the major ions, SO 2-, NO3, and NH~ and H +, at the higher elevations ( > 3000 m) (Duncan et al., 1986; Puxbaum, 1988). For smaller mountains (500-1000 m) in northwest Europe, orographic cloud frequently shrouds the summits and contains large concentrations of all major ions: SO 2-, NO 3, NH2, H +, CI- and Na +. This low-level cloud is readily scavenged by rainfall from higher levels, enhancing rainfall by the seeder-feeder process (Bergeron, 1965). As concentrations of the major ions
1355
1356
D. FOWLERet al.
in the feeder (lower) cloud are generally large compared with typical concentrations in rainfall, this process may lead to wet deposition on high ground exceeding the values predicted from a combination of regional rainfall chemistry measured at low-level sites and rainfall amounts. In the study reported here, the composition of precipitation and cloud water at a number of sites between 150 m a.s.l, and a hill summit 847 m a,s.1, in the northwest of England has been measured. The area chosen for this study includes a ridge lying northwest-southeast with the tops of individual hills at about 900 m, where annual rainfall amounts are in excess of 2000 mm, and a valley to the southwest of this line of hills at 150 m a.s.1, with an annual rainfall of approx. 900 mm. The summit of the specific hill selected for measurement (Great Dun Fell, GDF) at 847 m a.s.l, is in cloud for some part of 250 days every year, and the road between the valley site to the southwest and the summit provides access to a range of sites for rain and cloud water collectors. With winds from the southwest the seeder-feeder mechanism has been shown to be responsible for a substantial enhancement of rainfall amounts at the higher levels on these hills (Carruthers and Choularton, 1983). The initial objectives of this study were to obtain rainfall and cloud water composition measurements at a range of sites along the southwest slopes of G D F to show for individual rainfall events in southwesterly airflow: (a) whether the concentrations of major ions in rainfall are constant with altitude; (b) whether the presence of orographic cloud on G D F significantly influences the composition of rainfall at the higher altitudes; and (c) whether, in the long term, annual wet deposition of the major ions on the hilltops is larger than would have been predicted assuming constant rainfall composition with altitude.
METHODS
Rainfall was sampled using 20-cm diameter Pyrex glass funnels mounted 1.5 m above ground at each of eight locations between 250 m a.s.l, in the Eden valley and 847 m at G D F summit. The water collected in polypropylene bottles beneath each funnel. The sites were chosen to avoid, wherever possible, the influence of local discontinuities in the topography although the 'local' exposure of collectors varied considerably between the different collector sites. To provide more information on the capture efficiency of the rain collectors used, the drop size distribution and the total amount of rain was measured using a rain distrometer alongside the simple rain collectors at the summit and at 250 m. A ground-level Institute of Hydrology pattern rain collector (Institute of Hydrology, 1977) was also used at the summit. In addition, eight identical collectors were arranged (at 50-m intervals at 670 m
9oo
Transect through G.D.F
"E 800
=aS Summi:t.ElNI
e~"
1-8 Precipitation cotlec tars & Met instrument.s
~* 700
~ eS//eeellell el~
:
600.
I I
400' Eden V
! 200. ioo
;
;
;
o
i
;
,
i
t
e
-; ;
y
;
~
/o ,', ,2 ;3 ,,
Distance (kin}
Fig. 1. Schematic representation of GDF showing locations of sampling points.
a.s.1.) along a contour of the southwest side of GDF. These gauges provide an indication of variability in precipitation amount and ionic composition at one height, A further series of four collectors was placed at 30-m intervals at a height of 807 m a.s.1. The arrangement of rain and cloud water samplers on the hillside is ,shown diagrammatically in Fig. 1. Cloud water collection
Cloud water samples were collected at up to five different altitudes on the hill close to the location of precipitation samplers at the highest five levels. The conical shaped 'harp' wire collectors were strung with 550-/~m diameter ETFE-coated wire filaments, each separated at the top ring by 3 ram. The collection of cloud water is passive--droplets failing to follow the streamline of airflow around the strings are intercepted and run down the strings into a polypropylene collector. The efficiency of capture increases with droplet size and wind velocity. In practice, as droplets that have been collected run down the strings and grow by intercepting other drops, they may be blown off. It is therefore difficult to predict from theoretical work the collection properties of the gauges in field conditions. To provide a field calibration the passive cloud collectors were placed alongside a Knollenberg forward scattering spectrometer probe (FSSP) and the catch of the cloud water by the passive sampler was compared with the amount of cloud that would have passed through the vertical area cross-section of the cloud sampler. The FSSP provided cloud liquid water content and the mean drop size of the cloud water and a cup anemometer provided wind velocity. The periods of collection were between 20 and 200 min, with average wind speeds from 10 to 23 m s- 1 and mean cloud drop sizes from 3.5- to 6.9-/~m diameter. The collection efficiency ranged from 11 to 47% with a mean value (+ S.D.) of 28.9% +9.7%. All precipitation and cloud water collectors were washed by spraying with deionized water prior to a collection period. The duration of collection varied between 2 h and 3 days, but on average was between 5 and 12h.
Influence of altitude on rainfall composition at Great Dun Fell Rain and cloud water samples were stored at 4°C before analysis for SO 2- , NO~, C I - , K +, Na ÷, Ca 2+ , Mg 2+ and N H ~ by ion chromatography and for H ÷ from pH measurements. Precision in these measurements was generally between 1 and 4% and the sums of cation and anion concentrations expressed in equivalents agreed within 5%. The measurement of precipitation amounts was much less certain because of the exposed nature of many of the hillside collectors and the high wind speeds that are generally present at this site.
400
1357
"
• SO~-
&"
300-
~ NH~
200 •
p g ioo E3
0
200
48o
6~o
88o
Height (m)
RESULTS Fig. 3. Variation in deposition of major ions in rain, as a function of altitude, 4-5 December 1984.
