Indian Journal of Radio & Space PhysicsVol. 22, February 1993, pp. 50-55

Study of lidar and satellite data on stratospheric aerosols formed due toMt. Pinatubo volcanic eruption

P Ernest Raj & pes DevaraIndian Institute of Tropical Meteorology, Pune 411 008

Received 1 September 1992; revised received 15 December 1992

Lidar aerosol backscatter data of a few stations published in Bulletin of Global Volcanism Ne-twork (USA)" (Vol. 16, Nos 5-12, 1991) have been used to investigate various transport/dynamicalprocesses and also aerosol loading in the stratosphere due to the 15 June 1991 Pinatubo volcaniceruption. Some recently reported satellite observations have also been used to support the resultsobtained. Satellite data showed that in the zone extending 100N from equator, Pinatubo aerosols cir-cled the globe in 21 days, implying a mean easterly speed of 20-22 mls. Ground-based lidar obser-vations showed that aerosol cloud spread northward faster in low mid-latitudes and slowly in highermid-latitudes. The stratospheric aerosol loading due to Pinatubo was about 30 times the pre-erup-tion value at a low latitude station (Mauna Loa) and about 4 times at a higher mid-latitude station(Obninsk). The analysis further showed that the aerosol cloud due to Pinatubo was nearly 1.5 timesthan that produced due to the 1982 El Chichon volcanic eruption. The presence and morphology ofof the multiple stratospheric aerosol layers that appeared over Mauna Loa in the aftermath of Pina-tubo are discussed.

1 IntroductionThe physical origin of stratospheric aerosols is

nearly well understood, being mainly due to con-densation of sulphuric acid produced by the solarphotodissociation of carbonyl sulphide and sul-phur dioxide by UV light, injected into the stra-tosphere during volcanic eruptions. The stratos-pheric aerosol layer was one of the first atmos-pheric constituents to be measured by the lidartechnique I. Since then, lidar measurements havecontributed a substantial portion of the existingdata set on stratospheric aerosol behaviour. Thesemeasurements have documented the response ofthe stratospheric aerosol layer to several majorvolcanic injections, and have also helped to studythe dynamics by determining the vertical and latit-udinal structure of the quiescent or backgroundlayer.

Stratospheric aerosols are not normally a signi-ficant component of the earth's climate system be-cause their optical depth is usually quite small.However, after an explosive volcanic eruption, thenumber of stratospheric aerosols is greatly en-hanced and their optical depth can then be aslarge as that of tropospheric aerosols. Thus vol-canic aerosols can alter the climate by upsettingthe balance between sunlight entering the atmos-phere and infrared light leaving the atmosphere. It

is widely believed that volcanic eruptions that in-ject large quantities of sulphur dioxide into thestratosphere are responsible for increasing stra-tospheric temperatures by several degrees and forcooling tropospheric temperatures by several ten-ths of a degree for a year or two after the erup-tion-, The extent to which volcanic eruptions in-fluence surface temperature has been a matter ofcontroversy for decades. While some workershave shown evidence of volcanic influence on sur-face temperature+", there are others who questionthe efficacy of volcanoes in producing surfacecoolingv". In view of the fact that many historicalclimate changes are believed to have been causedby volcanic eruptions, it is important to make de-tailed measurements/studies of stratospheric aero-sol layers that are formed in the aftermath of avolcanic eruption.

