Long-term volumetric eruption rates and magma budgets

Authors


Abstract

[1] A global compilation of 170 time-averaged volumetric volcanic output rates (Qe) is evaluated in terms of composition and petrotectonic setting to advance the understanding of long-term rates of magma generation and eruption on Earth. Repose periods between successive eruptions at a given site and intrusive:extrusive ratios were compiled for selected volcanic centers where long-term (>104 years) data were available. More silicic compositions, rhyolites and andesites, have a more limited range of eruption rates than basalts. Even when high Qe values contributed by flood basalts (9 ± 2 × 10−1 km3/yr) are removed, there is a trend in decreasing average Qe with lava composition from basaltic eruptions (2.6 ± 1.0 × 10−2 km3/yr) to andesites (2.3 ± 0.8 × 10−3 km3/yr) and rhyolites (4.0 ± 1.4 × 10−3 km3/yr). This trend is also seen in the difference between oceanic and continental settings, as eruptions on oceanic crust tend to be predominately basaltic. All of the volcanoes occurring in oceanic settings fail to have statistically different mean Qe and have an overall average of 2.8 ± 0.4 × 10−2 km3/yr, excluding flood basalts. Likewise, all of the volcanoes on continental crust also fail to have statistically different mean Qe and have an overall average of 4.4 ± 0.8 × 10−3 km3/yr. Flood basalts also form a distinctive class with an average Qe nearly two orders of magnitude higher than any other class. However, we have found no systematic evidence linking increased intrusive:extrusive ratios with lower volcanic rates. A simple heat balance analysis suggests that the preponderance of volcanic systems must be open magmatic systems with respect to heat and matter transport in order to maintain eruptible magma at shallow depth throughout the observed lifetime of the volcano. The empirical upper limit of ∼10−2 km3/yr for magma eruption rate in systems with relatively high intrusive:extrusive ratios may be a consequence of the fundamental parameters governing rates of melt generation (e.g., subsolidus isentropic decompression, hydration due to slab dehydration and heat transfer between underplated magma and the overlying crust) in the Earth.

1. Introduction

[2] Despite the significant impact of volcanic systems on climate, geochemical cycles, geothermal resources and the evolution and heat budget of the crust, surprisingly little is known regarding the systematics of long-term rates of magma generation and eruption on Earth. Global rates of magma generation provide insight regarding the planetary-scale energy budget and thermal evolution of the Earth. Rates of magma generation and eruption are key factors affecting the petrological and geochemical evolution of magma bodies as well as eruptive styles due to the intrinsic coupling between magma recharge, fractional crystallization, wall rock assimilation and melt volatile saturation [Shaw, 1985; Spera et al., 1982]. Volcanoes and formation of intrusive bodies such as sill complexes have been suggested to play a role in global climate change [Svensen et al., 2004] and perhaps even trigger biotic extinctions. In addition, global rates of magmatism may have important implications for seismic energy release [Shaw, 1980] and the magnetic geodynamo by modulating heat transfer from the core-mantle boundary and the concomitant development of deep mantle plumes [Olson, 1994]. Rates of magmatism on Earth are also used in planetary research as analogues to constrain magmatic and thermal models. In summary, there is an exhaustive set of reasons for developing systematic knowledge regarding the rates of magmatism on Earth including the effects of magma composition and petrotectonic environment on volumetric rates.

[3] One of the key factors in understanding magmatism is a quantitative evaluation of the extent to which magmatic systems operate as open or closed systems. These alternatives have significantly different implications for magma evolution. However, the openness of magmatic systems is difficult to determine since there is no unambiguous way to track magma transport from the generation and segregation through the crust to volcanic output. On balance, many magma systems are thought to be open systems in that they receive additional inputs of heat and mass during magmatic evolution [Davidson et al., 1988; Fowler et al., 2004; Gamble et al., 1999; Hildreth et al., 1986; Petford and Gallagher, 2001]. Closed magmatic systems which exchange heat but little material with their surroundings (i.e., neither assimilation nor recharge is important) may be rather uncommon. What is more likely is that specific systems may behave as closed systems for restricted portions of their history [e.g., Singer et al., 1992; Zielinski and Frey, 1970]. It is important to note, however, for the olivine basalt-trachyte series at Gough Island where fractional crystallization appears dominant, Pb and Sr isotopic data indicates that assimilation of hydrothermally altered country rock and/or recharge of isotopically distinct magma has taken place [Oversby and Gast, 1970].

[4] In this paper, time-averaged volcanic output for periods >103 years are evaluated. Volcanic output rates for individual eruptions may vary wildly about some norm, but evidently settle to a representative “average” value when time windows on the order of 10 times the average interval of eruptions are considered [Wadge, 1982]. Crisp [1984] conducted a similar study of magmatic rates published between 1962 and 1982 and established some basic relationships between volcanic output and associated factors such as crustal thickness, magma composition, and petrotectonic setting. This work updates and extends that earlier compilation with 98 newly published volcanic rates and volumes from 1982–2004 for a total of 170 estimates (see auxiliary material Tables S1 and S2). We also endeavor to establish some scaling relationships based primarily on the compilation and some simple energy budget considerations with the goal of discovering possible systematic trends in the data.

2. Sources and Quality of the Data

[5] The data presented here are volumetric volcanic or intrusive rates published from 1962–2005, including data from the compilation by Crisp [1984] of rates published from 1962–1982 where these data have not been superseded by more recent studies. We have also reviewed the rate data presented by Crisp [1984] and corrected or removed several references as appropriate. Thus the data presented here is a completely updated compilation of volumetric rates of eruption.

[6] Most volcanoes have cycles of intense activity followed by repose. Comparing volcanic systems at different stages in their eruptive cycles can lead to erroneous conclusions, if the duration of activity is not long enough to average the full range of eruptive behavior over the lifetime of the volcano. The duration needed depends upon the individual volcano; longer periods are generally required for volcanic centers erupting more compositionally evolved magma due to lower eruption recurrence interval. Thus a period of ∼103 years may be a long time for a basaltic shield volcano (e.g., Kilauea, Hawaii) but captures only an insignificant fraction of one eruptive cycle at a rhyolitic caldera (e.g., Yellowstone, USA). Only long-term rates are considered in this study although this reduces the available data considerably. We have culled the data to include primarily those estimates over 104 years or longer, but have selected a few volcanic centers with shorter durations where the shorter time interval did not compromise the data quality (e.g., capturing several eruptive cycles, smaller volcanic centers, or similar reasons) in our judgment.

[7] Tables 1 and 2 show volcanic output rates for primarily mafic and silicic systems respectively. Output rates for volcanic systems (Qe) are determined by dividing volcanic output volume by the duration of the activity. For longer durations activity may not have been continuous. By use of density for different compositions [Spera, 2000] we can convert volume rate (Qe) to mass rate, which is probably the more fundamental parameter. Since density varies only slightly (basalt is ∼15% denser than rhyolite at the same temperature and pressure) compared to the uncertainty in the data and the original data is all reported in terms of volume, we use Qe exclusively in the rest of this study although mass rates are also given in Tables 1 and 2. Within each table, the rate estimates encompassing large areas, such as entire arcs or extensive volcanic fields, is presented separately from rates for individual volcanoes or smaller fields of vents. To remove ambiguity from the decision, a cutoff of 104 km2 was used to separate global data sets, typically involving compilations of several volcanoes themselves, from local data sets focused on individual volcanoes with a more constrained study area. However, we find that rates for entire arcs/fields when presented as km3/yr per 100 km are similar to those for individual volcanoes (Figure 1).

Figure 1.

Volumes and volcanism durations for locations in Tables 1 and 2 (see also auxiliary material Tables S1 and S2). The diagonal lines represent constant rates of volcanic output. The points are coded by color and shape to indicate lava composition by SiO2 content. Open symbols represent rates for arc and large areas (>104 km2), and solid symbols represent individual volcanoes and small volcanic fields (<104 km2).

