Eruption Data Criteria

Introduction and Cautions

Any attempt to assess the distribution of volcanism through time must take into account the variable definitions of the word eruption. We consider an eruption to consist of the arrival of solid volcanic products at the Earth's surface. This can be in the form of either the explosive ejection of fragmental material or the effusion of initially liquid lava. This definition excludes energetic, but non-ash-bearing steam eruptions. The ejection of fragmental material, however, does not require magmatic explosions producing fresh (juvenile) pyroclastics; phreatic explosions of greatly variable intensity are produced by the interaction of volcanically generated heat and near-surface water and can eject significant amounts of old material. Most eruptions in fact result from a combination of magmatic and non-magmatic processes and are referred to as phreatomagmatic.
The duration of eruptive events also influences eruption documentation. The word eruption has variously been applied to events ranging from an individual explosion to eruptive periods lasting up to hundreds of years. Quiescent periods are common during eruptions, and we have attempted to standardize eruption data by considering clearly linked events separated by surface quiet of up to three months to be part of the same eruption. This distinction is possible at volcanoes in more populated areas, but can be problematical in the case of scattered observations from travelers who witness an ongoing eruption from a remote volcano at separate times. Furthermore, the end of an eruption is often less dramatic than its start and therefore is often not documented; consequently many eruptions have only a start date. Further discussion of the uncertainties of eruption reporting and documentation can be found in Simkin and Siebert (1994).
Eruptions are documented in a wide variety of ways. The initial IAVCEI volcano catalogs were almost entirely restricted to historical eruptions documented at or near the time of their occurrence. Even historically documented eruptions are subject to vagaries such as the extent of monitoring, the proximity (and experience) of observers, and inclement weather that can inhibit observations. Tabular compilations are devoid of essential caveats and explanatory words and underscore the cautions necessary in interpreting these events.
Eruptions preceding human observation have been documented with a variety of techniques. These are shown by an alphabetical code in the table below. The dating methods range from radiometric procedures such as radiocarbon or fission track to tephrochronology, the careful study of the stratigraphic relationships of dated and undated tephra layers. Users should note in particular the distinction between uncorrected radiocarbon dates (C) and dates corrected for past variations in carbon isotopic ratios of atmospheric carbon dioxide (G). These dates are comparable (<100-150 years) for the last 2500 years, but begin to diverge to as much as 700-900 years for the last 4000 years of the Holocene. Some eruption reports, established in the volcanological literature such as the Catalog of Active Volcanoes of the World, have subsequently been found to be incorrect. Rather than delete these events, which could appear to be mistaken omissions, they have been flagged with an "X" to note that they have been discredited. These events (further distinguished by the fact that the eruption year is unbolded) should not be included in any eruption totals. Bolding is visible using most browsers when style sheets are enabled.
Caution is also necessary in the interpretation of historical eruption dates. In sparsely populated regions, reported eruption dates (even in recent years) may be that of major eruptive events likely to be noticed by distant observers, and minor preceding eruptive activity may go unreported. Even in populated regions, the likelihood that only major events are reported increases for events prior to the past few centuries.
Earlier historical eruption reports are further complicated by the great temporal and spatial variability in useage of calendars to document time. The Roman Julian calendar (referred to as the Old Style calendar) used in western Europe for more than 1500 years was surplanted by Papal decree in 1528 by the more precise Gregorian calendar (referred to as the New Style calendar). However, adoption of the New Style calendar was regionally variable and in some cases did not take place until the first part of the 20th century. The type of calendar used is rarely specified in the literature, and consequently it is often not known whether eruption dates are reported using Julian or Gregorian calendars (or even using regionally adopted lunar calendars). Japanese eruption dates from the Catalog of Active Volcanoes of the World have been converted to New Style dates, even for pre-1528 eruptions (Hayakawa 1996, pers. comm.); however, elsewhere the type of calendar used is often not known. Consequently we have not attempted to convert dates to the New Style calendar, but have accepted the dates used in earlier published papers or compilations. Users should note that Gregorian dates progressively diverge from Julian dates beginning with the adoption of the Julian calendar in the mid-1st century BC. Old-Style (Julian) dates are 3 days earlier during the 7th century, for example, and increase to 10 days earlier at the time of the adoption of the New-Style (Gregorian) calendar in the 16th century.

Area of Activity

The area of activity is listed for eruptions originating from known locations other than the central summit conduit. The names of flank vents and/or their location on the edifice are listed here when known. Comments (enclosed in parentheses) sometimes denote uncertainties in eruption validity or are used to identify the designated labels of specific tephra deposits.

