551.521.31 (Meteorological Office, Braeknell, Berks, E n g l a n d )
Absorption of Solar Radiation by Atmospheric Aerosol, as Revealed by Measurements at the G r o u n d 1 By
G. D. Robinson W i t h 2 Figures
Summary.~ R e c o r d e d values of global and diffuse solar r a d i a t i o n a~ the ground in cloudless conditions are c o m p a r e d w i t h values c o m p u t e d for a clean atmosphere, t a k i n g into account l%ayleigh scattering and a b s o r p t i o n b y t I 2 0 and Oa. The diffuse r a d i a t i o n allows an e s t i m a t e of the scattering b y aerosol; the global r a d i a t i o n allows an e s t i m a t e of the t o t a l a t t e n u a t i o n . Aerosol absorption is thus o b t a i n e d as the difference b e t w e e n the two estimates. A n a s s u m p t i o n of an angular d i s t r i b u t i o n for aerosol scattering is i n v o l v e d ; it is b r o a d l y justified b y appeal to aircraft observations. I n the course of t h e w o r k comparison is m a d e of t h e results of simple, single-scattering, a n d complex, multiple-scattering, c o m p u t a t i o n s of atmospheric transmission. F o r u r b a n atmospheres (Kew, near London, and Vienna) aerosol a b s o r p t i o n in excess of t h a t due to H 2 0 is found. F o r two island stations (Malta and Shetland) a b s o r p t i o n comparable w i t h t h a t due to H~O is found. F o r selected v e r y clear conditions at three high=level stations in Africa, and for m e a n conditions in the Antarctic, absorptions of two or three per cent are indicated. E x a m i n a t i o n o f records of global i l l u m i n a t i o n and of t h e direct solar r a d i a t i o n in t h r e e spectral bands at N e w suggests t h a t the t o t a l a t t e n u a t i o n and absorption b y aerosol are v e r y a p p r o x i m a t e l y non-selective there. These results are in b r o a d a g r e e m e n t - w i t h recent direct m e a s u r e m e n t s from aircraft. The m e t h o d is capable of wide application, as global and diffuse solar r a d i a t i o n are now recorded at m a n y stations. Zusammenfassung. l%egistrierte W e r t e der Global- u n d t t i m m e l s s t r a h l u n g a m B o d e n bei wolkenlosem laIimmel w e r d e n vergliehen m i t den Werten, die 1 D e d i c a t e d to Dr. W. MOtglKOFER on the occasion of his 7 0 t h birthday. 2*
20
G . D . ROBINSON:
filr eine reine Atmosph~ire unger Beriieksiehtigung der l%ayleigh-Streuung und der Absorption dureh t t 2 0 und O3 errechnet werden kbnnen. Die I-Iimmelsstrahlung gestattet eine Absch~itznng der Sgreuung du{'ch das Aerosol; die Globalstrahlung gibt eine Seh~tzung der gesamten Sehw~ehung. Die Absorption dutch das Aerosol wird daher als Differenz der beiden Abseh~tzungen erhalten. Fiir die Aerosolstreuung wurde eine Abh~ngigkeit yon der Sgreuriehtung angenommen; dies wird dureh Flugzeugbeobaehtungen gerechtfertigt. Es werden aueh die Ergebnisse aus einfachen (einmalige Streuung) und komplexen Berechnungsmethoden (mehrfaehe Streuung) der atmosph~rischen Durchlassigkeit verglichen. I n Stadtatmosphfiren (Kew und Wien) finder m an eine Aerosolabsorption, die die von H20 iibertrifft. An zwei Inselstationen (Malta und Shetland) ist die Aerosolabsorption yon der gleiehen Gr6Benordnung wie die yon H20. An ausgesuchten sehr klaren Tagen wxtrden an drei Hbhenstationen in Afrika Absorptionen yon zwei bis drei Prozent gefunden, ebenso unter durehschnigtlichen Bedingungen in der Antarktis. Die Untersuehnng der l~egisgrierungen d er globMen Beleuehtungsstiirke und der direkten Sonnenstrahlung in drei Spektralbanden in K ew ergab n~iherungsweise, dab dort die Gesamtsehw/~ehung und die Aerosolabsorption nieht selektiv sind. Diese l:Lesultate sind in guter Ubereinstimmung m i t neueren Messungen aus Flugzeugen. Die Methode kann eine weite Anwendung finden, da jetz~ Global- und tiimmelsstrahlung an vielen Stationen registriert werden. Rgsum~. Les valeurs enregistr6es de la radiation globMe aussi bien que du eiel au sol sent eomparges, dans le eas d'un eiel serein, avee les ehiffres ealeul~s pour une atmosphgre pure. Dans ee second eas, on a tenu eompte de la dispersion de Rayleigh et de l'absorption par t{20 et O3. Le rayonnement du eiel permet d'estimer la diffusion provoqu~e par l'a6rosol. La radiation globale permet en outre une estimation de l'exginetion regale du rayonnement. L'absorption par l'agrosol est obtenue par la diff6renee des deux estimations pr6e6dentes. On s'est servi d'une relation en fonetion de la direction du rayonnement pour ealeuler la dispersion due ~ l'a6rosol. Cette manigre de faire semble justifi6e au vu des observations d'avions. On compare en outre les r6sultats de la transmission de l'atmosphgre obgenus par des m6thodes de ealeuls simples (dispersion simple) ou complexes (dispersions multiples), Dans l'atmosphgre des villes (Kew et Vienne), l'absorption due ~ l'a6rosol d@asse eelle de la vapeur d'eau. A deux stations plae6es sur des ties (Malte et les Shetlands), l'absorption due ~a l'a6rosol est du m~me ordre de grandeur que eelle d'I-I20. On a enfin eonstatg, par des journ6es sp6eialement elaires, des absorptions de 2 ~ 3 % /~ 3 stations de montagne en Afrlque. I1 en va d'ailleurs de m6me en conditions normales dans l'Angaretique. Le d6pouillement des enregistrements de l'ingensig6 ]umineuse et de l'im solatton direet~ dans trois bandes speetrales a montr6, ~ Kew, clue la diminution gogale du rayonnement et l'absorption due s l'a6rosol n ' y song peu pr6s pas sdIeetives. Ces rgsultats correspondent tr6s bien ~ des mesures r6eentes faites .~ herd d'avions. La dire mgghode peut 6gre utitis6e sur une vaste 6ehelle vu que les radiations globale et du eiel song enregistr6es en de nombreuses stations.