Table 1. Rates and Volumes of Basaltic Volcanism
Location (Volcano Name)Duration, MyrExtrusive Volume, km3Volume Extrusion Rate Qe, km3 yr−1Mass Extrusion Rate, kg yr−1Bulk SiO2Petrotectonic SettingNotesReferences
Area < 104km2(Individual Volcanoes/Small Volcanic Fields)
Ascension1.500906.00E−051.49E+0948oceanic hot spotRough estimate of volumes from topography; rates constrained by a few K-Ar dates since 1.5 Ma.Gerlach [1990], Nielson and Sibbett [1996]
Auckland, New Zealand0.14021.07E−052.89E+07BContinental volcanic fieldVolume calculated from thickness and areal extent based on field mapping and boreholes for 49 volcanic centers and adjusted to DRE volume. Active for last 140 kyr based on K-Ar, thermoluminescence, and 14C dates.Allen and Smith [1994]
Bouvet0.700284.00E−054.59E+0848oceanic hot spotVery rough estimate of volume from island topography; active for the past 0.7 Myr.Gerlach [1990]
Camargo, Mexico4.641202.6E−057.02E+07BContinental volcanic fieldConstraints from K-Ar dates from 4.73 ± 0.04 Ma to 0.09 ± 0.04 Ma. Volume based on an area of 3000 km2 and average thickness of 40 mAranda-Gomez et al. [2003]
La Palma, Canary Islands0.1231251.0E−032.70E+0948oceanic hot spotDetailed field observations, mapping, and 39Ar/40Ar dating of uneroded Cumbre Viejo indicate activity since 123 ± 3 ka.Carracedo et al. [1999], Guillou et al. [1998]
Santo Antao, Cape Verdes1.750684.00E−051.08E+0848oceanic hot spotRates from main shield-building stage Cha de Morte volcanics deposited between 2.93 ± 0.03 and 1.18 ± 0.01 Ma (39Ar/40Ar ages) and field mapping.Plesner et al. [2002]
Coso, CA1.50024.31.60E−055.40E+1257continental volcanic fieldField mapping estimate of 23–25.5 km3 erupted between 4.02 ± 0.06 and 2.52 ± 0.05 Ma (K-Ar ages).Duffield et al. [1980]
Crater Flat, Nevada3.700103.7E−059.99E+07Bcontinental volcanic fieldComprehensive study of volcanism in southern Nevada. Volumes from topography and field mapping, ages constrained to last 3.7 Myr mainly by 40Ar/39Ar.Perry et al. [1998]
Edgecumbe0.607325E−051.35E+0853continental arcOldest age is K-Ar 611 ± 74 ka; youngest is radiocarbon date 4030 ± 90 a. Volume based on detailed geologic mapping and accounting for DRE volumes for lavas and pyroclastic deposits.Crisp [1984], Riehle et al. [1992]
Eifel Volcanic field0.69022.14E−065.79E+06Bcontinental volcanic fieldVolumes from field mapping. Activity 0.7 to 0.01 Ma based on 40Ar/39Ar and tephrachonology.Mertes and Schmincke [1985]
Erebus0.2504.00E−034.0E−031.08E+10BContinental arcVolumes based on topography and assumed erosion, age constraints from 39Ar/40Ar dates.Esser et al. [2004]
Erebus0.9501.20E−031.2E−033.24E+09BContinental arcRough estimate of volume for proto-Erebus from topography and age constraints from 40Ar/39Ar dates.Esser et al. [2004]
Fernando de Noronha12.000605.00E−061.35E+0748oceanic hot spotApproximate duration of volcanism 14 to 2 Ma.Gerlach [1990]
Gough2.42994.09E−051.10E+0848oceanic hot spotK-Ar ages from 2.55+0.51 to 0.13+0.02 Ma. Volume of basaltic shield does not include estimate of eroded volume prior to trachyte phase.Chevallier [1987], Maund et al. [1988]
Juan Fernandez1.000727.20E−051.94E+0848oceanic hot spotApproximate duration of volcanism 3 to 4 Ma. Ages constrained by K-Ar dates.Gerlach [1990]
Kilauea0.4200005.00E−021.35E+1148oceanic hot spotVolume estimate from drill hole stratigraphy and topography for the past 0.4 Myr, based on K-Ar ages.Quane et al. [2000], Dvorak and Dzurisin [1993]
Kluycheversuskoy0.0102752.70E−027.29E+10Bcontinental arcVolume based on cone is 275 ± 25 km3 and does not include ejecta beyond cone.Crisp [1984]
Kohala, Hawaii0.400140003.50E−029.45E+10Boceanic hot spotVolume estimate from bathymetry and topographic maps for the past 0.4 Myr, based on K-Ar ages.Crisp [1984]
Koolau, Oahu0.600209003.50E−029.45E+10Boceanic hot spotVolume estimate from bathymetry and topographic maps for the past 0.6 Myr, based on K-Ar ages.Crisp [1984]
La Gomera, Canary Islands1.43502.5E−046.75E+08Boceanic hot spotRates obtained for main shield building stage. Volume calculated from DEM and field mapping for period of 9.4–8.0 Ma (K-Ar dates and paleomagnetic stratigraphy).Paris et al. [2005]
Lunar Crater, NV5.7001001.70E−054.59E+07 Continental Volcanic fieldApproximately 100 km3 erupted over the past 5.7 Myr.Crisp [1984]
Mauna Kea, Hawaii0.300248008.30E−022.24E+1146oceanic hot spotVolume estimate from bathymetry and topographic maps for the past 0.3 Myr, based on K-Ar ages.Crisp [1984]
Mauna Loa, Hawaii0.004802.00E−025.40E+1048oceanic hot spotVolume estimate from bathymetry and topographic maps for the past 4 kyr, based on K-Ar ages.Lipman [1995]
Mauna Loa, Hawaii0.500425008.50E−022.30E+1151oceanic hot spotVolume estimate from bathymetry and topographic maps for the past 0.5 Myr, based on K-Ar ages.Crisp [1984]
Mt. Cameroon3.015005.00E−041.35E+0945continental hot spotRough estimate of volume for entire volcano, age constraints approximate activity from <3 Ma to present.Fitton and James [1986], Marzoli et al. [2000]
Ocate, New Mexico4.0862.1E−055.67E+07BContinental RiftLavas erupted from 4.8 to 0.8 Ma, based on whole rock K-Ar dates. Approximate volume from field mapping. Older deposits (5.5 to 8.1 Ma) are substantially eroded so these are excluded here.Nielsen and Dungan [1985]
Reunion2.00048002.40E−035.67E+1148oceanic hot spotBased on a volcanic rate of 2.4 ± 0.4 km3/kyr and geochronology from early K-Ar dates for building the island over the past 2 Myr.Gerlach [1990]
Ross4.00048001.20E−036.48E+0948oceanic hot spotBased on a volcanic rate of 1.2 ± 0.2 km3/kyr for building the island over an estimated period of 0–4 Ma.Gerlach [1990]
San Francisco Volcanic Field, AZ5.0005251.05E–042.84E+08BContinental Volcanic fieldVolume of small basaltic cinder cones and flows and larger silicic cones estimated from field mapping over the past 5 Myr based on K-Ar dates.Tanaka et al. [1986]
Sao Miguel, Azores0.004554.59.90E–042.67E+09Boceanic hot spotVolume of 4.5 km3, DRE, from field mapping erupted over the past 4.55 kyr based on 14C ages.Crisp [1984]
Servilleta Basalt1.02002.0E–045.40E+08BContinental riftEstimate of volume based on extent (200 km2) and average thickness of flows (50 m) which range from 10 to 175 m thick. Eruptions in middle-to-late Pliocene is 1.0 ± 0.5 Ma based on K-Ar dates.Dungan et al. [1986]
Skye, British Tertiary Igneous Province1.60023001.44E−033.88E+09BContinental riftEstimate of volume based on areal extent (1550 km2) times average thickness (1.5 km) of deposits. Volcanic activity between 60.53 ± 0.08 Ma and 58.91 ± 0.06 Ma based on U-Pb zircon ages.Fowler et al. [2004]
Springerville, Arizona1.83001.67E−044.51E+0848Continental Volcanic fieldVolumes from geologic mapping and borehole data. Eruptive period 0.3–2.1 Ma from K-Ar and magnetic polarity ages.Condit et al. [1989]
St. Helena8.0001922.40E−057.29E+0948oceanic hot spotBased on a volcanic rate of 0.24 ± 0.12 km3/Myr for building the island over an estimated period of 7–15 Myr.Gerlach [1990]
Tatara, Chile0.071223.1E−048.37E+08BAcontinental arcEstimates from field mapping, K-Ar and 40Ar/39Ar ages. Eruption time interval from 90 to 19 ka; units older than 90 ka are highly erodedSinger et al. [1997]
Tolbachik0.010696.90E−036.48E+08Bcontinental arcFrom rough estimate of average eruption rates over 2–10 ka and 2 ka to present.Crisp [1984]
 
