Vol. 11. pp 57% 588. Pergamon Press Ltd. 1980 Printed in Great Britain
J Aerosol Sci.
ATMOSPHERIC
AEROSOLS
AND
GLOBAL
CLIMATE
RUPRECHT JAENICKE Institute for Meteorology, University Mainz, West Germany
Abstract - The present day knowledge is summarized and the atmospheric aerosols are given in model size distributions. The optical properties relevant for climatological considerations are given. The most recent data about the stratospheric aerosol are given and the correlation with volcanic activities is discussed. Claimed man-made influences upon the global aerosols are critically discussed. While the influences of the stratospheric aerosol on global temperature seem to be confirmed the influence of the tropospheric aerosol remains speculative.
1. I N T R O D U C T I O N Two experiences should make it clear even to the nonprofessional observer, that the atmospheric aerosol might have an influence on the climate. Gore (1979) shows an impressive photograph of an approaching desert dust storm darkening the sky in the desert. Under these circumstances sun radiation cannot penetrate to the earth's surface and cooling and temperature decrease will follow. This is a direct consequence of the presence of dusts or aerosols. The second experience is a little bit more difficult to understand in its links to the atmospheric aerosols. The few visitors to the North American Yellowstone Park in wintertime have observed, with surprise, that most hot springs do not have cloud covers, despite the water supersaturation which should be present over these hot water surfaces in the extremely cold air. With the ignition of an ordinary match and the smoke it releases, the hot lakes suddenly become foggy. The fog drifts away as the match extinguishes and the air is fogless again. It can be explained as follows. Even in the presence of ample water vapor, clouds are formed only if aerosols are present. In the clean winter air of the Yellowstone Park, aerosols are present only in insufficient quantities and thus clouds are not formed. Only the smoke from a match triggers the formation of a cloud. These two experiences indicate the influence which the aerosol has on the radiation budget of the atmosphere and thus on climate. If the aerosol of the atmosphere is changed, the extinction of the sun radiation will be changed through the aerosol, directly, and indirectly through the change of clouds. In the following it will be shown how the atmospheric aerosol interacts with the radiation budget and thus our climate. Figure 1 shows, in a rough drawing, what the atmospheric aerosol might do to the incoming solar radiation. Two large aerosol bodies, the stratospheric and the tropospheric aerosol will weaken the solar radiation through scatter and absorption of energy. Absorption generally produces a warming of the aerosol body, while scatter returns part of the energy to the upward direction. These are the direct effects of the aerosols. Clouds are formed only in the presence of water vapor and aerosol particles, acting as cloud condensation nuclei (CCN). Because of the low concentration of water in the stratosphere (Heicklen, 1976) clouds are only formed under special circumstances in some rare cases (Stanford, 1977). Similar to aerosols, clouds absorb and scatter sun radiation. In Fig. 1 only the scatter is considered, because the absorption in cloud droplets is mainly caused by the CCN. These particles are present whether a cloud is formed or not. Thus the absorption of energy in clouds is considered to be part of the aerosol properties. However scatter is enhanced considerably through clouds. The scatter of the incoming radiation is described as albedo, whether for the whole system (planetary albedo) or for the individual subsystems (cloud-, aerosol-, and earth surface albedo). The scatter and absorption of radiation in aerosols and clouds depends on the optical properties of the bulk material, the size of the particles, and the concentration in the air. 577
578
RUPRECHT J 4ENI('KI!
\\,
~,
!i
\k /
/*'
STRATOSPHERIC
AEROSOLS
SCATTER
ABSORPTION
AND
TROPOSPHERIC AEROSOLS
CLOUDS
SCATTER + ABSORPTION
SCATTER
Fig. 1. Schematic drawing of the effects of aerosols on the radiation in the atmosphere. The stratospheric aerosol and the tropospheric aerosols scatter and absorb radiation. As an indirect effect in the troposphere the properties of clouds are changed. This effects more the scattering properties, because the absorption mainly occurs in the cloud condensation nuclei, present with or without clouds.
