Anda di halaman 1dari 8

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 87, NO.

C2, PAGES 1231-1238,FEBRUARY 20, 1982

The Volcanic Explosivity Index (VEI)' An Estimate of Explosive Magnitude for Historical Volcanism
CHRISTOPHER NEWHALL1 AND STEPHEN SELF2 G.
Departmentof Earth Sciences, DartmouthCollege,Hanover, New Hampshire03755

Knowledge the frequencies highly explosive, of of moderately explosive,and nonexplosive eruptions ould useful a variety volcano w be in of studies. Historical records generally are incomplete, however,andcontain very little quantitative fromwhichexplosive data magnitude be estimated. can Only the largest eruptions havea complete record backto the early 19thCentury; otherimportant explosivevents e wentunrecorded to about1960. prior Onlya handful theverybiggest of eruptions are represented thegeologic in record,soit will be impossible augment to historical records postfacto.A compositestimate the magnitude pastexplosive e of of eruptions, termedthe VolcanicExplosivity Index(VEI), isproposed a semiquantitative as compromise between poordataandtheneedin various disciplines evaluate record pastvolcanism. VEI hasbeenassigned over8000historic to the of The to
and prehistoric eruptions, a complete is available a companion and list in document.

1978), Smithsonian Center for Short-Lived Phenomena (Reports of Volcanic Eruptions on Event Cards, 1968-1975), In compilations of historic volcanism there is a need for an estimate of the scale or 'magnitude' of each eruption, and the Smithsonian Scientific Event Alert Network (SEAN Earthquakes are routinely reported with a Richter magni- Monthly Report, 1975-1981) provide valuable supplements tude, plus intensity estimates at different locations, but no to the Catalogue also in a text and table format. The most analogousreporting systemhas been instituted for volcanic recent addition to these sources, and one which includes eruptions. This presents a predicament for studies that information from the sourcesmentionedabove plus a comrequire some quantitativeor semiquantitative basisfor com- prehensive study of other literature, is another Smithsonian paringexplosiveeruptions.The problemhasbeenparticular- project, Volcanoesof the World, a regionaldirectory, gazetly acute in attempts to evaluate the role of historical volca- teer, and chronology of volcanism in the last 10,000 years nism on past climatic, particularly temperature, variation [Simkin et al., 1981]. This last source contains a more [e.g., Humphreys, 1940; Wexler, 1951, 1952; Lamb, 1970, complete chronology than any previous compilation and is format which allows for 1977;Mitchell, 1970;Bray, 1974;Schneiderand Mass, 1975; also in a flexible, computer-based easy data retrieval. Baldwin et al., 1976; Newell, 1976;Bryson and Goodman,

INTRODUCTION

1980; Hirschboeck, 1980;Robock, 1981]. The problem also arises in studies of freqency of various kinds of volcanic events, for use both in understanding fundamental controls of volcanism and in quantifying volcanic hazards. To be useful, studiesof historic volcanismneed: (1) up-to-date, readily manageablehistorical information about eruptions, includingdates and the nature of activity, and (2) a basisfor comparingthe scale or magnitudeof each type of activity. We describehere a simple schemefor estimatingexplosive
magnitude, with notes on its use and limitations.
2. PREVIOUS WORK

Compilationsof Volcanic Eruption Records for the Purpose of Comparison With Climate Records

Lamb [1970, 1977]usedatmospheric opacity,temperature


records, and volcanological information to estimate the amount of dust introduced in to the upper atmosphere by each of approximately 250 eruptions.When determinedfrom

Compilationsof VolcanologicalInformation

signed dvi/Emaxvalues, nor can the effects of nonvolcanic phenomena (e.g., dust storms) be easily filtered out. Of Lamb's 250 dvi estimates (excluding cumulative estimates for periods of several years), 5% are based on a changein radiation, 5% on temperature records, 12% on quantitative estmates of tephra volumes, 30% on Sapper's [1927] semiquantitative estimates of tephra volumes, and 48% on nonquantitative descriptionsof eruptions.Cronin [ 1971]referred to an unpublishedsurvey of explosive volcanismin the last two centuries, find Hirschboeck [1980] and Bryson and Now at the U.S. Geological Survey, Vancouver, Washington Goodman [1980] referred to another unpublished list in 98663. which they used volcanologic information to characterize 2Nowat the Department Geology, of Arizona StateUniversity, approximately 5000 eruptions as small, moderate, or great. Tempe, Arizona, 85281. This last list avoided the circular reasoning and limited, Copyfight 1982 the American by Geophysical Union. indeed often incorrect, eruption informationof someprevi-

Information about historic eruptionsis scatteredwidely throughoutthe historicaland geological literature. The Catalogueof Active Volcanoesand SolfataraAreas of the World (IAVCEI, Rome, Italy, 1950-1975)is an exceptionallyvaluable compilation of such information in a text and table format. Reports of G. Hantke (Ubersicht iber die vulkanischeTfitigkeit, in the Bulletin Volcanologique, volumes11, 14, 16, 20, and24, 1941-1962), Volcanological the Societyof Japan (Bulletin of Volcanic Eruptions,volumes1-14, 1961-

atmosphericopacity, his ratio of dust veil index/maximum extent (dvi/Emax) a realistically direct measureof the dust is injection. Unfortunately, dvi can only be evaluated in this way for the relatively small number of eruptionsidentifiable in atmospheric records. Early eruptions, small eruptions, and eruptions closely spaced in time cannot easily be as-

Paper number 1C1367.