As the lidar measurements of the stratosphericlayer are being well documented in recent years,attention is now being focussed on the behaviourand effects of the layer on atmospheric chemistry,radiative transfer and ultimately on the clim-ate of the earth. The perturbation to stratosphericaerosols caused by the massive erutpion of the E 1Chichon volcano (17.33°N, 93.2°W) in Mexico on4 Apr. 1982 has been very well documented andhas been studied by many workers around the

ERNEST RAJ & DEVARA: ERUPTION-INDUCED STRATOSPHERIC AEROSOLS 51

globe using the lidar technique (Refs. 11-14). TheJune 1991 eruption of Mt. Pinatubo, Philippines,is reported to have introduced large quantities ofaerosols into the stratosphere and many lidarstations allover the globe have made measure-ments of aerosol back scattering. Some papers de-scribing the early characterization of the stratos-pheric impacts of Pinatubo eruption have ap-peared in one of the recent issues of Geophys ResLeu (USA) (Ref. 15). Lidar data collected at fivestations for the period 1 July to 4 Dec. 1991 havebeen used in the present study to investigate var-ious transport/dynamical processes in the stratos-phere and aerosol loading due to the Pinatubovolcanic eruption. Some of the recently reportedsatellite observations have also been used to sup-plement the results.

2. Volcanic eruptions ofMt. PinatuboMt. Pinatubo (15.14°N, 120.35°E, summit ele-

vation 1745 m), Luzon Islands, Philippines, is re-ported to have last erupted some 635 years ago".In the eruption of 1991, renewed volcanic activitywas first observed in April-May 1991 which laterintensified by early June. S02 flux rose up to 5000metric tonnes per day (t/d) on 28 May. Initialstrong explosions were reported on 12 June and atephra column rose to about 20 km on that day.An explosion at 0555 hrs LT on 15 June fed a20-22 km ash column marking the onset of strongsustained activity that included the climactic ex-plosions and lasted until early hours of 16 June.Satellite data suggested that more than 95% ofthe aerosol cloud was produced during this 12 hphase!'. This stratospheric cloud expanded rapid-ly west-south-west, and by 1030 hrs LT on 16June, its leading edge had reached the Bangkokarea, more than 2000 km away. This travel timeimplies a mean easterly wind speed of 26.5 mls.Preliminary Nimbus- 7 satellite data showed a veryhigh concentration of S02 over a broad areacovering Indochina and south China sea, with atotal mass that appeared to be approximatelydouble that of the 1982 injection from EI Chi-chon. Light ash falls were reported in the south-ern Vietnam, Northern Borneo and Singaporeabout 20 h after the onset of strongest activity.

3 Results and Discussion3.1 Zonal and meridional transport or aerosols

Reflected solar radiation measurements fromthe advanced very high resolution radiometer(AVHRR) on NOAA-ll polar orbiting satelliteshowed the aerosol cloud'sdispersal as it movedgenerally west along the equator!". In the zone ex-

tending lOONfrom the equator, aerosols reachedthe longitude of Pinatubo (120"E) on 6 July, 21days after the paroxysmal eruption. This traveltime implies a mean easterly wind velocity of 20-22 mls during the three weeks following the er-uption. This speed is consistent with the stratos-pheric winds at about 30 mb level (about 24 kmaltitude). The 1982 El Chichon eruption clound(source latitude 17.33°N, about 2°N of Pinatubo)travelled around the globe in about 22 days.

The earliest reported detection of this stratos-pheric aerosol cloud due to Pinatubo by ground-based lidar was at Mauna Loa (19.5°N,130.35°W), Hawai, on 1 July 1991. Later on,several lidar stations in the low-middle and highlatitudes have detected the arrival of enhancedstratospheric aerosols over their locations frombackscatter information. Table 1 lists the detailsof 19 stations (mostly lidar) along with the datesof first detection of significant increase in stratos-pheric aerosol backscatter due to the Pinatubocloud. A few observations of visual effects, espe-cially from the southern hemisphere, are also in-cluded in the table to infer the meridional trans-port of aerosols. The dates of first detection atsome northern latitude stations (indicated with anasterisk mark in the table) are plotted againsttheir latitudes in Fig. 1. It is seen that the strato-spheric aerosol cloud seemed to have spread nor-thwards at faster rate initially in low mid-latitudes(at a rate of about 57 krn/day) and later slowly tohigh mid-latitudes (at a rate of about 44 kmlday).This can be inferred from the two best-fit lineshaving slightly different slopes shown in Fig. 1.