2. T R O P O S P H E R I C AEROSOLS People working in air chemistry usually consider the aerosol as a sum of handy particles, easy to deal with. People working with aerosols usually regard gaseous species as very handy, describable with only a few parameters like concentration, distribution in time and space, reaction constants. All these parameters are needed for the aerosol too. In addition it has been shown to be impossible to describe the aerosol sufficiently with one single concentration parameter (Jaenicke, 1978c). Within the troposphere, aerosol particles vary in radius from some 10 nm to several 100/~m. Of course, not all particle sizes are present in equal quantities, the aerosol exhibits a size distribution. This is of greatest importance, because only some of these particles do interfere with the radiation field. So practically no significant scatter can be observed from the so called Aitken particles smaller than 0.1/~m in radius. On the other hand, the largest particles of several hundred pm do scatter mainly in the forward direction, but their concentration is so low, that their influence mostly can be neglected. Within the troposphere, the aerosol varies widely from very clean locations, like the polar regions or the ocean areas or the upper troposphere, to places on the continents and in highly polluted areas. From a practical point of view however, we can distinguish 3 rather uniform aerosols on a global scale (Junge, 1963): continental aerosol, maritime aerosol, background aerosol. The continental aerosol is formed from sources on the continents and is confined in height to the average altitude of clouds ( ~ 5 km). As we shall see, cloud droplets form only around cloud condensation nuclei, which are aerosol particles with suitable properties. This way, particles are removed from the aerosol and only the residual aerosol penetrates the cloud layer to higher elevations (Dinger et al., 1970). The continental aerosol spreads out over adjacent oceans only as the aerosol residence time permits it. Over the oceans, a different aerosol is formed, depending on the source. The sea salt particles produced are extremely suitable as cloud condensation nuclei, thus they do not penetrate the average cloud layer (3 km) to higher altitudes. The remaining parts of the troposphere are filled with the rather uniform background aerosol. This aerosol, in its larger particles, is thus mainly an aged continental aerosol. As we will see in the discussion of the residence time, only a minor fraction of the smallest particles in this aerosol can originate from the continents. The majority must be produced within the background aerosol through transformations from the gas phase. Through atmospheric
A t m o s p h e r i c aerosols and g l o b a l climate
579
turbulence, the background aerosol is mixed with seaspray and is included in the maritime aerosol. If mixed with the continental aerosol, the continental aerosol only is 'diluted'. As a rough estimate, the continental aerosol fills 15Yo of the troposphere, the maritime aerosol 20% and the background 65%. Figure 2 shows model particle size distribution in these aerosols as derived from surface bound measurements. These measurements are not very numerous. Especially, the background aerosol is only covered in subsiding air masses. The presentation in this figure is different from previous ones, but not in opposition. For each aerosol number, surface and volume distributions are shown. This is done, because from a climatology point of view the knowledge of all three properties is required. The number distribution is important for all processes with condensation of water vapor, because some of the particles act as cloud condensation nuclei (Junge et al., 1971). In addition most other aerosol properties can be derived from the number distribution on the assumption of spherical aerosol particles. The surface or cross-section of aerosol particles is required for optical processes, like the scatter of radiation. In addition condensation, chemical surface reactions, and electrical conductivity of the air are influenced. Volume or mass distribution becomes important for the absorption of radiation and the chemical composition of the aerosol. The 3 distributions number, surface, and volume in Fig. 2 can easily be distinguished because the center of gravity for the number is in the Aitken range, for the surface in the radius range 0.1 to 1/~m, for the volume above 1/~m in radius. It can clearly be seen that in the number distribution only two aerosols can be distinguished, the continental and the background aerosol. Seaspray particles are very few in number and thus of minor importance. Continental and background aerosol differ greatly. In the surface distribution, the maxima of the 3 distributions are rather close together around 0.3/xm in radius. The distributions differ greatly in the Aitken range and above 1/~m in radius. In this range continental, maritime and background aerosol can be clearly distinguished. In the volume distribution the maxima vary over a range of a factor 10 for all 3 distributions. The variation in the Aitken range is small. It can clearly be seen, that seaspray affects only the range above several tenth micrometer with the center of gravity around several micrometer in radius. SURFACE
10 4_ NIIMA~R
..................
I lO-i~ •
~IFIIiii~ic
103
. i0-u
10-12
10 2`
101 . 10 -5
10-13
10-z
10-1 RADIUS
10 0 r,,um
101
10 2
--
Fig. 2. M o d e l size d i s t r i b u t i o n s of a t m o s p h e r i c a e r o s o l s : - . . r e m o t e c o n t i n e n t a l , maritime, - - b a c k g r o u n d . The h a t c h e d area represents the s e a s p r a y w h i c h a d d s to the b a c k g r o u n d aerosol to form the m a r i t i m e aerosol. T h e n u m b e r d i s t r i b u t i o n is given in dN/dlg r, cm - 3, the surface d i s t r i b u t i o n in dS/dlg r, cm 2 cm - 3, the v o l u m e in dV/dlg r, cm 3 c m - 3. In the n u m b e r d i s t r i b u t i o n the m a r i t i m e a n d b a c k g r o u n d aerosol h a r d l y can be distinguished, because the s e a s p r a y m a k e s u p only some 20 c m - 3. T h e h o r i z o n t a l bar i n d i c a t e the radius r a n g e m o s t i m p o r t a n t for c l i m a t o l o g i c a l c o n s i d e r a t i o n s as e x p l a i n e d in the paper. AS 11:5/6 J