0148-0227/82/001C- 1367501.00

1231

1232

NEWHALL AND SELF: VOLCANIC EXPLOSIVITY INDEX AND HISTORICAL VOLCANISM

.%o

NEWHALL AND SELF: VOLCANIC EXPLOSIVITY INDEX AND HISTORICAL VOLCANISM

1233

using barograph records from meteorological stations to measurethe atmosphereshockwave resultingfrom volcanic explosions, but this method has proven to be limited in application and difficult in interpretation. Walker [1980] has proposedfive complementary parameEstimating the 'Explosive Magnitude' of an Eruption ters for estimatingthe scaleof explosiveeruptions,namely, (1) magnitude(determinedfrom the volume of ejecta), (2) Previous attempts at quantifyingthe scale of eruptions intensity(volumeof ejectaper unit time, as determined from have centeredon estimatingthe thermal and kinetic energy the column height and calculated muzzle velocities), (3) involved in the eruption, as estimatedfrom volumesof lava dispersivepower (determinedby column height), (4) vioand pyroclasticdebris, columnheights,and ballistictrajec- lence (releaserate of kinetic energy,analogous intensity to tories of individual fragments [e.g., Sapper, 1927; Escher, but for instantaneous rather than sustained eruptions),and 1933; Tsuya, 1955;Hedervari, 1963; Yokoyama, 1956, 1957]. (5) destructivepotential (the extent of devastation,actual or The most widely adoptedscaleis that of Tsuya [1955], but it potential, caused by an eruption). For the estimate of can only be appliedwhen the volumeof ejectais known and magnitude, Walker proposesto use the scale of Tsuya even then includesboth pyroclasticejecta and lava flows. [1955]; for the other parameters,no scaleshave yet been Gorshkov [1960] proposeda different and novel approach proposed.
ous work, but unfortunatelyincludesin its categoryof 'great eruptions' both great eruptionswith potential climatic impact and many small eruptions with no likely climatic impact.

TABLE 2.

Numbersof ReportedEruptionsby VEI Category,Per Decade, From 1500to the Present


VEI

Decade

1500 1510 1520 1530 1540 1550


1560 1570 1580 1590 1600 1610 1620 1630 1640 1650 1660 1670 1680 1690

0 0 1 10 11 10
10 2 1 0 0 0 0 1 0 2 0 2 1 0

1700 1710 1720 1730 1740 1750 1760 1770 1780


1790 1800 1810 1820 1830 1840 1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 1970

5 1 5 4 1 3 5 6 7
16 4 3 12 17 23 27 30 33 38 19

34 26 32 31 25 51 69 104

10 10 10 12 11 12 21 22 21 20 23 27 26 25 28 26 20 20 22 20 20 24 23 21 20 20 12 8 11 12 18 23 22 13 13 12 14 23 29 42 36 35 63 46 37 76 76 110

12 14 26 23 7 7 9 7 18 20 16 11 4 11 23 18 21 47 35 32 29 26 42 42 47 74 96 108 111 105 122 129 154 172 182 213 207 185 225 227 306 231 299 256 261 345 287 304

1 4 5 8 6 7 11 7 13 10 7 8 5 8 14 14 15 12 15 20 5 9 7 13 2 2 6 9 6 7 5 8 11 11 11 17 10 13 18 12 21 19 13 23 30 40 81 67

0 1 1 1 0 0
1 1 1 3 1 0 0 2 2 0 2 2 2 2 1 1 2 0 2 1 5 2 4 3 1 4 2 0 2 2 1 4 5 1 8 7 3 3 3 5 7 5

1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 2 0 0 0 1 0 0 1 0 1 0 0 0 0 0 0 1 1 0 1 0 1 1 0 1 0 0 1 0 1 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 1 1 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Reporting has improved markedly in recent decades,especiallyfor smallereruptions.See note in text (section 3) regardingcountingconventions.