Absorption of Solar Radiation by Atmospheric Aerosol
21
1. Introduction I n a previous paper with limited circulation (BLACKWELL, ELDRIDGE and ROBINSON [3]) m y colleagues and I examined the records of solar radiation and illumination taken in cloudless conditions at Kew Observatory and concluded t h a t they could be interpreted only b y postulating considerable absorption in addition to t h a t due to atmospheric gases. We attributed this absorption, which appeared to be present both in the visible and infrared portions of the spectrum, to aerosol. Our argument was simple. The radiation outside the atmosphere is known with sufficient precision for our purpose and the attenuation b y atmospheric gases is calculable. The calculated radiation at the ground is considerably greater than that observed; the missing energy must either be absorbed or be scattered upwards. The radiation scattered downwards is measured and is itself generally less than the missing component, so there must be either scattering upward considerably in excess of t h a t downward, or absorption. We considered an excess of upward scattering a most unlikely explanation, and attempted to measure the upward scattering b y aircraft-mounted instruments. Several flights in the neighbourhood of Kew Observatory gave no useful results because of instrumental failure or the appearance of cloud, but one was partially successful and is referred to later (section 8). Aircraft observations elsewhere of diffuse reflectivity in hazy cloudless eonditions--RoBI~so~ [16], ROACH [15J--indicated t h a t upward scattering from aerosol was small, though they were not sufficiently precise to give numerical values. Moreover ROACH [15] and several Russian observers (see e. g. KO~DRATIEV [11]) found direct, evidence of absorption considerably in excess of the computed absorption of atmospheric gases. I have therefore re-examined the material for Kew Observatory, making some changes in the computation of scattered radiation (see section 3) and have used the same methods on material from several other observatories in widely different geographical situations; the sources are listed in the Appendix. The data, except for some from Kew examined in section 7, were extracted from tabulations of hourly means of the records of four types of instrument: pyranometers recording total solar irradiance of a horizontal surface ("global radiation"--G), and the irradiance of a horizontal surface due to scattered sunlight ("diffuse radiation"--D), a photometer recording the anMogous global illumination, and pyrheliometers recording the irradiance due to direct sunlight of a surface normal to the solar beam ("solar i n t e n s i t y " - - / ) . Details of these instruments and of methods of operation m a y conveniently be found in the I. G. Y. Instruction Manual [5], and various sources of error are listed by BLACKWELL, ELDRIDGE and RoBI~SO~ [3]. 2. Solar Radiation and Illumination Outside the Atmosphere. Pyrheliometric Scales The starting point of the work is the magnitude and spectral distribution of the solar radiation outside the atmosphere, and NICOLET~'S [14] specification of this is used. NICOLET expressed great confidence
22
G.D.
ROBINSON:
in his relative spectral distribution, but considered that the absolute values might be in error by about 5 per cent. A later assessment by F. S. JOHNSON [9] gives significantly the same solar constant but a spectral distribution differing from NICOLET'S at many points. The extraterrestrial illumination is less satisfactorily established. If Ix is the intensity of solar radiation at wavelength X, Yx the relative sensitivity at this wavelength of the standard eye in photopic vision (see e.g. LIST [13], p. 451) and K the maximum luminous efficiency, the illumination IL is K f Ix Yx d X. The photometric standard is unit area of a full radiator at the temperature of freezing platinum, so that K may be expressed in terms of the constants of P~A~CK'S formula and the platinum point temperature, and the illumination computed for any spectral distribution of radiation. With K = 6 8 0 1 m - W -1 and NICOLET'S radiation data the computed extraterrestrial illumination (illumination of a surface normal to the solar beam at mean solar distance) is 142 kl ; this is the value I use in this paper. The computed illumination is particularly sensitive to radiation of wavelengths around the maximum of Yx at X = 0.555 [_~where NICOLET'Sand JOHNSON'S tables differ. JOHNSON'S tables lead to a value of 150 kl for the extraterrestrial illumination. KARANDIKAI~ [10] measured the central luminance of the sun at various solar altitudes and extrapolated to zero air mass. Combining these observations with previously accepted results for limb darkening he arrived at a value of 131.5 kl for the extraterrestrial illumination. If he had used his own observations of limb darkening he would have found 134 kl. TOVSEu and HVLBERT [17] extrapolated an observation of normal illumination of 130.5 kl made on the stratosphere balloon Explorer I I to an estimated extraterrestrial value of 133.5 kl. The radiation records I have used are standardised with reference to one of three pyrheliometric scales--the International Pyrheliometrie Scale (I. P. S.) 1956, the Smiths0nian seMe of 19i3, or the uncorrected _lngstr6m scale. The relations between these scales are set out in the I. G. Y. Instruction Manual [5]. NICOLnT'Stabulations are not referred to any of these scales but are presumably the best estimate he could make of absolute values, and are here regarded as consistent with I . P . S . 1956. The illumination record used was standardised by means of substandard lamps referred in their turn to a nationM photometric standard, nominally the platinum point radiator. The sub-standards themselves have an uncertainty of about -~ 2 per cent relative to the standard. BLACKW~LL and POWELL [2] have discussed in detail the calibration of recording daylight photometers. 3. Computation o~ Gaseous Scattering and Absorption-the Model Atmosphere
Attenuation of solar radiation in the aerosol-free atmosphere occurs by absorption by the gases Os, C02 and H20, and by molecular scattering. To tile degree of accuracy useful in the present work COz absorption in
Absorption of Solar l~adiation by Atmospheric Aerosol
23