580
RLrF'RE('HT J ~,ENI('KE
It should be noticed, that the aerosol in all its 3 properties - number, surface, and volume covers 5 orders of magnitude in radius from 10- ~ #m to several 100/~m. From a climatology point of view, the important range can be condensed. Warner (1968) showed that in more than 97°/£ of all observations, the supersaturation in clouds was limited to 13.0. Following Junge et al. (1971) for a reasonable portion (10-100~o) of water soluble material present in the aerosol, this supersaturation activates only particles larger than several 0.01 #m as cloud condensation nuclei. This is the lowest particle size important for climatological considerations. For the upper particle size, one has to study the residence time of tropospheric aerosols. Jaenicke (1978a) has compiled the estimates available and concluded that the residence time of the aerosol should be seen as a function of particle size. Small particles are removed rather rapidly (in hours) because of their large mechanical mobility and their attachment to larger aerosol particles. The giant aerosol particles on the other side are removed quickly (in hours) because of the large sedimentation rate. The longest residence time is observed for particles of 0.1 #m to several micrometers in radius. Their residence time is limited mainly through wet removal and at the earth's surface is of the order of 2-5 days. In a rough estimate, Jaenicke (1980) could show, that because of the limited residence time, particles of 100 l~m are only of local influence, while particles of 10/~m can be transported on a regional scale. Particles smaller than several micrometer are transported intercontinentally. This discussion shows that from the point of view of climatology the important size range of the aerosol is condensed to roughly two orders of magnitude (several hundredth to several micrometer) or the range covered from the surface distribution in Fig. 2. This range is indicated with a horizontal bar. This range and the considerations about the residence time are of greatest importance if aerosol emission and production rates are discussed. Twomey (1977a) has stressed this fact. Especially, he points out that natural and man-made sources produce particles of different sizes. Indeed this is an important limitation, if data are gathered. We would like to see the production rates in the particle size range given above, but only few authors state the particle size range for their estimated source strength. Table 1 compiles the data available at present. Since the big effort in 1971, not much has been added. For this table, values have been scaled down to the interesting particle size range, if particle sizes are given in the original literature. This is most dramatic in the case of sea salt particles, thus only 180 Tg per year are derived from S mic (1971). For the total natural aerosol, Jaenicke (1980) has shown that a discrepancy in production of easily 1 : 1000 exists, if particles being subject to long range transport are compared with all particles. Sch/itz (1980) could show this for mineral aerosols too. Table 1 shows nicely that the data about man-made production rates are surprisingly close together for a large number of authors. The uncertainties in natural production are much larger. The contribution of man to the production of aerosols is estimated in the range of 5-30~o of the natural production. However, this figure should not be overinterpreted. From the data in Table 1, an average from all authors of 289 Tg per year can be calculated for the man-made contribution. On the other side, the natural contribution is 2237 Tg per year with a standard deviation of 53~o or 1179 Tg per year. That means, the estimated man-made production is much smaller than the uncertainties in the estimated natural production rates. From such figures, therefore, we cannot expect hints with some statistical significance about an increase of man's contributions to the atmospheric aerosols in recent years. Such hints can only be expected from measurements of the atmospheric aerosol using typical manmade tracers. For climatological considerations the optical properties of the atmospheric aerosol are important. The optical bulk properties are described with the complex index of refraction. The imaginary part of this index describes the absorption properties of the aerosol. Eiden et al. (1975) have compiled all data available up to that date. Within the wavelength range of the sun radiation, the real part n of this index is rather uniform around n = 1.53. This is not surprising, because a large variety of substances in the aerosol have a real index of refraction in this range. Minerals like rock salt (n = 1.544), calcspar 07 = 1.658), quartz (n = 1.544) show rather uniform data, but also substances like ammonium sulphate (n = 1.53) and
Table 1. Particle production summarized from various papers and scaled down to the size range indicated in Fig. 2, if possible. PJ
HD
SM
Others
500 250 25 5
1095 7-365 4 146
180" 60-300* 15-90* 3-150
1000-2000 60 360 4
780
1252-1610
258-720
1144-2444
335 60 75
37-365 600-620 182-1095
130 200 140-700 75-200
160 154-220
470 1250
819-2080 2071-3690
345-1100 603 1820
1319 2463-3763
Particles Converted sulfates Converted nitrates Converted HC
30 200 35 15
37- 110 110 23 27
6-54* 130 200 30-35 15-90
54- 126
Subtotal Total man-made Total man-made
250 280 22~o
160 196 270 9 - 7'Yo
175 325 181-379 30-21 'V,
270 324-396 13-11 '?/,~
Natural sources Seasalt Mineral dust Volcanoes Forest fires Biological material Subtotal Converted sulfates Converted nitrates Converted HC Subtotal Total natural
80
>
o ©
Man-made sources
Total natural Sources: PJ - Peterson et al. (1971), HD - Hidy and Bruck (1971), SM - Smic (1971). Others - from Bach (1976) including Jaenicke (1978b), Schiitz (1980) Values in Tg per a year or million of ton per a year.
* Those SM values are scaled down to particle radius smaller than 3 l~m.