Autumn 1984
During the a u t u m n of 1984 the first measurements of rainfall composition were made at the eight sites up the hill, as described above, on 10 occasions between 24 October and 5 December 1984. The total amount of rainfall over this period was 62 and 99 mm at the lowest and highest collectors, respectively. For individual events the variability between collectors on the hillside was large but a general increase in concentrations of the major ions towards the hill summit was observed for occasions with southwesterly airflow from the valley towards the summit. These features are illustrated by data for 4 and 5 December 1984. Rainfall amounts increased from 3 mm at an altitude of 433 m to 8 mm at the summit and concentrations of SO 2 , NO3, N H ~ and H ÷ all show an increase with altitude (Fig. 2). The concentrations of SO 2- and N O 3 increased between 433 and 847 m by a factor of 1.5. For NH~" the increase in concentration over the same height range is larger, about a factor of two. The acidity of rainfall also increases with altitude by about a factor of 3 between 433 and 847 m, but changes in H ÷ concentration are determined by the other ions in solution and may not be interpreted independently. On this occasion the concentrations of Na ÷, CI- and Mg 2÷, largely marine ions, did not change significantly between 433 and 847 m despite a large change in rainfall amount. The increases with height in rainfall a m o u n t and ionic concentrations lead to a large
increase in wet deposited SO 2-, N O 3, NH~-and H ÷ with altitude (Fig. 3). Similar increases in concentrations with altitude were observed for all other occasions during a u t u m n 1984 with airflow up the hillside from the west. Occasions with a strong southwesterly flow of maritime air generally show large concentrations of CI-, Na + and Mg 2÷ in rain and rather small SO 2-, N O 3, H ÷ and N H 2 concentrations. An example of such an event is shown in Fig. 4. On this occasion, the concentrations of all ions were larger at the hill summit. The precipitation amounts increased by about a factor of 2 between 347 and 847 m and CI- and Na ÷ concentrations increased by a factor of about 1.5, giving a three-fold increase in deposition (Fig. 5). Most of the other occasions show similar results of increasing concentration and deposition with increasing altitude, although the scatter on individual profiles is large. When the airflow was from the east, or with blocked flow, no increase in concentration or rainfall amount with altitude is observed. The average properties of the measurements during the a u t u m n of 1984 are provided in Fig. 6 for which the rainfall weighted mean concentrations of each ion over the 6 weeks of sampling have been plotted for each collector on the hill. The concentrations of N O 3, "
450
4-5.12 0 4
:2
. s~
p...l?
NO~
/
/
~ Go ~
A
375-
• No t
:L v
g 300-
40 o o
20
225"
200
Height ( m l
Fig. 2. Variation in concentrations of major ions in rain, as a function of altitude, 4-5 December 1984.
~
6~
8bo
Height (m)
Fig. 4. Variation in concentration of marine-derived ions in rain, as a function of altitude, 22-29 November 1984.
1358
D. FOWLERet al.
250.
E
200
o
the scavenging also enhancing rainfall amounts (Carruthers and Choularton, 1983).
/
Spring 1985
g t50. ~
100
~
50
O
400
6O0
800
Height(m)
Fig. 5. Variation in deposition of marine-derived ions in rain, as a function of altitude, 22-29 November 1984.
2410 - 5 1 2 8 4
80-
60-.
Cloud water composition measurements
.~ 4 0 -
U O,
2oo
Simultaneous measurements of cloud water composition in the orographic cap-cloud and rainwater composition were made during the spring of 1985 to show whether it was the scavenging of low-level cloud that was responsible for the increase in concentration of most ions in precipitation with altitude. Measurements were made between 16 April 1985 and 9 May 1985 and included seven occasions with significant amounts of precipitation at all levels with simultaneous samples of cloud water, using the passive collectors described earlier. As in the autumn of 1984, on all occasions with westerly or southwesterly airflow and the presence of a cap-cloud on GDF, concentrations of the major ions were larger at the hill summit than at lower elevations. The cloud water measurements provide essential additional information for the interpretation of these data.
3~o
,~bo
~o Height
e~o
r6o
a~o
(m)
Fig. 6. Average concentrations (rainfalJ weighted) of major pollutant ions in rain, as a function of altitude, for the autumn 1984 sampling period (24 October-5 December). SO 2-, H ÷, NH2, Na ÷, and C1- all increase with altitude, but only above 600 m are the changes significant. The concentrations increase by a factor of 1.5 between 530 and 847 m. This large increase, which seems from the measurements to be a common feature of the rainfall events at the site, is consistent with the scavenging of polluted low-level cap-cloud droplets by rain falling from a higher-level seeder cloud,
Measurements of cloud water composition were made at three sites (847, 808 and 753 m) and, in general, concentrations for all ions were larger in cloud water than in rainwater. The difference varied between a factor of 1.5 and 8. A much larger data set (including samples from spring and autumn 1985 and later sampling) shows (Fig. 7) that although the distributions of concentration (of SO~- in this example) overlap, the concentrations of ions in cloud water samples are larger on average than in rain by a factor of 2.5. A paired t-test for concentrations of SO~- in rainwater and cloud water showed that values differed significantly at the 0.1% level. For 11 precipitation and cloud water events at the hill summit in spring 1985, the mean ratios of cloud/rain concentrations (in #eq f - l) for major ions are shown in Table 1. The concentrations of any of these ions in cloud water is, however, strongly dependent on altitude. The
12
-- Rain CLoud
---
~o
~8
i ....3 •
c: 6¸
•
I t
I I
~° 4 2 0
,
5
38
i'--.
,
46
54 In SO~-(/..t.eq tO0 SOa2 -(,u.eq t-')
62
7
78
l-~)
20
49.5
2440
Fig. 7. Frequency distributions of sulphate concentrations in rain (solid lines) and cloud (dashed lines) for simultaneous measurements, 1985.
Influence of altitude on rainfall composition at Great Dun Fell Table 1. Ratio of concentrations in cloud/rain (~eq F- x) H+
3.9
NH: 2.4
¢12.6
No; so~2.8
2.0
Mean of 11 precipitation and cloud events at summit (847 m) during spring 1985. largest concentrations occur at cloud base and decrease with height as cloud liquid water content increases with altitude and this dilutes the solutes present. On many occasions, the lowest 100-200 m of the cap-cloud at G D F shows an adiabatic increase in liquid water content with height (Choularton et al., 1988), and in these conditions the change in concentration of major ions in cloud water with altitude may be predicted from a knowledge of cloud base. Figure 8 shows the change in concentrations for SO,2-, N O ; and CI- for the 17 April 1985, when concentrations of S O l - decreased from 238 to 141 #eq f - 1 between 755 and 847 m, a decrease of 40%. Reported cloud base measurements for the time of collection allow an estimate of the cloud liquid water content to be made for the three altitudes of the cloud water collectors, assuming the cloud to be adiabatic, and, using these data, the expected concentrations in solution may be estimated from the known concentration at the lowest sampling site. These are plotted in Fig. 8, and the Predicted change with altitude
3OO
so~oCt-_ [Measuredvotues
.
C
g
~
100"
7~o
86o
8~o
Height Qbove sea level(m)
Fig. 8. Variation in concentration of major anions in cloud, as a function of altitude, compared with that predicted from a calculation of adiabatic water content.
1359
agreement between measured and estimated values is good. A similar exercise for other occasions with known (and fairly constant) cloud base and cloud water concentrations also show good agreement. There is therefore a large vertical ionic concentration gradient in cloud water, with concentration decreasing with altitude. Rain falling from higher levels which does not interact with the cap-cloud may generally be sampled by the collectors below 500 m, and the results from autumn 1984 demonstrate that at these levels the concentration does not change with altitude. For the higher levels, the rain falls through a cap-cloud which itself contains a large vertical gradient in concentration of the ions present.
