Overall Schema group

Prepared for the Eruption Source Parameters
Overall Schema Workgroup
November, 2007
By Larry G. Mastin
USGS, Cascades Volcano Observatory

Download file

Introduction

The table in this document is a modification of Table 2 in the ESP workshop meeting notes, a list of well-studied eruptions that could be characterized for eruption source parameters. Our objective was to characterize these eruptions and categorize them into various eruption types, which we could use source parameters for each type as inputs to atmospheric models. The Eruption List working group was tasked with the job of assigning ESP’s to these well-studied eruptions. This is one product of that effort.

Table entries

Parameter estimations and their uncertainties

The compilation includes information on the following source parameters, which were estimated in different ways in different studies. The numbers all have caveats associated with them, which are explained below.

• Plume height. Plume height is estimated using the following methods:
  • Visual observation (“v” in Table 1), either ground-based or from pilot reports
  • radar-based observations (“r”)
  • satellite-based observations (“s”) which include the shadow method, brightness temperature, wind correlation, or CO₂-slicing
  • isopleth mapping from deposits (“i”) [Carey and Sparks, 1986].

Some major differences in height estimates have been reported using different techniques; for example a ground-based estimate of 5 km at Ruang volcano in 2002 was revised to ~20 km using satellite data [Tupper, et al., 2004]. Visual observations may also include the vapor-rich cloud top, which may not be reflective or visible with radar or satellite(?). These uncertainties suggest that we should establish some quality ranking of cloud height based on the estimation method . I am assuming for now that radar and satellite brightness temperature are perhaps the most reliable. It should also be noted that most accounts give maximum plume height, not average plume height. In a few cases (e.g Mount St. Helens, 1980), the change in plume height with time is documented in radar data, though this is rare.

• Erupted volume(dense-rock equivalent). Total erupted volume has typically been estimated by digging shovel pits at various locations through the deposit, plotting those locations on maps, and then drawing a series of contour lines or isopachs of equal deposit thickness that show the distribution of thickness. To integrate for volume, the log of thickness is plotted as a function versus square root of that area inside each isopach line, and best-fit lines are extended through the data points. The total estimate of volume depends in part on the type of best fit curve used (exponential versus power-law), and to what thickness the extrapolation is carried out [Pyle, 1989; Fierstein and Nathenson, 1992; Rose, 1993; Bonadonna, et al., 1998]. At Etna, Scollo et al. [2005] found more than a two-fold difference in volume of the 2001 tephra using exponential versus power law extrapolation methods (hopefully the uncertainty is not usually this high). In this table, I take volume estimates using whatever extrapolation methods were used by the author without attempting to recalculate for uniformity. The tephra volume is converted to erupted mass by multiplying by an average deposit density. Density estimates are not always made of these deposits, requiring some approximation (usually about 1,000-1500 kg/m³) to convert volume to mass. Dense-rock equivalent volume is the mass divided by a typical density of magma; 2500 kg/m³ in the case here. The uncertainties in these calculations imply an uncertainty of perhaps several tens of percent in total erupted volume.

• Duration.Duration estimates are based on visual observations, or, where available, seismic records. The latter is clearly more reliable, as ash-poor plumes have continued to rise for hours after the seismic end of an eruption. During the June 12, 1991 eruption of Pinatubo, for example, seismicity lasted only 38 minutes but the plume continued for nearly two hours [Wolfe and Hoblitt, 1996]. For the purposes of calculating average eruption rate I have sometimes chosen an “end time” of the eruption the time when the plume height or other activity dropped significantly (e.g. Hekla, 1947). If the plume height is proportional to the fourth root of eruption rate, as suggested by theoretical and empirical studies, then a factor-of-two decrease in plume height represents nearly an order of magnitude decrease in eruption rate.

• Eruption rate. The mass eruption rate in kg/s is calculated by dividing the erupted mass by the duration. In cases where the eruption rate changed significantly during the course of an eruption, especially for eruptions that continued for days (e.g. Hudson, 1991), I do not plot average plume height versus eruption rate.

• Grain-size distribution. Grain-size distributions in this table are given in median (md_φ) and standard deviation (σ_φ) of the φ size (φ=-log₂(d), where d is the diameter in mm). Few eruptions exist where attempts have been made to estimate a whole-deposit grain-size distribution [Murrow, et al., 1980; Bonadonna and Houghton, 2005; Scollo, et al., 2007]. Because of the lack of good data on grain size distribution, no serious attempt was made to compile these data. For the time being, therefore, we should rely on Adam and Bill Rose for grain-size data rather than the few entries in this table

Eruption type

One objective of the table was to divide eruptions by type, following Table 1 of the meeting notes. That table assigns three types each for silicic, andesite-to-basaltic-andesite, and basaltic eruptions; and within each of those types we would characterize “default”-type eruptions, plume/plumette type, discrete puffs, and so on. This results in a matrix with dozens of cells, which, in my opinion, could not be populated given the number of well-characterized eruptions. Therefore, in the table, I simply identify small, medium, and large eruptions of the three magma types given above. In column 3 of the table, under “type”, small, medium, and large silicic eruptions are indicated by abbreviations S1, S2, and S3, respectively, and color coded in red. Andesitic or basaltic-andesite eruptions are A1, A2, and A3 and purple; and basaltic eruptions B1, B2, and B3 in magenta. In assigning eruption size, I used the VEI, i.e. A1, A2, and A3 are VEI 3, 4, and 5+ respectively. Within each magma type, entries in the table are ordered in increasing eruptive volume. For purposes of assigning ESP, I suggest that we further simplify eruption types into basaltic types (B) and small, medium, and large silicic or intermediate eruptions (SI1, SI2, and SI3).

Eruptions chosen for the table

The eruptions chosen for this table are primarily those in Table 2 of the meeting notes. However for some entries in Table 2 I found no evidence of a documented eruption (e.g. Kliuchevskoy, 9/94), and for others (e.g. Karthala 2005, Ruang 2004, Galunggung 1983), there was no information on erupted volume. The latter eruptions were separated into a second section of the table. Plume height, erupted volume, and duration for some eruptions were also tabulated by Sparks et al. [1997, Table 5.1], and I used some of those entries after reviewing and throwing some out that did not appear appropriate. Finally, John Ewert suggested several eruptions that were not in either of these tables.