,.Ji
5~2
RUPRE('HT JAENICKI"
sulfuric acid (n = 1.44), Ammonium sulphate seems to be very important. Charlson et al. (1978) could show how closely the sulphate content is correlated with optical properties, an indication that the particle size range of interest to us is mainly determined from ammonium sulphate (Jaenicke, 1978c). The absorption coefficient obviously varies much more dramatically (Eiden et al., 1973) in the range n~c = 0.004-0.07. A value of 0.03 seems to be a good average (King, 1979). Of interest for energy considerations are the optical properties of the aerosol in the wavelength range of the terrestrial radiation, the radiation emitted from the earth's surface. In this far infrared the real part as well as the imaginary part of the complex index of refraction show selective character. Both have rather large anomalies around the wavelength 2 = 9 #m in the so called window range, n increases from 1.2 to 2.0 (Volz, 1972), and n~c from 0.01 to 0.3 (Fischer, 1975). Grassl (1973) compares the effect of the aerosol with that of atmospheric water vapor. Carlson et al. (1980) compiled the data for desert aerosols, which are comparable to the data given above. 3. T H E S T R A T O S P H E R I C A E R O S O L The atmospheric aerosol is produced from surface sources like the continents and the oceans and from gas-to-particle conversion within the aerosol body itself. This distribution of sources together with the transport, ageing, and residence time of the aerosol produces a certain spatial distribution. For the troposphere, this spatial distribution was discussed earlier. The situation in the stratosphere is different from the troposphere for several reasons. Let us first discuss the vertical distribution of the aerosol. The most recent measurements of Aitken nuclei (r < 0.1 #m) made by Podzimek et al. (1977) indicate a continuous concentration decrease from several hundred particles per cubic centimeter to some ten particles per cubic centimetre around 20 km altitude. Earlier, Junge et al. (1961) measured around 1 cm-~. To my knowledge this does not necessarily mean an increase in concentration over the years, because proof is lacking that both measurements are comparable. In their classical paper, Junge et al. (1961) could show the presence of a layer of large particles around 20km. This was confirmed in later years from a number of direct measurements (Friend, 1966 ; Miranda et al., 1973 ; Farlow et al., 1979 ; Hofmann et al., 1979). This layer of particles larger than 0.1 ~m in radius is located around 18 km over the equator and 12 km at 80°N. The concentration of large particles in this layer is of the order of several particles per cubic centimeter, but varies in time. In Fig. 3, a model of the size distributions of stratospheric aerosols is presented, based on the compilation in Rosen et al. (1978). Distributions by number, surface and volume are given as in Fig. 2. In contrast to any tropospheric aerosol, the stratospheric aerosol shows that the maxima of the number-, surface-, and volume distributions are very close together, around 0.25/~m in radius. This is an indication of a very monodisperse aerosol. This is in agreement with what must be expected from considerations about the residence time. In general, the residence time of aerosol particles is controlled from 3 factors: coagulation, sedimentation and wet removal. Coagulation is controlled by the particle size distribution and the particle concentration, sedimentation from the average altitude of the aerosol, and wet removal from the presence of water. Figure 4 shows the residence time as function of particle size using the equation of Jaenicke (1978a) and the most recent estimates from Pruppacher et al. (1978) and Hofmann et al. (1979). The equation of Jaenicke (1978a) contains a term for wet removal, which can be related directly to the water content of the atmosphere, and residence time curves for various altitudes can be calculated. In Fig. 4 they are compared with the most recent estimates for "suspended particulate matter in its entirety" (Pruppacher et al., 1978). In contrast to that, the residence time in Fig. 4 must be understood as that of individual particles. Figure 4 indicates that particles around 0.3 #m definitely have the longest residence time in the absence of wet removal, as is the case in the stratosphere. Their residence time is ten times longer than that of 0.1 or 1 #m particles. The picture is not changed drastically if we modify the coagulation branch and the sedimentation
Atmospheric aerosols and global climate 10-12
/~0 -7
101 PODZIMEK
~t ol
(1977)
1
T 10o JUNGE et
q, E o
z
583
al {1961)
~
/
/ /
SURFACE ~ - ~
[I
10-I
10-2 10-3
I 10-2
I 10-1 RADIUS
10-~3 ? E u u
VOLUME
10-I~ >
I 100
10-15 10I
r,)um
Fig. 3. The size distribution of the stratospheric aerosol as number distribution (dN/dlg r, cm-3), surface (dS/dlg r, cm 2 cm -3) and volume (dV/dlg r. cm 3 cm-3). The maxima are very close together around 0.25/am in radius to contrast to the tropospheric aerosols with a spread of up to 3 orders of magnitude (an immediate consequence of the residence time ?). The distribution in the Aitken range is still uncertain, but should tend toward the Junge et al. (1961) data - if the residence time in Fig. 4 is correct.
b r a n c h w i t h i n r e s o n a b l e limits. T h i s r e s i d e n c e t i m e o f r o u g h l y 1 y e a r m e a n s t h a t r e g a r d l e s s o f t h e p a r t i c l e size p r o d u c e d in or t r a n s p o r t e d t o t h e s t r a t o s p h e r e , o n l y p a r t i c l e s a r o u n d 0 . 3 / ~ m r e m a i n t h e r e for a l o n g e r p e r i o d . W e d i s c u s s e d earlier the i m p o r t a n c e o f t h e s e p a r t i c l e s for climatological considerations.