1234

NEWHALL AND SELF: VOLCANIC EXPLOSIVITY INDEX AND HISTORICAL VOLCANISM

destructiveness and/or (less frequently) dispersive power, violence, and energy release rate, dependingon which data 900 were available. Eruptions can be assigneda VEI on a scale 8oo ._o By Century of 0 to 8 (the maximum number of categories we could - 700 ByDecade "' 600 realistically distinguish),using one or more of the criteria in Table 1. Criteria are listed in this table in decreasingorder of reliability and discriminating power. The overlappingranges 400 between column heights, descriptive terms, and blast durations of successiveVEI values reflect interplay between the 200 z lOO intensity, magnitude, and rate of energy release during an I eruption. 0 100 300 500 760 960 11b0 13b0 100 17b0 19b0 From the viewpoint of those interested in the effects of Year A.D. aerosollayer, no corrections Fig. 1. Total numbers of reported eruptions (VEI 0-8) with eruptionson the stratospheric time, including reported eruptionsfor which a VEI has not been have been made at this stagefor latitude or elevationof the assigneddue to lack of data. Explanationsand implicationsof source vent. Although these parameters will clearly affect apparenttrendsare discussed the text. Seenote in text (section3) in the chancesof gasand dustbeinginjectedin the stratosphere regarding counting conventions. where it will have more potential effect on radiation, the The Volcanological Societyof Japanmadesuggestions for effect of a lower tropopause at high latitudes may be data to be gatheredat modern eruptionsand includedthose counterbalancedby the relatively restricted areal (latitudidata needed for estimating these parameters. However, nal) dispersion of aerosol at those same latitudes [Lamb, 1970; CadIe et al., 1976]. Neither does the VEI attempt to when one examines the historical record of volcanism, there distinguishbetween different types of injection to the atmoare but a handful of eruptions(less than 20 out of the more than 122 largest eruptions since 1500) for which the same sphere, e.g., injections with different amounts of sulfate data is available. A few historical reports contain some but aerosol, differentparticle size distributions,and so on. This not all of the necessarydata; most contain only a brief and information is only now becomingavailable for well-studied often ambiguousdescriptionof the eruptions, e.g., 'violent eruptions. Since the abundanceof sulfate aerosolis imporexplosions, 'terrifying darkness, 'flames and smoke, or tant in climate problems [Pollack et al., 1976; Self et al., simply 'active.' Some eruptionsreportseven in recent years have proven to be false, e.g., a reported major eruptionon 85PaganIsland (Marianas) on May 23, 1966,was later shownto 80have been a grassfire! This is a critical point: the historical record of volcanism contains very little information on 75which one can base a quantitative estmateof eruption scale. 70We believe that workers using Larnb's [1970] list are working with derivative numbersmuchmore quantitativethan the 65. basic data justify. Lamb noted the limitations of his esti60. mates, but many workers continue to use these estimates without due caution. Furthermore, deposits of all but a 55. [] 3
handful of the very biggesteruptionsare now eroded away [] 4
50. 1_>5 or mixed in with sedimentsbeyond the point of recognition. Thus it will be impossibleto build a quantitativedata base 45post facto for most eruptions.Certainly it will be impossible 40. to estimate post facto the five parametersof Walker [1980] for more than this handfulof exceptionally largeand/orwell35described eruptions.
1000 -

500

300

30-

3.

THE VEI:

A NEW INDEX OF THE SCALE


25-

OF EXPLOSIVEERUPTION

Given inadequacy of the historic volcanic record for estimating specific measures explosive of activity [cf. Walker, 1980]but need in variousdisciplines someevaluation for of pastvolcanism, havedesigned composite we a estimate of explosivity using all available data. We have restricted
ourselves to consideration of volcanologicaldata (no atmo-

20-

sphericdata), but have attempted stretchthat datawith a to knowledgeof volcanicbehaviorin many regions,so that it might reach far enoughback in time to be useful in both studies of climate change and studies of the recurrence frequenciesof varioustypes of volcanicactivity. The volcanicexplosivityindex(VEI) is a general indicator of the explosivecharacterof an eruption.It is a composite estimate of Walker's magnitude and/or intensity and/or

1500

1600

1700
Decade A.D.

1800

1900

Fig. 2. Numbers of reported VEI 3, 4, and greater than or equal to 5 eruptions since 1500. The reported frequency of moderately explosive eruptions (VEI 3) has increased sharply while that of larger, highly explosive eruptions (VEI greater than or equal to 4) has increasedonly slightly. See note in text (section 3) regarding counting conventions.

NEWHALL AND SELF: VOLCANIC EXPLOSIVITY INDEX AND HISTORICAL VOLCANISM

1235

1981], VEI's must be combined with a compositionalfactor


before use in such studies.