Inferences from the table

Plume height and eruption rate

An important question is whether one can reasonably use one observed source parameter, plume height for example, to infer others. Relationships between plume height and eruption rate have long been considered, both theoretically and empirically [Settle, 1978; Wilson, et al., 1978; Sparks, 1986; Woods, 1988; Sparks, et al., 1997]. In a uniform atmosphere of constant thermal lapse rate, Morton et al. [1956] suggested that plume height should be proportional to the fourth root of eruptive power (mass eruption rate times heat content of the magma). Sparks et al. [1997, Fig. 5.1] obtained a best-fit curve through empirical plume height-eruption rate observations of:

H= 1.67V^0.259

(1)

where H is plume height in kilometers and V is volumetric eruption rate (DRE) in cubic meters per second. The exponent lies remarkably close to the theoretical fourth-root relationship of Morton et al.

The Data

Figure 1 shows plume height versus eruption rate for the eruptions in this table. This compilation is more complete than previous ones [Settle, 1978, Table 1; Wilson, et al., 1978, Table 3; Sparks, et al., 1997, Table 5.1], and removes a few data points¹ that appeared in previous compilations but perhaps should not have been there. The data show a clear trend, with some obvious outliers above and below. Outliers below are several Mount St. Helens 1980 eruptions, the 1902 Soufrière St. Vincent eruption and the Redoubt eruption of 15 December 1989. Possible outliers above include the eruptions at Hekla in 1947 and Nevado del Ruiz in 1985. Outliers below the curve can occur if a significant fraction of the deposit results from coignimbrite ash, a process that Carey et al. [Carey, et al., 1990] suggested this as a possible explanation for the anomalously high volume of the May 18, 1980 tephra given its plume height. Pyroclastic flows also accompanied the 1989 Redoubt, 1902 Soufrière and Mount St. Helens June 12, 1980 eruptions (i.e. all the low-outliers), though it is not clear that pyroclast flows were unusually prominent relative to other eruptions. Outliers above the main trend can occur if the observers record the height of an ash-poor cloud above the main column; if erupted volume is underestimated or if duration is underestimated. At Hekla in 1947 Thorarinsson, [1949] considered most of the upper eruption cloud to be a vapor rich cap; however this also appeared to be the case during the March 8, 2005 eruption of Mount St. Helens [Mastin, 2007], a data point that doesn’t stand out as anomalously high. Plume heights at Mount Spurr, which are nearly upper outliers, were among the most accurate on the plot in terms of plume height (measured by radar), duration (measured seismically) and erupted volume (meticulously mapped by Neal et al. [1995]). In brief, I see no good reason to throw away outliers on the basis of observational data.

¹ Bezymianny’s 1956 eruptive plume was included in these three tables, but Belousov and Belosouva (2007) note that this was a coignimbrite cloud, not a Plinian eruption cloud. Also, Sparks et al. (1997, Table 5.1) includes plume height and eruption rate for four stratigraphic layers of the May 18, 1980 Mount St. Helens Plinian eruption, citing Carey et al. (1990) as the data source. Carey et al. (1990) however estimated eruption rate from plume height using a model, not the deposit volume and duration.

Best fit

The best-fit line through these data gives the following best-fit relationship between plume height H and eruption rate M is:

log₁₀(H)=Alog₁₀(M)+B

(2)

where best-fit values of A and B are 0.222 and -0.3833, respectively. Converted to volumetric flow rate V(DRE) for comparison with (1) , which is not significantly different from H= 2.35V^0.222 (1). The dotted red lines in Fig. 1 are error envelopes that enclose half the data points. From figure this one can see that attempts to estimate mass eruption rate from plume height are highly approximate. Within these error envelopes, for a given plume height, the uncertainty in mass eruption rate is more than a factor of five in either direction.

Comparison with theoretical curves

Theoretical curves by Wilson et al. [1978] and Woods [1988, Fig. 15] predict plume heights above most of the data whereas those by Woods [1995, Fig. 17a] and Plumeria [Mastin, 2007] in a dry atmosphere come closer to the best-fit curve. The few data points at eruption rates below about 106 kg/s do not show the spread in plume heights that are predicted in variable atmospheric conditions [Woods, 1993; Tupper, et al., 2007b], although the scarcity of data in this region prevent any meaningful conclusions from being drawn.

Erupted Volume and plume height

Carey and Sigurdsson [1989] show that plume height correlates with total erupted volume V. Using a different dataset we also find a correlation (Fig. 2), with R2=0.826 and a best-fit curve of:

H=6.38+25.64log₁₀(V)

(3)

with a standard error in plume height of about 5 km. The potential error in estimation of total erupted volume from this relationship is close to an order of magnitude on either side of the best fit curve.

Eruption duration

The duration in eruptions lasts from a few tens of minutes to hundreds of hours (271 for Cerro Negro 1995). There is no correlation between plume height and erupted volume (Fig. 3), although the longest-lasting eruptions tend to be those with the lowest eruption rates. The median time for eruptions in this table is 4.4 hours.

Figures

Figure 1

Plume height versus log eruption rate for eruptions in Table 1. Lines are theoretical curves from Wilson et al. [1978] (black dashed curve)

Figure 2

Plume height versus log total erupted volume (km³ DRE) for well-documented eruptions. Solid red line is the best-fit curve through the data. Red dashed lines represent one standard error above and below the best-fit relation.

Figure 3

Plume height versus log eruption duration in hours.

Figure 4

Histogram of the log of duration of eruptions in the eruption table.

Tables

Eruption source parameters for well-studied eruptions chosen by IAVWOPG for characterization for ESP, and other eruptions which have been well documented. Eruption types refer to small, medium and large silicic eruptions (S1, S2, S3 respectively), andesite or basaltic-andesite types (A1, A2, A3), and basaltic types (B1, B2, B3). Table entries are color-coded according to their eruption type; silicic (red), andesitic (brown), basaltic (magenta). Plume heights are annotated with a letter following the number, which indicates the method by which plume height was estimated: “v”=visual observation from the ground or an airplane; “r”=radar; “s”=satellite images; “i” = isopleth data.