2yr ~ \
108
1 yr
STRATOSPHERE
10 7 1m 106 .
l t~
TROPOPAUSE MIDDLE
Iw
T R O P O S P H E R ~
l0 S.
~ 10~. zw 103. o
//
~ 102. 101 .
I0 -t'
113-9
I0"2
10-'
I0 0
F~ADIUS r , # m --
ESTIMATES CALCULATED
AFTER
101
PRUPPACHER
HOFMANN
I0 2
,," ct o!
(1978)
et al ( 1 9 7 9 }
Fig. 4. Aerosol residence times for individual particles. While the function "below about 1.5 kin" is obtained from a large number of estimates (Jaenicke, 1978a), the free atmosphere functions are obtained by reducing the wet removal fraction depending on the water content of the atmosphere. These curves are in rather good agreement with the most recent estimates from collected data as indicated.
584
RUPRECHT J ~I!NI('KI
The source and the nature of these stratospheric aerosols was discussed in length (Rosen, 1971; Friend et al., 1973; Junge 1974; Lazrus et al., 1977). Most probably these particles consist of ammonium sulphate or sulfuric acid or some combination. The optical properties of the stratospheric aerosol are thus comparable to those of the tropospheric aerosol. Through volcanic activity, particles and gaseous precursors are injected directly into the stratosphere. In addition, in non-volcanic periods, we have to assume a constant flux of gaseous precursors of particles from the troposphere into the stratosphere (Junge, 1974; Hofman et al., 1979). From this point of view, a constant monitoring of the dust layer is of greatest importance. Figure 5 shows direct measurements of particles collected on filters (Castleman et al., 1974) or measured with an optical particle counter (Hofmann et al., 1979). The data are plotted for a fixed altitude (19 km), which does not necessarily reflect the altitude of maximum particle concentration (McCormick et al., 1978 ; Hofmann et al., 1979). Both sets of data have been adjusted roughly for comparison of particle mass per standard cubic meter (STP) of air. Both Castleman et al. (1974) and Hofmann et al. (1979) show a good correlation between stratospheric particle load and volcanic activity. Lidar measurements, using the backscatter of stratospheric aerosols became available in recent years (McCormick et al., 1978). Apart from their problem of being calibrated against an unknown particle concentration near the tropopause (assumed to be zero), they truly monitor the stratospheric dust layer. The decrease in stratospheric particle load after the October 1974 eruption of Fuego is clearly indicated. From January 1975 to January 1976 the signal was reduced to one third, indicating a stratospheric aerosol residence time of 1 year. Similar values are calculated in Hofmann et al. (1979) and are included in Fig. 4. 4. A E R O S O L AND C L I M A T E It was indicated above that the aerosol acts directly and indirectly upon the radiation budget of the atmosphere and thus upon climate. Grassl (1979) discusses the direct effects of the aerosol on radiation. The influence on the planetary albedo of the solar radiation depends mostly on the imaginary part of the complex index of refraction. In earlier papers (summarized in Bach, 1976) the absorption was overestimated. More realistic assumptions - as discussed above - lead to an overall increase in planetary albedo in the absence of clouds with increasing particle concentration. Over high surface albedo areas, the planetary albedo might be reduced. The consequence is a reduced heating rate of the earth's surface. But the aerosol itself is heated, because of its absorbing properties, in the order of 1 K per day, depending on the absorption coefficient. Harshvardhan et al. (1976) calculated a temperature increase up to 9 K for the stratospheric aerosol due to energy absorption. This additionally absorbed energy is emitted from the aerosol in the region of the terrestrial radiation. Grassl (1974) calculated a cooling rate of roughly 15~o in addition to the standard atmosphere cooling rate of 0.6 K d a y - 1 at the earth surface and 0.1 K d a y - 1 at 750 mbar. Because of the uncertainties in aerosol properties and spatial distribution, no estimates on the total effect of the aerosol are given. The aerosol acts indirectly on the climate through the formation of clouds. It was already discussed that a fraction of the aerosol particles act as cloud condensation nuclei. This is mainly effected from the chemical nature of the aerosol particles. In general, atmospheric aerosols and atmospheric water are connected in two ways. At relative humidities below 100~o, aerosol particles grow through absorption of water. At humidities above 100~, the formation of clouds is influenced. In studying the water uptake of aerosol particles, Winkler (1973) has improved the concept of the mixed nature of atmospheric aerosol particles. The water uptake mainly changes the scattering properties of the aerosol. H/inel et al. (1978) has calculated that for 70~0 r.h. 3 times more liquid water is attached on an anthropogenic influenced aerosol than on a remote continental aerosol. In clouds the aerosol, and thus the cloud condensation nuclei (CCN), act in three ways as Grassl (1978, 1979a) points out. It determines the optical depth, the shape of the scattering
Atmospheric aerosols and global climate
585
Z Z
u_
co 25 i --
"a
j
CASTLEMAN et ~
f
u_
~)974)
/r ~_
HOFMANN et at (I979)
125°[
•
~',20 i :~
-200 E
J
o
.150
a_
[ 5
x x
-50
×x ,i • . 01962
P
53 64
I.