Area > 104km2(Large Volcanic Fields/Arcs)
Baikal (Vitim)1050005.0E−041.35E+09Bcontinental riftRough estimate of volumes and rates constrained by very limited K-Ar dates. Eruptive period 16.6–6.6 Ma.Johnson et al. [2005], Crisp [1984]
Canary Islands20.0001500007.50E−031.08E+08Boceanic hot spotAge of oldest shield-building activity is uncertain. Oldest dated subaerial lava is 20 Ma (K-Ar) but oldest submarine activity could be 3X older.Crisp [1984], Hoernle and Schmincke [1993a]
Caribbean Plateau3.00040000001.33E+001.08E+0850ocean plateauVolume constraints from bathymetry and a drill hole. Eruptions from 91 to 88 Ma based on 40Ar/39Ar ages.Courtillot and Renne [2003], Sinton et al. [1998]
Central Atlantic Basalt Province4.00020000005.00E−015.94E+0950continental flood basaltRough estimate of volume based on extent including Eastern North America, western Africa and Europe. Eruption between 197 and 201 Ma from a variety of sources (40Ar/39Ar, U-Pb, biostratigraphy)Courtillot and Renne [2003]
Central Oregon p/100 km1.01631.63E−044.40E+08Bcontinental arc330 km3 from mafic volcanoes and 180 km3 from composite volcanoes for the past 1 Myr. Uncertainties 10%–30% of volume. Adjusted for 300 km length of arc.Sherrod and Smith [1990]
Central Oregon2.07503.75E−041.01E+0953continental arcVolume estimated from field mapping and probable eroded volume at 2150 km3 and adjusted for 300 km arc. K-Ar age dates cluster in the period from 4 to 6 Ma.Crisp [1984]
Columbia River Basalt (Grande Ronde Basalt)1.41486001.06E−012.86E+11Bcontinental flood basaltVolume estimate from field mapping and borehole thickness data. Lavas were erupted 17–15.6 Ma based on 40Ar/39Ar and K-Ar ages.Reidel et al. [1989]
Columbia River Basalt11.51743001.54E−024.32E+0650continental flood basaltVolume estimate from field mapping and borehole thickness data for 174,300 ± 27,900 km3. Lavas were erupted 17.5–6 Ma based on 40Ar/39Ar and K-Ar ages.Tolan et al. [1989]
Deccan Traps1.00020000002.00E+008.91E+0850continental flood basaltEruptive period 65–66 Ma. There is a fair amount of controversy over the duration of activity.Courtillot and Renne [2003]
E. Australia60.000200003.30E−042.59E+10Bcontinental flood basalVolume estimate from geologic mapping. Extrusion rate approximately constant for the past 60 Myr. Only slight erosion for deposits less than 37 Ma.Crisp [1984]
East Pacific Rise (p/100 km)1.00096009.60E−031.89E+1050ocean spreadingVolume calculated by multiplying area from spreading rate and given length, by extrusive layer thickness obtained from reflection seismic and stratigraphic mapping. Seismic Layer 2A assumed equal to extrusive layer thickness.Karson [1998], Harding et al. [1989], Becker et al. [1989]
Eastern Snake River Plain3.000210007.00E−037.02E+09Bcontinental hot spotRough estimate for the past 3 Myr, with no estimate of uncertainties.Crisp [1984]
Ethiopia-Yemen Traps1.50012000008.00E−012.70E+1250continental flood basaltEthiopian traps are 0.7 Mkm3 and Yemen traps are 1.2 Mkm3. Erupted from 29.5 to 31 Ma based on 40Ar/39Ar and magneto-stratigraphy.Courtillot and Renne [2003]
Emeishan1.00010000001.00E+002.21E+1050continental flood basaltRough estimate of volume from reconstruction of original area (5 Mkm2) and thickness (2 km). Ages from 40Ar/39Ar and U-Pb are somewhat uncertain. Stratigraphic age is 258 Ma.Courtillot and Renne [2003]
Faeroe Islands3.542001.20E−033.24E+0949oceanic hot spotVolume is rough estimate based on area of the islands, with 3 km thickness of lava pile. Unknown amount of original volume removed by glaciers, so volume is an underestimate. Ages based on 40Ar/39Ar and paleomagnetic dating.Crisp [1984], Riisager et al. [2002]
Great Rift, Snake River Plain, Idaho0.013312.40E−036.48E+0954continental hot spotUniform rate of 1.5E−03 km3/yr from 15 to 7 ka, then an increase to 2.8E−03 km3/yr from 7 to 2 ka. Lava flow area mapped and multiplied by average flow thickness to derive volumes for each of 38 individual flows. Age constraints from radiocarbon and paleomagnetic dating.Kuntz et al. [1986]
Hawaii-Emperor Seamounts73.60010800001.50E−024.05E+1048oceanic hot spotComprehensive assessment of volcano volumes using bathymetry. K-Ar ages.Crisp [1984]
Iceland0.0114844.4E−021.19E+1150oceanic hot spotVolumes of postglacial eruptions over the past 10–12 kyr, constrained by radiocarbon and tephrachronology.Crisp [1984]
Iceland1.000200002.00E−025.40E+10Boceanic hot spotBased on total volume of Iceland divided by age of volcanism.Crisp [1984]
Ireland5.50020003.60E−049.72E+08BContinental riftEruptive period from 66 Ma to 60.5 ± 0.5 Ma estimated from K-Ar and 40Ar/39Ar ages. Volume estimated from the total area of 4000 km2 and average thickness of 0.5 km.Crisp [1984]
Juan de Fuca Ridge (p/100 km)1.00090009.00E−032.43E+1050ocean spreadingBased on stratigraphic mapping of the extrusive layer thickness where exposed by faults, and the average spreading rate.Karson [1998]
Kamchatka0.25087003.50E−026.48E+10Bcontinental arcEstimate of erupted volume between 0.85–0.6 Ma corrected for DRE.Crisp [1984]
Kamchatka (basalts)0.850206602.40E−029.45E+10Bcontinental arc20660 ± 300 km3 erupted over the past 0.08 Myr (from mainly radiocarbon ages). Overall rate calculated from a detailed assessment of rates as a function of time and type of volcanism.Crisp [1984]
Kamchatka (basalts)0.08061407.70E−022.08E+11Bcontinental arcRate calculated for just the basaltic volume erupted over the past 0.08 MyrCrisp [1984]
Karoo-Farrar6.00025000004.16E−011.12E+1250continental flood basaltThe volume is highly uncertain due to erosion, and the duration from 184 Ma to 178 Ma is based on 40Ar/39Ar ages.Courtillot and Renne [2003], Jourdan et al. [2005]
Kenya (basalts)2.500155806.20E−039.45E+09Bcontinental riftVolume from geologic mapping and pre-erosional estimates for the past 2.5 Myr, based on K-Ar and 40Ar/39Ar ages.Crisp [1984]
Kenya (basalts)4.500230005.10E−031.38E+10Bcontinental riftVolume from geologic mapping and pre-erosional estimates for 7–2.5 Ma, based on K-Ar and 40Ar/39Ar ages. Uncertainties larger for older deposits.Crisp [1984]
Kenya (basalts)11.000390003.50E−031.67E+10Bcontinental riftVolume from geologic mapping and pre-erosional estimates for 23–12 Ma, based on K-Ar and 40Ar/39Ar ages. Uncertainties larger for older deposits.Crisp [1984]
Kenya and Tanzania2.500404701.60E−024.32E+10Bcontinental riftVolume (40,470 ± 7680 km3) from geologic mapping and pre-erosional estimates for the past 2.5 Myr, based on K-Ar and 40Ar/39Ar ages.Crisp [1984]
Kenya and Uganda11.0001083759.80E−032.65E+10Bcontinental riftVolume (108,375 ± 23,125 km3) from geologic mapping and pre-erosional estimates for 23 to 12 Ma, based on K-Ar and 40Ar/39Ar ages.Crisp [1984]
Kerguelen Archipelago11.000990009.00E−035.94E+0850oceanic hot spotRough estimate of volume for island based on topography and bathymetry for 29.26 ± 0.87 to 24.53 ± 0.29 Ma, based on 40Ar/39Ar ages.Nicolaysen et al. [2000]
Kerguelen Island26.00057202.20E−042.43E+1048oceanic hot spotBased on a volcanic rate of 0.22 ± 0.01 km3/kyr for building the island over an estimated period from 27 Ma to 1 Ma.Gerlach [1990]
Kerguelen-Rajmahal-Bunbury10.00060000006.00E–011.62E+1250continental flood basaltGrouped based on geochemical similarity. Erupted at 119–109 Ma.Courtillot and Renne [2003]
Madagascar4.00044000001.10E+002.97E+1250continental flood basaltRough estimate of volume from area and thickness; erupted from 86–90 Ma based on 40Ar/39Ar dating.Courtillot and Renne [2003]
Mid-Atlantic (p/100 km)1.00015001.50E–034.05E+0950Ocean spreadingVolume calculated by multiplying area from spreading rate and given length, by extrusive layer thickness obtained from reflection seismic. Seismic Layer 2A assumed equal to extrusive layer thickness.Tolstoy et al. [1993], Tucholke et al. [1997], Hooft et al. [2000]
Midland Valley, Scotland27.50064002.3E–046.21E+08Bcontinental riftHighly uncertain estimate of volume (4800 to 8000 km3) erupted over 25–30 Ma during Carboniferous Period. Volume and age range estimates from Crisp [1984] corrected here.Crisp [1984]
Modoc Basalt5.00025505.10E−047.59E+08Bcontinental arcVolume estimate from 1950 to 3150 km3 based on field mapping. K-Ar ages from 5 to 10 Ma.Crisp [1984]
Ninetyeast Ridge40.00072000001.80E−011.08E+1048oceanic hot spotRough estimate of volume for whole ridge based on topography and bathymetry for 38 to 83 Ma, based on 40Ar/39Ar ages.Duncan [1991], Nicolaysen et al. [2000]
Ninetyeast Ridge40.000240000006.00E−014.86E+1148oceanic hot spotEstimate for whole ridge based area and thickness of volcanic extrusive layer interpreted from seismic profiles.Grevemeyer et al. [2001]
North Atlantic Tertiary Volcanics 13.00020000006.67E−011.62E+1250continental flood basaltRough volume estimate for the period of 62–59 Ma based on an extensive suite of dates including 40Ar/39Ar and U-Pb ages.Courtillot and Renne [2003]
North Atlantic Tertiary Volcanics 23.00020000006.67E−011.62E+1250continental flood basaltConservative estimate of extrusive volume for the period of 57–54 Ma based on 40Ar/39Ar and U-Pb ages. Some estimates include ∼9.9 Mkm3 of total igneous material.Courtillot and Renne [2003]
Ontong-Java Plateau2.00060000003.00E+001.80E+1250ocean plateauTotal, corrected volume of extruded lavas at 122 ± 1 Ma. Total volume of plateau is 44.4 Mkm3Courtillot and Renne [2003], Neal et al. [1997]
Parana-Etendeka0.60023000003.83E+002.16E+0850continental flood basaltVolume calculated from surface area (1.5 Mkm2) and average thickness of flows (36 km). Total duration of trap volcanism is constrained to 0.6 ± 1 Myr at 133 ± 1 Myr ago by 40Ar/39Ar dates.Courtillot and Renne [2003], Renne and Basu [1991]
Parana 13.000900003.00E−021.04E+1350continental flood basaltFirst phase of basaltic volcanism. Rough volume estimate from average flow thickness and extent. Ages constrained by numerous 40Ar/39Ar dates.Stewart et al. [1996]
Parana 22.0002600001.30E−012.16E+1150continental flood basaltSecond phase of basaltic volcanism. Rough volume estimate from average flow thickness and extent. Ages constrained by numerous 40Ar/39Ar dates.Stewart et al. [1996]
Parana 32.0004200002.10E−018.10E+1050continental flood basaltThird phase of basaltic volcanism. Rough volume estimate from average flow thickness and extent. Ages constrained by numerous 40Ar/39Ar dates.Stewart et al. [1996]
Parana 43.000300001.00E−023.51E+1150continental flood basaltFourth phase of basaltic volcanism. Rough volume estimate from average flow thickness and extent. Ages constrained by numerous 40Ar/39Ar dates.Stewart et al. [1996]
Rio Grande Rise37.000111000003.00E−012.70E+1048oceanic hot spotVolume estimated from bathymetry and subsidence.Gallagher and Hawkesworth [1994]
San Juan Mountains (basalts)19.80010205.10E−053.24E+09BContinental Volcanic fieldVolume estimated from field mapping, includes estimate of eroded volume. K-Ar ages range from 23.4 to 3.6 Ma.Crisp [1984]
Siberia32.0007500002.70E−031.73E+10Bcontinental flood basaltVolume highly uncertain: 575,000 to 1,150,000 km3 erupted from 248 to 216 Ma.Crisp [1984]
Siberian Traps1.00030000003.00E+002.43E+0950continental flood basaltVolume from reconstruction of original total area of 4 Mkm2 and preserved average flow thickness. Erupted during 1 Myr at about 250 Ma based on magnetic stratigraphy, U-Pb, and 40Ar/39Ar ages.Courtillot and Renne [2003], Renne and Basu [1991]
Tasmania44.5004009.40E−068.10E+12Bcontinental riftDuration from 47.5 ± 7.5 to 2.55 ± 2.45 Ma.Crisp [1984]
Tibesti Massif64.00026804.20E−052.54E+0755continental hot spotRough estimate of total volume from 65 to 0.95 ± 0.85 Ma.Crisp [1984]
Zitacuaro-Valle de Bravo1.000242.40E−051.13E+0850continental arcVolume estimated from detailed field mapping of uneroded young volcanics broken out by lava type and corrected to DRE. A total of 25 km3 of basalts were erupted in shields, cones, and tephra along 50 km of arc length. Extensive suite of dates from 40Ar/39Ar, K-Ar, and radiocarbon. Rates only for volcanics of past 1.0 Myr.Blatter et al. [2001]
Table 2. Rates and Volumes of Silicic Volcanism
LocationDuration, MyrExtrusive Volume, km3Volume Extrusion Rate Qe, km3 yr−1Mass Extrusion Rate, kg yr−1SiO2 Wt%Petrotectonic SettingNotesReferences
Area < 104km2(Individual Volcanoes/Small Volcanic Fields)
Alban Hills, Italy0.5612905.2E−041.33E+09AContinental arcGeologic map. Some ages from thermoluminescence. Period of eruptions 580 ka to 19 ka. Not corrected for DRE. Unknown amount of erosion.Chiarabba et al. [1997]
Asama0.030371.20E−038.61E+08Aoceanic arc37 ± 7 km3 erupted over past 0.03 MyrCrisp [1984]
Avachinsky, USSR0.0601001.70E−051.62E+08BAcontinental arcRough estimate excluding ejecta beyond cone.Crisp [1984]
Ceboruco-San Pedro0.880.58.05E−52.05E+08Acontinental arcVolume determinations 80.5 ± 3.5 km3 from field mapping, digital topography, and orthophotos. Only minor erosion. Age from numerous 40Ar/39Ar dates.Frey et al. [2004]
Ceboruco-San Pedro0.160.46.04E−41.54E+09Acontinental arcVolume determinations from field mapping, digital topography, and orthophotos. Only minor erosion. Age from numerous 40Ar/39Ar dates.Frey et al. [2004]
Clear Lake, California2.050733.50E−052.81E+0964Continental Volcanic FieldFor period from 2.06–0.01 Ma. Volume includes estimate of eroded material.Crisp [1984]
Coso, California0.42.45.7E−061.34E+07RContinental Volcanic FieldGeologic mapping estimate of 0.9 km3 of basalt and 1.5 km3 of rhyolite erupted over past 0.4 Myr based on K-Ar ages.Bacon [1982]
Davis Mountains, Texas1.515251.0E−032.35E+09RContinental Volcanic FieldDetailed field mapping and 40Ar/39Ar ages from 36.8 to 35.3 Ma. No DRE correction applied, as deposits have low porosity. The actual total volume may be as high as 2135 km3, if buried lava flows over full extent of area suggested.Henry et al. [1994]
Fuji0.011888.00E−034.59E+08BAoceanic arcVolume estimated from detailed field mapping for eruptions over past 11 kyr (tephrachronology)Togashi et al. [1991]
Gran Canaria, Canary Islands0.6002003.30E−047.76E+0869oceanic hot spotDuration is an estimate for the age range 14.1–13.4 Ma, based on K-Ar and 40Ar/39Ar for the Mogan Formation.Crisp [1984], Freundt and Schmincke [1995], Hoernle and Schmincke [1993b]
Hakone0.4001503.70E−044.34E+09Aoceanic arcVolume 150 ± 50 km3 erupted over past 0.4 MyrCrisp [1984]
Kaimondake0.00341.30E−032.04E+10Aoceanic arcVolume 3.7 ± 1.2 km3, eruptions from 4–1 ka.Crisp [1984]
Kaiser Spring, Arizona3.27.52.34E−065.50E+06RContinental Volcanic FieldRough estimate of volumes for rhyolite domes and flows based on field mapping. No estimate for basalt flows. Eruptive activity from 8.8 to 12 Ma based on K-Ar dates.Moyer and Esperanca [1989]
Kuju0.01596.00E−041.53E+09Aoceanic arcVolumes from field mapping and average thickness of flows, and isopach maps of tephras corrected for DRE. Age constraints from 14C and tephrachronology. Rates estimated at 0.4–0.7 km3/kyr.Kamata and Kobayashi [1997]
Long Valley, CA0.767009.21E−042.16E+9RContinental Volcanic FieldDRE-corrected volume from geological mapping. Activity since 0.76 Ma based on 40Ar/39Ar dates.Hildreth [2004], McConnell et al. [1995]
Mazama (Crater Lake), Oregon0.3401012.97E−046.98E+08RContinental arcAge constraints from 14C and K-Ar dates. Volume estimates from field mapping and estimates for erosion are 40–52 km3 for Mt. Mazama precaldera and 51–59 km3 erupted during caldera formation. A detailed chronology of eruptive events.Bacon [1983], Bacon and Lanphere [1990]
Misti0.11276.56.8E−041.73E+09AContinental arcExtensive geologic field mapping and ages from 40Ar/39Ar, plagioclase thermoluminescence, and 14C chronology. Volume of stratocones estimated at 70–83 km3. Not enough information to calculate DRE volume.Thouret et al. [2001]
Mt. Adams0.5203035.8E−041.48E+09Acontinental arcVery detailed assessment of rates as a function of time and type of eruption. Extensive set of whole rock K-Ar ages. Comprehensive geologic mapping with estimates of areas and potential range of pre-erosional thickness for 124 map units. Earliest ages are 940 ka, but most volume was erupted since 520 ka.Hildreth and Lanphere [1994]
Mt Baker1.2901611.25E−047.14E+0960continental arcVery detailed assessment of rates as a function of time from 80 K-Ar and 40Ar/39Ar dates for past 1.29 Myr. Estimate for pre-erosional volume is 161 ± 56 km3 from detailed field mapping.Hildreth et al. [2003a]
Mt Griggs0.292351.21E−045.87E+1154oceanic arcTotal erupted volume estimated from detailed field mapping as 35 ± 5 km3 since 292 ± 11 ka.Hildreth et al. [2003b]
Mt Katmai0.890707.0E−046.38E+1061oceanic arcTotal erupted volume estimated at 70 ± 18 km3 since 89 ± 13 ka, although most volume has erupted since 47 ka.Hildreth et al. [2003b]
Mt Mageik0.093303.3E−047.05E+0962oceanic arcTotal erupted volume estimated from detailed mapping and pre-erosional estimates for the past 93 kyr.Hildreth et al. [2003b]
Mt. St. Helens0.040791.98E−035.04E+09AContinental arcVolumes include estimates of main cone, flank flows, pyroclastic flows, and rough estimate of pre-1980 tephra volumes corrected for DRE.Sherrod and Smith [1990]
Oshima0.025331.50E−031.82E+08Aoceanic arcTotal volume from cone 33 ± 11 km3 over the past 25 ± 15 kyr.Crisp [1984]
Ruapehu, New Zealand0.253001.20E−033.06E+09AContinental arcVolume estimate includes cone (146 km3) plus reworked deposits on ring-plain. Age constraints from 40Ar/39Ar for eruptions since 250 ka.Gamble et al. [2003]
Sakurajima0.014251.80E−032.55E+12Aoceanic arcTotal volume from cone 25 ± 5 km3 over the past 14 ± 1 kyr.Crisp [1984]
Santorini0.067456.72E−041.58E+09ROceanic arcRough estimates of volumes from detailed field mapping and reconstruction Minoan eruption and Skaros shield. Detailed chronology from radiocarbon, K-Ar, and 40Ar/39Ar dates indicates activity since 360 ka, and building Skaros shield began at 67 ± 9 ka.Druitt et al. [1999]
Shiveluch0.20010005.00E−031.12E+09Acontinental arcRough estimate, does not include ejecta beyond cone.Crisp [1984]
Sierra la Primavera0.068345.00E+041.48E+08Rcontinental arcMagma volume based on field mapping of 320 km2 study area, erupted over the period 0.095 ± 0.005 Ma to 0.0275 ± 0.0025 Ma from K-Ar ages.Crisp [1984]
Soufriere Hills – South Soufriere0.174261.5E−043.83E+08Acontinental arcRate given includes DRE correction and assumed submarine deposits. Only minor erosion. Age from numerous 40Ar/39Ar dates indicate eruptions since 174 ± 3 ka.Harford et al. [2002]
Taupo0.027732.76E−036.35E+0974continental arcEruptive volume calculated from detailed stratigraphic record and tephra isopach maps since the volcano's 26.5 ka caldera-forming eruption. Rates constrained by 14C dates.Sutton et al. [2000]
Taupo0.61070441.10E−024.59E+09Rcontinental arcIncludes area of pyroclastic sheets (20,000 km2) from Taupo vents. Careful assessment of volumes from geologic mapping, erupted over the period 1.13 to 0.51 Ma. Ages are K-Ar, 14C, and fission track.Crisp [1984]
Taupo (recent)0.0503507.00E−031.18E+10Rcontinental arcIncludes Taupo and Okatina volcanic centers. Volume determined from geologic maps and tephra isopach maps, corrected for DRE, over past 50 kyr (14C ages).Crisp [1984]
Terceira, Azores0.0235.462.40E−046.48E+0760oceanic hot spotMinimum volume estimated from geological maps and corrected for DRE. Eruptions over past 23 kyr constrained by 14C dates.Crisp [1984]
Tequila, Mexico1.01281.28E−043.26E+08AContinental arcK-Ar dating, detailed geologic mapping, and digital elevation models provide minimum-maximum estimates for 49 eruptive units. Overall volume uncertainty is 128 ± 22 km3.Lewis-Kenedi et al. [2005]
Timber Mountain, Nevada2.840001.43E−033.36E+09RContinental Volcanic FieldVolume from large tuff deposits corrected to DRE. Duration from 12.8 Ma to ∼10 Ma from 40Ar/39Ar sanidine and K-Ar. Volume for other minor volcanism in area not reported.Bindeman and Valley [2003], Farmer et al. [1991]
Tongariro0.275602.2E−045.61E+08Acontinental arcBased on a rough estimate of volume of the volcanic cone and K-Ar ages.Hobden et al. [1999], Hobden et al. [1996]
Trident Volcano0.142221.54E−041.18E+1762oceanic arcTotal eruptive volume of 22 ± 3 km3 since 142 ± 15 ka based on K-Ar dates.Hildreth et al. [2003b]
Tungurahua0.00231.50E−039.18E+1055–65continental arcVolumes from most recent edifice in volcanic complex based on estimates from field mapping and topography, erupting since 2215 ± 90 years ago based on detailed tephrachronology and 14C dates. Volumes include distal tephras and are adjusted to DREHall et al. [1999]
Twin Peaks, UT0.39123.1E−057.29E+0776Continental Volcanic FieldMinimum volume of 12 km3 of silicic magma erupted from 2.74 ± 0.10 Ma to 2.35 ± 0.08 Ma, based on K-Ar ages.Crecraft et al. [1981], Evans et al. [1980]
Valles1.332652.00E−042.59E+10Rcontinental hot spotVolume constraints from mapping. Age constraints from K-Ar dates from 1.43 ± 0.09 to 0.1 MaCrisp [1984]
Volcan San Juan0.034601.78E−031.65E+10Rcontinental arcCombined volumes of main cone and adjacent satellite cone, from topographic since 33.75 ± 1.8 ka. Volume estimates include very detailed isopach maps of tephra and DRE corrections.Luhr [2000]
Yatsugatake0.1715133.00E−033.92E+08Aoceanic arcEstimate of volumes from tephras and rough estimate of volcano volumes. Age constraints from tephrachronology for Younger Yatsugatake, past 171 kyr.Oishi and Suzuki [2004]
 