Appendix: Notes on the various eruptions

Cerro Negro

13 April 1992.
The 1992 eruption is described in Hill et al. [1998] as having erupted ~2.35x10¹⁰ kg of tephra and sending a plume to 7-7.5 km height. According to the Smithsonian Institution monthly reports, the eruption began at 2320 on April 9 and ended at 1800 on April 12, for an average eruption rate of 1x10⁵ kg/s. Hill et al. also give the grain size distribution as md_φ=-0.1, σφ=1.4.

November-December 1995.
The November 19-December 2, 1995 eruption is also described in Hill et al. as having a total erupted mass of ~1.3x10⁹ kg (using a tephra volume of 1.3x10⁶ m³ and density of 1200 kg/m³ ). The activity started on 19 November at 11:45 and ended at seismically at 6:30 on December 2, giving a total of about 270 hours. According to Hill et al. [1998, p. 1232], “by November 29, the eruption sustained a continuous tephra column and ash began to accumulate at Leon. . . . The continuous phase of the eruption ceased on December 2 at 6:30 and was marked by a return of seismicity to near background levels.” Regarding plume height they say “Incandescent ejecta columns commonly reached altitudes of 200–400 m above the cone rim and were accompanied by roiling black clouds of cooled basaltic tephra. A maximum incandescent column height of 810 ± 20 m above the cone was measured at 17:03 on December 1.” Although not mentioned in the text, in Table 2 he gives the observed column height for this eruption as 2-2.5 km. The column height for the 1992 eruption (7 km) is also given in this table. Estimates of the total erupted mass (1.3x10⁹ kg) and duration (~12 days, 18 hours) yields an exceptionally low eruption rate of 1.1x10³ kg/s or about half a cubic meter per second.

El Chichon

March-April 1982
Carey & Sigurdsson [1986, Table 3] tabulate the volumes (DRE) of the 1982 tephra deposits based on mapping the tephra distribution and extrapolating thicknesses out to 1 micron using the method of Rose et al. [1973b]. These values (0.3, 0.39, 0.40 km³ respectively) are significantly greater than estimates (0.19, 0.36, 0.31 km³) of Gutierrez-Coutino et al. [1983], who did no such extrapolation. Carey and Sigurdsson give the durations of the 3/29, 4/4 AM, and 4/4 PM events as 5, 4, and 7 hours respectively. Using their volumes and durations I get mass eruption rates for these eruptions of 4.2x10⁷, 6.8x10⁷, and 4.0x10⁷ kg/s, which differ somewhat from their numbers in Table 3 (3.5, 6.0, and 4.0x10⁷ kg/s respectively). Estimates of eruptive column height are not easily determined. Carey and Sigurdsson inferred that each eruption extended above 18 km elevation based on ENE transport of ash. They estimated a column height of 20, 24, and 22 km for the eruptions of 3/29, 4/4 AM, and 4/4 PM, respectively, by matching their deposit distribution with a numerical model. They also argued that this was consistent with particle concentration measurements in the atmosphere, made after the eruption was over, which showed maxima at 20-30 km elevation [Carey and Sigurdsson, 1986, p. 137]. Duration, plume height and eruption rate for these eruptions were taken from Table 5.1 in Sparks et al. [1997], which cites Carey and Sigurdsson [1986] for the original data. Eruptions B and C started at 0139 and 1110 GMT, respectively, on April 4 1982.

Etna

October 2002
The October 2002 eruption was on the IAVWOPSG list as a “well studied eruption”, although I’ve found no published reports on it. The GVN monthly report notes that, “On the morning of 28 October the south fissure developed at least three explosive vents and a 100-200 m high lava fountain, 200 m downslope, fed lava flows more than 2 km toward the uninhabited area of Monte Nero degli Zappini.” “Sustained release of high-pressure gas fed a voluminous SE-directed ash plume that reached to more than 5 km altitude.” By 30 October the accumulation of pyroclasts had built a spatter rampart ~200m long and 30 to 40 m high. On 31 October, ashfall from this event affected localities NE of the volcano, including Reggio di Calabria, whose airport had to be closed. By 1 November the activity had ceased on the northern side but the S2700 upper vent entered a phase of sustained explosive activity resembling a small subplinian column. The activity continued into December. Ash emission from the 2,750 m cinder cone significantly declined on 17 December, allowing the local airport of Catania to reopen. Volcanic tremor amplitude showed a marked decrease on 27 January and on 28 January at 2240 it returned to background levels, signaling the end of the eruption.

July-August 2001
Scollo et al. [2007] describe the chronology of a Strombolian phreatomagmatic eruption that started on 19 July at 1700. Ash fall continue until sometime on 24 July (no time given). In the text they note plume heights of 3-3.5 km a.s.l. on 20 July, a “fairly stable ash column rising 3.5-4 km a.s.l.” (the vent elevation is 2570 m), on 21-22 July “pulsing jets” to 5 km a.s.l.”, and on 24 the ash plume dropped at about 4.5 km a.s.l. before ending. From these I take the average daily plume height of 3.25, 3.75, 3.75, 5, 4.5 = ~4 km, or ~1.5 km above the vent. Assuming the eruption ended in mid-day on 24 July, the duration was about 115 hours. They mapped the deposit and estimate deposit mass using both an exponential and power-law fits to the thickness-area data, yielding 1.02x10⁹ and 2.31x10⁹ kg, respectively, the latter of which is perhaps closer to reality due to the tendency of exponential curves to underestimate deposit volume. Total grain size distribution was estimated using a Voronoi method and gives a Gaussian distribution centered on 2.0 Phi with a standard deviation of 1.9 Phi (Scollo et al., 2007). The average eruption rate would therefore have been about 2.31x10⁹ kg/(115 hrs * 3600 s/hr) = 6x10³ kg/s. The volumetric flow rate has been estimated by Behncke and Neri [2003] for the same period and yield values between 12-14 m³s^-1. Assuming a density of the deposit of 1321 kg m^-3 (proposed by Scollo et al., 2007), the mass flow rate was about 1.7e04 kg s^-1. Note that the density of the mixture at the vent should have been used to calculate the mass flow rate from the volumetric flow rate.