the presence of appreciable quantities of It20 may be neglected. I have used two approximate methods of computing the attenuation. The first, which I shall call the single scattering method, was employed by BLASKWELL, ELD~IDG~ and ROBI~SO~ [3]. I supposed the whole Os layer concentrated at the outer limit of a plane atmosphere and calculated the absorption for an ozone amount of 0.2 em S. T. P. and three values of solar altitude h (sin -1 0.8, sin -1 0.4, sin -1 0.2), using the absorption coefficients of LIS~ ([13], p. 429) supplemented at wavelengths greater than 0.6 ~ by those of C~AId ([4], p. 23). I then computed the attenuation by molecular seattering of the residual radiation, using the l~ayleigh scattering eoeffieients tabulated by L~sT ([13], p. 431), and supposed half the radiation seattered from the beam to reach the ground as diffuse radiation. To this point the computation included the spectral detail given by NICOnF~T i.e. bands of width 0.01 ~ for X < 0.75 ~, 0.05 for 0 . 7 5 [ ~ < X < 1 . 5 0 ~ , 0.10~ for 1 . 5 0 ~ < X < 2 . 0 0 [ z and 0.50 for 2.00 ~ < k. I also computed by the same method the absorbed and scattered illumination, using the function Yx and summing over all X. I then estimated the absorption due to H20 by means of KI~BALn'S curve of transmission integrated over all wavelengths (LIST, [13], p. 437). Since H~0 absorption is negligible for X < 0.7 ~, and the infra-red Rayleigh scattering is small, I considered the whole of the depletion by I-IsO to be eoneentrated in the direct solar beam. The computation so far contains no allowance for the effect of surface albedo. In the present work I have estimated this effect by multiplying the computed global radiation by the surface albedo to give the reflected radiation, arbitrarily supposing this to be concentrated in a beam at angle sin -1 0.5 to the horizon, finding the scattering from this beam by interpolation in the integrated l~ayleigh scattering computations, and adding half the radiation scattered from the reflected "beam" to hag the radiation scattered from the direet beam to give the diffuse radiation D. This procedure is not appreeiably sensitive to relevant variations in H20 content. The final estimate of global radiation was taken as the sum of D and the flux per unit surface area of direct radiation (G = D ~ - ] sin h). An analogous procedure, using the photometric albedo of the surface, yields global and diffuse illumination. Multiple scattering is an important feature of radiative transfer in the atmosphere, and DEII~MENDJIAN and SSK~aA[Z] published results of a computation of global and diffuse radiation in a plane Rayleigh atmosphere by CgAND~AS~XgAa'S methods, using NICOSET'S specification of the extraterrestrial radiation and ineluding the effect of surface albedo. They tabulate values of G and D for solar altitudes h = sin -1 1, h = sin -1 0.6 and h = sin -1 0.1 and surface albedos 0.0, 0.25 and 0.8. DAv~ and SEKERA [6] extended these tabulations to allow for absorption by O3, by the approximate method of considering the Os concentrated at the upper limit of the atmosphere. They used the 03 absorption coefficients due to VmRoT:x. They tabulate values of G and D, integrated over the spectrum, for ozone amounts 0.25 and 0.50 cm S.T.P., albedo 0.0 and 0.8, and solar
2zi
G.D. ROBINSON:
altitudes h = sin 1 0.6, sin -1 0.4, sin -1 0.2 and sin-1 0.1. (Both these papers contain computations for lower solar altitudes which are not of interest in the present work.) I have used these tabulations as the basis of a second method (which I will refer to as the multiple-scattering method) of computing G and D for an aerosol-free atmosphere containing H20, making the necessary interpolations for required Values of h (or air mass m), ozone amount and Mbedo by graphical methods which do scant justice to the precision and complexity of the calculations of SEKEI~A and his colleagues but which seem adequate for the present purpose. Some of the stations considered are at considerable altitudes, and extrapolation to m < 1 was necessary. DAv~ and SEKERA tabulate only the integrated values of G and D. I have made estimates for two narrow bands of wave-length by reducing the tabulated values of DEIRM~DJIA~ and S~KERA ([7], Tables I V and V) by the appropriate ozone absorption, and interpolating in their Fig. 3 to the required Mbedo. As in the single scattering method the whole of the water vapour absorption from KIM~ALL'S curve is subtracted, as the last, stage of the computation, from the direct solar beam. Table 1 compares results of the single and mnltiple scattering methods for a dry atmosphere for values of the parameters used later in this paper. I n this Table the italic entries are computed values, roman entries are interpolations, of the ratios of G and D to the extraterrestriM flux per unit horizontal area, I0 sin h. The most remarkable feature of Table 1 is l~he relatively small differences between the results of the single-scatter and multiple-scatter methods in the context of the enormous difference in complexity of the two sets of computations. In this paper I have used the results of the multiple scatter method as first choice when comparing radiation measurements with computations, and the single scatter method when dealing with illumination. The differences between the results of the methods are not significant in comparison with other uncertainties of measurement and computation. 4. Estimation of the Upward Scattering by Aerosol The depletion of the global radiation by absorption and by molecular scattering upwards having been computed, t h a t due to upward scattering from aerosol must be estimated. BLACKWELL,ELD~IDGE and R o m ~ s o ~ [3] equated dou-nward and upward scattering. WALDRAM[18] has published measured angular distributions of scattered radiation, and numerical integration of these is possible for any solar position, though the variability of WALI)~AM'S results and the doubtful reliability of measurements in this difficult field discourages undue elaboration. Table 2 gives the results of a very rough estimate. This treatment ignores possible effects of multiple scattering by aerosol. I do not think these are serious. Some unpublished calculations by Mr. D. B. LA~E of the back scattering of isotropie radiation incident
Absorption of Solar Radiation by A~mospherio Aerosol
25
~3
I
I
~L:L
E E
~L r
~q
9
39
26
G.D. ROBInSOn:
on a l a y e r of " l a r g e " w a t e r droplets show t h a t it does not become equal to t h e f o r w a r d s c a t t e r i n g untile the o p t i c a l thickness of t h e l a y e r is 5. Table 2. Estimated Ratio o/ Downward to Upward Scattering by Aerosol Solar Altitude sin_ 1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
i ' 1.5 ! 2.0
2.5
3.5
5
6
8
i0
12
I
Ratio of downward to upward scatter
The ratios of forward to backward scatter listed in Table 2 are lower than RoAcH's [15] estimates of this parameter over the English Channel.