Area > 104km2(Large Volcanic Fields/Arcs)
Aleutians p/100 km extrusive only3.5003501.00E−045.61E+09Aoceanic arcDetailed study of volumes in the past 3–4 Myr indicate a total of 4700–10000 km3 over an arc length of 2100 kmCrisp [1984]
Aleutians p/100 km int+ext80.0002720003.50E−038.93E+09Aoceanic arcEstimated from 80 Ma age of arc and approximate 34 km3/Myr/km volume excess of the arc over oceanic crustCrisp [1984]
Altiplano-Puna Volcanic Complex1061206.12E−041.56E+09AContinental arcNo uncertainties given. Duration for the past 10 Myr.Francis and Hawkesworth [1994]
Cascade Range p/100 km2.019009.5E−042.42E+09Acontinental arcEstimated for 0–2 Ma Quaternary volcanics in Northern California to southern British Columbia and 8 to 11 km3/km/Myr extrusion rate.Sherrod and Smith [1990]
Central Andes Volcanic Zone p/100 km10.00037403.74E−049.54E+0860continental arcVolume estimated from volcano edifice volumes (1113 total) for the past 10 Myr. No correction for erosion or distal tephras.Francis and Hawkesworth [1994]
Central Andes Volcanic Zone p/100 km1.0001591.59E−044.05E+0860continental arcVolume from edifices active during the past 1 Myr (246 total). Somewhat uncertain; ages based partly on geomorphological evidence.Francis and Hawkesworth [1994]
Central Oregon p/100 km4.000124253.10E−033.09E+0657continental arcVolume estimated from field mapping. K-Ar ages cluster in periods from 16 to 14 Ma and 11 to 9 Ma. Adjusted for length of arc of 200 km.Crisp [1984]
Central Oregon p/100 km14.00050003.60E−023.0E+0762continental arcConstrained as minimum volume from field mapping of Oligocene-Early Miocene deposits with K-Ar ages from 34 to 20 Ma. This estimate has more uncertainty than other entries for Central Oregon. Adjusted for length of arc of 200 km.Crisp [1984]
Chon Aike, South America35.000230006.57E−041.51E+09RContinental riftGeologic mapping and detailed stratigraphy of 9 provinces to derive a volume estimated from areal coverage and thickness of representative units. Eruptive activity between 188 ± 1 Ma and 142 ± 4 Ma constraints from K-ArPankhurst et al. [1998]
East Nicaragua p/100 km0.1351871.38E−03 60continental arcTotal arc length is 130 km. Volume is an estimate of volcano volumes, and is not corrected for erosion or distal tephras. Age constraints from tephrachronology.Patino et al. [2000]
El Salvador p/100 km0.22921.5E−033.52E+0960continental arcVolume estimate of 736 km3 for 252 km arc length for the past 0.2 Myr. Volume is an estimate of volcano volumes, with no correction for erosion or distal tephras.Patino et al. [2000]
Ethiopia4.000281257.00E−038.93E+07Rcontinental riftVolume of 28,125 ± 5625 km3 erupted from 5.5 to 1.5 Ma.Crisp [1984]
Guatemala p/100 km0.0841401.7E−034.34E+0960continental arcVolume estimate of 163 km3 for 116 km arc length. Volume is an estimate of volcano volumes, and is fairly uncertain with no correction for erosion or distal tephras. Age constraints from tephrachronology.Patino et al. [2000]
Japan0.25020208.10E−031.02E+10Roceanic arcMinimum volume estimate for the past 0.25 MyrCrisp [1984]
Japan23.0001094004.76E−039.52E+09Roceanic arcVolume estimated at 109,400 ± 21,900 km3 based on cones of recent volcanoes, flow thickness maps, and average thickness of older volcanic deposits. Duration updated from Crisp [1984] to conform to Neogene timescale.Crisp [1984]
Kamchatka (silicic)0.0808401.10E−021.02E+0964continental arc840 ± 165 km3 erupted over the past 0.08 Myr. Detailed assessment of rates as a function of time and type of volcanism.Crisp [1984], Erlich and Volynets [1979]
Kenya (phonolites)2.500281251.20E−024.95E+08Rcontinental riftVolume (28,125 ± 9375 km3) from geologic mapping and pre-erosional estimates for the past 2.5 Myr, based on K-Ar and 40Ar/39Ar ages.Crisp [1984]
Kenya (silicic)2.50046401.90E−039.44E+08Rcontinental riftVolume (4640 ± 930 km3) from geologic mapping and pre-erosional estimates for 13.4 to 11 Ma, based on K-Ar and 40Ar/39Ar ages.Crisp [1984]
Kenya (silicic)4.500206004.60E−039.64E+09Rcontinental riftVolume (20,600 ± 4100 km3) from geologic mapping and pre-erosional estimates for 7 to 2.5 Ma, based on K-Ar and 40Ar/39Ar ages.Crisp [1984]
Kurile Islands0.0703004.30E−031.90E+1058oceanic arcNot corrected for porosity. No details on uncertainty.Crisp [1984]
Kurile Islands p/100km int+ext83.0003925004.75E−033.32E+0958oceanic arcVery rough estimate for the past 83 Myr.Crisp [1984]
Lesser Antilles p/100 km0.01033.0E−047.65E+0855oceanic arcDetailed volume estimates for individual islands with age constraints from tephrachronology. Volumes adjusted for 100 km of arc length.Wadge [1984]
Lesser Antilles p/100 km0.100404.0E−041.02E+0955oceanic arcDetailed volume estimates for individual islands with age constraints from tephrachronology. Volumes adjusted for 100 km of arc length.Wadge [1984]
Marianas p/100km5.00063701.25E−031.10E+10BAoceanic arcVery rough estimate for an arc length of 620 km, where an estimated 39,500 km3 DRE magma was erupted since 5 Ma. Volumes derived from topography, volcanic deposits, and volcaniclastic sediments.Crisp [1984]
Michoacan-Guanajuato4.00010902.73E−046.95E+0853continental arcVolume of cinder cones and shields for the past 4 Myr, based on K-Ar dates. calculated assuming symmetrical truncated cones and average flow thickness times mapped area of flow, all corrected to DRE based on vesicularity assumptions. Air fall ash volume was assumed 7.73 times the volume of the associated cone. No corrections made for erosion or burial.Hasenaka [1994]
Michoacan-Guanajuato0.0430.57.6E−041.94E+0953continental arcSame as above except volume of young cinder cones only. Volcanism since 40 ka based on constraints from K-Ar and 14C.Hasenaka [1994], Hasenaka and Carmichael [1985]
Mongollon-Datil, New Mexico17.0400002.35E−035.53E+09RContinental volcanic fieldVolume estimates based on area (20,000 km2) times average thickness of deposits (2–3 km). No correction for DRE or erosion. Period of activity estimated as 37–20 Ma, with some minor activity later. Over 50% of volume estimated as rhyolite, with a range of compositions for the remainder.Davis and Hawkesworth [1995]
North Costa Rica p/100 km0.14414.41E−033.19E+1060continental arcTotal arc length is 100 km. Volume estimated from volcano cone volumes, with no correction for erosion or distal tephras. Age constraints from radiometric dates.Patino et al. [2000]
North Queensland, Australia105.00065506.30E−051.12E+09Rcontinental flood basaltPeriod of eruptions was 345 to 240 Ma.Crisp [1984]
San Juan Mountains4.700400008.51E−032.17E+1060Continental Volcanic fieldVolume estimated from field mapping and includes estimate of eroded volume. K-Ar ages span 34.7 Ma to 30 Ma.Lipman [1984]
Sierra Madre Occidental p/100 km15.000489003.26E−033.8E+08Rcontinental riftRough estimates of volumes, from reconstructed areal coverage and thickness from detailed stratigraphy, of main field (393 km3) and surrounding smaller fields over a total arc length of 1200 km. Ages constrained loosely to 38–23 Ma with some smaller eruptions as late as 16 Ma.Aguirre-Diaz and Labarthe-Hernandez [2003]
Twin Peaks-Cove Fort-Black Rock-Mineral Mtn2.320773.30E−056.50E+09RContinental Volcanic FieldVolume 77 ± 4 km3 erupted from 2.74 ± 0.1 to 0.42 ± 0.07 Ma based on K-Ar ages.Crisp [1984]
Whitsunday, Australia371000002.7E−036.19E+09Rcontinental riftVolume is order-of-magnitude estimate based on outcrop and calculated volume of volcaniclastic sediment. Volcanic and intrusive activity occurred from 132 to 95 Ma, but the main phase of activity was 120 and 105 Ma.Bryan et al. [1997]
Yellowstone2.10049802.37E−035.25E+0974continental hot spotDetailed study of 71,000 km2 region from many years of fieldwork, geologic mapping, volcanic stratigraphy breaking out 3 large volcanic cycles. The estimated volume of all known rhyolites is estimated at 4730 km3 and basalts 250 km3. No formal errors are given. Age constraints from an extensive suite of K-Ar and 40Ar/39Ar dates from 2.16 ± 0.04 to 0.070 ± 0.002 Ma.Christiansen [2001]
Zitacuaro-Valle de Bravo p/100 km1.000115.61.16E−042.04E+0960–65continental arcVolume estimated from detailed field mapping of uneroded young volcanics broken out by lava type and corrected to DRE. Total of 57.8 km3 of andesite and dacite was erupted in cones, flows, and tephra. Arc length is 50 km. Extensive suite of dates from 40Ar/39Ar, K-Ar, and radiocarbon. Rates only for volcanics of the past 1.0 Myr.Blatter et al. [2001]
Zitacuaro-Valle de Bravo p/100 km1.00011.41.14E−057.76E+07BAcontinental arcVolume estimated from detailed field mapping of uneroded young volcanics broken out by lava type and corrected to DRE. Total of 5.7 km3 of basaltic andesite was erupted in cones, flows, and tephra. Arc length is 50 km. Extensive suite of dates from 40Ar/39Ar, K-Ar, and radiocarbon. Rates only for volcanics of the past 1.0 Myr.Blatter et al. [2001]