Fuego

14 September 1971
In describing the 1971 eruption, Bonis and Salazar say “Volcan Fuego erupted violently at 2:45 PM, September 14, 1971, without any forewarning and stopped at 3:00 AM the next day, for a total eruption period of about twelve hours. The activity consisted of strong eruptions of basaltic ash and scoria, and hot ash flows that rushed down the barrancos (canyons) on the volcano slopes. The eruption clouds reached over 10,000 m in height.” Rose et al. [Rose, et al., 1973a] estimate minimum tephra volume of 6x10⁷ m3, not extrapolating beyond ~60 km distance. I convert this to DRE volume by dividing by 3.

Galunggung

4 June 1982
The eruptive sequence in question lasted from April to December 1982. Two incidents in June were climactic: one on June 4 that sent ash to nearly 20 km altitude [Gourgaud, et al., 1989], and another on June 24-27 with a maximum plume height of about 13.5 km, slightly below the tropopause elevation (GVN website). Information on plume height and eruption duration are better constrained than erupted volume during the climactic phases. For example, Gourgaud et al. [1989] give only the total volume erupted in Phase 2 (mid-May to October 1). Therefore the eruption rate cannot be adequately determined. Similarly, the Smithsonian web site and Gourgaud et al. give only the days of the start and end of the June 24- 27 phase, not the time of day when the eruption started and ended, limiting the accuracy of the duration.

Grímsvötn

1 November 2004
The ash cloud from the 1 November 2004 eruption was used for a model intercomparison by the International Airways Volcano Watch Operations Group (IAVWOPSG), the results of which are reported on 8 March 2005 in their Working Paper IAVWOPGSG/2-WP/21. Freysteinn Sigmundsson and others have also written online reports of this eruption of this eruption which are available at the Nordic Volcanological Institute web site. According to that paper, the volcano erupted through the ice cap on the evening of 1 November 2004 at approximately 22.50 UTC. “The tremor was steady for the initial 15 hours of the eruption. After that it was pulsating and declining.” The eruption produced no visible lava flows and was at least partly phreatomagmatic. It sent a continuous plume to 9 km altitude with maximum heights of 12-14 km. Tephra fell in small quantities in inhabited areas of north and northeast Iceland. Sigmundsson notes that “The eruption follows a pattern similar to previous eruptions in 1983 and 1998, with probably less than 0.1 km³ of magma erupted.” This is the only estimate of erupted volume.

Hekla

29 March 1947.
Hekla is the most active volcano in Iceland. Its eruptions frequently start out silicic and end more mafic, with most eruptive products having 59-65% SiO₂ [Thorarinsson, 1949]. The 1947 eruption produced two tephra types: a brownish-gray tephra (63-61% SiO₂) and finer-grained brownish-black tephra (56-57% SiO₂). Isopach maps of the deposit were converted to volumes by extrapolating them using an exponential relationship of log thickness versus distance. Thorarinsson [1949, Table 5] estimates the total erupted volume (DRE) of tephra as 4.5x10⁷ m³ and the total mass as 1.20x10⁸ tons (1.2x10¹¹ kg). Of this, Thorarinsson (p. 60) estimates that 8.5x107 tons was brownish-gray and 3.5x10⁷ was brownish-black ash. Eyewitness accounts (p. 54) suggest that the brownish gray tephra was erupted during about the first half hour whereas most of the brownish-black ash was erupted in the following half hour. The eruption went on for “some hours” (p. 63) but after 07:10 it “reduced rather rapidly (p. 63) and by 08:00, 1:10 after the onset of the eruption, the cloud had dropped from 24-30 km to about 10 km (Plate XV). These durations and volumes are the basis for Einarsson’s estimate (p. 65) of eruption rates of 17,000 and 6,000 m³/s DRE (4.2x10⁷ and 1.5x10⁷ kg/s) for the brownish-gray and brownish-black ash, respectively. A photo taken from an airplane (Plate IV) shows a plume reaching to about 27 km elevation by 06:59, 18 minutes after the eruption onset; later photos were apparently used to infer a maximum plume height near 30 km at about 7:08 (Plate XV; Hekla’s summit elevation is 1,491 m). Thorarinsson however states (p. 54) “I am of the same opinion as Einarsson that the vast cap of volcanic cloud above the tropopause consisted mainly of vapor.” These eruption rates were included in Thorarinsson [1968] and tabulated in Wilson et al. [1978, Table 3].

5 May 1970
The 1970 magma was basaltic andesite (53.66% SiO₂). Information on column height, eruption rate and duration taken for the 1970 eruption from Thorarinsson and Sigvaldason [1971]; Here’s what Thorarinsson and Sigvaldason [1971] say about the Plinian phase: “The initial phase of the eruption lasted two or three hours and was characterized by a very vigorous and continuous fountain activity, or rather a continuous, although somewhat spasmodic uprush of lava and tephra. At 22:10 the tephra-vapour column had reached its max. height, nearly 16.000 m. At 24:00 its height was 12.000 m and at 01:00 it was 7.500 m. The tephra production during the two hours of the main tephra fall averaged about 10.000 m³/sec and the total production of tephra was about 70 million m³, or 45 million tons.” Thorarinsson does not say exactly how the plume height was determined, but clear weather allowed for many eyewitness observations during the eruption.