We have estimated VEI's for over 8000 eruptions and

presentthese in Simkin et al. (in press).A summaryof this data is given in Table 2, and graphical summariesof the numbersof eruptionsin each VEI category are shown in Figures 1 and 2. A list of those eruptions with VEI 4 or greateris given in Table 3, for comparison with other lists of major explosiveeruptions.In order to providean overview of activity in any given year or decade, eruptions that continuedfrom one year into the next are countedonce for each year they continued. For different purposes, it is desirable to count a long-continued eruption only once [cf.
Sirekin et al., 1981].

strongerthan a VEI of 2, but was poorly documented, would likely have been assigneda conservativeVEI of 3, but now with better documentationit might be assigneda VEI of 4. Due to this effect, it is sometimes difficult to distinguish earlier (e.g., 1700-1950)VEI 2 and 3 eruptionsfrom presentday VEI 3 and 4 eruptions, respectively.
4. APPLICATIONS

It is beyond the scopeof this paper to attempt to evaluate the frequency of various volcanic events or to discussthe possible effects of volcanism on climate (or vice versa: see Rampino et al. [1979]). We would like, though, to show

Two conventionshave been used in estimatingVEI which might lead to erroneousor unsupportable conclusions unless they are recognized: 1. Whenever an explosion (magmatic or phreatic) has been indicated without any further description, a default VEI of 2 has been assigned(for early eruptions, see the second convention, below). There are, as a result, many more VEI 2's than any other VEI value. This is partly an artfact of the default assumption,creatinga relatively wide range of characteristicsfor a VEI = 2 eruption, partly an artifact of reporting(in which very small eruptionsmay not be reported), and perhapsalso an actual relative abundance of VEI = 2 eruptions over all others.
2. A correction has been made for the inevitable de-

briefly how the VEI can assistin evaluatingthe completeness of the historicalrecord and thereby assistin assessing the validity of studiesutilizing this record. The simplest way to assess the completeness of the historicalrecord is to preparehistograms recordederupof tions through time (Figure 1). It is immediately obviousthat the numberof recordederuptionshasincreased dramatically
with time. Sirekin et al. [1977, 1981] discuss the various

historical factors affectingthe reporting of volcanic eruptions.

The question then becomes, has there been a real increase

creasein the quality and completeness data back in time. of There is a certain time in every region before which the completenessof the data decreasesmarkedly [see also Lamb, 1970] and thus reports which have survived from before that time may indicate relatively more important eruptions.For example,an explosion reportedfrom Melanesia in 1500 A.D. is likely to have been stronger than an explosionreportedfrom the sameregiontoday. To counteract such an imbalance, all VEI values 1 through 4 prior to a certaindate in eachregionhave beenupgraded 1 VEI unit by unless there is other contradicting information. For the purposesof this study, estimates were made of the current frequencyof eruptionsin each regionand of the date before which the reported frequency was much lower than at at present. In general the date of 1700 A.D. has been used, except in Iceland, Japan, Greece, and Italy (1000 A.D.), parts of Central America and Mexico, including Volcfin Fuego and Volcan Popcat6petl (1500 A.D.) and TongaKermadec-Samoa, Melanesia, and Kamchatka (1800 A.D.). A correctionwill be necessary Aleutian eruptions,f and for when enough data are collected to estimate VEI values. VEI's which have been upgradedare flaggedin the work by Sirekin et al. [1981] with a plus sign. This correctionallows for better comparison individualeruptions,but the reader of is cautionedagainstusingupgradedvaluesin a studyof VEI through time, because it will introduce an apparent maximum in VEI 3 and 4 eruptionsjust before the switchover, (e.g., just before 1700 worldwide). At the other end of the scale, eruptions in the past few decadeshave been relatively well-documented.When documentation is better, eruptionsof that type which once were given a default VEI = 2 can now be given their proper VEI value, and the generaleffectis to increasethe numberof VEI = 3 eruptions.Likewise, an eruptionwhich was known to be

in the frequency of volcanic eruptions or has improved reporting simply increased the number of reported eruptions? Certainly the latter seemsmore likely, and the VEI can help us test this hypothesis, because the larger an eruption,the greaterlikelihoodthat it would be reportedand survive in the annals of history. Consider first a breakdown of eruptions through recent centuries by VEI (Table 2, Figure 2). The number of relatively small, unremarkable eruptions (VEI = 1, 2) has increaseddramatically in recent decades,while the number of large, highlyexplosive,memorableeruptions(VEI = 3, 4, 5,) has only increasedslightlyover the past two centuries. This is the patternone wouldpredictif the overallincrease is simply an artfact of improved reporting.
A variety of statistical methods have been used to test for

completenessin earthquake catalogues[cf. Knopoff and Gardner, 1969; Stepp, 1973; Lee and Brillinger, 1979; and
others], and the same methods can be used with VEI's for

historic volcanic eruptions. As one simple guide to the completenessof eruption records, assume that eruptions have a Poissondistributionwith mean frequenciesin each VEI category equal to the frequencies reported for the 1970's. Thus we assumea mean frequency of 104/decade for VEI 0's, 110/decadefor VEI l's, 304/decadefor VEI 2's, 67/ decadefor VEI 3's, and 82/decadefor VEI's greaterthan or
equal to 3. The standard deviation (sd) for each 'mean' is simply the square root of each value, and the 95% confidence interval for each estimate is the mean _+ 1.96 sd. For