[8] A large amount of uncertainty is associated with inferring volcanic rates from unobserved eruptions. In the tables, a “Notes” field contains information about the methods used to derive the estimates and uncertainties that were available in the original literature, but in many cases no formal uncertainties were reported. Generally the rates reported here should be taken as order-of-magnitude estimates although in some cases the uncertainties may be as small as a factor of two. The extrusive rate often depends on the duration considered; therefore data for one volcanic center measured over different durations are included in Tables 1 and 2. The period of volcanism may also be important since eruptions from further in the past may have experienced more erosion, partial burial, or be more difficult to accurately date.

[9] Sources of error reported in the original publications, as well as most unquantified unreported error, mainly arise from estimating (1) the thickness of the volcanic deposits, (2) the age of lavas, or (3) amount of erosion. Less significant potential sources of error are uncertainty in the conversion from volume to dense rock equivalent (DRE) volume, and uncertainty in the area covered by deposits. One may attribute some of the variance in rates to error introduced by comparing volcanic systems at different scales. For example, the volcanic output rate over continuous lengths of oceanic arcs and ridges is expected to be higher than small individual volcanoes. The arcs and ridges are divided into unit volcano lengths of 100 km based on the spacing of volcanoes in arcs [de Bremond d'Ars et al., 1995]. Petrologic and tectonic factors are also reported for each volcanic system where data are available include lithic type or bulk wt% SiO2 of erupted magma, and petrotectonic setting. Rock names are given for the dominant magma type associated with each area simplified in one of the following categories: basalt, basaltic andesite, andesite, rhyolite. The mode wt% SiO2 reported here is the mode of erupted products by volume reported within the given period for that volcanic system. Petrotectonic setting groups the systems into six categories based on crustal type, oceanic or continental, and association with a plate boundary type; convergent, divergent, or intraplate.

3. Volcanic Rates and Regimes

3.1. Rates of Eruption

[10] Eruption rates are examined on the basis of dominant lithology and petrotectonic setting. Rock type affects many factors related to flow behavior such as viscosity, temperature, and pre-eruptive volatile content. Thus it may be an important control on eruption rate. Petrotectonic setting most strongly reflects the magma generation process, but is also a way to qualitatively look at the effects of crustal thickness.

[11] The effect of magma composition on eruption rate is assessed by broadly grouping the lavas from a volcanic area into one of four categories based on the dominant SiO2 of the reported rock compositions: basalt, basaltic andesite, andesite, or rhyolite. Many important physical properties of lava are a function of SiO2 content, as well as being easy to measure and widely reported, making this a useful tool for first-order comparison of volcanoes. The rock type dominant in the area is assigned as the rock type to represent the entire area. For example, Yellowstone is assigned as rhyolite on the basis of its repeated large caldera-forming eruptions even though a small amount of basalt leaks out between large-volume paroxysmal rhyolite eruptions. Where bimodal volcanism is equally balanced by volume or a change in rock type has occurred at some point in the eruptive history of the volcanic system, the basaltic (Table 1) and silicic (Table 2) volumes and rates of eruption are reported separately. For example, recent Kamchatka eruptions are split into andesites (Table 2) and basalts (Table 1).