Rainfall composition Results from the autumn 1984 measurements of rainfall amount and composition at different altitudes show considerable scatter, and there appeared to be particular problems at collectors Nos 7 and 8 close to the hill summit. The high wind speed and exposed locations led to an aerodynamic sorting of the raindrop size spectrum, with many small drops being carried over or around the funnel and only large drops entering. As the chemical composition of rain is expected to vary with drop size in these conditions (the large drops having smaller concentrations), this will lead to a systematic underestimate of the concentrations of major ions in rain at the higher levels. To provide information on the variability of rainfall amount and composition at one altitude, eight collectors were placed in two parallel lines each with four collectors, 50 m separating each collector in the line and 20 m separating the lines, at the 670-m contour on the hill. During spring 1985, eight rain events were measured, varying between 0.7 and 11.0 mm, and the variability in concentrations of SO,2-, CI- and NO3 ions was typically 10% (average coefficient of variation). Variability in rainfall amount was rather larger at about 20%. The results of these measurements are summarized in Table 2. The variability at higher, more exposed, collectors will no doubt be greater, but practical problems such as finding the equipment on the hill in cloud prevented further extensive work of
Table 2. Measurements of rainfall and SO,2-, CI- and NO 3 concentrations at 675 m a.s.l, on the southwest slopes of GDF using eight collectors, each separated by 40 m. The additional column provides the average rainfall amounts at 250 and 810 m Date 16-17.4.85 19-22.4.85 25-26.4.85 29.4.85 29-30.4.85 30-1.5.85 2-5.5.85 5 6.5.85
Time 1800~900 1200-1500 1800-1300 1000-1600 160(01000 1800-1000 1800-1600 1600-0600
Mean rain (mm) -+S.D. 1.63+0.13 5.88+3.16 0.70+0.15 7.15--+2.79 10.98_+1.93 3.76-+0.98 3.55-+1.17 4.58+1.43
SO]--+S.D. 106+18 151+23 245+67 92+20 34+3 43--+5 350-+30 196-+69
CI--+S.D.
NO3_+S.D.
89+18 688+141 518+332 261+144 16+ 13 128_+37 276-+62 74+49
47+3 54_+20 96_+21 23_+5 10_+ 1 9_+2 183+21 139+51
Rain (mm) 17 1.02-3.63 7.80-4.77 0.32~).32 2.80-1.40 4.61 13.37 1.43~..20 0.32~..62 2.90~.27
1360
D. FOWLER
this kind. The collector at 823 m frequently caught less rain than the summit collector during the autumn 1984 measurements and a series of four collectors was placed along a southwest-northeast line at approx. 30-m intervals to provide a better estimate of both amount and composition of rain at this level. Whenever possible, the mean value for all collectors at a given height was used to estimate ion concentrations and rainfall amounts. Two precipitation events during the spring of 1985 which showed contrasting chemical properties provide examples of the range of cloud and rain compositions that is frequently encountered at GDF. The composition of precipitation which fell during 20 and 21 April 1985 was dominated by marine ions Na ÷, CI- and Mg 2+ (these contributing 75% of the ions present, expressed in p e q d - t ) . The concentrations of NO 3 and SO~- were in the range 15-70 and 70-250 #eq f - 1 , respectively and all ions in precipitation showed a marked increase in concentration with altitude (Fig. 9). For SO~- and NO3the increase was almost a factor of 3 while for C1- it was closer to a factor of 2. The differences in concentration of all ions between cloud water and rainwater at the summit were (relative to other occasions) quite small, less than a factor of 2. The 900-mbar trajectories of the air prior to precipitation showed an essentially northerly trajectory so that from the originally maritime trajectory the air crossed only northern Britain before reaching G D F (Fig. 10a). With such a trajectory there is little opportunity for the composition of cloud and rain to become dominated by the ions derived from the pollutants SO2 and NOx. By contrast, the precipitation on 5 May 1985 was associated with a 900-mbar trajectory taking the air over the major British source regions of SO2 and NO~ before reaching G D F (Fig. 10b). On this occasion, the concentrations of all ions in rain increase with altitude, the increases being typically a factor of 4 (Fig. 11). The concentrations of SO,2- and NO 3 in cloud water at the summit are large (600-900 peq d - 1) and exceed concentrations in rain at the summit by
et al. Trajectory f o r 9 0 0 mb at 3 h i n t e r v a l s 204.85 (a)
Trajectory for 9 0 0 mb at 3h intervals
5585
(b)
Fig. 10. (a) 900-mbar trajectories of air masses arriving at GDF 20-21 April 1985 (6-h intervals); (b) 900mbar trajectories of air masses arriving at GDF 5 May 1985 (3-h intervals). Trajectories were calculated by Dr K. J. Weston, University of Edinburgh, Department of Meteorology.
1600, 5-6.585
1400
•
SO~-
Ctoud
NH~ • NO~oct
o
g ¢
1300 •
600
19-22 4 85 200
I IO0-
0
®
=L 900" g
Rain
CLoud
E 700. E
oCt• SO2o NH~
e
g 500-
• NO~
E
• 1 o
4oo
6~
8oo
16oo
Height above sea tevet (m)
~loRain
~ b ~ ' 6
CLoud Ro in
~ . . ~ _ . ~ Ctoud
100"
4~
2o0
Rain
Fig. 11. Variation in concentration of major ions in cloud and rain, as a function of altitude, 5-6 May 1985.
300 -
20O
CLoud
o
~
6~
8oo
Height above lea Level (m)
Fig. 9. Variation in concentration ofmajor pollutant ions in cloud and rain, as a function of altitude, 19-22 April 1985.
about a factor of 3. The cloud/rainwater concentration differences are much larger close to cloud base (factor of about 6). The differences between these two occasions, and the link with trajectories, suggests that cloud and
1361
Influence of altitude on rainfall composition at Great Dun Fell rainwater chemistry at G D F is largely controlled by the source of the air and the scale of sources of major air pollutants that lie along the trajectory. There is no evidence in any of the measurements to date that the mechanism of enhancement of SO 2- and N O ~ deposition at this site is peculiar to local properties of the site. The valley upwind is a rural area with few sources of air pollutants and the only feature which appears to be responsible for the large increase in deposition is the presence of orographic cloud, which effectively transforms particulate pollutants from sub-#m aero-
sols into cloud droplets which are then scavenged relatively efficiently by falling rain.
Average change in rainfall composition with altitude The whole data set for rainfall and cloud water composition at different altitudes from the spring of 1985 provides average concentrations at the hill summit that exceed the concentration in the valley by at least a factor of 2 for most ions (Table 3). The average data are plotted for CI-, SO 2 , N O 3, H + and N H ~ i n Fig. 12a-e. The larger rainfall at the higher levels
(b)
(a)
40.
ange in ave.ogeconoentr=,on o, H+.ith
I00
attitud
Change in overage concentration of NH,~ /
3O
•
with a
t
t
~
eo
.~-20
"
~
i I0-
60
o (..)
20
200
300
400
600
500
700
800
Height above sea Level (m)
300
400
500
600
700
800
900
Height above sea level (m)
(c) (d)
I00
45 ¸ Change in average concentration of NO~ A 40.
[I 80
o- 3530o~'25 . o
g 4o
~
0
20o~ 15. 8
20
I0.