From SAGE-II atmospheric extinction measure-ments during June-August 1991, it was reportedthat Pinatubo aerosols were confined to tropics(15°S to 25°N) up to about 40 days after the 16June eruption 15.17. Reflected solar radiation datafrom NOAA-II satellite also showed no signifi-cant migration into mid-latitude regions till theweek ending 1 August.

Ground-based measurements of stratosphericaerosols made at several stations, covering a widerange of latitudes, have also been reported (asshown in Table 1) and they yield spatial distribu-tion of aerosol cloud in the aftermath of Pinatuboeruption. Thus the ground-based lidar data alongwith satellite observations provide an insight intothe zonal and meridional transport of aerosolcloud in the stratosphere after volcanic injection.

3.2 Stratospheric aerosol loadingLidar-derived integrated aerosol backscattering

(total backscatter, expressed in steradians - !) data

52 INDIAN 1 RADIO & SPACE PHYS, FEBRUARY 1993

Table 1 - Details of some ground-based observing stations that reported the signatures of the Pinatubo aerosol cloud 17

S.No. Station Geographic Date of first Observationallocation detection technique19.5"N 130.4"W 01.07.91 Ruby lidar33.7"N 155.6"E 07.07.91 Nd-YAG lidar36.2"N 140.1"E 15.07.91 Nd-YAG lidar37.1"N 76.3"W 03.08.91 Ruby lidar39.9"N 105.4"W 28.08.91 Visual effects40.0"N 105.3"W 27.07.91 Lidar40.8"N 111.8"W 28.07.91 Ruby lidar41.0oN 78.0"E 13.07.91 Nd-YAG lidar41.0"N 105.8"W 30.07.91 Lidar47.5"N 11.8"E Early August Nd-YAG lidar52.0"N 106.5"W 18.08.91 Backscattersonde55.0"N 38.0oE 14.08.91 Nd-YAG lidar69.3"N 16.0oE 18.09.91 Nd-YAG lidar74.7"N 95.0"W 09.10.91 Backscattersonde82.5"N 62.5"W 02.12.91 Backscattersonde25.9°S 28.2°E Early August Visual effects35.3"W 148.1°E Early August Visual effects38.0"W 145.0"E 19.07.91 Visual effects66.7"W 140.0"E 20.07.91 Lidar

1 Mauna Loa*2 Wlruoka*3 Tsukuba*4 Hampton"5 Coal Creek Canyon6 Boulder"7 Salt Lake City*8 Teplocluchenka"9 Laramie*

10 Garmisch Partenkirchen"11 Saskatoon"12 Obninsk*13 Andoya Rocket Range*14 Resolute"15 Alert16 Irene17 Canberra18 Melbourne19 Antarctica

(Dumont d'Urville base)*The dates of first detection of enhanced stratospheric aerosols at these stations have been plotted against their latitudes inFig. 1.

80

z

~L 60wCl=:l /-t-

;::: 40o ___

ID •

<t ~ .-l <t

Z20 0:: •-' -~

0 1 I

JUNE JULY AUG. SEP. OCT.

199\

Fig. 1-The dates of first detection of Pinatubo cloud byground-based lidar systems at different northern latitudes.

published in Bulletin of Global Volcanism Net-work (Ref. 17) for the period April to December1991 for five stations, namely, Mauna Loa,Hampton, Salt Lake City, Teplocluchenka andObninsk have been plotted in Fig. 2. It may benoted here that the backscatter information of thefirst three stations is for the Ruby lidar wave-length of 0.69 Jim and that of the last two stationsis for the Nd-YAG wavelength of 0.53 Jim. Gen-erally, aerosol scattering decreases with increasein wavelength and this type of spectral depend-ence in the case of Pinatubo aerosol opticaldepths on different days has been reported by