Code Before Date (Eruption Uncertainty and Dating Techniques)

Volcanologists have used a wide range of procedures to date the prehistorical eruptions that are critical in determining the geologic history of a volcano. When an eruption date in this compilation is not historical, the dating technique used is
shown by a letter code immediately preceding the start year. By far the most commonly used techniques are radiocarbon dating (corrected and uncorrected) and tephrochronology. These and other techniques are shown in the table below and then
briefly described, along with associated age uncertainties.

A "?" before the eruption date denotes uncertainty about the validity of the eruption. This is applied, for example, to common reports that a volcano was "smoking," which could denote either simple steam emission or ash-bearing eruption plumes. The year columns of valid eruptions are bolded (visible using most browser settings) to distinguish them from the unbolded dates of both uncertain and discredited eruption reports. In some cases eruptive activity was observed from a distance, without clear indication of which volcano it originated from. These events are attached to the most probable volcanic source, but a "@" precedes the date to indicate uncertainty about the source of the eruption.

Some events--although once established in the volcanological literature, such as the CAVW--have since been discredited. These are included in none of our eruption totals, but great effort is often invested in proving a reported eruption to be
false, and we thought it better to retain these "non-events"--in a form that allows easy identification (and removal)--rather than have them appear to readers of earlier compilations as mistaken omissions. Discredited eruptions are
flagged by an "X" before the event date. Both discredited and uncertain eruptions can be further distinguished in the eruption table by the absence of bolding of dates.

Code Before the Date

- = BC date ? = eruption itself uncertain @ = eruption locality uncertain X = discredited eruption A = anthropology C = carbon-14 (uncorrected) D = dendrochronology (tree ring) E = surface exposure F = fission track G = carbon-14 (corrected)

H = hydration rind - glass  I = ice core  K = K-Ar  L = lichenometry  M = magnetism  N = thermoluminescence  R = Ar-Ar  S = SOFAR (hydrophonic)  T = tephrochronology  U = Uranium-Series V = varve count

Dating Techniques

A = "ANTHROPOLOGY." Eruption dates carrying this designation include native legend (or "traditional" dates) or dates obtained from the age of human artifacts or structures buried or entrained in tephra layers
or lava flows. Some are entered without a date uncertainty code (e.g., Bedouin legends of a 640 AD Arabian eruption), while other uncertainties range to 50 years (the "11th century" eruption of Mexico's
Michoacán-Guanajuato), but all should be treated with some caution, recognizing the human ability to misremember an undocumented date. Still other dates have been obtained by anthropologists but entered in our file under the dating
technique used (commonly 14C).

C = "¹⁴C," or UNCORRECTED RADIOCARBON. This is the most common dating technique used for prehistorical eruptions. The technique is based upon the 1951 discovery that wood and other organic matter contains minute amounts of carbon's radioactive isotope (of atomic weight 14). When the organism dies, however, its radioactive carbon is no longer replenished and the proportion of ¹⁴C in its carbon begins to decrease by radioactive decay. Because this decay rate is accurately known, careful laboratory measurement of the ¹⁴C/¹²C ratio in prehistoric wood can accurately date that wood's death. Although the half-life of ¹⁴C is about 5568 years and its initial concentration is only one part in a trillion (10¹²) parts of ¹²C, ages to 100,000 years are now being successfully measured.

Radiocarbon dates are normally expressed in years BP ("before present"), and we have followed the standard convention of treating 1950 as "present" (unless otherwise stated) in converting to calendar year dates. Some uncertainty in radiocarbon dates is guaranteed by analytical error and the fact that the ¹⁴C decay rate is known only to within 30 years. Most authors combine these and other factors in a single uncertainty, or "±" value, after each radiocarbon date presented. We then accept the author's reported date and attach the appropriate uncertainty code upon entry to our file. Many eruptions' radiocarbon dates in this compilation would have been "historical" if they had taken place in southern Italy where the written record extends to 1500 BC. The "uncorrected" adjective applied to this technique is important: note the distinction between uncorrected radiocarbon dates (C) from corrected radiocarbon ages (G) discussed below under "corrected radiocarbon."

D = "DENDROCHRONOLOGY." The annual character of tree rings was first noticed by the ancient Greeks, and precise chronologies have been developed from comparison of tree-ring growth patterns. Eruptions frequently
perturb the growth cycle of nearby trees, and comparison of narrowed tree ring intervals from affected trees with regional tree-ring chronologies allows the precise dating of volcanic eruptions. Distant but long-lived trees bear frost rings from
known historical eruptions, and the hope is strong that tree-ring chronologies will help establish a detailed record of the planet's largest eruptions.