5. 500
400
500
600
700
800
900
200
3()0 4C~0 500
600 7(~0
800
Height above sea fever(m)
Height above seo tevet (m)
(e) 180-
Change in overage concentration of CL with
160~
140-
.._~ ~2 0 .~ I 0 0 -
8060t~ 4020200
300
400
500
600
.700
800
900
Height obove sea lever (ml
Fig. 12. Average concentrations (rainfall weighted) of major ions in rain, as a function of altitude, for the spring 1985 sampling period (16 April-9 May). (a) H ÷, (b) NH2, (c) SO~-, (d) NO 3 , (e) CI .
1362
D. FOWLERet al. Table 3. Ratio of concentration at summit/ valley for major ions and rain amount H+
2.9
NH4 3.1
C12.9
NO~- S O l -
2.3
2.2
Rain 2.0
Twenty precipitation events at GDF with measurements at eight levels (244-847m) showing increase in concentration between valley and summit, 1984--85. (typically by a factor of 2) with the larger concentrations leads to wet deposition of most ions at the summit exceeding that at the valley site by a factor of at least 4. The widely applied technique for predicting deposition on hills from measurements at lower elevations is therefore likely to be seriously in error whenever hill cloud is common. Although only a small proportion of the annual rainfall events have been sampled in this study, the sampled events showing the increase in concentration with altitude were all from the west or southwest. These wind sectors provide the bulk of annual precipitation. The west-facing hills in western Britain from Wales to the northwest of Scotland, and the coastal mountains of Norway, are commonly shrouded by orographic cloud so that for these regions, which are already the areas of largest wet-deposited SO 2-, N O 3 and H + ions, the existing published values are underestimates. The work presented here may not be readily applied to other hills. We do not know whether the effect of 'seeder-feeder' modification of wet deposition patterns is limited just to the westfacing hills, nor have we shown from measurement the effects of different hill shapes on the process. Such effects have so far been restricted to modelling studies (Hill et al., 1988). These studies show that the effects observed here would be expected for most hills on which orographic clouds are a common feature, but that the rate of change of composition with altitude depends strongly on hill shape and atmospheric conditions. It is not possible, therefore, to extrapolate to other areas to gauge the magnitude of the underestimate in wet deposition. To provide the necessary basis for such estimates, more needs to be known about the average composition of cloud water at different locations and the mechanisms of wet deposition over extensive upland regions, to show whether the effect observed at G D F is a local phenomenon.
Acknowledoements--This work was supported by the U.K. Department of the Environment and the Natural Environment Research Council and forms a part of a larger collaborative study between ITE, UMIST, AERE (Harwell) and the University of East Anglia. The help from colleagues in all of these organizations and K. J. Weston of Edinburgh University is acknowledged.
REFERENCES
BAPMoN (1986) Global atmospheric background monitoring for environmental parameters. Precipitation Chemistry, Vol. II. WMO/TD No. 116. Barrett C. F., Atkins D. H. F., Cape J. N., Fowler D., Irwin J. E., Kallend A. S., Martin A., Pitman M. J., Scriven R. A. and Tuck A. F. (1983) Acid deposition in the United Kingdom, 72pp. Warren Spring Laboratory, Stevenage. Barrett C. F., Goldsmith A. L., Hall D. J. and Irwin J. F. (1985). The variation of precipitation composition with altitude: a feasibility study. LR 534 (AP) Warren Spring Laboratory, Stevenage. Bergeron T. (1965) On the low-level redistribution of atmospheric water caused by orography. Proc. Int. Conf. Cloud. Phys. Tokyo, 1965, pp. 96-100. Carruthers D. J, and Choularton T. W. (1983). A model of the feeder-seeder mechanism of orographic rain including stratification and wind drift effects. Q. JI R. Met. Soc. 109, 575-588. Choularton T. W., Gay M. J., Jones A., Fowler D., Cape J. N. and Leith I. D. (1988). The influence of altitude on wet deposition. Comparison between field measurements at Great Dun Fell and the predictions of a seeder-feeder model. Atmospheric Environment 22, 1363-1371. Duncan C. L., Welch E. B. and Ausserer W. (1986). The composition of precipitation at Snoqualmire Pass and Stevens Pass in the central Cascades of Washington State. Water Air Soil Pollut. 30, 217-229. Hill T. A., Jones A. and Choularton T. W. (1988). Modelling sulphate deposition onto hills by washout and turbulence. Q. Jl R. met. Soc. (in press). Institute of Hydrology (1977). Selected measurement techniques in use at Plynlimon experimental catchments. Institute of Hydrology, Wallingford, Rep. 43. Puxbaum H. (1988). The chemistry of wet deposition in the Austrian Alps. In Acid Deposition Processes at High Elevation Sites (edited by Unsworth M. H. and Fowler D.). NATO-ARI, Edinburgh (in press). The Royal Society (1984). The ecological effects of deposited sulphur and nitrogen compounds. Phil. Trans. R. Soc. Ser. B, 305, 255-577. Semb A. and Dovland H. (1986). Atmospheric deposition in Fenno-Scandia: characteristics and trends. Water Air Soil Pollut. 30, 5-16. Sweden (1972) Sweden's Case Study for the UN Conference on the Human Environment, 1972: Air Pollution Across National Boundaries. Stockholm.
0004-6981/88 $3.00+0.00 Pergamon Press plc
THE INFLUENCE OF ALTITUDE ON RAINFALL COMPOSITION AT GREAT DUN FELL
D. FOWLER,J. N. CAPE, I. D. LEITH Institute of Terrestrial Ecolog2~,Bush Estate, Penicuik, Midlothian EH26 0QB, U.K.
T. W. CHOULARTON,M. J. GAY and A. JONES Department of Pure and Applied Physics, University of Manchester Institute of Science and Technology, P.O. Box 88, Manchester M60 1QD, U.K. (First received 19 May 1987 and received for publication J 1 January 1988)
Abstract--The influence of altitude on rainfall composition and wet deposition has been investigated at Great Dun Fell in northern England. Measurements of rainfall at eight altitudes between 250 and 850 m on the western slopes of the hill show marked changes in both amount and composition when orographic cloud is present and a west or southwest wind is blowing.On average (20 precipitation events from autumn 1984 and spring 1985), the rainfall at the summit (847 m) exceeded that at 250 m by a factor of 2, and concentrations of SO2-, NO3, CI-, NH,~and H ÷ were larger at the summit by factors of between 2.2 and 3.1. Thus, wet deposition at the summit was larger than at 250 m by about a factor of 5. The concentrations of major ions in orographic cloud at 847 m exceeded concentrations in rain by a factor of between 2.0 and 3.9. A large change occurred in the concentrations of major ions in orographic cloud with altitude, decreasing with increasing altitude from cloud base. Such changes could generally be explained by the expected dilution as cloud liquid water content increased adiabatically. When the wind was from the east or with blocked flow no increase in concentration or rainfall amount was observed. Key word index: Wet-deposition, orographic enhancement, cloud chemistry.