5. Depletion of the Global and Diffuse Radiation The s t a r t i n g p o i n t was a t a b u l a t i o n of h o u r l y m e a n values of global a n d diffuse r a d i a t i o n in cloudless conditions, either on a n i n d i v i d u a l d a y , or as m e a n values over a p e r i o d of d a y s short enough for t h e solar a l t i t u d e a t a given hour to be t a k e n as constant, or as m e a n values r e l a t e d to solar altitude. The t a b u l a t e d values were expressed as a f r a c t i o n of t h e corresponding e x t r a t e r r e s t r i a l r a d i a t i o n a n d c o m p a r e d w i t h t h e global a n d diffuse r a d i a t i o n c a l c u l a t e d for t h e m o d e l clean a t m o s p h e r e w i t h approp r i a t e a m o u n t s of ozone a n d w a t e r v a p o u r . The excess of diffuse r a d i a t i o n over t h a t c o m p u t e d for t h e m o d e l was t e n t a t i v e l y a t t r i b u t e d to d o w n w a r d s c a t t e r i n g b y aerosol, a n d t h e corresponding u p w a r d s c a t t e r i n g r e a d from Table 2. The deficit of ~he o b s e r v e d global r a d i a t i o n , r e l a t i v e to t h a t c o m p u t e d for the model, less this u p w a r d scattering, must, in t h e absence of s y s t e m a t i c errors of m e a s u r e m e n t or c o m p u t a t i o n , be a t t r i b u t e d to absorption. M e a s u r e m e n t s of global a n d diffuse solar r a d i a t i o n are k n o w n to be p r o n e to s y s t e m a t i c errors which v a r y w i t h solar e l e v a t i o n a n d azim u t h , a n d to be p a r t i c u l a r l y unreliable a low solar elevation. The effect of this was m i n i m i s e d b y considering only occasions for which the solar elevation, h, was g r e a t e r t h a n sin 1 0.3. 'Diffuse' r a d i a t i o n was recorded in all cases used here b y means of a p y r a n o m e t e r e q u i p p e d w i t h a shadow ring s u b t e n d i n g a n angle of a b o u t 10 ~ a t the thermopile, a n d s t a n d a r d i s e d b y reference to m e a s u r e m e n t s of " d i r e c t " r a d i a t i o n which included r a d i a t i o n s c a t t e r e d f o r w a r d in a cone of angle a b o u t 10 ~ (see e. g. I G Y I n s t r u c t i o n M a n u a l [5]). This does n o t i n v a l i d a t e t h e m e t h o d for d e t e r m i n i n g a b s o r p t i o n - - t h e f o r w a r d s c a t t e r e d r a d i a t i o n e x c l u d e d f r o m t h e "diffuse r a d i a t i o n " as m e a s u r e d is i n c l u d e d in the "direct radiation" as measured. "Total attenuation" and "scattering" are u n d e r e s t i m a t e d on a strict i n t e r p r e t a t i o n ; it is of course c o n v e n t i o n a l to include some small angle f o r w a r d - s c a t t e r e d r a d i a t i o n in p u b l i s h e d m e a s u r e m e n t s of direct solar r a d i a t i o n s i m p l y because t h e r e is no reliable w a y of excluding it. Surface albedo was t a k e n as 0.15 a t all stations considered e x c e p t H a l l e y Bay. A value of 0.25 m i g h t be more a p p r o p r i a t e a t W i n d h o e k ,
Absorption of Solar Radiation by Atmospheric Aerosol
27
but the computations are not very sensitive to Mbedos of this magnitude. The much larger surface albedo at Halley Bay is difficult to estimate and very significant. The albedo of the snow surface itself has been determined on many occasions in the Antarctic and results between 0.8 and 0.9 have been quoted for snow which has never been subjected to melting. In the present context ~the albedo of a very wide area surrounding the station is required, and Halley Bay in the summer is close to stretches of open water, the effect of which on Mbedo is spectacularly shown in photographs of cloud cover. The Mbedo assumed in Table 3 is 0.65. This was chosen after examining the diffuse radiation record on one or two very clear days, and assuming that on these days Rayleigh scattering only was operative. (This was found by DAVE and SEI~I~A [6] to hold for certain of LILJEQUIST'S[12] observations at Maudheim (71 ~ S, 11 ~ W), and I have confirmed (unpublished) that records recently made at Argentine Island (65 ~ S, 64 ~ W) contain occasional days when it appears to be true.) Ozone amounts were chosen as a compromise between known mean values for the date and place and those vMues for which absorption computations were available. Water content was taken from the mean values presented b y BA~!~O:N and STEELE [1], except for Halley Bay, where it was estimated from radio-soundings. I n forming the ratio of measured to extraterrestrial radiation attention was paid to the pyrheliometric scale in which the tabulations were expressed, and the solar constant was corrected to true solar distance. This is of course not possible with m u c h precision for the long period m e a n values, but in these eases the extraterrestrial radiation was weighted towards the appropriate summer values, by - - 1 . 5 per cent for Kew and Vienna, and b y + 2 . 5 per cent for Halley Bay. t~esults of application of this method are shown in Table 3. The data sources are set out in the Appendix. Tables 3 a, 3 b and 3 c are based on long period mean values for cloudless occasions--values of G and D for identical periods at Halley Bay and Kew Observatory, and, at Vienna, of D and I sin h ( = G--D) for different periods. Only the tabulated record was available to me from the stations Lwiro, Windhoek and Pretoria and I chose those days with particularly low values of diffuse radiation to be sure of having cloudless conditions. Results for these stations are therefore not representative of mean conditions but refer to days with minimum aerosol scattering. In the ease of Malta a few days which were known to be cloudless and which had high radiation were chosen. Aerosol scattering on these days was probably less than the average for cloudless days. For Lerwick records on all known cloudless days were used.