Other paleobotanical techniques can also be useful to volcanology. The famous eruption resulting in Oregon's Crater Lake, for example, is dated only to 50 years (nearly 6000 years ago), but careful study of pollen associated with its
volcanic ash in a far-away Montana bog shows that the eruption began in the autumn and apparently continued for at least 3 years. Analysis of annual layers in Irish peat bogs is revealing detailed records (including fine particles of volcanic
ash) from Icelandic eruptions, and leaf impressions under Japanese ash layers are dating prehistoric eruptions to the exact season of the year. In New Zealand, insect remains preserved by the famous Taupo eruption of the second century AD have
shown that the eruption took place in the early afternoon. The application of biology to eruptive deposits holds great promise for unraveling the recent histories of many volcanoes.

E = "SURFACE EXPOSURE" This relatively new technique measures the exposure ages of rocks at the earth's surface to cosmic-ray production. Cosmic-ray production rates are dependent on both altitude and latitude, but if local cosmogenic helium (3He) production rates can be determined, helium isotopes can be used to determine the ages of rocks exposed to the surface since their formation. Chlorine isotopes (³⁶Cl) have also been used to date young volcanic rocks. Careful sampling is required, but surface exposure procedures such as these have been used to date late-Pleistocene and Holocene lava flows. The technique has somewhat larger uncertainties (many hundreds to more than a thousand years) than ages from calibrated radiocarbon dating (see below). It has been relatively infrequently used but is useful where organic material for radiocarbon dating is unavailable and correlates fairly well with radiocarbon ages where both techniques have been used.

F = "FISSION TRACK." Another relatively new technique depends upon the natural spontaneous fission decay of uranium. The resulting heavy fission particles leave minute damage tracks in volcanic glass that can be
revealed by chemical etching of a cut and polished surface. The number of tracks per unit area, counted microscopically, is proportional to the age of the glass (for any given uranium content) and can therefore provide eruption dates. Although
the technique is capable of better accuracy, one fission track date included here--1000 BP from Canada's Mount Edziza--carries the largest uncertainty in the VRF: 6000 years!

G - ¹⁴C, or "CORRECTED RADIOCARBON." Careful radiocarbon dating has been done on selected portions of long-lived bristlecone pine trees that can be independently dated by tree-ring techniques. This work shows generally
close agreement between the two methods for the last 2,500 years, but they then start to diverge until "true" tree-ring dates exceed radiocarbon dates by about 900 years for specimens at the limit of the tree-ring time scale (about
7500 years ago). The reason for this divergence is apparent variation in past content of atmospheric radiocarbon. When a corrected date is available we have proceeded it by the letter G (which can be thought of, mnemonically, as a slightly
altered C), but many published dates are not accompanied by all the information required for an accurate correction, and we have not applied a correction factor to uncorrected dates in our file. The mixing of uncorrected radiocarbon dates with a
growing number of calendar dates can be very misleading to readers who do not pay attention to the letter code in front prehistorical dates. It is imperative that readers be aware of the significant age difference between C and G dates: to
100-150 years during the last 2500 years rising to 900 years in the early Holocene.

H = "HYDRATION RIND." Obsidian flows are formerly-molten liquids that cooled too quickly to permit growth of the crystals that make up most volcanic rocks. The resulting glass is unstable and gradually decomposes by
the addition of moisture from the atmosphere. The thickness of the hydration rind on an obsidian flow surface is proportional to the time that it has been exposed to the atmosphere, and this thickness has been used to date 10 flows in our file,
mainly from Oregon's Newberry Caldera and California's Mono Craters. Uncertainties are large (several hundred to more than a thousand years) for this technique.

I = "ICE CORE." The far-traveled aerosol of major eruptions eventually settles to the earth's surface, leaving a chemical trace in glaciers and ice caps that grow by annual accumulation of snow. Cores through these
annual layers then provide an important record of past volcanism that can extend, as with the new cores from Greenland, over 250,000 years. Whereas tree ring studies give unequivocal link to volcanism only if close to the source, strong sulphate
layers are formed in the ice of both polar regions by major historical eruptions, and similar [even larger] layers in prehistoric portions of the core point clearly to volcanism with global distribution as the sulphate source. This gives the
exciting potential of establishing a complete chronology of large eruptions, but the difficulty lies in determining what volcano was the source of a specific sulphate layer.

