Vulcanologia

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 nonexplosiveeruptions 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 representedthegeologic 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 Richtermagni- 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 acomrequire 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, particularlytemperature, 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,

INTRODUCTION1980; 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) abasisfor 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 amountof 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...