6. Depletion of the Global and Diffuse Illumination The method used was the same in principle as that set out in 5. The same ratio of downward to upward scattering by aerosol was used as for the total radiation. This is unlikely to be the ease but any attempt at differentiation would be quite arbitrary. The single scattering method was employed for computations of the global and diffuse illumination.
~, D. I~OBINSON:
28 o
o
o
ddddddddd
dddddddddd
ddddddddd
co
dddddddddd
o-~ ~ o ~
ddddddddd ~dd~odddd~ ~e
e
d
c~
~ d d d d d d d ~
ddddddddd @
ddddddddd
o
~ d
..~.~
d
d
~
d
d
d
d
~
o ~ o ....,
~ ~
~
5
ddddddddd
d
ddddddddd
%-
~o ~
00 d
4
J
I
o
o
c~
d d c ~ d d d d d d
o
Q~
.~ = ~~: ~
eeeedeeee
,.4
.,,
c~
m
d
d
o ~ o,,@ . o ~-~c~ ~
.,_
g.
o
oo~
~iI ~ . ~ ~.o ~
,4
O , -.@
~o ~
~
~
m e l e e
~~
~NN
Absorption of Solar Radiation by Atmospheric Aerosol
,4-~
o
"~
~.~
~
d
N
m
d
o
E
~
~'~
29
30
G. D. ROBINSON:
o
g d d d d d d d d ~
~ m m m
~ ~mm_m
~3
~5
Diffuse illumination was not recorded at K e w , but the flux through a normal surface of the direct solar b e a m in the wavelength range 0.55 ~ to 0.64 ix (I0.6) is available for a period. The relative luminous efficiency of the standard eye is sharply peaked with a m a x i m u m at X = 0.555 ~, so the optical properties of the atmosphere in the range X = 0.55 po to X = 0._64 ~ m a y be expected to be v e r y similar to its photometric properties; the ra~io of recorded to extraterrestrial direct solar illumination was therefore represented by the ratio of 10.6 sin h to the appropriate extraterrestriM value. This ratio was subtracted from the ratio of global to extraterrestrial illumination, and the result was t a k e n to be the corresponding ratio for the diffuse illumination. The photometric albedo of the surface at K e w was t a k e n as 0.1. Results os application of the m e t h o d to long term m e a n values of illumination at K e w are shown in Table 3 d.
d d d d d d d d d
7. Attenuation of the Direct Solar Radiation
%
u~
%
o~- ~ 2
~.~
~
Measurements are available on certain cloudless days at K e w of the intensity of direct solar radiation on a surface normal to the solar beam, for the whole solar radiation and for that part of it passed by eertMn glass filters, including one with an
equivalent sharp cut-off at a wavelength of 0.64 ~z. Computations of transmission of the model atmosphere were m a d e by the single scattering method for the whole radiation and for
Absorption of Solar Radiation by Atmospheric Aerosol
31
radiation of wavelength greater than and less than 0.64 ~. The excess of attenuation in the real atmosphere over that in the model is thus found for k < 0.64 ~ and X > 0.64 ~, but in the absence of measurements of diffuse radiation in either wavelength range this excess attenuation cannot be separated into components due to scattering and absorption. This separation is possible for the total solar radiation, using the same method as in section 5, determining the excess downward scatteiing from the :observed total diffuse radiation and the computed downward ]Rayleigh scattering corrected for the effect of surface Mbedo, and reading the excess upward scattering from Table 2. The method thus yields excess total attenuation in two wavelength ranges corresponding approximately to the ultraviolet plus visual and the infra-red respectively, and excess scattering and absorption of the whole solar radiation for the same periods. ]Results of the application of this method are set out in Table 4, in which all entries are expressed as fractions of the appropriate total extraterrestrial solar radiation except the bracketed values in column 11, which are expressed as fractions of the extraterrestrial solar radiation in the relevant spectral band and allow a comparison of the specific attenuation in these bands. Except for observations on 5 August 1955 mean hourly values are tabulated, but on this day an attempt was made to synchronise ground and aircraft observations, and the tabulated values are means fo~ a few minutes centred on the times quoted. Instrumental questions bearing on the accuracy of observations of this type are discussecl by ELDRIDGE and ]RoBINSOn, but it should be noted that the occasions listed in Table 4 have been selected; they are in most cases associated with a calibration of the instruments and the measurements are of higher quality than the mean values of radiation on a horizontal surface used elsewhere in this paper. The measurements detailed in Table 4 may be summarised by grouping them into two sets, one with air masses in the range 1.2 to 1.6, the other with air masses 2.6 and 3.0; the result is shown in Table 5. Some further information on transmission of the direct solar beam at Kew is summarised in Fig. 1. ]Recordings were made for a period of two years of tile direct: solar beam transmitted by a-filter with equivalent sharp cut off at about 0.55 ~. Combining this record with that obtained simultaneously with the 0.64 ~ filter allows a determination of the transmission in the wavelength range 0.55 ~ < X < 0.64 ~.. The intensity transmitted grouped according to solar altitude in ranges of sin - i 0.1 < h < < sin i 0.2 etc. and expressed as a fraction of the extraterrestrial radiation, is plotted in Fig. 1 for all occasions on which it could cdnfidently be said that cloud was not interfering with transmission, range and median value being shown for each group. The broken line shows the transmission in this range computed by the single scattering method for the model atmosphere with 0.2 cm O3. This material does not allow a n y division of attenuation into absorption and scattering, but so far as total attenuation is concerned Fig. 1 is compatible with Table 5.
]~LAOKVCELL,
32
v
v
v
v
~
~
v
v
v
~
~b
.-~,~
]
~o
~|
~c~
~
:~
oo
oo
o=
~c~
~c~
~
~o~
c~c~
VA
VA
VA
VA
VA
-4 .