INTRODUCTION
Contemporary interest in the chemical composition of rain was largely stimulated by Scandinavian studies in the 1950s and 1960s which linked the acidification of rainfall with the long-range transport of S compounds in the atmosphere over Europe (Sweden, 1972). Research during the last 15 yr into the meteorological and chemical processes involved in the atmospheric transport and chemical conversion of gaseous pollutants has identified the major processes controlling the long-range transport of S compounds in the atmosphere over Europe (Royal Society, 1984). Monitoring studies have improved the quality and the quantity of data to describe the spatial patterns of wet deposition (BAPMoN, 1986). Within northwest Europe and Scandinavia the areas receiving the largest amounts of wet-deposited acidity are the highrainfall regions of southwest Norway and western and northern Britain (Semb and Dovland, 1986; Barrett et al., 1983). Precipitation chemistry networks in Europe and North America have been designed to obtain regionally representative measurements of the average composition of precipitation. To obtain the wet deposition fields from these data, the much more extensive data on precipitation amount from meteorological and hydrological precipitation samplers have been combined with the data from precipitation chemistry networks. In general, annual rainfall
weighted average concentrations of SO ] - , H +, NO 3 and NH2 have been shown to vary much less than precipitation amounts. There are, however, areas of uncertainty, in particular the tops of hills, where rainfall amounts and wet deposition are larger and where for practical reasons few measurements are made. For estimates of wet deposition on the high ground in northwest Britain, Scandinavia and continental Europe, for which representative sites are not available, the composition of precipitation has been assumed to remain constant with altitude. As the areas experiencing the largest wet deposition are the upland districts which have been a focus of interest for freshwater acidification in Britain and Scandinavia, a measurement study of the variation in the composition of precipitation with altitude is of considerable importance. Recent measurements of the ionic composition of rainfall at different elevations in the Alps of France, in Austria and in western North America show generally smaller concentrations of the major ions, SO 2-, NO3, and NH~ and H +, at the higher elevations ( > 3000 m) (Duncan et al., 1986; Puxbaum, 1988). For smaller mountains (500-1000 m) in northwest Europe, orographic cloud frequently shrouds the summits and contains large concentrations of all major ions: SO 2-, NO 3, NH2, H +, CI- and Na +. This low-level cloud is readily scavenged by rainfall from higher levels, enhancing rainfall by the seeder-feeder process (Bergeron, 1965). As concentrations of the major ions
1355
1356
D. FOWLERet al.
in the feeder (lower) cloud are generally large compared with typical concentrations in rainfall, this process may lead to wet deposition on high ground exceeding the values predicted from a combination of regional rainfall chemistry measured at low-level sites and rainfall amounts. In the study reported here, the composition of precipitation and cloud water at a number of sites between 150 m a.s.l, and a hill summit 847 m a,s.1, in the northwest of England has been measured. The area chosen for this study includes a ridge lying northwest-southeast with the tops of individual hills at about 900 m, where annual rainfall amounts are in excess of 2000 mm, and a valley to the southwest of this line of hills at 150 m a.s.1, with an annual rainfall of approx. 900 mm. The summit of the specific hill selected for measurement (Great Dun Fell, GDF) at 847 m a.s.l, is in cloud for some part of 250 days every year, and the road between the valley site to the southwest and the summit provides access to a range of sites for rain and cloud water collectors. With winds from the southwest the seeder-feeder mechanism has been shown to be responsible for a substantial enhancement of rainfall amounts at the higher levels on these hills (Carruthers and Choularton, 1983). The initial objectives of this study were to obtain rainfall and cloud water composition measurements at a range of sites along the southwest slopes of G D F to show for individual rainfall events in southwesterly airflow: (a) whether the concentrations of major ions in rainfall are constant with altitude; (b) whether the presence of orographic cloud on G D F significantly influences the composition of rainfall at the higher altitudes; and (c) whether, in the long term, annual wet deposition of the major ions on the hilltops is larger than would have been predicted assuming constant rainfall composition with altitude.
METHODS
Rainfall was sampled using 20-cm diameter Pyrex glass funnels mounted 1.5 m above ground at each of eight locations between 250 m a.s.l, in the Eden valley and 847 m at G D F summit. The water collected in polypropylene bottles beneath each funnel. The sites were chosen to avoid, wherever possible, the influence of local discontinuities in the topography although the 'local' exposure of collectors varied considerably between the different collector sites. To provide more information on the capture efficiency of the rain collectors used, the drop size distribution and the total amount of rain was measured using a rain distrometer alongside the simple rain collectors at the summit and at 250 m. A ground-level Institute of Hydrology pattern rain collector (Institute of Hydrology, 1977) was also used at the summit. In addition, eight identical collectors were arranged (at 50-m intervals at 670 m
9oo
Transect through G.D.F
"E 800
=aS Summi:t.ElNI
e~"
1-8 Precipitation cotlec tars & Met instrument.s
~* 700
~ eS//eeellell el~
:
600.
I I
400' Eden V
! 200. ioo
;
;
;
o
i
;
,
i
t
e
-; ;
y
;
~
/o ,', ,2 ;3 ,,
Distance (kin}
Fig. 1. Schematic representation of GDF showing locations of sampling points.
a.s.1.) along a contour of the southwest side of GDF. These gauges provide an indication of variability in precipitation amount and ionic composition at one height, A further series of four collectors was placed at 30-m intervals at a height of 807 m a.s.1. The arrangement of rain and cloud water samplers on the hillside is ,shown diagrammatically in Fig. 1. Cloud water collection
Cloud water samples were collected at up to five different altitudes on the hill close to the location of precipitation samplers at the highest five levels. The conical shaped 'harp' wire collectors were strung with 550-/~m diameter ETFE-coated wire filaments, each separated at the top ring by 3 ram. The collection of cloud water is passive--droplets failing to follow the streamline of airflow around the strings are intercepted and run down the strings into a polypropylene collector. The efficiency of capture increases with droplet size and wind velocity. In practice, as droplets that have been collected run down the strings and grow by intercepting other drops, they may be blown off. It is therefore difficult to predict from theoretical work the collection properties of the gauges in field conditions. To provide a field calibration the passive cloud collectors were placed alongside a Knollenberg forward scattering spectrometer probe (FSSP) and the catch of the cloud water by the passive sampler was compared with the amount of cloud that would have passed through the vertical area cross-section of the cloud sampler. The FSSP provided cloud liquid water content and the mean drop size of the cloud water and a cup anemometer provided wind velocity. The periods of collection were between 20 and 200 min, with average wind speeds from 10 to 23 m s- 1 and mean cloud drop sizes from 3.5- to 6.9-/~m diameter. The collection efficiency ranged from 11 to 47% with a mean value (+ S.D.) of 28.9% +9.7%. All precipitation and cloud water collectors were washed by spraying with deionized water prior to a collection period. The duration of collection varied between 2 h and 3 days, but on average was between 5 and 12h.