the various scales of eruptions these confidence intervals are, respectively, 104 +_20/decade (VEI 0), 110 _+21/decade (VEI 1), 304 +_34/decade(VEI 2), 67 +_ 16/decade(VEI 3), and 82 _+ 18/decade(.VEI >_3). By this guide, reportingas seen in Table 2 is incompletefor VEI O's and l's prior to about 1970, for VEI 2's prior to about 1950 or 1960, and for VEI 3's prior to about 1960. Eruptions with VEI's of 4 or greater are too infrequent to be covered by this guide, but a visual examination of Figure 2 indicatesthat reporting for these larger eruptionshas been more or lesscompletesince the late 1800's. Clearly theseare not rigoroustests, because

1236

NEWHALL

AND SELF: VOLCANIC EXPLOSIVITY INDEX AND HISTORICAL VOLCANISM

TABLE 3.

ChronologicalList of Largest Explosive Eruptions Since 1500, VEI Greater Than or Equal to 4
Latitude, degrees +63.98 -0.65 +46.20 +37.77 + 13.74
-7.93 + 1.22

Volcano
Hekla Cotopaxi St. Helens Agua de Pau San Salvador
Kelut Galeras

Longitude, degrees -19.70 -78.43 -122.18 -25.47 -89.29


+ 112.31 - 77.30

Date
July 25, 1510 June-, 1534 --, 1540 __+ 70 June 28, 1563 - -, 15757
--, 1586 Dec. 7, 1590

Elevation, m 1491 5897 2549 948 1850


1731 4482 3332 5389 1491 4800 805 1281 3124 1140 1320 4794 3914 725 1024 1850

VEI
4 4 5 4 4?
4 4 4? 4 4 4 4 4 4? 4 5? 4 4? 5 5 4?

Raung Ruiz Hekla Huaynaputina Furnas Vesuvio Llaima Komaga-take Awu Guagua Pichincha Quilotoa Usu Tarumai San Salvador Gamkonora Tongkoko Iwate Chikurachki Hekla Long Island Fuji Chirpoi group Katla

-8.12 +4.88 +63.98 + 16.58 +37.77 +40.82 -38.70 +42.07 +3.67 -0.17 -0.85 +42.53 +42.68 + 13.74 + 1.38 + 1.52 +39.85 +50.33 +63.98 -5.36 +35.35 +46.52 +63.63

+ 114.04 -75.37 - 19.70 -70.87 -25.32 +14.43 +71.70 + 140.68 +125.50 -78.60 -78.90 + 140.83 + 141.38 -89.29 + 127.52 + 125.20 +141.00 + 155.47 -19.70 +147.12 +138.73 + 150.88 -19.03

--, 1593 March 12, 1595 June 3, 1597 Feb. 19, 1600 Sept. 3?, 1630 Dec. 16, 1631 Feb.-, 1640 July 31, 1640 Jan. 4, 1641 Oct. 27, 1660 Nov. 28, 1660 Aug. 16, 1663 Aug. 6, 1667 --, 1671 May 20, 1673 --, 1680 Feb. 29, 1686 --, 1690 __+ 10 Feb. 13, 1693 --, 1700 --+100 Dec. 16, 1707 - -, 17127 May 11, 1721

1635 1149
2041 1817 1491

4 4
4 4 4

1304 3776 624 1363

6 4 4? 4

OraefajSkull
Tarumai Oshima-o-shima Cotopaxi Katla Jorullo
Peteroa Jorullo Komagatake

+64.00
+42.68 +41.50 -0.65 +63.63 +19.03
-35.25 + 19.03 +42.07

-16.65
+ 141.38 +139.37 -78.43 -19.03 -101.67
-70.57 - 101.67 + 140.68

Aug.3, 1727
Aug. -, 1739 Aug. 23, 1741 Nov. 11, 1744 Oct. 17, 1755 Sept. 9, 1759
Dec. 3, 1762 - -, 1764 - -, 1765

2119
1024 714 5897 1363 1330
4090 1330

4
5 4 4 5 4
4 4

Hekla Cotopaxi Papandayan Raikoke Asama Lakagigar Asama Alaid San Martin Pogromni St. Helens
Soufrire

+63.98 -0.65 -7.32 +48.25 +36.40 +64.07 +36.40 +50.80 + 18.58 +54.57 +46.20 +13.33 + 13.25 -8.25 +71.08 +19.42 +42.53 -7.25 +54.75 + 12.98 +63.98 +2.37 +50.33 +56.78 +42.07 +2.37 +48.87 +64.42 +65.03

-19.70 -78.43 +107.73 + 153.25 + 138.53 - 18.25 + 138.53 + 155.50 -95.17 -164.70 - 122.18 -16.18 + 123.68 + 118.00 -8.17 -103.72 + 140.83 + 108.05 - 163.73 -87.57 - 19.70 -76.38 + 155.47 +161.58 + 140.68 -76.38 + 154.17 - 17.33 - 16.75