65
I
I.
66 6.7 ~ 66
I.
69
I
70
E
7!
.72
17 3
;~
74
75
XxXxx
x×~x'~ xx?;Xl
76 .77 78
80
i
Fig. 5. The stratospheric aerosol content around 19 km altitude. A good correlation can be seen with volcanic eruptions, as indicated (both for the Northern hemisphere). function, and the absorption of radiations in the clouds. Typical CCN-concentrations are 400 cm - 3 over continents, 100 c m - 3 over oceans slightly decreasing at higher altitudes. If we discuss the influence on climate, we have to look on the effects of increased C C N concentration and the residence time of CCN. An increase in CCN-concentration usually results in increased optical depth of the clouds, flattening of the scattering function (Grassl, 1979) and increased life time of clouds (Smic, 1971). Twomey (1977c) shows that increased optical thickness usually first increases the cloud albedo, but with further increasing optical thickness, the cloud albedo decreases. The direction of the effect of increasing pollution thus is still open to question. To my knowledge, no studies are available about the change of cloud lifetime due to pollution and the immediate effects on planetary albedo. Smic (1971) discusses the possibility that the cloud cover might increase due to aerosol pollution, but Twomey (1977b) points out that the water content of clouds and thus the presence of clouds is determined from the dynamics of the atmosphere. It would go beyond the frame of this paper, to discuss a change in dynamics of the atmosphere because of changed aerosol properties. Concerning the atmospheric residence time for CCNs, we must keep in mind their size range of 0.01-0.1 #m (Junge.et al., 1971). Following Fig. 4, a residence time slightly shorter than the m a x i m u m residence time of aerosols must be expected. This means, any influence of cloud formation from increased CCN-concentration must probably be expected more on a regional scale than on a global scale. 5. O B S E R V E D A N D C L A I M E D T R E N D S IN T H E A E R O S O L R E L E V A N T TO T H E G L O B A L C L I M A T E Do we have at present any indications that the aerosol has changed on a global scale? From our previous discussion it is obvious that' any change of climatic significance must occur in the aerosol size range with longest residence time and with suitable optical properties. C o b b et al. (1970) report a decrease in atmospheric electrical conductivity over the N o r t h Atlantic since 1900. They stated a twofold increase in fine particle content. A more detailed calculation based on Jaenicke (1978c) results in Fig. 6, where the size distributions of Fig. 2 are compared to calculated distributions from the turn of the century. The change hardly influences the total particle concentration and the total particle mass, but mainly the particle surface. This property is not monitored, but is of greatest importance for climatological considerations. The effect of this aerosol surface increase on the atmosphere has not yet been estimated.
586
RUPRECHT
JAENICKE
I
]0 C
SURFACE
NUMBER
VOLUME
:
10-~o
10-" I
10 3 . /" /
'7 E w
/ I ?
J
S _~ i0 L
//
,Z 10-~2
z i/
I !
1010-3
10-2
I
10-I RADIUS r,,um
i0°
i01
10-13 10z
I
Fig. 6. Comparison of present day tropospheric aerosol distributions (see Fig. 2) with those calculated for the turn of the century in the maritime aerosol ( ) (after Cobb et al., 1970). The horizontal bar indicates the radius range effecting the electrical conductivity of the atmosphere. The change was measured over the North Atlantic. The same fraction was subtracted from the remote continental aerosol ( - - - ) , because this particle size range exhibits the longest residence times (Fig. 4) and therefore might influence both aerosols.