Valero and Pilewskie'", However, the backscatter-ing data used here could not be corrected for thewavelength dependence as such correction fac-tors, 'varying from day to day, were not available.Data for the months of April and May (pre-Pinat-ubo eruption period) is available only for MaunaLoa. It can be seen from the figure that in thepost-eruption period up to December 1991, Mau-na Loa showed the highest stratospheric loadingand the other four mid-latitude stations showedmore or less the same values. To examine the in-crease in stratospheric aerosol content, the inte-grated backscattering values for the period 16June-4 December have been averaged at each sta-tion to represent the value for post-volcanic erup-tion period. The pre-eruption data of integratedbackscatter at each station are taken according tothe availability of the data and the same is shownin Table 2. The periods considered to obtain theaverage pre-eruption value are given in paren-theses for each station. From the table it can beseen that the pre-eruption values represent fairlythe background integrated backscatter uncontami-nated by volcanic aerosols. The stratospheric aer-osol loading in the 5 to 6 months' period follow-ing the Pinatubo eruption decreased with increas-ing latitude. The ratios of post-eruption to pre-er-uption backscatter also confirm that the effective

ERNEST RAJ & DEVARA: ERUPTION-INDUCED STRATOSPHERIC AEROSOLS

-I•..U) I-MAUNA LOA(!) 2-HAMPTONz 3- SALT LAKE CITY-a:

4-TEPLOCLUCHENKAILlI- 5-0BNINSKI- -3ex 10 ....:/>;uU)

~uexIn-oJ0U)

0 -4a: 10 5ILlex0ILlI-exa:(!)

ILl .;I- -5 2z 10

APR. MAY JUNE JULY AUG. SEP.

1991

2III

\ I\ I\ I'V

5

1991

J J A S 0 ~ D

NOV.OCT.

53

..Jc(u

0·3 ~a..o

IIIu III

0·2~ ~.1&1 lCl: Ua.. -U)l:0·' 0 •..::Ii•..c(

DEC.

Fig. 2 +Lidar-derived integrated aerosol backscattering between April and December 1991 at five stations (Inset shows the .monthly mean atmospheric optical thickness values measured over Pacific Ocean from 30"N to 20·S).

Table 2 - Lidar-derived mean integrated aerosol backscatter at five stationsStation Mean integrated aerosol Ratio of

backscatter (x 10-3Sr- I) post-eruption topre-eruptionbackscatter

MaunaLoa 32.5

Hampton

Pre-eruption0.077

(July 90-June 91)

0.070(Nov. 90)

Salt Lake City

Teplocluchenka 0.123(Nov.-Dec. 89)

0.099(Aug. & Sep. 89Oct.-Dec. 90)

Obninsk

Post-eruption

2.500

1.143 16.3

enhancement in the stratosphere due to the vol-canic injection decreased away from the equatorin the first few months after the eruption. It is fur-ther seen that at Mauna Loa which is latitudinallyclose to Pinatubo, the stratospheric aerosol load-ing was increased by as much as 30 times thebackground value. Monthly mean integrated opti-cal depth data for Mauna Loa 19,20 for.pre-El Chi-chon eruption period (April 1981 to March 1982)and post-eruption period (April 1982 to March1983) are taken and the ratios are obtained as ex-plained above. It is found that the increase in aer-

0.530

0.712 5.8

0.486 4.9

osol loading was about 21 times during 1982 EIChichon eruption. Thus the preliminary resultssuggest that the aerosol cloud due to Pinatubowas massive and larger (approximately 1.5 times)than that due to EI Chichon. Stowe et al." haveused the mean tropical Pacific AOT (atmosphericoptical thickness) values derived from NOAA sa-tellite and have shown that post-Pinatubo AOTvalues are 1.6 to 1.8 times than those of the 1982EI Chichon. Monthly mean AOT values measuredover Pacific Ocean (300N-200S) from June to De-cember 1991 have heen taken and plotted (inset

54 INDIAN J RADIO & SPACE PHYS, FEBRUARY 1993

in Fig. 2). It is observed that the mean tropicalPacific AOT reached a maximum in August.