.
.
.
I .
.
VA
.~
~
o0
.
~
~
~
~
~
~
Absorption of Solar Radiation by Atmospheric Aerosol
~.~
~
v
d~ VA
~.~
v
v
dd~ VA
Arch. l~et. Geoph. :Biokl. B. Bd, 12, H. 1
~,~
v
v
dd~ VA
(sO
v
~
d~ VA
33
~1
v
v
dd~ VA
3
34
O.D. ROBINSON:
Table 5. A t t e n u a t i o n o/ Direct Solar Radiat,ion at K e w , A d d i t i o n a l to T h a t o / t h e M o d e l Atmosphere, Expressed as a Fraction o/ the Extraterrestrial R a d i a t i o n i n the Appro?grs Spectral R a n g e
(Summary of Table 4) I
Jl l!
Total Attenuation k < 0.64 X > 0.64 ,411 X
Scatter All k
Absorption All ),
i. Air mass 1.2 to 1.6 Minimum
0.187
0.082
0.165
0.036
0.103
Mean
0.247
0.181
0.204
0.064
0.140
5Iaximum
0.324
0.218
0.257
0.085
0.175
0.236
0.283
0.084
0.199
ii. Air mass 2.6 to 3.0 Mean
[
0.349
8. 0 b s c r v a t i o n s f r o m Aircraft As m e n t i o n e d earlier, a t t e m p t s to synchronise aircraft m e a s u r e m e n t s using the i n s t a l l a t i o n described b y I%OBI~SO~ [16J with observations on the ground at K e w Observatory were unsuccessful. Air n a v i g a t i o n reguA/?" H3s~ - - ~
-dd I
-/,o i I
Fig.
1.
Attenualdon
'4'
of
the
direct
solar
radiation
in
the
wavelength
range
0.55 ~z < X < 0.64~x by the cloud-free atmosphere at Kew Observatory-- summary of observations between September 1947 and April 1949 lations greatly restricted the n u m b e r of a t t e m p t s which could be made, a n d i n s t r u m e n t a l failures i n the aircraft occurred on the very few days when flying was possible. However on 5 A u g u s t t955 a record was o b t a i n e d from the d o w n w a r d p o i n t i n g solarimeters only. The readings are set out i n Table 6. I n the absence of m e a s u r e m e n t s of the d o w n w a r d r a d i a t i o n a n d of information on the variation of surface albedo near Kew, no precise quantitative use can be made of Table 6. It can however he used to show
35
A b s o r p t i o n of Solar iRadiation b y A t m o s p h e r i c Aerosol
t h a t the upward scatter from aerosol, assumed to be all below 20,000 ft, is very probably not more t h a n 0.02 of the extraterrestrial radiation. The corresponding entry in Table 5 is 0.009, based on the assumptions of Table 2. Table 6. Global Radiation and Upward Di]/use Solar Radiation, Kew, 5 August 1955, Expressed as a Fraction o/ the Extraterrestrial Global Radiation Time (L. A.T.)
Aircraft tIeJght
Upward radiation
Global radiation at surface
1030 1135 1200
1,750 ft 9,750 ft 19,750 ft
0.092 0.117 0.135
0.669 0.676 0.657
9. Discussion of Results 9.1. P r e v i o u s
Measurements
of A b s o r p t i o n
by Aerosol
I t is of course, not novel to suggest t h a t atmospheric aerosol scatters solar radiation. The existence of considerable absorption is however less thoroughly established, indeed the possibility is often ignored. WALDRA~ [18] made direct measurements of total attenuation and integrated scattering in heavy industrial haze ; by difference he inferred t h a t absorption and scattering were roughly equally responsible for the total attenuation. These very remarkable direct measurements do not appear to have been repeated; they would in a n y case be very difficult in the much lower atmospheric opacities with which the meteoro]ogist is normally concerned. The other direct measurements of atmospheric absorption have been made using aircraftmounted solarimeters. I~OACH[15] found absorptions of up to 20 per cent for air masses 1.2 to 2 over the English Channel. Several investigations within the U.S.S.1%. are quoted by KONDtCATIEV[ / / ] - - n o t a b l y those of KASTROV, LvovA and S~LYAKov--which showed aerosol absorption roughly equal to t h a t computed for 1-I20, and others by BEnEZI~A and by SAVrKOVSKY which showed no aerosol absorption. The facts t h a t absorption due to aerosol often occurs, t h a t it is not confined to industrial areas, and that it is sometimes absent, seem to be established by these observations. 9.2. G e n e r a l
Features
of t h e
Present
Evidence
I have shown that at Kew total attenuation of the direct solar beam, of the direct solar beam in a restricted spectral region 0.55 ~ < ~ < 0.64 ~, of the global radiation and of the global illumination is on the average about twice that due to Rayleigh scattering, absorption b y H20, and absorption by 03 in a clean atmosphere. The diffuse solar radiation (integrated over all speotral regions) is considerably less than would be expected if the extra attenuation were entirely due to scattering. Results derived from measurements of direct and diffuse radiation at Vienna are in close agreement with those at Kew. So far as Kew and Vienna are concerned the evidence 3*
36
G.D.
ROBINSON:
points to an approximately grey attenuation of mean magnitude about 20 per cent for the range of solar altitude 20 ~ to 60 ~ made up of 2 parts absorption and 1 part scattering. This can hardly be attributed to any cause other than atmospheric aerosol. I now further examine the result from these and other observing stations on the basis of this hypothesis. This is conveniently done by considering the variation of attenuation with change of solar altitude. 9.3.