Eiden (1979) has calculated for a maritime aerosol in the surface layer of the atmosphere a solar heating rate of 3.5 K d a y - 1. If we assume the heating rate being fully effective in the discussed size range, the heating rate around the turn of the century would be 1.5 K d a y - 1 for a maritime aerosol only. This is a crude estimate and we cannot decide if it comes even close to the truth. Prospei'o et al. (1977) have measured an increase in the mineral dust content transported over the North Atlantic from the Sahara. They claimed to observe the first proof of a manmade influence on climate because of a possible connection with the Sahelian drought. It could be shown (Jaenicke et al., 1978b) that an increase did not occur. Depending on the global circulation, the dust transport from the Sahara is meandering, so at any given location the mineral dust content varies widely. Twomey (1977a) reports about CCN concentration measurements between 1968 and 1973 on the Australian east coast. During this period of time obviously the scatter of the measurements increased, but the average remained more or less unchanged. This period might be too short for any further conclusions. Turbidity measurements from the surface of the earth use the direct radiation of the sun. Thus they contain the variations of the aerosol and possible changes of the sun radiation. Jaenicke et al. (1978a) calculated the atmospheric turbidity for the period 1890-1940 and could show a good correlation with the volcanic dust veil index (DVI) of Lamb (1970). It is surprising to see that atmospheric turbidity in volcanic active periods is mainly determined from the stratospheric dust rather than from the aerosol in the boundary layer. This is even true for the anthropogenic influence upon the aerosol in a major city like Copenhagen. Similar conclusions have been drawn from Hammer (1977). He studied the variations in the electrical conductivity of dated ice cores buried in Greenland ice domes. The observed variations are closely correlated to the above mentioned DVI. These measurements cover the period 1770-1970. He argues that if the stratospheric aerosol consists of water-soluble sulfur compounds, their removal must be mirrored in the precipitation and thus buried in the ice cores. Obviously his assumption was correct. For the industrial period he did not find any increasing trend. This is not necessarily in contrast to the observation of Cobb et al. (1970) as
Atmospheric aerosols and global climate
587
Fig. 6 reveals, because the conductivity is proportional to the aerosol volume (or mass) in the ice. Thus the change reported from Cobb et al. (1970) is in another size range as the nonchange reported from H a m m e r (1977). It is only recently (Hoffmann et al., 1979) that anthropogenic influence on the stratospheric aerosol has been claimed. A comparison of two volcanic quiet periods 20 years apart (1960-1980) seems to give evidence of an increase of 9 ~ per year. However, we must keep in mind that the stratosphere had probably not yet returned to a background level. So, also, the data of H ofmann et al. (1979) show no indication of the Augustine eruption early in 1976. In addition, it is not certain if the earlier measurements of Junge et al. (1961) are of comparable quality to support the conclusion of a man-made increase in the stratospheric aerosol concentration. The volcanic variation of the stratospheric dust layer has always attracted researchers to look for connection with climatic variations. There have been papers neglecting any influence (Landsberg et al., 1974), but more recently (Mass et al., 1977; Robock, 1978) we see increasingly evidence of a connection between global temperature and volcanic dust. Robock (1978) used a numerical climate model to test the influence of various parameters on the global temperature. While he did not find any significant correlation of temperature and anthropogenic parameters like heat, CO2 and aerosols, he found a good correlation of r = 0.92 between the DVI and 5 year averages of global temperature. The general shape of the temperature observations was well simulated, while details remained undetected. So the cooling after the late 1800-eruptions could be explained as well as the general warming in the eruption free period after 1920. Robock (1978) concludes that volcanic dust seems to have been an important cause of climate change during the past 100 years. Dansgaard (1980) showed for an extended period including the little ice age a good correlation between the ice core electrical conductivity and a Northern hemisphere temperature index. 6. C O N C L U S I O N S During the past years all fields of research connected to the question "Aerosol and Climate" have made considerable progress. Models of the aerosol size distribution and spatial distribution are at hand, first optical properties have been measured with actual samples and models are available for calculating heating and cooling rates. We still lack a general model, taking into account all effects simultaneously and resulting in quantitative estimates of global cooling or heating rates if the aerosol is changed due to man's activities. This paper shows how important it is to monitor relevant properties of the aerosol and what properties are relevant for the global climate. Most important seems to be the monitoring of the stratospheric aerosol of volcanic origin. For short term climate prediction it seems necessary to develop prediction models for volcanic emissions. Acknowledgement--This paper is based on a lecture given at the NATO Advance Study Institute on "Climatic Variations and Variability: Facts and Theories" and will be published in the forthcoming proceedings, Edited by A. Berger, by Reidel Publishing Company, Dordrecht, Holland.