[12] Basalts exhibit a wider range of eruption rates than other rock types, ranging from <10−5 km3/yr to >1 km3/yr (Figure 1). Basaltic systems in general show both short-term and long-term changes in eruption rates especially in long-lived systems (e.g., Hawaii [Dvorak and Dzurisin, 1993; Vidal and Bonneville, 2004]). More-silicic rock types, the rhyolites and andesites, have a more limited range of eruption rates than basalts. Long-term rates for silicic eruptions range from <10−5 km3/kyr to 10−2 km3/yr (Table 2 and Figure 1). Among the major rock type groups we have used here, the mean and variance of Qe decreases as the amount of silica increases. In Figure 1, this trend is apparent as the basalts form a wide field of values whose mean is 10−2 km3/yr while andesites and rhyolites form a much narrower band of values around 10−3 km3/yr. The flood basalts form a small cluster of values above 1 km3/yr on Figure 1, outside of a more uniform field of values for all compositions, and seem to form a distinct group. Therefore flood basalts were not considered with the rest of the basalt rates when comparing to other compositions to avoid skewing the results. With flood basalts removed, basaltic eruptions still have an order-of-magnitude higher average rate (2.6 ± 1.0 × 10−2 km3/yr) than basaltic andesites, andesites and rhyolites. Average rates for andesites (2.3 ± 0.8 × 10−3 km3/yr) and rhyolites (4.0 ± 1.4 × 10−3 km3/yr) are also significantly different, although not as distinct as the difference between basalts and these two groups.

[13] The effect of petrotectonic setting on eruption rate is assessed by grouping the volcanoes by the main differences in magma genesis based on plate tectonic theory. In contrast to lithology, petrotectonic setting lends itself to grouping into categories (Figure 2). Volcanoes at convergent plate boundaries are arcs, divergent plate boundaries are rifts or spreading ridges, and intraplate volcanoes are so-called hot spots. Also included is a separate designation of volcanic fields (continental volcanic fields) for areas characterized by areally distributed volcanism of primarily small (<1 km3), monogenetic cones. These fields tend to occur in regions that are difficult to classify by traditional plate tectonic theory such as slab windows (e.g., Clear Lake, CA) or continental extension (e.g., Lunar Crater, NV). In order to also assess the role of crustal thickness/composition, the petrotectonic settings are further subdivided into volcanoes erupting through continental or oceanic crust. The exceptions are oceanic plateaux, the flood basalt equivalent for oceanic crust. Reliable data are so sparse for plateaux that we have grouped oceanic and continental flood basalts in Figure 2.

Figure 2.

Volcanic rates grouped by petrotectonic setting for all locations in Tables 1 and 2. Shaded boxes represent the range of one standard deviation from the mean rate. The black bars show the minimum and maximum rates for each setting. For all settings, mean Qe is skewed toward high values, which may imply a natural upper limit set by magma generation but no lower limit.

[14] Flood basalts have the highest single Qe value and mean Qe of any volcanic system on Earth (Figure 2). In this respect, flood basalts form an exceptional group unlike the other forms of terrestrial volcanism. In contrast, the continental volcanic fields have the lowest single and mean Qe of any group. A very wide range of eruption rates have been reported for oceanic hot spots that overlaps significantly with oceanic arcs and ocean spreading ridges. Although the mean Qe appears higher for oceanic hot spots than other classes of oceanic volcanism, the two-tailed t-test indicates that Qe for all groups of oceanic volcanism are not statistically different. When grouped by petrotectonic setting, Qe from continental areas tend to be lower on average than for oceanic areas, however the range of output rates for any one setting overlaps all other settings (Figure 2). Crisp [1984] noted a similar pattern of higher eruption rates in oceanic settings although found no specific value of crustal thickness that acted as a filter threshold. All of the volcanoes occurring in oceanic settings fail to have statistically different mean Qe and have an overall average of 2.8 ± 0.5 × 10−2 km3/yr. Likewise, all of the volcanoes on continental crust also fail to have statistically different mean Qe and have an overall average of 4.4 ± 0.8 × 10−3 km3/yr, excluding flood basalts. A two-tailed t-test for means indicates that oceanic and continental Qe are statistically different. This implies that crustal thickness, as the overarching contrast between oceanic and continental lithosphere, exerts some control on volcanic rates. Flood basalts also form a distinctive class of volcanism with an average Qe (9 ± 2 × 10−1 km3/yr) two orders of magnitude larger than the range of any other class (Figure 2).

3.2. Intrusive:Extrusive Ratios

[15] The average and range of intrusive:extrusive (I:E) volume ratios for different petrotectonic settings are useful in estimating hidden intrusive volumes at other locations and perhaps on other planets [Greeley and Schneid, 1991]. However, I:E ratios are difficult to estimate and rarely published because the plutonic rocks are either buried or the volcanic rocks are eroded, or the relationship between the volcanic and plutonic rocks is uncertain. Seismic, geodetic, and electromagnetic techniques can reveal the dimensions of molten or partially molten regions under a volcano. Likewise, the sulfur output by magma degassing can be used to estimate the volume of the cooling magma [Allard, 1997]. However, the size of the molten magma reservoir at one time in a longer history may not be a good indicator of the total intrusive volume. Likewise, broad constraints on intrusive volume based on petrologic modeling of the fractional crystallization of a parent basalt are not considered because they will always calculate lower bound on intrusive volume, because such calculations based on extrusive rocks cannot account for strictly intrusive events. Better estimates of total intrusive volume can sometimes be obtained by seismic or gravity measurements of buried plutons. Another way to determine I:E ratios is to compare geographically related volcanic and plutonic sequences. Three such determinations were made in this compilation for the Andes, the Bushveld Complex, and the Challis Volcanic Field-Casto Pluton. However, in each of these cases it is uncertain how well linked extrusive and intrusive rocks are in fact. Despite this uncertainty, we proceed with an analysis if for no other reason than to highlight that this issue has received so little attention.

[16] Previous studies have reported a wide range of I:E ratios from 1:1 to 16:1 [Crisp, 1984; Shaw et al., 1980; Wadge, 1980]. Shaw [1980] hypothesized that the I:E ratio would be higher where crustal thickness is greater, up to 10:1. This makes sense since magma traveling greater path lengths through thicker continental crust has longer to cool and dissipate energy. In addition, mean crustal densities are closer to typical magma densities compared to the mantle (i.e., positive buoyancy forces are likely smaller for magma in the crust compared to magma in the mantle). Subsequently, Wadge [1982] made the argument based on steady state volcanic rates and indirect calculations of intrusive volume that less evolved systems have I:E ratios as low as 1:1.5 for basaltic shields on oceanic crust and up to 1:10 for rhyolite calderas on continental crust. Crisp [1984] presented 14 ratios but did not find any strong connection between magma composition and I:E ratio.

[17] The I:E ratios in this compilation encompass a wide range of values but fails to show any systematic variations with eruptive style, volcanic setting, or total volume (Table 3). While some well-known basaltic shields do have I:E ratios of 1:1 to 2:1, the oceanic ridges have considerably higher ratios of at least 5:1. The range of estimates goes as high as 34:1 at Mount Pinatubo, and 200:1 for the Coso Volcanic Field. Conversely, the I:E ratios at calderas may be much lower than 10:1. Yellowstone has a fairly well constrained I:E ratio of 3:1. Continental magma systems that have had detailed geophysical investigations tend to have magma chamber volume estimates comparable to the total erupted volume, as noted by Marsh [1989]. A ratio of 5:1 could be viewed as common to most magmatic systems when the considerable uncertainty is considered. Ratios higher than 10:1 are uncommon in our data set. When volume of magma involved in crustal “underplating” or magmatic addition to the lower crust is also counted, much higher ratios of intrusive:extrusive activity sometimes result (Ninetyeast Ridge [Frey et al., 2000], Coso [Bacon, 1983]) but other times do not (Aleutians [Kay and Kay, 1985], Marquesas [Caress et al., 1995]).

Table 3. Intrusive:Extrusive Ratios
VolcanoIntrusiveExtrusiveRatioMethodReferences
  • a

    Values that include crustal underplating.

Aleutians1073–1738 km3/km627–985 km3/km1:1–3:1aSeismic and crystallization of Hidden Bay Pluton and related extrusivesKay and Kay [1985]
Bushveld-Rooiberg, South Africa1 × 106 km33 × 105 km33:1Stratigraphic mapping. Cr and incompatible trace element analyses indicate that the total magma volume intruded was approximately 1 × 106 km3.Cawthorn and Walraven [1998], Schweitzer et al. [1997], Twist and French [1983]
Central Andes, Peru9–29 × 104 km32.25 × 104 km33:1–12:1Extrusive from geologic mapping. Intrusive from mapping and gravity.Francis and Hawkesworth [1994], Haederle and Atherton [2002]
Challis Volcanic Field, Idaho3.5 × 103 km34 × 103 −2.8 × 104 km3>1:1–8:1Very uncertain; field and stratigraphic mapping; extrusive converted to DRE using 75% porosity; total plutonic thickness unknownCriss et al. [1984]
Coso Volcanic field, California2.8 km3/Myr (basalt) 5.4 km3/Myr (rhyolite)570 km3/Myr1:200a 1:100aExtrusive from geologic mapping for the past 0.4 Myr; intrusive rate based on current heat flow and estimates of local tectonic extension.Bacon [1983]
East Pacific Rise7 km0.5–0.8 km5:1–8:1Seismic; stratigraphic mapping.Detrick et al. [1993], Harding et al. [1993], Karson [2002]
Etna, Italy (1 Ma)3 × 102 km31 × 102 km33:1Seismic (estimate for ∼0.1 Ma).Allard [1997], Hirn et al. [1991]
Italy (since 1975)0.6 km35.9 km310:1SO2 flux 1975–1995 AD. 
Hawaiian-Emperor Seamount Chain5.9 × 106 km31.1 × 106 km36:1aExtrusive from topographic maps; intrusive from flexural models and seismic, averaged over the past 74 Myr.Bargar and Jackson [1974], Vidal and Bonneville [2004]
Iceland5 km20–40 km4:1–8:1Seismic.Bjarnason et al. [1993], Darbyshire et al. [1998], Menke et al. [1998], Staples et al. [1997]
Kerguelen Archipelago9.9 × 104 km32.75 × 106 km328:1aSeismic.Nicolaysen et al. [2000]
Kilauea, Hawaii9 × 10−2 km3/yr5 × 10−2 km3/yr2:1Drill hole stratigraphy; ground deformation; geologic mapping.Dvorak and Dzurisin [1993], Quane et al. [2000]
Long Valley, California7.6 × 103 km37.5 × 102 km310:1Rough estimate from seismic tomography, stratigraphic mapping, drill holes, and gravity.Hildreth [2004], McConnell et al. [1995], Weiland et al. [1995]
Marquesas Islands6.2 × 105 km33.3 × 105 km32:1aSeismic.Caress et al. [1995]
Mauna Loa, Hawaii8 × 101 km31.1–2.4 × 102 km3>1:1–3:1Stratigraphic mapping, for the 1877–1950 time period.Klein [1982], Lipman [1995]
Mid-Atlantic Ridge5.5–7 km0.5–1.5 km5:1–10:1Seismic.Hooft et al. [2000]
Miyake, Japan4 km31.5 × 10−1 km33:1Geodetic modeling; SO2 emissions.Kumagai et al. [2001]
Mull Volcano, Scotland1.3 × 104 km37.6 × 103 km32:1Stratigraphic mapping.Walker [1993]
Ninetyeast Ridge7–8 km3–4 km2:1Seismic.Grevemeyer et al. [2001], Nicolaysen et al. [2000]
Pinatubo, Philippines60–125 km33.7–5.3 km311:1–34:1Seismic, stratigraphic mapping.Mori et al. [1996], Wolfe and Hoblitt [1996]
San Francisco Mountain, Arizona94 km3500 km36:1Geologic mapping, estimated amount of eroded material included, and seismic low-velocity body with a volume of 300–700 km3.Tanaka et al. [1986]
Twin Peaks, Utah290–430 km340–43 km35:1–9:1Geologic mapping, gravity and thermal modeling.Carrier and Chapman [1981], Crecraft et al. [1981], Evans et al. [1980]
Yellowstone6.5 × 103 km31.89 × 104 km33:1Seismic; stratigraphic mapping.Christiansen and Blank [1972], Clawson et al. [1989], Miller and Smith [1999]