Variation
of t h e A t t e n u a t i o n
with Solar Altitude
Fig. I shows that at Kew the median value of total attenuation in the wavelength range 0.55 ~ % ~ % 0.64 ~z varies with air mass according to Lambert's law. (The fact that attenuation in the air mass range
5 to 10 seems to be less than the Lambert's law value is almost certainly due to rejection of occasions with high opacity owing to the difficulty of excluding the possibility of cloud on these oceo~sions.) If the atmospheric aerosol reatIy acts as a grey body attenuator then a similar exponential law should be followed for the total radiation and for broad wavelength bands. The exponential relation would appear to hold only approximately because of the simultaneous occurrence of non-grey Rayleigh scattering and selective absorption and because of multiple scattering and scattering-absorption processes. The relation would also be masked by any diurnal variation in the amount or nature of the aerosol. I n Fig. 2 the extra absorption and scattering disclosed i a Table 3 are plotted against air mass. Three types of relation appear. The group of "clean" stations, Halley Bay (mean of cloudless conditions) and Lwiro, Pretoria and Windhoek (selected occasions of minimum diffuse radiation) can be considered to show the expected exponential relation of both "absorption" and "scattering" with air mass. The magnitudes are quite small--one to three per cent per air mass. The technique used hardly seems capable of establishing attenuations of this order at a single observing station, but the similarity of the results for four localities, two entirely different climates, and two independent observing and standardising routines strongly supports the explanation here proposed. The urban stations, Kew and Vienna do not show an exponential relation for either "absorption" or "scattering", though the three sets of data, the results of three independent observing and standardising techniques, are in surprising agreement. I can offer no convincing explanation. A systematic diurnal variation of atmospheric pollution would be expected, but has not been established. (The generally observed variation of pollution of near-surface air is largely eontrolled b y upward mixing and is therefore irrelevant.) I t should be noted in this connection that neither Kew nor Vienna is subject to excessively heavy industrial pollution. At Kew for example the visibility in cloudless conditions with the sun's altitude greater than 20 ~ averages about 15 km, corresponding to a transmission of about 0.88 per kin.
Absoi-ption of Solar Radiation by Atmospheric Aerosol
37
The maritime stations, Lerwick and Malta, use the same observing and standardising technique as Kew. Fig. 2 discloses an increase in "absorption" with decrease in air mass, but a roughly exponential relation
/
A/) ~ P'3.ss g
i
3
i
i
Ag$oppflOi7
/i~eS'iffl/LiinlFl3llOil X'?~ Rad/3i/'os?
• 0
o,/ 0
i
i
+
Jc3/lep/)~#
-gi
~
H311ejI~Lr
9
[w/fo
+
Fre/af/8
•
Ik/~/la
o
-O~l
Jc3/?er/'r@ -0,/
i
i
++
,
Absarpl/on
i
-0,I
-0i Fig. 2.
Logarithmic plots of aerosol absorption and scattering (from Table 3)
for "scattering". For corresponding solar heights the magnitude of the absorption is of the same order as that observed by RoAoE in his aircraft measurements over the English Channel. Again I can offer no explanation for the variation of attenuation with solar altitude, other than a considerable systematic diurnal variation in the quantity or nature of the aerosol concerned.
38
G.D. ROBIZqSOX: 9.4.
Significance
of t h e R e s u l t s
Studies of the heat balance of the earth and atmosphere (e. g. those of HOUGJ~TON[8]) have used estimates of the global radiation rather than measurements, for the very good reason t h a t measurements have not been available. These estimates have in general made adequate allowance for scattering by aerosol (turbidity), but absorption b y aerosol has either been neglected or given a magnitude considerably less than t h a t suggested in this paper or by the measurements from aircraft which I have mentioned. Estimates of zonal heat balance would not however be greatly changed if the allowance for aerosol absorption were increased: this absorption must occur mainly in the lowest one or two kilometers of the atmosphere and any amendment of the estimates would merely transfer a proportion of the heat source from the surface to this layer. I t must be remembered, when considering any climatological or dynamical repercussions o~ the aerosol absorption, that ~he conditions here investigated m a y be far from typical of the weather at the station concerned. Completely cloudless days m a y be frequen~ at Malta, but they are so unusual at Lerwick t h a t only five were found in six years record. Furthermore, selection of days at the three African stations was made from the radiation tabulations, by taking a few days with minimum diffuse radiation. For these stations we would therefore expect to have investigated conditions with minimum aerosol scattering. I t m a y not be out of place to mention a consequence of aerosol absorption of more practical application. For a variety of purposes it is interesting to calculate the magnitude and angular distribution of the diffuse radiation from a ]oealised source within the atmosphere. If, as is suggested here in the case of urban atmospheres, the absorption is an appreciable factor in the total attenuation, it greatly modifies the result for moderate and large optical thickness. 10. Conclusions
I conclude, in spite of m y inability to explain certain features of the records, that measurements of solar radiation at the ground in cloudless conditions indicate the frequent and geographically widespread occurrence of atmospheric absorption additional to that attributable to the gases H20 and 03. This absorption is correlated with the scattering associated with atmospheric aerosol. I t is, in very rough approximation, nonselective. I t is observable, though small, even in the Antarctic, and is present in the clearest conditions at high-level stations on the African plateau. I t is often quite large, of the order 10 per cent, at some island stations remote from major sources of atmospheric pollution, and is even larger in the suburbs of large towns. These deductions f r o m surface measurements are compatible with recently published aircraft observations. The method I have used can be applied at any station at which global and diffuse solar radiation is recorded, so the possibility exists of a fairly complete world-wide survey. At present this cannot be made
Absorption of Solar Radiation by Atmospheric Aerosol
39
b y a single investigator, since it is necessary to use only m e a s u r e m e n t s i n cloudless conditions, a n d the usual forms of publication, even those a d o p t e d d u r i n g the I. G. Y., do n o t allow these to be separated.
Acknowledgements I a m greatly i n d e b t e d to colleagues at K e w Observatory a n d at the Meteorological Research Flight, F a r n b o r o u g h , during the period 1947-1955, p a r t i c u l a r l y to Mr. M. J. BLACKWELL a n d Mr. R. H. ELDI~IDGE. This paper is published b y permission of the Director General of the Meteorological Office, Bracknell, E n g l a n d .