Influence of altitude on rainfall composition at Great Dun Fell Rain and cloud water samples were stored at 4°C before analysis for SO 2- , NO~, C I - , K +, Na ÷, Ca 2+ , Mg 2+ and N H ~ by ion chromatography and for H ÷ from pH measurements. Precision in these measurements was generally between 1 and 4% and the sums of cation and anion concentrations expressed in equivalents agreed within 5%. The measurement of precipitation amounts was much less certain because of the exposed nature of many of the hillside collectors and the high wind speeds that are generally present at this site.
400
1357
"
• SO~-
&"
300-
~ NH~
200 •
p g ioo E3
0
200
48o
6~o
88o
Height (m)
RESULTS Fig. 3. Variation in deposition of major ions in rain, as a function of altitude, 4-5 December 1984.
Autumn 1984
During the a u t u m n of 1984 the first measurements of rainfall composition were made at the eight sites up the hill, as described above, on 10 occasions between 24 October and 5 December 1984. The total amount of rainfall over this period was 62 and 99 mm at the lowest and highest collectors, respectively. For individual events the variability between collectors on the hillside was large but a general increase in concentrations of the major ions towards the hill summit was observed for occasions with southwesterly airflow from the valley towards the summit. These features are illustrated by data for 4 and 5 December 1984. Rainfall amounts increased from 3 mm at an altitude of 433 m to 8 mm at the summit and concentrations of SO 2 , NO3, N H ~ and H ÷ all show an increase with altitude (Fig. 2). The concentrations of SO 2- and N O 3 increased between 433 and 847 m by a factor of 1.5. For NH~" the increase in concentration over the same height range is larger, about a factor of two. The acidity of rainfall also increases with altitude by about a factor of 3 between 433 and 847 m, but changes in H ÷ concentration are determined by the other ions in solution and may not be interpreted independently. On this occasion the concentrations of Na ÷, CI- and Mg 2÷, largely marine ions, did not change significantly between 433 and 847 m despite a large change in rainfall amount. The increases with height in rainfall a m o u n t and ionic concentrations lead to a large
increase in wet deposited SO 2-, N O 3, NH~-and H ÷ with altitude (Fig. 3). Similar increases in concentrations with altitude were observed for all other occasions during a u t u m n 1984 with airflow up the hillside from the west. Occasions with a strong southwesterly flow of maritime air generally show large concentrations of CI-, Na + and Mg 2÷ in rain and rather small SO 2-, N O 3, H ÷ and N H 2 concentrations. An example of such an event is shown in Fig. 4. On this occasion, the concentrations of all ions were larger at the hill summit. The precipitation amounts increased by about a factor of 2 between 347 and 847 m and CI- and Na ÷ concentrations increased by a factor of about 1.5, giving a three-fold increase in deposition (Fig. 5). Most of the other occasions show similar results of increasing concentration and deposition with increasing altitude, although the scatter on individual profiles is large. When the airflow was from the east, or with blocked flow, no increase in concentration or rainfall amount with altitude is observed. The average properties of the measurements during the a u t u m n of 1984 are provided in Fig. 6 for which the rainfall weighted mean concentrations of each ion over the 6 weeks of sampling have been plotted for each collector on the hill. The concentrations of N O 3, "
450
4-5.12 0 4
:2
. s~
p...l?
NO~
/
/
~ Go ~
A
375-
• No t
:L v
g 300-
40 o o
20
225"
200
Height ( m l
Fig. 2. Variation in concentrations of major ions in rain, as a function of altitude, 4-5 December 1984.
~
6~
8bo
Height (m)
Fig. 4. Variation in concentration of marine-derived ions in rain, as a function of altitude, 22-29 November 1984.
1358
D. FOWLERet al.
250.
E
200
o
the scavenging also enhancing rainfall amounts (Carruthers and Choularton, 1983).
/
Spring 1985
g t50. ~
100
~
50
O
400
6O0
800
Height(m)
Fig. 5. Variation in deposition of marine-derived ions in rain, as a function of altitude, 22-29 November 1984.
2410 - 5 1 2 8 4
80-
60-.
Cloud water composition measurements
.~ 4 0 -
U O,
2oo
Simultaneous measurements of cloud water composition in the orographic cap-cloud and rainwater composition were made during the spring of 1985 to show whether it was the scavenging of low-level cloud that was responsible for the increase in concentration of most ions in precipitation with altitude. Measurements were made between 16 April 1985 and 9 May 1985 and included seven occasions with significant amounts of precipitation at all levels with simultaneous samples of cloud water, using the passive collectors described earlier. As in the autumn of 1984, on all occasions with westerly or southwesterly airflow and the presence of a cap-cloud on GDF, concentrations of the major ions were larger at the hill summit than at lower elevations. The cloud water measurements provide essential additional information for the interpretation of these data.
3~o
,~bo
~o Height
e~o
r6o
a~o
(m)
Fig. 6. Average concentrations (rainfalJ weighted) of major pollutant ions in rain, as a function of altitude, for the autumn 1984 sampling period (24 October-5 December). SO 2-, H ÷, NH2, Na ÷, and C1- all increase with altitude, but only above 600 m are the changes significant. The concentrations increase by a factor of 1.5 between 530 and 847 m. This large increase, which seems from the measurements to be a common feature of the rainfall events at the site, is consistent with the scavenging of polluted low-level cap-cloud droplets by rain falling from a higher-level seeder cloud,
Measurements of cloud water composition were made at three sites (847, 808 and 753 m) and, in general, concentrations for all ions were larger in cloud water than in rainwater. The difference varied between a factor of 1.5 and 8. A much larger data set (including samples from spring and autumn 1985 and later sampling) shows (Fig. 7) that although the distributions of concentration (of SO~- in this example) overlap, the concentrations of ions in cloud water samples are larger on average than in rain by a factor of 2.5. A paired t-test for concentrations of SO~- in rainwater and cloud water showed that values differed significantly at the 0.1% level. For 11 precipitation and cloud water events at the hill summit in spring 1985, the mean ratios of cloud/rain concentrations (in #eq f - l) for major ions are shown in Table 1. The concentrations of any of these ions in cloud water is, however, strongly dependent on altitude. The
12
-- Rain CLoud
---
~o
~8
i ....3 •
c: 6¸
•
I t
I I
~° 4 2 0
,
5
38
i'--.
,
46
54 In SO~-(/..t.eq tO0 SOa2 -(,u.eq t-')
62
7
78
l-~)
20
49.5
2440
Fig. 7. Frequency distributions of sulphate concentrations in rain (solid lines) and cloud (dashed lines) for simultaneous measurements, 1985.