K = "K-Ar", Potassium-Argon dating. One of the most widely used methods of geochronometry. Like radiocarbon dating, it depends upon the relative proportions of parent (40K) to daughter (40A) isotopes, and the
well-established half-life of that constant decay. It has been used to date rocks approaching the age of the earth [4.5 x 109 years], but is rarely used on materials younger than 100,000 years. The technique has been applied to some Holocene
dates, but the associated uncertainties are large (often several thousand years).

L = "LICHENOMETRY." The slow but rather regular growth rate of lichens on a lava flow surface has been used to date two eruptions on Penguin Island, Antarctica [1683 and 1905 AD]. The technique is useful for
establishing relative ages on young lava flows, but absolute ages require accurate baseline growth rates, under comparable conditions of-- climate and substrate, that are rarely available over more than a century.

M = "MAGNETISM." When lava cools from its molten state, it often retains an accurate "memory" of the earth's magnetic field at that time. Secular variation, or historical wander of the earth's
magnetic poles, has been large enough that careful study of -a lava's magnetic "memory" may reveal its approximate date of cooling. Most dates carry uncertainties in the 25-150-year range. The accuracy of the technique decreases
greatly for events older than a few thousand years, and the oldest eruption in this compilation dated by magnetics--Oregon's Mount Bachelor around 5800 BC--carries a ~750-year uncertainty.

N = "THERMOLUMINESCENCE" dating depends on the effects of radioactive decay (like the Fission Track technique) rather than direct counts of isotopic ratios. Some electrons freed during decay are trapped in crystal defects, and laboratory heating frees them, with light being produced in the process. The amount of light depends, in part, on the age of the crystal. This technique is much used by archeologists, but has uncertainties often larger than those from radiocarbon dating. This dating technique was referred to in previous compilations by the dating method code "U," a letter now used for Uranium-series dating.

R = "ARGON-ARGON" (⁴⁰Ar/³⁹Ar) dating was first developed in the late 1980s. It offers greater precision than typical K-Ar and requires much smaller amounts of material. During stepwise heating, a spectrum of apparent ages is shown by changing isotopic ratios (reflecting contamination) until reaching a plateau representing the crystal's true age. The technique is particularly useful for relatively young materials, and is bringing new order to geologic time scales over the past few tens of millions of years. Holocene dates can have uncertainties of several thousand years, but stratigraphically consistent ⁴⁰Ar/³⁹Ar ages have successfully been obtained for Holocene volcanic rocks.

S = SOFAR, or submarine "Hydrophone" detection. Explosive eruptions on the sea floor send out shock waves through the water in much the same way that earthquakes send shock waves through the solid earth's crust.
The velocities are slower, about 5300 km/hr, but they travel for long distances through the SOFAR channel (a layer of water within 1200 m of the surface) and their arrival times at submarine hydrophones can be used to locate the eruption in the
same way that seismologists locate earthquake epicenters. Study of hydrophone records from observed submarine eruptions has shown features characteristic of volcanism, and when these features appear on records from more remote parts of the sea
floor they have been used to locate and to date (often to the hour and minute) volcanism that would otherwise have been completely missed.

Although the quiet, nonexplosive effusion of lava that typifies most seafloor volcanism is difficult to detect by hydrophones, earthquake swarms commonly accompany these more gentle eruptions in places such as Hawaii and Iceland, and such swarms
from submerged seamounts have been interpreted as submarine eruptions. We have included several volcanoes because of earthquake swarms (STATUS entered as "Seismicity") and the fresh glass dredged from their submerged summits. However,
the earthquake swarms might represent magma movement without eruption, so we have preceded these dates with a question mark rather than a symbol representing a "seismic" dating technique.

T = "TEPHROCHRONOLOGY." Aristotle used the Greek word for ash, "tephra," in describing an eruption on the island of Vulcano. Because modern volcanologists define "ash" as particles smaller than 2 mm
in diameter, a broader term is useful for describing material of all sizes explosively ejected by volcanoes. In 1944, Sigurdur Thorarinsson proposed the word "tephra" for this purpose and it is widely accepted today. Tephra from large
explosive eruptions may be distributed over enormous distances, forming a distinctive layer that later proves useful as a "marker" horizon dating nearby layers of sediment. Careful mapping of layers throughout a volcanic area can
develop a relative sequence of overlapping ash layers. When some of these ash layers are dated, either historically or by some other technique, then dates (generally with large uncertainty) can be assigned to the intervening layers in this
relative sequence. The technique is a broad one, embracing a variety of field geologic and stratigraphic methods, and we have used this designation to cover prehistoric dates for which our source specified no technique. Uncertainties are often
large (hundreds to a few thousand years), and those dates without listed uncertainties should likewise be treated with caution.