Integrated backscatter data at Mauna Loa from16 June (as can be seen in Fig. 2) showed an ap-parent oscillatory pattern till the middle of Sep-tember. To investigate this in more detail, the databetween 1 July and 30 September were taken andthe linear trend was removed using a straight linefit. The detrended curve showed more clearly thepresence of a 15-18 day period wave pattern inthe lidar-derived integrated backscatter data atMauna Loa in the post-Pinatubo period. It is re-ported'? that the optical depth data between 28June and 28 July 1991 revealed such a wave pat-tern that propagated coherently westward. Satel-lite data have also shown a series of bands of aer-osol layers moving westward with a frequency of16-18 days and this was suggested to be due tothe non-uniform longitudinal mixing of inhomog-eneities in the aerosol cloud which were beingtransported around the globe". These bands ofaerosol layers may have resulted in the observedwave pattern in the lidar backscatter data at Mau-naLoa.

3.3 Eruption-induced aerosol layersThe lidar data from Mauna Loa for the period

1 July to 25 Sep. 1991 suggested the presence ofmultiple layers in the stratosphere which started

appearing/forming on different days and continu-ed to persist for several days. As many as six dis-tinct layers, sometimes with 2 to 3 layers appear-ing simultaneously for few days, could be seen inthe above period. The altitude of each layer peakwas picked up and plotted for different days andis shown in Fig. 3 to examine the vertical redis-tribution of the layers. These curves have beennumbered from 1 to 6 in the figure for identifica-tion. It is observed that the stratospheric layerthat appeared from 3 to 10 July (No.2) wasnarrow (about 1 km thick) but the subsequentlayers (Nos 3, 4, 5 and 6) were 1-4 km thickon formation. However, the peak scattering ra-tios of all the layers on formation were nearly ofthe same order. The layers which were narrow onformation subsequently widened (vertically) ondescent, may be due to diffusion and sedimenta-tion. McCormick et al." from ground-based lidarobservations at Hampton in the aftermath of ElChichon have also found that the altitude of thepeak of stratospheric layers initially centredaround 25 km, moved downward to 15-20 kmand the layers broadened in width as the materialaged. In this study all the layers appeared initiallyin the 22-25 km altitude range and then des-cended down to altitudes between 17 and 20 kmup to the period reported here. The layer thick-ness was on an average 2 km (ranging from 0.6 to

LAYER PEAK25 3 (MAUNA LOA)

4

wCl:J 21t-

t-..J

<{

19

17

5 10 15 20 25 30 4

July AUQust September

9 14 19 24 29 3 8 13 18 23

Fig. 3-Time-height variation of stratospheric aerosol layers as observed by Mauna Loa lidar during 1 July-25 Sep. 1991

ERNEST RAJ & DEVARA: ERUPTION-INDUCED STRATOSPHERIC AEROSOLS 55

3.9 km) on formation and thickened to about 5km (ranging from 0.9 to 12.9 km) as they propa-gated downwards.

It is also seen from Fig. 3 that the descent ofthe layers appears to be exponential, initially withfaster descent rates (as large as 0.86 km/day). Theaverage descent rates of the layers correspondingto the curves numbered 1 to 6 are, respectively,0.44, 0.53, 0.43, 0.17, 0.16 and 0.14 km/day.Thus the first three layers which formed around1-2 July had higher descent rates compared to thelayers that formed later. This suggests that thelarger sized aerosol particles might have des-cended down quickly due to sedimentation. Bysuperposing the initial dates of all the sixcurves shown in Fig. 3, it is found thatthe aerosol layers which formed gener-ally around 25 km altitude took nearly 30 days toreach a stable state (descent rate = 0). This corre-sponds to a mean descent rate of 0.14 kmlday.Thus the lidar aerosol backscatter data providean insight into the formation, dissipation and des-cent of stratospheric aerosol layers in the post-volcanic eruption period.