17 August 1980.
The 1980 Hekla Plinian eruption is well described in Gronvold et al. [1983]. It began about 12:30 that day and the main Plinian phase lasted 5-6, though Gronvold et al. say that most of the tephra was erupted in the first two hours; thus I give the total duration of the eruption as roughly 5 hours for aviation purposes (perhaps it should be less for calculating average eruption rate). Lava flows issued simultaneously with the tephra, apparently from other vents along the fissure. The plume reached a maximum height of 15 km (presumably above the vent), as judged apparently by eyewitness accounts and photos of the plume [Gronvold, et al.,1983, Figs. 8-11] and most of the downwind transport of tephra was thought to have been transported in a high-velocity atmospheric layer between 7 and 12 km elevation. Total volume of the tephra blanket is given as 58 million m³, which I convert to DRE volume by dividing by three. Total mass is calculated assuming a density of 2500 kg/m³.

26 February 2000
The 2000 Hekla eruption has a wealth of satellite images from the NOAA polar orbiting weather satellites. According to the GVN monthly report, “On 26 February 2000 the WSW-trending, elongated Hekla volcano erupted. A fissure 6-7 km long opened along the SW flank of the Hekla ridge, from which a discontinuous curtain of lava erupted starting at 1819. Just a few minutes later, at 1825, an ash plume reached a height of 11 km and was carried N by light winds. Based on the tremor amplitude the eruption reached peak intensity in the first hour of activity, then gradually declined.” Tremor plots show that tremor increased rapidly and reached a maximum at 1850, then decreased until about 0700 on 27 February and became steady. From this I give an eruption duration of about 13 hours. The monthly report also notes that “During a transit flight on 28 February a SOLVE (SAGE III Ozone Loss and Validation Experiment) mission with an instrument-laden DC-8 aircraft flew through the plume shortly after the eruption ~11.3 km NNE of Iceland at 76°N and 5°W, just off the Greenland coastline. The plume extended up to ~13 km altitude, well into the lower stratosphere.” This monthly report says that the preliminary estimate of lava production was 0.11 km³. There was no estimate of tephra volume.

Hudson

12 August 1991
The duration of this eruption is complicated due both to the poor weather and the on-off history on August 12-15. Scasso et al. [1994] report that “The largest eruption commenced at about 1200 hours Chile local time on 12 August. Bad weather prevented aerial observations on this day and most of 13 August. At about 2000 hours on 13 August an eruption plume extending 1200 km to the southeast became visible on AVHRR (NOAA 9 and 11) and GOES satellite images. The plume became disconnected from the volcano at about 1200 hours on 14 August and continuous eruption began again at about 2000 hours. . . . By 1100 hours on the 15^th the plume had become disconnected from the volcano.” This adds up to about 31 hours of eruption, but not continuous. Scasso et al., infer a total tephra volume of 7.6 km³ and average density of 1000 kg/m³, suggesting a total erupted mass of 7.6x10¹² kg (3.0 km³ DRE), erupted in about 31 hours, giving an average eruption rate of about 7x10⁷ kg/s. Scasso also notes that the grain-size distribution of the tephra is bimodal with a small mode around φ=6 or 7. The grain-size distribution however is determined from tephra fall deposits and may not be relevant to the GSD within ash clouds. Although Scasso et al. say that bad weather on 12 August prevented aerial observations, aerial photos of the plume on 12 August, 14:30, are shown, taken above the cloud cover, in Naranjo et al. [Naranjo S., et al., 1993, Fig. 13] Naranjo et al., also show a table giving the plume height on August 12, 13, 14, 16, and 16 as 12, 16, 18, 16, and 2 km respectively.

Karthala

24 November 2005
Karthala Volcano in the Komoros Islands, northwest of Madagascar, is a basaltic shield volcano. An eruption on 24-25 November 2005 produced ash clouds. According to the GVN monthly report, Charles Holliday analyzed a NASA Terra MODIS image taken at 0710Z (1010 local time) and estimated a cloud top at ~11.6 km altitude, with E-W dimension of ~280 km. (The summit elevation is 2,361 m, making the plume height ~10.1 km above the summit). Fred Prata processed both MODIS and AIRS images of the 25 November eruption, taken at 0710Z, and found a total SO₂ mass in the cloud of ~2.8 kilotons and a grain size distribution of about 4+1 micron. Analysis of an AIRS images taken at 2223Z yielded about 2.0 kilotons. Prata used the MODIS image to estimate the 25 November eruption's mass loading. This resulted in an estimate of fine ash amounting to 83 kilotons (kt) in the grain-size ranges indicated. Prata goes on to say that "assuming my fine ash loading of 83 kt is right and (big assumption now) this represents ~ 1% of the total erupted mass, then the volume of erupted ash would be ~ 0.006 km³. This suggests a VEI ~ 2. If the 1% estimate is robust (I've seen this quoted in Bill Rose's work) then the fine ash estimates from remote sensing may be quite helpful [in] assessing the 'size' of an eruption. There may be some inconsistency between the small erupted volume (if correct) and the plume height. That is, if the erupted mass is really something like 8,300 kt and it was erupted over a 24-hour period, the mass eruption rate would have been about 100 kg/s, roughly four orders of magnitude lower than predicted for an 11 km high plume in a dry atmosphere. On the other hand, if the vent elevation was 2 km and the atmosphere were warm and moist, the plume height could have been boosted but probably would still have been much lower for a 100 kg/s eruption rate. But plume height, erupted volume, and duration are all very poorly constrained for this eruption.

Manam

24 October 2004.
According to the GVN monthly report, the 24 October 2004 eruption emanated from the southern crater starting after 0800. “It persisted throughout the morning and early afternoon, peaking between 1000 and 1100. At 1400 the eruption’s intensity decreased slightly. Later that day it continued at a reduced level with moderate explosions and subcontinuous low rumbling and roaring noises.” From this account a duration of at least seven hours was assigned. The plume height, judged by satellite temperatures, was estimated at 15 km (Tupper et al. [2007a] estimated 17-18.5 in from post-analysis of cloud temperature). This was the beginning of an eruptive sequence lasting several months in which 7,900 people were evacuated and eventually relocated. The most vigorous phase of that eruption occurred on 27 January 2005, with a plume height of 21-24 km MSL. I’ve found no information on deposit volume or grain size.