April 5, 1766 April 4, 1768 Aug. 11, 1772 --, 1778 May 9, 1783 June 6, 1783 July 26, 1783 - -, 1793 March 2, 1793 --, 1795 --, 1802 April 27, 1812 Feb. 1, 1814 April 5, 1815 --, 1818 Feb. 15, 1818
March 9, 1822 Oct. 8, 1822 - -, 1825 June 20, 1835

1140 1491 5897 2665 551 2550 500 2550 2339 1550 2286 2549

4 4 4 4 4 4 4 4 4 4 4 4

(St. Vincent) Mayon Tambora Beerenberg Colima Usu Galunggung Isanotski Cosiguina Hekla Purace Chikurachki Sheveluch Komagatake Purace Sinarka GrimsvOtn Askja

Sept. -, 1845
Dec.-, 1849 Dec. -, 1853 Feb. 17, 1854 Sept. 25, 1856 Oct. 10, 1869 - -, 1872 Jan. 8, 1873 March 29, 1875

1178 2462 2851 2277 3960 725 2168 2480 859 1491 4590 1817 3395 1140 4590 934 1719 1510

4 4 7 4 4 4 5? 4 5 4 4 4 5 4 4 4 4 5

NEWHALL

AND SELF: VOLCANIC EXPLOSIVITY INDEX AND HISTORICAL VOLCANISM

1237

TABLE 3.

(Continued)

Volcano
Suwanose-Zima

Latitude, degrees
+29.53

Longitude, degrees
+ 129.72 --,

Date
1877

Elevation, m
799

VEI
4

Cotopaxi Nasu Krakatau Augustine Tungurahua


Tarawera

-0.65 +37.12 -6.10 +59.37 -1.47


-38.23

-78.43 +139.97 + 105.42 -153.42 -78.45


+176.51

June 25, 1877 July 1, 1881 Aug. 26, 1883 Oct.6, 1883 June 11, 1886
June 10, 1886

5897 1917 300 1227 5016


1111

4
4
6
4
4

Bandai
Suwanose-Zima Dona Juana

+37.60
+29.53 + 1.52

+140.08
+129.72 -76.93

July 15, 1888


Oct. 2, 1889 Nov. 13, 1899

18i9
799 4250

4
4

Mt. Pel6e
Soufri6re

+14.82
+13.33

-61.17
-61.18

May2, 1902
May 6, 1902

1397
1178

(St. Vincent)

Mt. Pel6e
Santa Maria Thordarhyrna Ksudach Tarumai Taal Novarupta (Katmai) Colima Sakurazima Agrigan Tungurahua Katla Manam Puyehue Raikoke Komaga-Take Kliuchevskoi
Fuego Quizapu
(Cerro Azul) Rabaul Kliuchevskoi Sarychev Hekla Lamington Ambryn Bagana Mt. Spurr

+ 14.82
+ 14.75 +64.27 +51.83 +42.68 + 14.00 +58.28 +19.42 +31.58 + 18.77 - 1.47 +63.63 -4.10 -40.58 +48.25 +42.07 +56.18
+ 14.48

-61.17
-91.55 -17.60 + 157.52 + 141.38 + 121.00 -155.17 -103.72 +130.67 + 145.67 -78.45 -19.03 +145.06 -72.10 +153.25 +140.68 + 160.78
-90.88

May 8, 1902
Oct. 24, 1902 May 28, 1903 March 28, 1907 April 12, 1909 June 27, 1911 June 6, 1911 June 20, 1913 June 12, 1914 April 9, 1917 April 4, 1905 Oct. 12, 191.8 Aug. 11, 1919 Dec. 13, 1921 Feb. 15, 1924 June 17, 1929 March 25, 1931
Jan. 21, 1932

1397
2700 1660 1079 1024 300 2285 4100 1118 965 5016 1363 1300 2240 551 1140 4850 3763

6
4
5
4
4
6
4?
4
4
4
4
4
4
4
4
4

-35.67 -4.27 +56.18 +48.10 +63.98


-8.94 - 16.25
-6.14

-70.77 + 152.20 +160.78 +153.20 - 19.70


+148.17 + 168.08 +155.19

April 10, 1932 May 29, 1937 Jan. 1, 1945 Nov. 9, 1946 March 29, 1947
Jan. 15, 1951 Sept. -, 1951 Feb. 29, 1952