Influence of altitude on rainfall composition at Great Dun Fell Table 1. Ratio of concentrations in cloud/rain (~eq F- x) H+
3.9
NH: 2.4
¢12.6
No; so~2.8
2.0
Mean of 11 precipitation and cloud events at summit (847 m) during spring 1985. largest concentrations occur at cloud base and decrease with height as cloud liquid water content increases with altitude and this dilutes the solutes present. On many occasions, the lowest 100-200 m of the cap-cloud at G D F shows an adiabatic increase in liquid water content with height (Choularton et al., 1988), and in these conditions the change in concentration of major ions in cloud water with altitude may be predicted from a knowledge of cloud base. Figure 8 shows the change in concentrations for SO,2-, N O ; and CI- for the 17 April 1985, when concentrations of S O l - decreased from 238 to 141 #eq f - 1 between 755 and 847 m, a decrease of 40%. Reported cloud base measurements for the time of collection allow an estimate of the cloud liquid water content to be made for the three altitudes of the cloud water collectors, assuming the cloud to be adiabatic, and, using these data, the expected concentrations in solution may be estimated from the known concentration at the lowest sampling site. These are plotted in Fig. 8, and the Predicted change with altitude
3OO
so~oCt-_ [Measuredvotues
.
C
g
~
100"
7~o
86o
8~o
Height Qbove sea level(m)
Fig. 8. Variation in concentration of major anions in cloud, as a function of altitude, compared with that predicted from a calculation of adiabatic water content.
1359
agreement between measured and estimated values is good. A similar exercise for other occasions with known (and fairly constant) cloud base and cloud water concentrations also show good agreement. There is therefore a large vertical ionic concentration gradient in cloud water, with concentration decreasing with altitude. Rain falling from higher levels which does not interact with the cap-cloud may generally be sampled by the collectors below 500 m, and the results from autumn 1984 demonstrate that at these levels the concentration does not change with altitude. For the higher levels, the rain falls through a cap-cloud which itself contains a large vertical gradient in concentration of the ions present.
Rainfall composition Results from the autumn 1984 measurements of rainfall amount and composition at different altitudes show considerable scatter, and there appeared to be particular problems at collectors Nos 7 and 8 close to the hill summit. The high wind speed and exposed locations led to an aerodynamic sorting of the raindrop size spectrum, with many small drops being carried over or around the funnel and only large drops entering. As the chemical composition of rain is expected to vary with drop size in these conditions (the large drops having smaller concentrations), this will lead to a systematic underestimate of the concentrations of major ions in rain at the higher levels. To provide information on the variability of rainfall amount and composition at one altitude, eight collectors were placed in two parallel lines each with four collectors, 50 m separating each collector in the line and 20 m separating the lines, at the 670-m contour on the hill. During spring 1985, eight rain events were measured, varying between 0.7 and 11.0 mm, and the variability in concentrations of SO,2-, CI- and NO3 ions was typically 10% (average coefficient of variation). Variability in rainfall amount was rather larger at about 20%. The results of these measurements are summarized in Table 2. The variability at higher, more exposed, collectors will no doubt be greater, but practical problems such as finding the equipment on the hill in cloud prevented further extensive work of
Table 2. Measurements of rainfall and SO,2-, CI- and NO 3 concentrations at 675 m a.s.l, on the southwest slopes of GDF using eight collectors, each separated by 40 m. The additional column provides the average rainfall amounts at 250 and 810 m Date 16-17.4.85 19-22.4.85 25-26.4.85 29.4.85 29-30.4.85 30-1.5.85 2-5.5.85 5 6.5.85
Time 1800~900 1200-1500 1800-1300 1000-1600 160(01000 1800-1000 1800-1600 1600-0600
Mean rain (mm) -+S.D. 1.63+0.13 5.88+3.16 0.70+0.15 7.15--+2.79 10.98_+1.93 3.76-+0.98 3.55-+1.17 4.58+1.43
SO]--+S.D. 106+18 151+23 245+67 92+20 34+3 43--+5 350-+30 196-+69
CI--+S.D.
NO3_+S.D.
89+18 688+141 518+332 261+144 16+ 13 128_+37 276-+62 74+49
47+3 54_+20 96_+21 23_+5 10_+ 1 9_+2 183+21 139+51
Rain (mm) 17 1.02-3.63 7.80-4.77 0.32~).32 2.80-1.40 4.61 13.37 1.43~..20 0.32~..62 2.90~.27
1360
D. FOWLER
this kind. The collector at 823 m frequently caught less rain than the summit collector during the autumn 1984 measurements and a series of four collectors was placed along a southwest-northeast line at approx. 30-m intervals to provide a better estimate of both amount and composition of rain at this level. Whenever possible, the mean value for all collectors at a given height was used to estimate ion concentrations and rainfall amounts. Two precipitation events during the spring of 1985 which showed contrasting chemical properties provide examples of the range of cloud and rain compositions that is frequently encountered at GDF. The composition of precipitation which fell during 20 and 21 April 1985 was dominated by marine ions Na ÷, CI- and Mg 2+ (these contributing 75% of the ions present, expressed in p e q d - t ) . The concentrations of NO 3 and SO~- were in the range 15-70 and 70-250 #eq f - 1 , respectively and all ions in precipitation showed a marked increase in concentration with altitude (Fig. 9). For SO~- and NO3the increase was almost a factor of 3 while for C1- it was closer to a factor of 2. The differences in concentration of all ions between cloud water and rainwater at the summit were (relative to other occasions) quite small, less than a factor of 2. The 900-mbar trajectories of the air prior to precipitation showed an essentially northerly trajectory so that from the originally maritime trajectory the air crossed only northern Britain before reaching G D F (Fig. 10a). With such a trajectory there is little opportunity for the composition of cloud and rain to become dominated by the ions derived from the pollutants SO2 and NOx. By contrast, the precipitation on 5 May 1985 was associated with a 900-mbar trajectory taking the air over the major British source regions of SO2 and NO~ before reaching G D F (Fig. 10b). On this occasion, the concentrations of all ions in rain increase with altitude, the increases being typically a factor of 4 (Fig. 11). The concentrations of SO,2- and NO 3 in cloud water at the summit are large (600-900 peq d - 1) and exceed concentrations in rain at the summit by
et al. Trajectory f o r 9 0 0 mb at 3 h i n t e r v a l s 204.85 (a)
Trajectory for 9 0 0 mb at 3h intervals
5585
(b)
Fig. 10. (a) 900-mbar trajectories of air masses arriving at GDF 20-21 April 1985 (6-h intervals); (b) 900mbar trajectories of air masses arriving at GDF 5 May 1985 (3-h intervals). Trajectories were calculated by Dr K. J. Weston, University of Edinburgh, Department of Meteorology.
1600, 5-6.585
1400
•
SO~-
Ctoud
NH~ • NO~oct
o
g ¢
1300 •
600
19-22 4 85 200
I IO0-
0
®
=L 900" g
Rain
CLoud
E 700. E
oCt• SO2o NH~
e
g 500-
• NO~
E
• 1 o
4oo
6~
8oo
16oo
Height above sea tevet (m)
~loRain
~ b ~ ' 6
CLoud Ro in
~ . . ~ _ . ~ Ctoud
100"
4~
2o0
Rain
Fig. 11. Variation in concentration of major ions in cloud and rain, as a function of altitude, 5-6 May 1985.