Miyakejima

18 August 2000.
The eruptive sequence from 26 June the end of August 2000 involved several phreatic and phreatomagmatic eruptions, and the formation of a caldera in the summit of the volcano whose volume was several times larger than the erupted volume. The largest eruption, on August 18 from ~1704 to 2030, was described in the GVN monthly report; its plume height was measured as at least 15 km by lidar and ground-based photos (the summit elevation is 815 m). Tupper et al. [2004, Table 1] give a best plume-height estimate of 16.5+/-1 km based on brightness temperature (16 km), radar (16 km), and lidar (17.5 km). Nakada et al. [2005, Table 1] estimate the eruptive mass during this eruption to have been about 8.5x10⁹ kg, of which about 40% was juvenile [Geshi, et al., 2002], suggesting a mass discharge rate of about 2.8x10⁵ kg/s. Nakada et al. also measured grain-size distribution of some deposits from this eruption but were not thorough enough in their samples to give a total GSD. Houlie et al. [2005] were able to study the plume structure through tomographic analysis of the residuals of GPS radio signals. The monthly report indicates that the eruption started at 1700 but does not say when it ended. An aircraft ash encounter was indicated in the monthly report as follows: “According to articles by the Associated Press and Reuters, white clouds rising to 8 km above the summit were encountered by a commercial airline pilot who was in route from Guam to Narita airport in Tokyo. The plane, which was flying over the island of Miyake shortly after the eruption, later landed safely at Narita. Aviation contacts later revealed that while in flight a commercial airliner encountered airborne ash and underwent a dual-engine flame-out, but managed to land safely. The airliner sustained ~$4 million (US dollars) in damage.”

Nevado del Ruiz

13 November 1985
Naranjo et al. [1986] describe the 1985 eruption as follows in the abstract: “A small Plinian eruption of the Nevado del Ruiz volcano in Colombia ejected 3.5 x 10¹⁰ kilograms of mixed dacite and andesite tephra on 13 November 1985, with a maximum column height of 31 kilometers above sea level.” They estimate that the duration of the eruption was about 20 minutes, from 9:09 PM on 13 November until about 9:30 PM. They estimate the column height from isopleth data (the volcano’s peak is at 5.3 km elevation). Regarding erupted volume they say: “The volume of tephra within the 1-mm isopach of the fall deposit is 2.9x10 m³. By analogy with the May 1980 fall deposit of Mount St. Helens, in the United States, where 25% of the total deposit was outside the 1-mm isopach (10), we estimate total erupted volume of 3.9x107 m³. With a bulk density of 900 kg/m³, the fall deposit represents a total erupted mass of 3.5x10¹⁰ kg”.

Nyamuragira

25 July 2002
On July 25, 2002, Nyamuragira, a high-K basaltic shield volcano, emitted a fissure eruption with 200- to 200-m high lava fountains and 6-7 km long lava flows starting around 1310. According to the GVN monthly report, helicopter flights on August 1 and 3 revealed that the eruption continued “at a high rate”. The end of the eruption is not well described but apparently continued through the first week in August. Several satellites detected volcanic clouds rising from this eruption, including Earth Probe TOMS, METEOSAT. The TOMS satellite images detected an SO₂ cloud as far west as 800 km from Nyamuragira. The TOMS satellite did not however detect any ash cloud associated with the eruption. No tephra deposits are mentioned in the monthly report, nor is there any mention of the plume height other than the 100-200m height of the lava fountains.

Pinatubo

15 June 1991
The duration and total erupted volume during this eruption are subject to a lot of uncertainty. Observations suggest that the umbrella cloud expanded for about five hours on June 15, 1991 [Koyaguchi, 1996], although Koyaguchi [1996] notes that some expansion could have continued after the eruption stopped, and suggests 3 hours as a more reasonable duration. Sparks et al. [1997, Table 5.1] use 1.3 hours as a duration, citing Koyaguchi [1994] as their source. Koyaguchi [1996] also suggests that the rate of expansion of the umbrella cloud was consistent with a total erupted volume (DRE) of 2-10 km³. Lynn et al [1996] suggest a volume on the smaller side of this range, perhaps, 3.4-4.4 km³ DRE, based on tephra mapping and extrapolation of the volume that fell into the South China Sea. The GVN web site suggests a total tephra volume of 11+5 km³ (not DRE), presumably based on the work of Lynn et al. Dividing by 2 yields ~5-8 km³ DRE, or 1-2x10¹³ kg. Over a period of ~3 hours this yields an average eruption rate of 1-2x10⁹ kg/s. Self et al. [1996] note that plume height reached a maximum of 40 km, although in Table 1 he uses the more approximate estimate >35 km as an overall statement of plume height during the climactic phase.

12 June 1991, 0851.
According to Hoblitt [1996, Table 1], “Abrupt onset of intense sustained tremor and visual observations of rapidly expanding eruption column. Weather radar indicated tephra-column height >19 km [current summit elevation is 1,461 m]. Aerial observations at 0950 showed continuing tephra production from vent and pyroclastic density current in O’Donnell River Drainage.” Hoblitt et al., note that activity at the vent had declined to a low level by 1040. Seismic data provide the best constraint on duration: Wolfe and Hoblitt [1996, p. 10] say “Interpretation of the seismic records indicate that this event lasted about 38 min., although field observations suggest a duration of at least an hour.” Lynn et al. [1996] estimate a tephra volume of 0.014 km³ for this eruption. Assuming a deposit density of about 1000 kg/m³, this would give an erupted mass of 1.4x10¹⁰ kg and volume (DRE) of 0.0056 km³. Assuming that most of the tephra erupted in the first hour this would give an average eruption rate of about 6x106 kg/s.

13-14 June, 1991, 0841
This time period included four vertical eruptions with minor pyroclastic flows, continued dome growth, and heavy emission of ash [Hoblitt, et al., 1996]. Plume heights were well tracked by radar and rose to >19 km during each event; moreover the duration of each event was well known, but Ash deposits from this period could not be separated out into individual eruptions [Lynn, et al., 1996], making it impossible to estimate total erupted volume. The total volume of the layer that included all of these deposits [layer B of Lynn, et al., 1996] had a total bulk volume of 0.17 km³.