U = "URANIUM-SERIES." Several dating techniques utilize Uranium-series disequilibrium ratios. The Uranium-Thorium disequilibrium series is often used to date carbonate materials such as speleothems, travertines, corals, deep sea sediments, bones, teeth, peat, or evaporites. More complex applications of this technique have also been applied to volcanic rocks. ²³⁰Thorium, part of the ²³⁸Uranium decay series, has a half-life of about 75,000 years, in comparison to the half-life of ²³⁸Uranium of 4,470,000,000 years. When the amounts of Uranium and Thorium isotopes are compared, an estimation of the age of an object can be obtained. This technique has been applied to volcanic rocks as young as the end of the Pleistocene and the beginning of the Holocene and has relatively large uncertainties (from hundreds to a few thousand years) during these time intervals. Other Uranium-series nuclides have shorter half-lives. ²²⁶Ra-²³⁰Th ratios have been used to date eruptions during the mid-Holocene, and ²¹⁰Po-²¹⁰Pb ratios have been applied to eruptions as young as a few decades or less.

V = "VARVE COUNT." Seasonal changes affect the sediment accumulation in many small lakes, particularly where the spring melting of ice provides an annual layer of coarse sandy particles to the lake floor in alternation
with the finer clay deposited through the rest of the year. These layers, or varves, can later be counted to establish the date for a layer of volcanic ash in their midst. Like tree rings and ice-core layers, these annual layers provide very
accurate dates under ideal conditions and careful work, but uncertainty increases with age and non-ideal conditions. Few dates in the file carry stated uncertainties (up to several hundred years). The sediments of Turkey's Lake Van provide a
remarkable record--16 eruptions since 8104 BC--of nearby Nemrut volcano, but uncertainties are not listed.

Code After the Date (Date Uncertainty)

Codes after dates denote uncertainties about the date itself. When the date is known only to the year or month, the following columns are left blank. Letter codes are used when the size of the dating uncertainty is known. This allows us to deal with eruption dates known only between two observations ("after July 10 but before July 24" would be shown as 0717g). Frequently used codes include a "t" in the year column (± 50); a 17th century eruption would appear as 1650t. A "p" in the month column (± 30) likewise would be used for an eruption known only to have begun in July or August. Larger uncertainties (such as those accompanying radiocarbon dates) may not exactly match one of the codes below; in these cases the closest available letter is used.

A ">" symbol after the year or the day indicates that the eruption was continuing as of that date. There is substantially less interest in documenting the end of an eruption than either its beginning or its most vigorous phases; consequently many eruptions (even in recent years) are listed as "continuing." The waning stages of an eruption are often not considered noteworthy, and unreported eruptive activity may occur after the departure of observers from an isolated volcano. When an eruption is reported to be continuing on one date, and on a later date activity is observed to have ceased, the mid-point of the range is entered as the stop date (along with the appropriate uncertainty range). If the time between these observations is long, however, the eruption is generally listed as continuing on the date of the last observation. All these factors emphasize the substantial caution necessary when interpreting eruption stop dates.

Code ±Years ±Days a 1 1 b 2 2 c 3 3 d 4 4 e 5 5 f 6 6 g 7 7 h 8 8 i 9 9 j 10 10 k 12 12 m 14 15 n 16 20 >After date listed	Code ±Years ±Days o 18 25 p 20 30 (1 mo) q 25 45 r 30 60 (2 mo) s 40 75 t 50 90 (3 mo) u 75 120 v 100 150 w 150 180 (6 mo) x 200 270 y 300 365 (1 yr) z 500 545 * 1000 730 (2 yr) ? Date uncertain(no data) <Before date listed
EXAMPLES: 1731< = on or before 1731 1731a = between 1730 & 1732 1731 1105d = between Nov 1 & 9	1750t = 18th century 1790j = late 18th century 1778 02 ?= February (?) 1778

Eruptive Characteristics

Twenty common eruptive characteristics designated by the IAVCEI originators of the Catalog of Active Volcanoes of the World are shown in these tables. The reported presence of a particular characteristic is shown by an "X" in the appropriate column, a "?" marks uncertain occurrence, and a "-" indicates that this characteristic was not reported. This tabular format allows quick visual inspection of the occurrence of a particular eruptive characteristic, but the quality of eruption reporting is highly variable, and the absence of an "X" does not necessarily mean that this characteristic did not occur.