3810 229 4850 1497 1491


1780 1334
1702

5
4?
4
4
4

4
4

Nilahue
Bezymianny Agung

+61.30 -40.35
+56.07 -8.34

-152.25 -72.07
+ 160.72 +115.50

July 9, 1953 July 26, 1955


March 30, 1956 March 17, 1963

3375 400
2800 3142

4
4

5
4

Sheveluch Taal

+56.78 + 14.00

+161.58 + 121.00

Nov. 12, 1969 Sept. 28, 1965

3395 300

4
4

Kelut
OldoinyoLengai Awu Fernandina Tiatia
Fuego

-7.93
-2.75 +3.67 -0.37 +44.35
+14.48

+ 112.31
+35.90 +125.50 +91.55 + 146.25
-90.88

April26, 1966
Aug. -, 1966 Aug. 12, 1966 June11, 1968 July 14, 1973
Oct. 10, 1974

1731
2880 1320 1495 1822
3763

4
4
4
4

PloskyTolbachik Augustine
Bezymianny

+ 55.93 +59.37
+56.07

+ 160.47 - 153.42
- 160.72

July 6, 1975 Jan. 22, 1976


Feb. 11, 1979

3085 1227
2800

St. Helens

+46.20

-122.18

May 18, 1980

1920

4
4
4
5

Notes-Volcano nameis the name generally used the volcano for whichprodu, the eruption, ced not
the name of the individual vent or the eruption itself. Positive latitude is north; negative, south. Positive longitude is east; negative, west. Dashes in the date column indicate unknown months and days.Elevationis from sealevel; If the ventwasvery different height(e.g., SantaMaria, 1902)from in the volcano itself, the vent heightis given.

the actual frequencies are not known; rather, these and the
complete), two explanationsare possible.Either the process simple histogramsserve as guidesto completenessrelative
is not Poissonian, or there was actually a deterioration in to recent records.
reporting. Given the overall suspectnature of the historical Where decadeswith incomplete reporting follow decades
record, plus reasonable historical explanations for most of with what appears to be complete reporting, as for VEI 2
these cases[Simkin et al., 1981]we concludethat reporting eruptions in the 1900's (possibly complete) and 1910's (in-
has deteriorated on occasion and that for volcanism over the

1238

NEWHALL

AND SELF: VOLCANIC EXPLOSIVITY INDEX AND HISTORICAL VOLCANISM

volcanoes accordingto barograms,Bull. Volcanol., 23, 141-144, past few centuriesat least the Poisson distribution remains 1960. the most likely. We do not deny that there may be special Hammer, C. U., H. B. Clausen, and W. Dansgaard,Greenland ice peaksor lulls in the record,especially specific in regions, but sheet evidence of post-glacialvolcanismand its climatic impact, do note that in most casesof apparentchange,reportingis Nature, 288, 230-235, 1980. too incompleteto know whether the changeis real or not. Hedervari, P., On the energy and magnitudeof volcanic eruptions,

To usethe VEI in assessing validityof volcanological the


or climatologicalstudies,one must ask whether the historical record of volcanism is reasonablycompletefor the time spanand purposes each study.A simpletest suchas that of described above enables one to determine the period for which the recordis complete(or, strictlyspeaking, which for the record is as completeas it has ever been). For example, the set of eruptsions which needs to be consideredfor volcano-climate studies is probably the set with VEI's greater than or equal to 3, from 1755to the present.From the above, the reporting of VEI = 3 eruptions is incomplete prior to 1960, and therefore the data set in consideration is incomplete prior to about 1960.
5. SUMMARY

Bull. Volcanol., 25, 373-385, 1963. Hirschboeck, K. K., A new world wide chronology of volcanic

eruptions, Palaeogeogr. Palaeo'Climatol. Palaeoecol., 22329,


241, 1980.

Humphreys, W. J., Physics of the Air, McGraw-Hill, New York,


1940.

Knopoff, L., and J. K. Gardner, Homogeneous catalogues of earthquakes, Proc. Natl. Acad. Sci. U.S.A., 63,1051-1054, 1969. Lamb, H. H., Volcanic dust in the atmosphere;with a chronology

and assessment its meterological of significance, Phil. Trans.R.


Soc. London, 266, 425-533, 19702

Lamb, H. H., Supplementaryvolcanicdust veil assessments, Clim.


Monit., 6, 57-67, 1977. Lee, W. H. K., and D. R. Brillinger, On Chinese earthquake history--An attempt to model an incomplete data set by point processanalysis, Pure Appl. Geophys., 117, 1229-1257, 1979. Mitchell, J. M., Jr., A preliminary evaluation of atmospheric pollution as a cause of the global temperature fluctuation of the last century, in Global Effects of EnvironmentalPollution, edited by F. Singer, pp. 139-155, D. Reidel Dordrecht, Netherlands,
1970.