Quizapu

10 April 1932
Hildreth and Drake [1992, p. 112] estimate that about 4.05 km3 (DRE) of magma was erupted in about 18 hours during this event on April 10-11, 1932. They cite a photograph of the volcanic plume from which the plume height of 27-30 km was estimated. This height estimate agrees well with the 25-32 km estimate obtained from isopleth data. Hildreth and Drake include grain-size data but make no attempt to estimate the total grain-size distribution of the deposit.

Rabual

19 September 1994
The September 1994 eruption is described in detail in the GVN monthly report. In that report, plume height was estimated as at least 18 km based on space shuttle observations and perhaps as high as 30 km from ground-based observations. The eruption began at about 0600 on 9/19 and appeared to continue for days. The ash cloud elevation was ~12 km on 9/20 at 1532, 7.5 km on 9/21 at 1832, and 7.5 km on 9/22 at 1230. I have found no studies giving the total erupted volume.

Redoubt

15 December 1989
The largest eruption of the 1989-1990 sequence was December 15. Miller and Chouet [1994] note pilot reports indicating that the plume rose >12 km ASL (the summit elevation is 3108 m). Miller and Chouet give the seismic duration of an event starting at 1015 as 40 minutes, Scott and McGimsey [1994, Table 1] give the total duration of the three events that day as 62 minutes, and the total mass of tephra erupted that day as 15.0-27.3x10⁹ kg (5.8-10.5x10⁶ m³ DRE). These numbers were used to derive the mass eruption rate above.

Reventador

3 November 2002
The GVN monthly report shows a chronology of activity on 11/3- 4/2002. The main phase of the eruption began at 0912 and extended until about 0100 on 11/4. The GVN report indicates that the plume height at 0912 reached 16-17 km above the cone, presumably as judged by eyewitnesses (Fig. 3 of this report is a photo of the plume taken at 0912). At least five significant pf’s were produced. The GVN report also cites J.L. Le Pennec as estimating that 2.82x10⁸ m³ of pyroclastic material was erupted; the report does not indicate whether this was DRE or tephra; presumably some significant fraction of this material was incorporated into the pf’s rather than the ash cloud. Assuming that it was original tephra volume with a density of 1000 kg/m3 and that it was all incorporated into the ash cloud would give a maximum erupted mass of 2.8x10¹¹ kg (0.12 km³ DRE) erupted over ~16 hours, for an average eruption rate of 1.2x10⁷ kg/s. I was unable to find any peer-reviewed publications giving erupted volume or grain-size distribution.

Ruang

25 September 2004
The GVN monthly report for 10/2002 notes primarily satellite imagery analyzed by the Darwin VAAC showing an ash cloud drifting westward toward Borneo and Sumatra, and ground-based estimates of a 5-km high plume; however the monthly report of 2/2004 says “The 25 September 2002 eruption of Ruang was, according to the Darwin VAAC, the largest in Indonesia in many years, and was well observed by satellite sensors.” Tupper et al. [2004] describe satellite observations in detail. It notes a cloud height of ~20 km, a pyroclastic flow to the SE that damaged about 1.6 km2, and says the main eruption began at 0340 UTC on 8/25 and went until about 2240 UTC. Tupper and Kinoshita [2003] explain the large discrepancy between the real-time plume-height estimate of 5 km and the later satellite-based 20- km estimate as due to the fact that the observatory, located 3 km from the volcano, is not a good location for ground-based observers to estimate plume height. The exact location of the eruption on the volcano was not firmly established, but was presumed to be “Crater II”, where the 1949 eruption originated. The GVN monthly report notes that the eruption commenced at 1140 and that by 1210 the activity “subsided enough to observe glowing material on the E flank.” This is the basis for the 0.5 hour duration given in the table. No information on erupted volume or grain-size distribution is available.

Ruapheu

17 June 1996
Excellent descriptions of the June 17, 1996 eruption are published in the GVN monthly report and in journal publications. This is the classic “weak plume” that is modeled by Bonadonna et al. [2005]. The maximum plume height of 8.5 km asl was reached in the distal portion of the first plume, already detached from the vent [Prata and Grant, 2001] (Ruapehu’s summit elevation is 2,797 m). Plume height varied with time, being greatest between 0830 and 1300, and between 1500 and 1700 respectively. The GVN Monthly Report of 5/1996 indicates that “around 1500 the volcano started to erupt every 10-15 minutes. A significant Strombolian eruption during 2100-2200 was characterized by loud detonations and sprays of glowing rocks ejected above the crater.” Some evidence of eruptive activity persisted at least through the morning of June 18. The stated “duration” of this eruption depends on what one defines as an end. For purposes of plume dynamics perhaps sometime in the evening, say, 12 hours or so after the eruption began. Bonadonna et al. use a total erupted mass of 5x10⁹ kg in their modeling, which converts to 2x10⁶ m³ DRE at a density of 2500 kg/m³. Dividing this mass by 12 hours gives an average eruption rate of 1x10⁵ kg/s. Bonadonna and Houghton [2005] used a variety of methods to estimate the total grain-size distribution of the deposit from the samples. Their best estimate (Technique C in Table 1, which uses the Voronoi Tessellation method to weight each sample by the map area of a cell it occupies) gives md_φ=-0.8, σ_φ=2.43. However this is the GSD of the deposit that fell out of the ash cloud, not the ash cloud itself.

Santa Maria

25 October 1902
The eruption of 1902 is described by Rose [1972] and Sapper [1904], and data on eruption rate and plume height are tabulated in Sparks et al. [1997, Table 5.1] and Wilson et al. [1978, Table 3]. Sapper gives the total erupted volume as 5.5 km³, presumably of tephra. Using a slightly different method of estimation, Rose et al. [1973a, Table 7] estimate about 4.3 km³ of tephra. Rose [1972, p. 3] also says that “from most descriptions the activity seemed to be highly concentrated in the first 24-36 hours.” Assuming a bulk density of 1500 kg/m3, I obtain ~3.3 km³ DRE, or 8.2x10¹² kg; and an average mass flow rate over 24-36 hours of 6-10x10⁸ kg/s. Rose cites Anderson [1908a] for the plume height estimate, saying “On October 25, Captain Saunders aboard the S.S. Newport off the Pacific coast estimated the height of the cloud at 27-29 km using a sextant.” Santa Maria volcano is 3,772 m height, making the plume about 23-25 km above the summit. A second observation Carey and Sparks [1986] note a second observation published in Anderson [1908b], which gives the column height as 48 km (45 above the vent, presumably). Carey and Sparks use ispopleth data to estimate a column height of about 34 km.