Eruptive characteristics are shown in five groups of four. The first four characteristics relate to vent location, and note activity originating from the central vent and/or from flank vent(s). Some eruptions may originate from long fissures cutting the summit or flanks of the volcano. These may be either radial to the central conduit or parallel to regional tectonic trends. The second four characteristics relate to interaction with water, and document submarine eruptions (and their occasional formation of new islands), subglacial eruptions, and those from crater lakes. The third group covers tephra-related processes, such as explosive eruptions, the formation of pyroclastic flows and surges (hot glowing avalanches--sometimes referred to as nuées ardentes--that can move down slopes at hurricane velocities), phreatic explosions, and fumarolic activity. The fourth group documents processes related to lava extrusion, and includes lava flows, lava lakes (molten lakes over submerged vents that may keep lava circulating for years), lava domes (the extrusion of viscous lava that accumulates around the vent), and lava spines. The last group documents the impact of eruptions on humans, and notes the occurrence of fatalities, damage to land, property, etc., as well as the formation of often destructive mudflows (also referred to by the Indonesian term lahar) and tsunamis. Mudflows directly associated with glacier outbursts (often known by the Icelandic term jökulhlaups) are identified by a "J" rather than the "X" used to indicate other characteristics.

Place C = Central crater eruption E = Flank (excentric) vent R = Radial fissure eruption F = Regional fissure eruption Water S = Submarine eruption I = New island formation G = Subglacial eruption C = Crater lake eruption Tephra E = Explosive N = Pyroclastic flows P = Phreatic explosions F = Fumarolic activity

Lava F = Lava flow(s) L = Lava lake eruption D = Dome extrusion S = Spine extrusion Damage F = Fatalities D = Damage (land, property, etc) M = Mudflows (lahars) T = Tsunami (giant sea waves) Symbol Key X = recorded ? = uncertain - = not recorded

The 20 standardized eruptive characteristics of the IAVCEI volcano catalog displayed in the "Eruptive History (table)" format are supplemented by three additional eruptive characteristics in the "Eruptive History (expanded)" format. These are "Caldera collapse," "Evacuation," and "Debris avalanche(s)." Caldera collapse is restricted to caldera formation by magma chamber collapse and is not used for large horseshoe-shaped avalanche calderas formed by sector collapse of the volcanic edifice. The latter can often be distinguished by the occurrence of the "Debris avalanche(s) characteristic that accompanies these edifice failures, although sometimes this characteristic is attached to smaller slope failures.

Particular caution should be used in the interpretation of several eruptive characteristics. Although the Catalog of Active Volcanoes of the World volumes distinguished phreatic explosions from "normal" explosions, many historical reports are inadequate to distinguish magmatic from phreatic explosions, and the presence of an "X" in the Explosive column should not be taken as an indication of magmatic eruptions. Solfataric or fumarolic activity accompanies most eruptions, but the Fumarolic column has been mostly restricted to cases where original accounts are unclear as to whether explosive activity or only fumarolic activity occurred. The formation of lava spines was included by the Catalog of Active Volcanoes of the World compilers in part as a response to the spectacular 311-m-high spine that temporarily formed during the 1902 Pelée eruption in the West Indies. However, this typically minor process accompanying the growth of lava domes is often not documented, and this column does not reflect the actual frequency of spine formation.

VEI (Volcanic Explosivity Index)

The reported size, or "bigness," of historical eruptions depends very much on both the experience and vantage point of the observer. To meet the need for a meaningful magnitude measure that can be easily applied to eruption sizes, Newhall and Self (1982) integrated quantitative data with the subjective descriptions of observers, resulting in the Volcanic Explosivity Index (VEI). It is a simple 0-to-8 index of increasing explosivity, with each successive integer representing about an order of magnitude increase. Criteria for VEI assignments are shown in the table below, which is followed by examples of eruptions in different VEI size classes. VEI assignments have been updated from those in Newhall and Self (1982) and Simkin and Siebert (1994).

VEI  Tephra Volume(km)     Example

 0   Effusive              Masaya (Nicaragua), 1570
 1   >0.00001           Poás (Costa Rica), 1991
 2   >0.001             Ruapehu (New Zealand), 1971
 3   >0.01              Nevado del Ruiz (Colombia), 1985
 4   >0.1               Pelée (West Indies), 1902
 5   >1                 Mount St. Helens (United States), 1980
 6   >10                Krakatau (Indonesia), 1883
 7   >100               Tambora (Indonesia), 1815
 8   >1000              Yellowstone (United States), Pleistocene

Codes associated with the VEI

A "*" following a VEI indicates that there are two or more VEI assignments for that eruption, as in the common example of one or more short, paroxysmal eruptions preceded by lower level activity. The "*" follows the maximum VEI recorded between the indicated start and stop dates, and alerts the user to the fact that more information on the eruption exists.