As the stratospheric lidar data for severalstations and also satellite optical depth data in thepost-Pinatubo period continue to come, furtherdetailed studies are planned to understand thevarious dynamical processes taking place in thestratosphere. Using global climate model to makea preliminary estimate of Mt. Pinatubo's climateimpact, Hansen et alP assume the aerosol opticaldepth due to Pinatubo eruption to be nearly twiceas great as due to the EI Chichon eruption, andforecast a dramatic but temporary break in recentglobal warming trends. Their simulations indicateintense aerosol cooling late in 1992. Thus thepreliminary results reported in this paper empha-size the fact that such a unique set of measure-ments help for a better assessment of the short-and long-term impacts of the volcanic eruption in-duced stratospheric aerosols on the earth's radia-tion budget.

AcknowledgementThe authors are thankful to the Director, IITM,

and to Dr A S R Murty for their encouragement.The valuable comments and suggestions of the

referees are gratefully acknowledged. The assist-ance from Mr S Sharma and Mr G Pandithurai isalso acknowledged.

References1 Fiocco G & Grams G W, J Atmos Sci (USA), 21 (1964)

323.2 Toon 0 B & Pollack J B, The Stratospheric Aerosol Lay-

er, edited by R C Whitten (Springer-Verlag, New York),1982,121.

3 MitchellJ M, Global Effects of Environmental Pollution,edited by S F Singer (Springer-Verlag, New York), 1970,139.

4 Yamamoto R, Swashima T & Hoshiai M, J Meteorol SocJpn(Japan), 53 (1975) 482.

5 Schneider S H & Mass C, Science (USA), 190 (1975)741.

~. Harshvardhan & Cess R D, Tellus(Sweden), 28 (1976) 1.7 Bryson R A & Goodman B M, Science (USA), 207

(1980) 1041.8 Schonwiese C D, Arch Meteorol Geophys & Bioclimatol

Ser A (Austria), 30 (1981) 1.9 Deirmendjian D, Adv Geophys( GB), 16 (1973) 267.

10 Rampino M R & Self S, Quaternary Res (USA), 18 (1982)127.

11 McCormick M P, Swissler T J, Fuller W H, Hunt W H &Osborn M T, Geofis Int(Mexico), 23 (1984) 187.

12 Jager H & Camuth W, Geophys Res Lett (USA), 14(1987)696.

13 Kent G S & McCormick M P, Opt News (USA), 14(1988) 11.

14 Hayashida S, Sasano Y & likura Y, J Geophys Res (USA),96 (1991) 15469.

15 Geophys Res Lett( USA), 19 (1992) 149.16 Wolfe J & Self S, Structural Lineaments and Neogene Vol-

canism in Southwestern Luzon: The Tectonic and Geolog-ic Evolution of Southwest Asian Seas and Islands, editedby D E Hayes, Part 2, American Geophysical Union Mo-nogr., 27 (1983) 157.

17 Bulletin of Global Volcanism Network (Smithsonian Insti-tution, Washington DC, USA), 16, 1991.

18 Valero F P J & Pilewskie P, Geophys Res Left (USA), 19(1992) 163.

19 DeLuisi J, DeFoor T, Coulson K, Fernald F & Thorne K,Lidar observations of stratospheric aerosol over MaunaLoa observatory: 1974-1981, NOAA Data Rep ERLARL-4, 1984, pp. 107.

20 DeLuisi J, DeFoor T, Coulson K & Fernald F, Lidar ob-servations of stratospheric aerosol over Mauna Loa obser-vatory: 1982-1983, NOAA Data Rep. ERL ARL-5, 1985,pp.7R.

21 Stowe L L, Strong A E & Durkee P A, E 0 S (USA), 72(1991)95. .

22 Stowe L L, Carey R M & Pellegrino P P, Geophys ResLett(USA).19(1992) 159.

23 Hansen J, Lacis A, Ruedy R & Sato M, Geophys Res Lett(USA), 19 (1992) 215.