Soufrière of St. Vincent

7 May 1902
The 7 May 1902 eruption of Soufrière of St. Vincent was documented by Anderson and Flett [1903]. Descriptions of the course of events that day help constrain the eruption duration. Deposits, 94% of which were tephra fall, were studied by Carey and Sigurdsson [1978]. Carey and Sigurdsson cite Anderson and Flett as saying that “the height of the Soufrière eruption cloud was surveyed by Major Hodder on St. Lucia as 15 km” (the summit elevation is 1220 m). Carey and Sigurdsson estimate the volume of tephra fall (DRE) as 0.14 km³. Like the Mount St. Helens eruptions, this one has an anomalously low plume height for the calculated eruption rate, raising the question of whether co-ignimbrite ash clouds contributed significantly to the tephra sheet. Carey and Sigurdsson [1978] start their paper by saying “The 7 May 1902 eruption of the Soufrière on St. Vincent has become a textbook example of the production of pyroclastic flows from a vertical eruption column (Hay, 1959). It therefore seems possible that the elutriation of pyroclastic flows could have played a significant role.

Spurr

1992 eruptions
For these eruptions, Neal et al. [1995, p. 68] give erupted volumes. Eichelberger et al. [1995, p. 11-12] give the start and end times of the eruptions, and an Eos article [Alaska_Volcano_Observatory, 1993] reports radar data on plume height.

St Helens

18 May 1980.
The information on the May 18, 1980 eruption is taken mostly from Professional Paper 1250. The plume heights above sea level shown in Harris et al. [1981, Fig. 190], measured from the hours of 0900 to 1700 range from 13.4 km at 1300 to 19.2 km at 1700. The average of 12 measurements in this time period is 15.5 km. The vent elevation during the eruption was roughly 2,000 m; hence I give the plume height as roughly 13.5. Sarna-Wojcicki et al. [Sarna-Wojcicki, et al., 1981] estimated a total tephra volume of about 1.1 km³ and an uncompacted bulk density of 450 kg/m³, yielding an approximate erupted volume of 0.2-0.25 km3 DRE (4.9x10¹¹ kg). (Carey and Sigurdsson [1982] however estimate >0.55 km³ tephra.) Over the 9-hour-long eruption this gives an average a mass eruption rate of 1.7-2.1x10⁷ kg/s. Fierstein and Nathenson [1992, Table 3] used a slightly different method for calculating tephra volume and obtained 1.19 km³. The grain –size distribution was estimated by Adam Durant (personal communication) as part of his Ph.D. Sparks et al. [1997, Table 5.1] tabulate column height-eruption rate data for units B1 through B4 [Waitt and Dzurisin, 1981; Criswell, 1987]of the May 18, 1980 deposit. The data come from Carey et al. [1990], who derived column heights from isopleths. The isoplethderived column heights agreed reasonably well with the radar data [Harris, et al., 1981]. Carey et al. then used column height and the model of Sparks [1986] to estimate eruption rate. He found that the eruption rate predicted by these column heights underestimated the total mass of the tephra sheet by about 2.5 times (1.8x1011 versus 5.1x10¹¹ kg), and inferred from this that most of the deposit resulted from elutriation of pyroclastic flows. The entries in Table 5.1 of Sparks et al. [Sparks, et al., 1997] are derived from the Sparks [1986] model, not from measured deposit volume and eruption time, and therefore they are not included in this table.

Late May through July 1980
Erupted volumes for the other eruptions taken from Sarna- Wojcicki et al. [1981]; plume heights taken from radar measurements in Harris et al. [1981, Table 36]. Harris gives the height of the May 25 eruption as 12.2 km, the heights of four pulses of the June 12 eruption as 15.2, 10.7, 9.8, and 10.7 km; and three heights of pulses of the July 22 eruption as 13.7, 14.5, 8.7 km. For the latter two eruptions I take the average of the heights and subtract the vent elevation (~2 km). The June 12 eruption produced a highly irregular tephra-fall distribution map (Fig. 345 of Sarna-Wojcicki et al. [1981]), which was attributed to variations in the wind vector with elevation. All of the eruptions of Mount St. Helens appear to give lower plume heights for the given eruption rate than other eruptions, raising the question whether post-May 18 tephra deposits contained a significant fraction of co-ignimbrite as ash Carey et al. [1990] infer for the May 18 deposit. Christiansen and Peterson [1981, p. 26] note that “Several new ash flows issued from the vent crater and descended the volcano’s north flank during the June 12 activity.” During the July 22 eruption they also note that “Pumiceous ash flows were erupted on the north flank of the volcano, especially during the second and third eruptive bursts.” The July 22 eruption does not however appear to have an anomalously low eruptive plume for the inferred eruption rate.

8 March 2005
The duration, plume height, and average eruption rate of the May 8, 2005 eruption are reported in Mastin [2007].

Tungurahua

14 July 2006
Tungurahua is an andesite-dacite volcano in the Andes that started an eruptive sequence in 1999. Typical eruptions from this volcano are Strombolian. The eruption on July 14-15, 2006 was the strongest during this eruptive period. Seismic records posted on the GVN monthly report suggest that the eruption started about 2230Z and had greatly subsided by about 0300Z, though periodic explosions continued throughout the rest of 15 July. Table 10 in that monthly report indicates that the plume reached an altitude of 15 km (it’s not clear whether this is above the 5,023-m high vent or above sea level). There was intense lava fountaining and moderate scoria fall to the west. I’ve found no information on erupted volume.