AppendixEData Sources 1. Kew. The data on which I based Tables 3 b and 3 e were supplied to me by Mr. M. g. B~ACKWE~L, who extracted them from unpublished tabulations of hourly means and from the original records. Some of the material concerning global and diffuse radiation was used by him in a paper with restricted circulation (BLAc~WELL, M. J., Five Years Continuous Recording of Total and Diffuse Solar 1%adiation at Kew Observatory. M. I%. P. No. 895, London 1954). The general nature and the accuracy of the observations are discussed in this paper. The illumination measurements for Table 3 d were also supplied by Mr. BLACKWELL The original tabulations of illumination have been subjected to considerable correction, in the manner detailed by BLACKWELI~ (BLACKWEIm, M. J., Five Years Continuous 1%ecording of Daylight Illumination at Kew Observatory. M. 1%. P. No. 831, London 1953), and BLACKVCELI~ and POWELL [2]. The data on transmission in the spectral range 0.55 bt < X < 0.64 bt which I have used in Fig. i, and to derive the diffuse component of illumination in Table 3 e, were computed from unpublished tabulations of the direct solar radiation passed by filters, and are mainly the work of Mr. R. H. ELDItIDGE. I extracted the 1948-1949 data in Table 4 from manuscript tabulations at Kew. 2. Halley Bay. A very complete tabulation of the radiation observations at Halley Bay is in preparation (The 1%oyal Society I. G. Y. Antarctic Expedition, Halley Bay 1955-1959, Vol. III, London [in the press]), and will include a separate Table for the cloudless occasions used here in Table 3 a. A preliminary summary has been published by MAoDowAIm (MAoDOWALL,J., Prec. 1%oy. See. A 256, 149-192 [1960]). I am indebted to Mr. J. MAoDowAIm and to the Royal Society for access to this material prior to publication. 3. Vienna. I took the data for Table 3 d from a n undated reprint sent to me by the author, the ]ate Dr. F. SAI:a~EI~ER,fJbersicht fiber die Strahlungsverhgltnisse an der Zentralanstalt ftir Meteorologie a n d Geodynamik, Wien. - - Sonderdruek aus der Forsehungsgelneinschaft ffir GroSstadtprobleme (biok]imatische Gruppe). 4. Lwiro, Pretoria, Windhoek. I extracted these data from the Quarterly Radiation Bulletin of the W. M. O. Regional Association 1 Working Group on Radiation, which is prepared and issued by the Weather Bureau of South Africa. I had no indication of cloudless sky other than the tabulations of diffuse radiation, so selected some days when this was a m i n i m u m for all hours, and roughly symmetrical about noon. 5. Lerwick, Malta. These data were extracted for me from manuscript tabulations made at the stations concerned, available in the Meteorological Office, Bracknell, Berks, England.
40
G.D. I%OBII~SOIV:Absorption of Solar Radiation by Atmospheric Aerosol References
1. BANNO~, J. K., and L. P. STEELE: Average Water-Vapour Content of the Air. Geophys. Mere., No. 102, Met. Office. London, 1960. 2. BL~0~:WELL, M. J., and D. B. B. FOWELL: On the Development of an I m p r o v e d Daylight Illjamination l~ecorder. M . R . P . No. 988. London, 1956. 3. 13T,A0~:WELL,M. J., R. II. ELD~IDOE, and G. D. ROmNSOX: Estimation of the Reflection and Absorption of Solar Radiation by a Cloudless Atmosphere from Recordings at the Ground, with Results for I(ew Observatory. M . R . P . No. 894. London, 1954. 4. CRAIG, 1%. A. : The Observations and Photochemistry of Atmospheric Ozone and Their Meteorological Significance. Met. Monogr., Amer. Met.. Soe. 1, No. 2 (1950). 5. C. S. A. G. I., I. G. u Instruction ManuM, Ft. I V - - R a d i a t i o n Measurements. Annals of I. G. Y. 6, 367-466 (1957). 6. DAv~, J. V., and Z. SEKER~: Effect of Ozone on the Total Sky and Global Radiation Received on a ttorizontal Surface. J. Met. 16, 211-212 (1959). 7. DEIR~ENDJIA2~, D., and Z. SEKER~: Global Radiatiou Resulting from Multiple Scattering in a Rayleigh Atmosphere. Tellus 4, 382-398 (1954). 8. Ilov~To~, H. G. : On the Annual Heat Balance of the Northern Hemisphere. J. Met. II, I-9 (1954). 9. JoHxsosr, F.S.: The Solar Constant. J. Met. II, 431-439 (1954). 10. KA~Am'DIKA~, R.V.: Luminance of the Sun. J.O.S.A. 45, 483-488 (1955). 11. KO~DXA~IEV, K. YA.: Some Problems of Aetinometry in the Free Atmosphere, in "Reports at the Symposium on Radiation in Vienna, August 1961". 12. LILJEQUIST, G. I~. : Energy Exchange of an Antarctic Snow Field, Part I. Oslo, 1956. 13. LlS~r, R. J. (ed.): Smithsonian Meteorological Tables, 6th Ed. Washington, 1949. 1~. NICOLET,~{. : Sur la d6termination du flux ~nerg@tique du rayonnement extraterrestre du soleil. Arch. Met. Geoph. Biokl. 13 3, 299-219 (1951). 15. RoAch, W. T. : Some Aircraft Observations of Fluxes of Solar Radiation in the Atmosphere. Quart. J. Roy. 5~et. Soc. 87, 346-363 (1961). 16. R o m ~ s o ~ , G. D.: Some Observations from Aircraft of Surface Albedo and the Albedo and Absorption of Cloud. Arch. Met. Geoph. 13iokl. B 9, 28-41 (1958). 17. TousEY, R., and E. 0. HULBEtCT: Brightness and Polarization of the Daylight Sky at Various Altitudes Above Sea Level. J . O . S . A . 37, 78-92 (1947). 7.8. WALDRAM,J. M. : Measurement of the Photometric Properties of the Upper Atmosphere. Quart. J. Roy. Met. Soe. 71, 319-336 (1945).