3.3. Repose Time Between Volcanic Events

[18] A major discriminant in the behavior of volcanic systems is their frequency of eruptions through time. Most basaltic volcanoes erupt small volumes of lava frequently whereas continental calderas erupt great volumes of silicic magma infrequently. At Hekla, Thorarinsson and Sigvaldason [1972] noted a positive relationship between repose length and the silica content of the initial lavas erupted following the repose. Data from 17 volcanic centers in Table 4 selected to span a wide range of SiO2 content define an exponential relationship between repose time and SiO2 content in the lava (Figure 3). The volcanic centers in Table 4 were chosen to span a range of SiO2 compositions for sequences of at least three eruptions.

Figure 3.

Repose interval between the end of one eruption and the start of the next, and range of SiO2 content of lavas for locations in Table 4. Error bars represent the high and low values of the data. The points represent the mean repose interval and the middle of the SiO2 range. The solid line represents the best fit to a least squares regression for an exponential equation which yields trepose = 10−6* exp(X/2.78). The e-folding factor of 2.78 indicates that repose time increases by a factor of ∼3 for each ∼3 wt% increase in silica.

Table 4. Repose Times at Selected Volcanic Centers
VolcanoRepose Time Avg, yearsRepose Min, yearsRepose Max, yearsNumber of Reposes in Recordwt% SiO2 minwt% SiO2 maxReferencesNotes
Colima804813835661Luhr and Carmichael [1980]Four cycles of activity ending with ash flow eruptions since 1576 AD.
Etna40.1100704750Tanguy [1979], Wadge [1977]Constrained by historical records from 1536 to 2001 AD.
Fogo, Cape Verde20194274042Doucelance et al. [2003], Trusdell et al. [1995]Constrained by historical records from 1500 to 1995 AD.
Fuego10010150604955Martin and Rose [1981]Constrained by historical records since 1500 AD; eruptions occur in clusters of activity.
Izu-Oshima6813190235357Koyama and Hayakawa [1996]Detailed syncaldera and postcaldera eruptive history from tephra and loess stratigraphy; reposes since caldera formation.
Katla461380204650Larsen [2000]Last 11 centuries; constrained by historical records.
Kilauea0.80.110464850Klein [1982]Constrained by historical records from 1918 to 1979 AD.
Mauna Loa50.120344850Klein [1982]Constrained by historical records from 1843 to 1984 AD.
Mt Adams1500005000032000035764Hildreth and Fierstein [1997], Hildreth and Lanphere [1994]Major cone building episodes since 500 ka.
Mt St Helens860050001500066367Doukas [1990], Mullineaux [1996]From 40 ka to present, major eruptive cycles only.
Ruapehu30000100006000055565Gamble et al. [2003]Constrained by 40Ar/39Ar ages.
Santorini300001700040000125871Druitt et al. [1999]For major explosive volcanism since 360 ka. Both 40Ar/39Ar and K-Ar ages for older units, radiocarbon ages for younger.
Taupo2000206000287276Sutton et al. [2000]Post-Oruanui eruptions from 26.5 ka to present.
Toba37500034000043000036877Chesner and Rose [1991]Reposes between tuff-forming eruptions since 0.8 Ma.
Valles33500032000035000036975Doell et al. [1968], Heiken et al. [1990]Reposes based on eruption of Bandelier and pre-Bandelier tuff, and collapse of Toledo and Valles calderas.
Yatsugatake32000100008500055363Kaneoka et al. [1980], Oishi and Suzuki [2004]Plinian eruptions since 0.2 Ma. Tephrochonology and radiocarbon ages.
Yellowstone70000060000080000037579Christiansen [2001]Considers major tuff-forming eruptions.

[19] The minimum, maximum, and mean repose time for an eruption sequence is presented along with the minimum and maximum SiO2 content for the corresponding suite of compositions erupted from a “single” center. Repose time is determined by the interval between the end of one eruption and the start of the next. Measuring repose time is somewhat subjective because what may count as a repose at one volcano may not be considered as a repose elsewhere. Closely observed volcanoes (e.g., Etna or Kilauea) have reposes reported on a scale of days but on older or more silicic volcanoes (e.g., Santorini or St. Helens) have their eruptive periods divided into major eruptive units separated by thousands of years. We have tried to determine repose period as the length of time between eruptions of a characteristic size for that volcano. For example, at Santorini reposes between the Kameni dome-forming eruptions are much shorter than the major ashfall eruptions [Druitt et al., 1999]. This example also highlights the potential for bias toward the Recent with shorter repose times for smaller eruptions that are not preserved in the long-term geologic record. For these reasons, the reposes between major eruptions are considers whereas the “leaking” of minor volumes of lava between major eruptions is not considered in this study.

[20] The exponential relationship between SiO2 content and repose time is mainly determined by basaltic shields and rhyolite calderas. For volcanoes in the andesite-dacite range, the data jump from short repose intervals to longer repose at ∼60% SiO2 (Figure 3). While composition is unlikely to be the exclusive control on repose time, more error is likely to emerge in the 60–70% SiO2 range due to difficulties in dating the eruptions of complex stratocones, the dominant constructional volcanic morphology for intermediate compositions. Measuring the repose periods at stratocones and calderas requires high resolution stratigraphy and precise ages over several millennia to smooth out the short-timescale volume/frequency relationship [Wadge, 1982]. These data are very limited but are becoming more available recently with improvements in geochronological methods [Hildreth et al., 2003a]. If the maximum SiO2 in the system controls the repose period then the fit parameter of the exponential equation improves slightly (R2 = 0.69 to 0.73).

[21] There are several reasons to expect repose time to increase as silica increases. Direct melting of mantle produces basaltic compositions, and more evolved compositions require time for fractional crystallization and assimilation. Higher silica compositions also have greater melt viscosity, requiring additional excess pressure to erupt [Rubin, 1995] and, in that sense, are far less mobile. More viscous magmas are more likely to suffer “thermal death” compared to less viscous magmas. A few studies have already pointed out a positive correlation between eruptive volume and repose interval [Cary et al., 1995; Klein, 1982; Wadge, 1982]. The magma storage time, based on rock geochronometers from crystal ages and from crystal size distribution analysis (CSD), tends to increase exponentially as SiO2 and stored magma volume increase [Hawkesworth et al., 2004; Reid, 2003]. These observations are all consistent with the idea that longer magma storage times allow time for that, in turn, results in longer repose periods associated with higher silica content magmas.

4. Discussion

4.1. Upwelling and Magma Production Rate Limits

[22] Factors that might influence volcanic rates and intrusive:extrusive ratios are local crustal thickness, tectonic setting (magnitude and orientation of principal stresses), magma composition, and melt generation rate in the source region. For 170 examples, long-term volcanic output rate varies from 10−5 to 1 km3/yr. Only flood basalts attain the highest Qe, above 10−1 km3/yr, while various volcanoes with the lowest measured Qe, below 10−5 km3/yr, seem to have very little in common (Figure 1). Tectonic setting, but not magma composition, affects volcanic rates. Continental crust reduces the average Qe to 4.4 × 10−3 km3/yr from 2.8 × 10−2 for oceanic crust.

[23] The output rates all show a strong skewness with long tails toward low Qe values suggesting that an upper limit may exist (Figure 2). Furthermore, although there is essentially no lower limit to volcanic rates in that magma supplied from depth may intrude but never erupt, or dribble out slowly, this is not usually the case. Most volcanoes have a Qe above 1 × 10−3 km3/yr. This result was also found empirically by Smith [1979] and Crisp [1984]. Hardee [1982] derives a simple analytic solution showing that this critical Qe of ∼10−3 km3/yr represents a “thermal threshold” where magmatic heat from the intrusion tends to keep a conduit open and begin formation of a magma chamber. We infer that long-term volcanism is unlikely to occur without an open magma conduit to supply and focused melt delivery. This threshold value is dependent on intrusive rate, not volcanic output rate. The I:E ratios found are somewhat lower than the often cited 10:1 ratio, and suggest that an I:E ratio of ∼5:1 may be regarded as a better average value. Nevertheless, this suggests that, using the Qe values present here as data for the Hardee [1982] model, virtually all of the volcanic systems in Tables 1 and 2 meet the requirements for conduit wall rock meltback and magma chamber formation.

[24] It is perhaps surprising that given the large differences in eruptive style and melt generation mechanisms (e.g., isentropic decompression, triggering by metasomatic introduction of volatiles or mafic magma underplating) in different tectonic settings an aggregate view of volcanic rates exhibits such a small range of variation, by and large. The similarity of the rates leads us to speculate that a magma upwelling rate limit is set within the mantle at a value near 1 km3/yr, with magma generation being subject to greater variances based on the local composition of the mantle being melted. In this view, flood basalts represent systems with low I:E ratios and form when a large fraction of mantle-generated magma reaches the surface. The upper limit on magma generation may be controlled by the subsolidus upwelling rate within the upper mantle of 0.01–0.1 m/yr, and this may explain the upper limit of magma generation due to isentropic decompression [Asimow, 2002; Verhoogen, 1954].