The precedingdiscussion has a dual purpose.It is meant to introduce the VEI as a useful tool for comparing the

relative explosivity of historic eruptionsand to illustrate Newell, R. E., Factors governing tropospheric mean temperature, Science, 194, 1413-1414, 1976. applications of the VEI to studies utilizing the historical
record of volcanism. In these applications,one messageis clear: the historical record is generally incomplete.The VEI can help detect incompleteness reportingbiasesand can and help one select subsetsof the historicalrecord suitablefor
Pollack, J. B., O. Toon, C. Sagan,A. Summers,B. Baldwin, and W. Van Camp, Volcanic explosion and climatic change:A research assessment,J. Geophys. Res., 81, 1071-1083, 1976. Rampino, M. R., S. Self, and R. W. Fairbridge, Can rapid climate change cause volcanic eruptions?, Science, 706, 826-829, 1979. Robock, A., A latitudinally dependentvolcanic dust veil index and its effect on climate simulations, submitted to Palaeogeogr.
Palaeoclimatol. Palaeoecol., 1981.

eachstudy.We wouldlike to be moreoptimistic propose and


a fully quantitative VEI, but we concludethat for the present a more immediate need is to improve the raw quantitative data that we gather on eruptions. For historic explosive

Sapper, K., Vulkankunde, Engelhorn Verlag, Stuttgart, 1927. Schneider, S. H., and C. Mass, Volcanic dust, sunspots and volcanism will have to turn not only to conventional temperature trends, Science, 190, 741-746, 1975. we historic reports and the near-volcanogeologicrecord, but Self, S., M. R. Rampino, and J. J. Barbera, The possibleeffectsof large 19th and 20th Century volcanic eruptions on zonal and also toward new techniques (such as ice core analysis, cf. hemisphericsurface temperatures,J. Volcanol. Geotherm. Res., Hammer et al. [ 1980]). An even more promisingdirection is in press, 1981.

toward better reporting in the future. More quantitative Simkin,T., L. Seibert,L. McClelland,and W. G. Melson,Historic globalvolcanism: Eruptiondurations, frequencies characterand descriptionsof future eruptions will aid not only in their
classification, but also in suggestingthe best parameters to istics, paper presented the IAVCEI GeneralAssembly,Int. at

searchfor in the recordsof past eruptions.

Assoc. of Volcanol. and Chem. of the Earth's Inter., Durham, England, 1977.

Acknowledgments. We thank Robert W. Decker, who suggested G. Newhall, and J. Latter, Volcanoes the World, Hutchinson of that a simple scale for eruptionsbe devised, and we thank Tom Ross, New York, in press, 1981. Simkin, Michael Rampino, Richard E. Stoiber, and Lee Siebertfor Stepp,J. C., Analysis completeness the earthquake of of sample in their help throughout this project.Computer time wasprovidedby Puget Sound area and its effect on statistical estimates of earththe Kiewit ComputationCenter, Dartmouth College.This work was quakehazard, NOAA Tech. Rep. ERL 267-ESL30, 16-28, U.S. performed under NASA Grant NSG5145. Departmentof Commerce,Washington,D.C., 1973.
REFERENCES

Simkin, L. Seibert, McClelland, G. Melson, Bridge, T., L. W. D. C.

Baldwin, B., J. B. Pollack, O. B. Toon, C. Sagan,and A. Summers, Stratosphericaerosolsand climate change,Nature, 263, 551-555,
1976.

Tsuya, H., Geological and petrological studies volcanoFuji, 5, of TokyoDaigaku JishinKenkyusho lho, 33, 341-382, 1955. Walker,G. P. L., The Taupopumice: Product themostpowerful of known(Ultraplinian)eruption?, Volcanol.Geotherm. J. Res., 8,
69-94, 1980. Wexler, H., On the effects of volcanic dust on insolation and

.Bray, J. R, Volcanism and glaciation during the past 40 millenia,

changes, Science, 207, 1041-1044, 1980. in 1, Cadle, R. D., C. S. Kiang, and J. F. Louis, The global scale Yokoyama,I., Energetics activevolcanoes, TokyoDaigaku Jishin Kenkyusholho, 34, 185-195, 1956. dispersionof the eruption cloudsfrom major volcanic eruptions, Yokoyama,I., Energetics activevolcanoes, in 2-3, Tokyo Daigaku J. Geophys. Res., 81, 3125-3132, 1976.
Jishin Kenkyusholho, 35, 75-107, 1957. Cronin, J. F., Recent volcanism and the stratosphere,Science, 172,

847-849, 1971.
Escher, B. G., On a classificationof central eruptions accordingto gaspressureof the magmaand viscosityof the lava, Leidse Geol. Meded., 6, 45-50, 1933. Gorshkov, G. S., Determination of the explosion energy in some

weather, 1, Bull. Am. Meteorol. Soc., 32, 10-15, 1951. Nature, 252,679-680, 1974. Bryson, R. A., and B. M. Goodman,Volcanicactivity and climatic Wexler, H., Volcanoesand world climate, Sci. Am., 186, 74-80, 1952.

(Received June 1, 1981' revised August 24, 1981' accepted August 24, 1981.)

Anda mungkin juga menyukai