A "?" accompanies those VEIs that were particularly difficult to assign, and those that are based on purely circumstantial evidence. For example, a VEI of 1? might have been assigned to an undescribed eruption because a nearby contemporaneous eruption received sufficient historical comment to confidently assign a VEI of 2. When there was simply no evidence on which to base a VEI, this column has normally been left empty (20% of the eruptions in our file).

A "+" following a VEI indicates an eruption volume in the upper third of the range for that particular VEI designation. It shows those eruptions known to be larger than most others sharing the same VEI numeral, but its absence does not necessarily indicate a relatively small event. The designation is used only for VEIs > 4, volume data permit adding it to only 22 events globally, but it is helpful to identify the obviously larger events in volume ranges that span a full order of magnitude.

A very few eruptions, mostly before 1500 AD, have been upgraded by one VEI unit with the assumption that early in the historical record only relatively large eruptions would have been documented. These are shown by a "^" following the VEI.

Eruptions associated with caldera collapse are normally large (probably VEI >4), and those for which data are lacking to assign a specific VEI are indicated by a "C" in the VEI column. Likewise, Plinian eruptions in the absence of more quantitative data are marked with a "P" in the VEI column.

Eruptions that were definitely explosive, but lack other descriptive information to assess their magnitude, have been assigned a default VEI of 2, that of "moderate" eruptions. Conversely, other eruptions in which substantial tephra volumes were accumulated over long periods of time and/or much of the tephra volume was in near-vent cone construction, have been downgraded by one VEI unit.

Eruptive Volumes

Accurate measurement of eruptive volumes requires careful field work and is often subject to unresolvable uncertainties. Consequently, volume information is available for only a small proportion of eruptions. Volume data is displayed in two different formats. Because of space constraints in the Eruptive History table view, only the order of magnitude (in cubic meters) of calculated lava and/or tephra volumes is displayed. An entry of 8/9 under the L/T header, for example, indicates an eruption with 10⁸ m³ of lava and 10⁹ m³ of tephra. Because only the exponent is displayed, this means that the eruption volume may be nearly 10 times larger than shown. The Eruptive History expanded view, in contrast, has room to display the full volume data, including in some cases uncertainty ranges. It is important to note that tephra and lava volumes are listed without correction for vesicularity (the void space occupied by air bubbles, or vesicles), the extraneous fragments of older rock included accidentally in the deposit, or compaction of ash layers with time. The tephra volumes displayed, therefore, are bulk tephra volumes, and not Dense Rock Equivalents (DRE), or volumes of new magma erupted.

Data Sources

The volcano and eruption data of this digital version of Volcanoes of the World (Siebert and Simkin, 2002-) are updated from its hardcopy predecessor (Simkin and Siebert, 1994) and originate from more than 3500 references. These references are accessible in this website through both regional and volcano-specific listings. The basic building block of the Smithsonian's volcano database is the Catalog of Active Volcanoes of the World (CAVW), a series of regional volcano catalogs published by IAVCEI beginning in 1951. In order to more easily locate these important compilations (which contain many primary references not listed in our compilation), these IAVCEI regional catalog references are bolded in our regional and volcano-specific listings.

The listings appearing here are not intended to be a comprehensive bibliography of references for a particular volcano or region, but represent those references that are cited as the sources of the volcano and eruption data in Volcanoes of the World. Several other global compilations have been helpful: among them are IAVCEI data sheets of post-Miocene volcanoes (1975-80), Volcano Letter reports of the U S Geological Survey from 1926-1955 (compiled in Fiske et al., 1987), independent compilations by Latter (1975) and Gushchenko (1979), and a caldera compilation by Newhall and Dzurisin (1988). Major sources of eruption data subsequent to or supplementing the CAVW can be found in a series of annual summaries by Gustav Hantke published between 1939 and 1962 (mostly in the IAVCEI publication Bulletin of Volcanology), and annual eruption compilations by the Volcanological Society of Japan (1960-96) and Smithsonian Institution reports (since 1968) in various formats, compiled in McClelland et al., (1985) and in the Activity Reports section of this website (Venzke et al., 2002-). The data sources referenced focus almost exclusively on Holocene volcanism and emphasize papers on volcanic stratigraphy and physical volcanology. Abstracts are typically not referenced unless they contain significant data not in other sources. As with the Georef bibliographic database, diacritical marks are not used.