4.2. Openness of Magmatic Systems

[25] The volcanic output rate and repose periods between eruptions gives us some basic constraints on the behavior of magma systems as open or closed systems. We have noted the empirical correlation of repose period and magma silica content. That is, a repose interval can be roughly predicted on the basis of either mean or maximum SiO2 wt% of the eruptive composition. What constraints can be put on storage time in volcanic systems from purely thermodynamic considerations?

[26] A volcanic system can be crudely modeled as a magma storage zone in the crust and a volcanic pile at the surface (Figure 4). Four processes affect the volume of magma in the storage reservoir or magma chamber: eruption (Qe) and solidification (Qs) remove magma from the system, while recharge (QR) and crustal assimilation (QA) add magma to the system. When a volcano acts as a closed system (one that receives no input of mass or heat via advected hot magma) all of the magma erupted remains molten for the duration of volcanic activity under consideration. In such a system, crystallization can occur due to the loss of heat or volatiles from the magma body to its colder surroundings but the extent of crystallization must be insufficient to preclude eruption. One way to approach this problem is to assume that volcanoes act as closed systems during repose periods between eruptions and treat each eruption as the result an isolated batch of magma supplied by recharge in a single event and stored until eruption.

Figure 4.

Cartoon of a simplified volcanic system representing storage, and the processes affecting the volume of magma available for eruption. At some depth below the volcano, a volume of magma is stored in a liquid/crystal mush magma chamber. Inputs to the system are by recharge, a function of the magma upwelling rate, and assimilation of host rock. Outputs are by eruption or solidification of the magma by cooling within the magma chamber. A closed volcanic system in this context is one that receives no input.

[27] Simple heat transfer considerations based on Stefan cooling of magma permit a first-order test of the hypothesis that a volcanic system is a closed system. If we know the eruption rate (Tables 1 and 2), and assume a closed system with respect to mass and heat recharge, the magma in storage will solidify at a rate specified by Stefan cooling. Using data for volcanic output rate of individual eruptions and repose time between eruptions gathered for several volcanic centers at a wide range of eruptive compositions, a simple 1-D Stefan cooling model [Carslaw and Jaeger, 1959] can be applied to estimate solidification times t (years) in a spherical magma volume of V (km3)

display math

where κ is the thermal diffusivity, λ is the solution to the transcendental equation

display math

where L is the latent heat of fusion (J kg−1), cp is the isobaric specific heat capacity (J kg−1 K−1), and ΔT is the temperature difference between the ambient external temperature and the liquidus of the melt phase. The thermal diffusivity is calculated as

display math

where ρ is magma density (kg m−3) and K is magma thermal conductivity (J/kg m s). Values for the various constants are taken from Spera [2000] for gabbro, granodiorite, and granite melts. This very basic approach permits a first-order look at the issue of cooling as a constraint on magma system longevity and openness. Heat calculations for lens or sill-like geometries alter the results by a factor of 2–4 [Fedotov, 1982]. Consideration of hydrothermal cooling would tend to enhance cooling rates so that the lifetime of a given volume of magma presented here is always an upper limit on cooling times. A more complex model is not justified given the order-of-magnitude estimates used as input.

[28] If we consider that volcanoes act as closed systems only between two successive eruptions, the solution to the 1-D Stefan Problem described above allows us to examine the thermal viability of the volcanic system given the repose period and the volume of magma involved (Figure 5). A closed system, in this context, means that one batch of magma is intruded at some time and stored until the eruption. Thus a maximum “storage time” for a batch of magma in the shallow plumbing system of a volcano can be estimated (Figure 6). The solidification time is determined as the time for a volume of magma to completely solidify as calculated from equation (1). The volume of magma is assumed to be five times the DRE volume of the eruption following the repose period based on the average I:E ratio from the data in Table 3. The assumption of complete solidification puts an upper limit on the time necessary to cool the magma enough to prevent eruption.

Figure 5.

Cartoon depicting a time sequence of a simplified volcano that is closed to mass and advected heat between individual eruptions. The arrows indicate mass inputs and outputs. Eruptible magma is represented by the white oval, the lath pattern is cooling and crystallizing magma, and the stipple is country rock. Time t0 shows the preexisting conditions, while the sequence begins with the eruption at t1 which removes the eruptible magma from the magma chamber. Recharge occurs at t2. Cooling during the storage period, shown in t3, is the interval between recharge and eruption (t2 − t4). There must be enough magma left at t4 to equal the known volume of eruption. The repose period, as calculated for Figure 6, is the interval t1 − t4. This model assumes that magma is fed into the system in isolated batches, as discussed in the text.

Figure 6.

Openness of selected volcanic centers with well-constrained eruptive volumes and repose intervals based on simple 1-D Stefan analysis. Each point, color coded by volcano, represents the repose interval between preceding an eruption and the solidification time for the erupted volume to completely crystallize before erupting. The volume of magma in storage is taken from the intrusive:extrusive ratio in Table 4 or assumed to be 5:1 if unavailable. The colored lines represent cutoff values for the amount of time magma may spend cooling and crystallizing in storage compared to the repose period. Points that plot below the line demand thermodynamically open systems that experience magma recharge prior to eruption. Points above the line may be closed in the sense that multiple eruptions could come from the same batch of magma without additional input. Note that this does not require that these volcanoes act as closed systems.

[29] Only a handful of volcanoes have been studied well enough to be able to estimate both volume and timing of eruptions over many eruptive cycles. The long, detailed records of eruptions at Mauna Loa [Klein, 1982] and Etna [Tanguy, 1979; Wadge, 1977] are used as examples of basaltic volcanoes, and the regular eruptive pattern at Izu-Oshima for the past 103 years [Koyama and Hayakawa, 1996; Nakamura, 1964] makes the volumes of individual eruptions more clear. Toba [Chesner and Rose, 1991] and Yellowstone [Christiansen, 2001] are two calderas with a high quality record of multiple major eruptions. A few other examples from volcanoes with shorter, but still well-documented, records are also used with data from sources cited in Table 2.

[30] Whether the magma would solidify, and thus require the volcano to be an open system, depends on the magma storage time. Estimates of magma storage times from various crystal-age geochronometers are available at a range of volcanic centers and suggest that magma storage period, like repose, is a function of silica content of the magma [see Reid, 2003, and references therein]. Storage time from crystal ages for basaltic systems are generally longer or equal to repose, while storage times for andesites and rhyolite systems are slightly shorter than or equal to repose. On the basis of this information, we can draw a set of lines for different fractions of storage to repose time representing the limits for volcanoes that may be thermally closed systems between eruptions (Figure 6).

[31] The repose time between eruptions at large calderas (Yellowstone, Long Valley, and Toba) can be more than 10 times greater than the storage time and the volcanoes are still required to be open systems in this analysis (Figure 6). The basaltic systems (Etna, Mauna Loa, and Oshima) are required to be open systems in this analysis only if magma is stored more than 10–100 times longer than the repose period (Figure 6). A few outliers for Etna with extremely short eruption reposes arguably may be the same eruption, but it is easy to see why these might be from “closed” systems on the timescales presented.

4.3. Heat Flux Associated With Magma Transport

[32] Rates of magmatism may be translated into excess heat flows for specific magmatic provinces to obtain estimates of advected heat via magmatism at regional scales over magmatic province timescales. For mafic eruption rate Qe and an I:E ratio of ℜ, the volumetric rate of magma flow into the crust is ℜQe. The excess heat power H (J yr−1) associated with magma transport from mantle to crust is

display math

where ΔT is the temperature difference between the magma and local crust, L is the enthalpy of crystallization (250–400 kJ/kg dependent on magma composition), ρ is magma density, cp is the isobaric heat capacity of the magma, and Tliquidus − Tsolidus is the liquidus to solidus temperature interval.

[33] As an example, consider the Skye subprovince of the British Tertiary Igneous Province (BTIP). For the estimated volume eruption rate of 2 × 10−3 km3/yr averaged over ∼1600 km2 area of Skye, the average excess heat flow is ∼3.5 × 107 J/m2/yr (1.1 W/m2). This excess heat flux is more than an order of magnitude greater than the average terrestrial global heat flux 0.09 W m−2. These estimates are consistent with a crustal thickening rate of ∼5 km/My and a background (regional) heat flux of 10–15 times the global average during 60–53 Ma. We conclude that the volume flux of magma in the active years of this part of the BTIP focused heat flow about an order of magnitude above the background at the regional scale for ∼5 Ma. The regional energy/mass balance estimate appears consistent with inferences drawn from geochemical modeling that point to significant magma recharge during magmatic evolution at Skye [Fowler et al., 2004].

[34] The excess heat power divided by the area affected by volcanism can be compared to the average terrestrial heat flux to the area. The heat power into the crust due to magmatism is therefore approximately 1017 J/yr for an overall average eruption rate taken from Table 1 of 10−2 km3/yr for ∼1000 km2 of arc or ridge and I:E ratio of 5. Thus typical values for the “average” magmatic system, 101 W/m2, exceed the global terrestrial background value of 10−1 W/m2 by two orders of magnitude.

5. Conclusions

[35] The 170 long-term estimates of volcanic output rate compiled from literature references from 1962–2004 corroborate much of the previously published information about magmatic systems but also reveal a few surprises. Long-term volcanic rates are higher for basaltic volcanoes than andesitic and rhyolitic volcanoes taken as a group. Oceanic hot spots, arcs, and ridges have an average volcanic output rate of 10−2 km3/yr while continental arcs and hot spots have an average output rate of 10−3 km3/yr, implying that thinner crust/lithosphere is associated with higher volcanic rates on average but not systematically.

[36] For the small number of volcanic systems where adequate data exist (Table 3), the I:E ratio is most commonly less than 10:1 with 2–3:1 being the most commonly occurring value, and a median value of 5:1. On the basis of the data compiled here, there is little indication that composition is strongly or systematically associated with I:E ratio. We conclude only that further work needs to be done on this important topic.

[37] In contrast, composition and repose period between eruptions (end to next start) are strongly linked. We found that an exponential relationship between repose period and silica content of the magma provides a satisfactory fit to the data.

[38] Purely on the basis of thermal considerations, volcanic systems must be open to recharge of magma between individual eruptions, except for the most frequently erupting basaltic volcanoes. The fact that basaltic systems are indeed open magmatic systems can be demonstrated by other means [e.g., Davidson et al., 1988; Gamble et al., 1999; Hildreth et al., 1986].

Acknowledgments

[39] The authors would like to thank Arwen Vidal, Yanhua Anderson, and Joesph Goings for tracking down some of the data that went into the tables. Some of the work that went into this paper was carried out at and for the Jet Propulsion Laboratory, California Institute of Technology, sponsored by the National Aeronautics and Space Administration. Support from NASA, NSF, and the DOE for magma transport research at UCSB is gratefully acknowledged. We thank M. R. Reid, C. R. Bacon, and R. S. J. Sparks for their very thorough and thoughtful reviews.

Ancillary