300 -
20O
CLoud
o
~
6~
8oo
Height above lea Level (m)
Fig. 9. Variation in concentration ofmajor pollutant ions in cloud and rain, as a function of altitude, 19-22 April 1985.
about a factor of 3. The cloud/rainwater concentration differences are much larger close to cloud base (factor of about 6). The differences between these two occasions, and the link with trajectories, suggests that cloud and
1361
Influence of altitude on rainfall composition at Great Dun Fell rainwater chemistry at G D F is largely controlled by the source of the air and the scale of sources of major air pollutants that lie along the trajectory. There is no evidence in any of the measurements to date that the mechanism of enhancement of SO 2- and N O ~ deposition at this site is peculiar to local properties of the site. The valley upwind is a rural area with few sources of air pollutants and the only feature which appears to be responsible for the large increase in deposition is the presence of orographic cloud, which effectively transforms particulate pollutants from sub-#m aero-
sols into cloud droplets which are then scavenged relatively efficiently by falling rain.
Average change in rainfall composition with altitude The whole data set for rainfall and cloud water composition at different altitudes from the spring of 1985 provides average concentrations at the hill summit that exceed the concentration in the valley by at least a factor of 2 for most ions (Table 3). The average data are plotted for CI-, SO 2 , N O 3, H + and N H ~ i n Fig. 12a-e. The larger rainfall at the higher levels
(b)
(a)
40.
ange in ave.ogeconoentr=,on o, H+.ith
I00
attitud
Change in overage concentration of NH,~ /
3O
•
with a
t
t
~
eo
.~-20
"
~
i I0-
60
o (..)
20
200
300
400
600
500
700
800
Height above sea Level (m)
300
400
500
600
700
800
900
Height above sea level (m)
(c) (d)
I00
45 ¸ Change in average concentration of NO~ A 40.
[I 80
o- 3530o~'25 . o
g 4o
~
0
20o~ 15. 8
20
I0.
5. 500
400
500
600
700
800
900
200
3()0 4C~0 500
600 7(~0
800
Height above sea fever(m)
Height above seo tevet (m)
(e) 180-
Change in overage concentration of CL with
160~
140-
.._~ ~2 0 .~ I 0 0 -
8060t~ 4020200
300
400
500
600
.700
800
900
Height obove sea lever (ml
Fig. 12. Average concentrations (rainfall weighted) of major ions in rain, as a function of altitude, for the spring 1985 sampling period (16 April-9 May). (a) H ÷, (b) NH2, (c) SO~-, (d) NO 3 , (e) CI .
1362
D. FOWLERet al. Table 3. Ratio of concentration at summit/ valley for major ions and rain amount H+
2.9
NH4 3.1
C12.9
NO~- S O l -
2.3
2.2
Rain 2.0
Twenty precipitation events at GDF with measurements at eight levels (244-847m) showing increase in concentration between valley and summit, 1984--85. (typically by a factor of 2) with the larger concentrations leads to wet deposition of most ions at the summit exceeding that at the valley site by a factor of at least 4. The widely applied technique for predicting deposition on hills from measurements at lower elevations is therefore likely to be seriously in error whenever hill cloud is common. Although only a small proportion of the annual rainfall events have been sampled in this study, the sampled events showing the increase in concentration with altitude were all from the west or southwest. These wind sectors provide the bulk of annual precipitation. The west-facing hills in western Britain from Wales to the northwest of Scotland, and the coastal mountains of Norway, are commonly shrouded by orographic cloud so that for these regions, which are already the areas of largest wet-deposited SO 2-, N O 3 and H + ions, the existing published values are underestimates. The work presented here may not be readily applied to other hills. We do not know whether the effect of 'seeder-feeder' modification of wet deposition patterns is limited just to the westfacing hills, nor have we shown from measurement the effects of different hill shapes on the process. Such effects have so far been restricted to modelling studies (Hill et al., 1988). These studies show that the effects observed here would be expected for most hills on which orographic clouds are a common feature, but that the rate of change of composition with altitude depends strongly on hill shape and atmospheric conditions. It is not possible, therefore, to extrapolate to other areas to gauge the magnitude of the underestimate in wet deposition. To provide the necessary basis for such estimates, more needs to be known about the average composition of cloud water at different locations and the mechanisms of wet deposition over extensive upland regions, to show whether the effect observed at G D F is a local phenomenon.
Acknowledoements--This work was supported by the U.K. Department of the Environment and the Natural Environment Research Council and forms a part of a larger collaborative study between ITE, UMIST, AERE (Harwell) and the University of East Anglia. The help from colleagues in all of these organizations and K. J. Weston of Edinburgh University is acknowledged.
REFERENCES
BAPMoN (1986) Global atmospheric background monitoring for environmental parameters. Precipitation Chemistry, Vol. II. WMO/TD No. 116. Barrett C. F., Atkins D. H. F., Cape J. N., Fowler D., Irwin J. E., Kallend A. S., Martin A., Pitman M. J., Scriven R. A. and Tuck A. F. (1983) Acid deposition in the United Kingdom, 72pp. Warren Spring Laboratory, Stevenage. Barrett C. F., Goldsmith A. L., Hall D. J. and Irwin J. F. (1985). The variation of precipitation composition with altitude: a feasibility study. LR 534 (AP) Warren Spring Laboratory, Stevenage. Bergeron T. (1965) On the low-level redistribution of atmospheric water caused by orography. Proc. Int. Conf. Cloud. Phys. Tokyo, 1965, pp. 96-100. Carruthers D. J, and Choularton T. W. (1983). A model of the feeder-seeder mechanism of orographic rain including stratification and wind drift effects. Q. JI R. Met. Soc. 109, 575-588. Choularton T. W., Gay M. J., Jones A., Fowler D., Cape J. N. and Leith I. D. (1988). The influence of altitude on wet deposition. Comparison between field measurements at Great Dun Fell and the predictions of a seeder-feeder model. Atmospheric Environment 22, 1363-1371. Duncan C. L., Welch E. B. and Ausserer W. (1986). The composition of precipitation at Snoqualmire Pass and Stevens Pass in the central Cascades of Washington State. Water Air Soil Pollut. 30, 217-229. Hill T. A., Jones A. and Choularton T. W. (1988). Modelling sulphate deposition onto hills by washout and turbulence. Q. Jl R. met. Soc. (in press). Institute of Hydrology (1977). Selected measurement techniques in use at Plynlimon experimental catchments. Institute of Hydrology, Wallingford, Rep. 43. Puxbaum H. (1988). The chemistry of wet deposition in the Austrian Alps. In Acid Deposition Processes at High Elevation Sites (edited by Unsworth M. H. and Fowler D.). NATO-ARI, Edinburgh (in press). The Royal Society (1984). The ecological effects of deposited sulphur and nitrogen compounds. Phil. Trans. R. Soc. Ser. B, 305, 255-577. Semb A. and Dovland H. (1986). Atmospheric deposition in Fenno-Scandia: characteristics and trends. Water Air Soil Pollut. 30, 5-16. Sweden (1972) Sweden's Case Study for the UN Conference on the Human Environment, 1972: Air Pollution Across National Boundaries. Stockholm.