REFERENCES Bach, W. (1976) Rev. Geoph. Space Phys. 14, 429. Carlson, T. N. and Benjamin, S. G. (1980) J. Atmos. Sci. 37, 193. Castleman, A. W., Munkelwitz, H. R. and Manowitz, B. (1974) Tellus 26, 222. Charlson, R. J., Covert, D. S., Larson, T. V. and Waggoner, A. P. (1978) Atmos. Environ. 12, 39. Cobb, W. E. and Wells, H. J. (1970) J. Atmos. Sci. 27, 814. Dansgaard, W. (1980) In Climatic Variations and Variability: Facts and Theories (Edited by Berger, A.) Reidel, Dordrecht. Dinger, J. E., Howell, H. B. and Woijeciechowski, T. W. (1970) J. Atmos. Sci. 27, 791. Eiden, R. and Eschelbach, G. (1973) Z . f Geophys. 39, 189. Eiden, R. (1979) Proc. Man's Impact on Climate (Edited by Bach, W., Pankrath, J. and Kellogg,W.) p. 115, Elsevier, Amsterdam. Farlow, N. H., Ferry, G. V., Lem, H. Y. and Hayes, D. M. (1979) J. Geophys. Res. 84, 733. Friend, J. P. (1966) Tellus 18, 465. AS H : 5 / 6
K
588
RUPRECHT JAENICKE
Friend, J. P., Leifer, R. and Trichon, M. (1973) d. Atmos. Sci. 30, 465. Fischer, K. (1975) Appl. Opt. 14, 2851. Gore, R. (1979) Nat. Geogr. 1$6, 586. Grassl, H. (1973) Tellus 25, 386. Grassl, H. (1974) Beitr. Phys. Atmos. 47, Grassl, H. (1978) Strahlun 9 in getriibten Atmosphiiren und in Wolken, Hamburger Geophysikalische Einzelschriften 37. Grassl, H. (1979) Proc. Man's hnpact on Climate (Edited by Bach, W., Pankrath, J. and Kellogg, W.) p. 229, Elsevier, Amsterdam. Hg.nel, G. and Bullrich, K. (1978) Beitr. Phys. Atmos. 51, 129. Hammer, C. U. (1977) Nature 270, 482. Harshvardhan and Cess, R. D. (1976) Tellus 28, 1. Hidy, G. M. and Brock, J. R. (1971) Proc. 2 int. Clean Air Congr. 1088. Heicklen, J. (1976) Atmospheric Chemistry, Academic Press, New York. Hofmann, D. J. and Rosen, J. M. (1979) On the Background Stratospheric Aerosol Layer. NSF Report No. AP-56. Jaenicke, R. (1978a) Ber. Bunsenges. Phys. Chem. 82, 1198. Jaenicke, R. (1978b) Pageoph. 116, 283. Jaenicke, R. (1978c) Atmos. Environ. 12, 161. Jaenicke, R. (1980) Ann. New York Acad. Sci. 338, 317. Jaenicke, R. and Kasten, F. (1978a) Appl. Opt. 17, 2617. Jaenicke, R. and Sch~tz, L. (1978b) d. Geophis. Res. 83, 3585. Junge, C. E., Chagnon, C. W. and Mason, J. E. (1961) d. Meteor. 18, 81. Junge, C. (1963) d. Rech. Atmos. 1, 185. Junge, C. and McLaren, E. (1971) d. Atmos. Sci. 28, 382. Junge, C. (1974) Proc. Int. Conf. Structure, Composition and General Circulation of the Upper and Lower Atmospheres and Possible Anthrop. Perturbations, Melbourne. Vol. I, 85. Lamb, H. H. (1970) Phil. Trans. R. Soc. 266, 425. Landsberg, H. E. and Albert, J. M. (1974) Weatherwise 27, 63. Lazrus, A. L. and Gandrud, B. W. (1977) Geophys. Res. Letters 4, 521. Mass, C. and Schneider, S. H. (1977) d. Atmos. Sci. 34, 1995. McCormick, M. P., Swissler, T. J., Chu, W. P. and Fuller, W. H. (1978) Y. Atmos. Sci. 35, 1296. Miranda, H. A.. Dulchinos, J. and Miranda, H. P. (1973) Stratospheric Balloon Aerosol Particle Counter Measurements. Final Report AFCRL Contract No. F 19628-73-C-0138. Peterson, J. T. and Junge, C. E. (1971) Man's Impact on the Climate. (Edited by Matthews et al.) p. 310, MIT Press, Cambridge, Massachusetts. Podzimek, J., Sedlacek, W. A. and Haberl, J. B. (1977) Tellus 29, 116. Prospero, J. M. and Ness, R. T. (1977) Science 196, 1196. Pruppacher, H. R. and Klett, J. D. (1978) Microphysics of Clouds and Precipitation. Reidel, Dortrecht. Robock, A. (1978) d. Atmos. Sci. 35, 1111. Rosen, J. M. (1971) d. appl. Meteon. 10, 1044. Rosen, J. M., Hofmann, D. J. and Singh, S. P. (1978) J. Atmos. Sci. 35, 1304. Schiitz, L. (1980) Ann. New York Acad. Sci. 338, 515. SMIC (1971) Inadvertent Climate Mod!fication. Report of the Study of Man's Impact on Climate, MIT Press, Cambridge, Massachusetts. Twomey, S. (1977a) Atmospheric Aerosols, Elsevier, Amsterdam. Twomey, S. (1977b) Proc. Symp. Radiation in the Atmosphere. Garmisch-Partenkirchen, 171. Twomey, S. (1977c)d. Atmos. Sci. 34, t149. Volz, F. (1972) Appl. Opt. 11,755. Warner, J. (1968) J. Rech. Atmos. 3, 233. Winkler, P. (1973) d. Aerosol. Sci. 4, 373.