References are linked directly to data in our Volcano Reference File. This sometimes results in apparently incorrect citations in lists of data sources for a volcano or a region. Discussion of another volcano or eruption (sometimes far from the one that is the subject of the manuscript) may produce a citation that is not at all apparent from the title. Alert readers will note a backlog of uncited references for publications in recent years, which we will continue to address.

Regional Maps

Volcano locations are shown in two symbol sizes, with the smaller triangles representing volcanoes with uncertain Holocene eruptions. Red triangles on each map mark volcanoes of that region; yellow triangles indicate volcanoes of other regions. The physiology of the world and regional maps on this web site originates from two data sets, plotted using ER Mapper. Subaerial topography uses the GTOPO30 data set of the U S Geological Survey, and submarine topography originates from satellite altimetry data (Smith and Sandwell, 1997) of sea-surface topography, which mimics that of the sea floor.

Volcano Images

Volcano photos by Smithsonian scientists are supplemented by many other images by volcanologists from the U.S. Geological Survey and other organizations around the world. Photographers are acknowledged with individual photo credits, and their collective contributions have greatly helped to give a visual footprint to the world's volcanoes and their eruptions. Photo galleries for volcanoes show volcano morphology images first, followed by eruption images linked to the start date of the eruption. For each eruption (which may have lasted for multiple years), an image with a summary caption appears first, followed by additional images for that eruption in chronological order.

References

CAVW Editors (1951-1975). Catalog of Active Volcanoes of the World. Rome: International Association of Volcanology and Chemistry of the Earth's Interior, 22 volumes.

Fiske R S, Simkin T, Nielsen E A (eds) (1987). The Volcano Letter. Washington, DC: Smithsonian Inst Press, 1536 p (Reprinting of 1926-1955 issues of the U S Geological Survey's Hawaiian Volcano Observatory).

Gushchenko I I (1979). Eruptions of Volcanoes of the World: A Catalog. Moscow: Nauka Pub, Acad Sci USSR Far Eastern Sci Center, 474 p (in Russian).

Hantke G (1939-62). Übersicht über die Vulkanische Tätigkeit. Eruption summaries published in the Zeitschrift Deutsche Geologie Gesellschafft in 1939 and the Bulletin of Volcanology in 1951, 1953, 1955, 1959, and 1962.

IAVCEI (1973-80). Post-Miocene Volcanoes of the World. IAVCEI data sheets, Rome: Internatl Assoc Volc Chem Earth's Interior.

Latter J H (1975). The history and geography of active and dormant volcanoes. A worldwide catalogue and index of active and potentially active volcanoes, with an outline of their eruptions. Unpublished manuscript.

McClelland L, Simkin T, Summers M, Nielsen E, and Stein T C (eds.) (1989). Global Volcanism 1975-1985. Prentice-Hall and American Geophysical Union, 653 p.

Newhall C G, and Dzurisin D (1988). Historical unrest at large calderas of the world. U S Geol Surv Bull, 1855: 1108 p, 2 vol.

Newhall C G, and Self S (1982). The volcanic explosivity index (VEI): an estimate of explosive magnitude for historical volcanism. J Geophys Res (Oceans & Atmospheres), 87: 1231-38.

Siebert L, and Simkin T (2002-). Volcanoes of the World: an Illustrated Catalog of Holocene Volcanoes and their Eruptions. Smithsonian Institution. Global Volcanism Program Digital Information Series, GVP-3, (http://www.volcano.si.edu/world/).

Simkin T, and Siebert L (1994). Volcanoes of the World, 2nd edition. Geoscience Press in association with the Smithsonian Institution Global Volcanism Program, Tucson AZ, 368 p.

Smith W H F, and Sandwell D T (1997). Global seafloor topography from satellite altimetry and ship depth soundings. Science, 277: 1957-1962.

U S Geological Survey (2002). GTOPO30. Land Processes Distributed Active Archive Center (LP DAAC), U S Geol Surv EROS Data Center http://edcdaac.usgs.gov.

Venzke E, Wunderman R W, McClelland L, Simkin, T, Luhr, J F, Siebert L, and Mayberry G (eds.) (2002-). Global Volcanism, 1968 to the Present. Smithsonian Institution, Global Volcanism Program Digital Information Series, GVP-4 (http://www.volcano.si.edu/reports/).

Volcanological Society of Japan (1960-96). Bulletin of Volcanic Eruptions, no 1-33 [Annual reports issued 1 to 3 years after event year, published since 1986 in the Bulletin of Volcanology].