Skip to main content
U.S. flag

An official website of the United States government

Filters

Learn More

The map displays volcanoes, earthquakes, monitoring instruments, and past lava flows.

Get Help

Earthquake Plots
To view an earthquake cross-sectional plot, click "Draw Plot" to start drawing on the map. Once the plot area is selected, click on "View Plot" to view chart or "Clear Plot" to start over.

Results
Filters

Learn More

The map displays volcanoes, earthquakes, monitoring instruments, and past lava flows.

Get Help

Earthquake Plots
To view an earthquake cross-sectional plot, click "Draw Plot" to start drawing on the map. Once the plot area is selected, click on "View Plot" to view chart or "Clear Plot" to start over.

Settings

x

Map Legend

x

Map Legend

Results

Volcano Hazards Program

Find U.S. Volcano

There are about 170 potentially active volcanoes in the U.S. The mission of the USGS Volcano Hazards Program is to enhance public safety and minimize social and economic disruption from volcanic unrest and eruption through our National Volcano Early Warning System. We deliver forecasts, warnings, and information about volcano hazards based on a scientific understanding of volcanic behavior.

Learn more about NVEWS

Volcano Activity Notifications

Ash-rich plume rises out of Halemaʻumaʻu Crater, Kilauea Volcano Hawaiʻi

USGS Volcano Observatories release regular updates and notifications to communicate increases or decreases in volcanic activity and explain unusual or hazardous circumstances. 

Get Latest Updates

Volcano Hazards Assessments

Mount Rainier seen from Puyallup, Washington

Long-term hazards assessments identify locations at risk from volcanic hazards. They are released as high-resolution hazards-zonation maps with accompanying explanations, which help guide eruption preparedness plans.

View Assessments

U.S. Volcano Observatories

Volcano Observatory staff monitor, research, and issue formal notices of activity for volcanoes in assigned geographic areas. Scientists also assess volcano hazards and work with communities to prepare for volcanic eruptions.

Learn More

News

Volcano Watch — The 1969 Maunaulu eruption: 12 lava fountaining episodes

Volcano Watch — The 1969 Maunaulu eruption: 12 lava fountaining episodes

Photo & Video Chronology — January 12, 2026 — Kīlauea episode 40

Photo & Video Chronology — January 12, 2026 — Kīlauea episode 40

Volcano Watch — Hau’oli Makahiki Hou: a round-up of fireworks from Kīlauea’s fountains

Volcano Watch — Hau’oli Makahiki Hou: a round-up of fireworks from Kīlauea’s fountains

View All

Publications

The magmatic-hydrothermal system of the Three Sisters volcanic cluster, Oregon, imaged from field gravity measurements The magmatic-hydrothermal system of the Three Sisters volcanic cluster, Oregon, imaged from field gravity measurements

From 2019 to 2024, gravity surveys were conducted at the Three Sisters volcanic cluster (TSVC), measuring 246 gravity sites using a spring relative gravimeter. We calculated the residual Bouguer anomaly and identified three main zones with negative anomalies, ranging from −4 to −8 mGal, located southwest and west of South Sister, within an area that has been uplifting for the past two...
Authors
Helene Le Mevel, Nathan Lee Andersen, Annika E. Dechert, Josef Dufek
By
Volcano Hazards Program, Volcano Science Center

The anatomy and lethality of the Siberian Traps large igneous province The anatomy and lethality of the Siberian Traps large igneous province

Emplacement of the Siberian Traps large igneous province (LIP) around 252 Ma coincided with the most profound environmental disruption of the past 500 million years. The enormous volume of the Siberian Traps, its ability to generate greenhouse gases and other volatiles, and a temporal coincidence with extinction all suggest a causal link. Patterns of marine and terrestrial extinction...
Authors
Seth D. Burgess, Benjamin A. Black
By
Volcano Hazards Program, Volcano Science Center

Mitigation of human cognitive bias in volcanic eruption forecasting Mitigation of human cognitive bias in volcanic eruption forecasting

Modern operational eruption forecasting methods rely heavily on human judgment in the face of uncertainty and are thus susceptible to myriad cognitive biases and errors by the scientist-forecasters. Recent developments in the behavioral sciences have elucidated cognitive biases across a wide spectrum of human behaviors and found ways to mitigate them. These insights have led to...
Authors
Heather M. Wright, J. D. Pesicek, Stephen A. Spiller
By
Volcano Hazards Program, Volcano Science Center
View All
Was this page helpful?

The Formation of Silicate-Melt Inclusions

Types of Inclusions

The common terminology for categorizing aqueous or carbonic fluid inclusions becomes unwieldy when applied to silicate MI, particularly those in volcanic rocks. For example, most non-silicate fluid inclusions are described as: 1) primary, 2) secondary, and 3) pseudosecondary (Roedder 1984). Primary inclusions contain any phase present at the time of crystal growth. Secondary inclusions contain phases that enter crystals along fractures (after primary crystal growth has ceased) and then are trapped as the fractures heal. Pseudosecondary inclusions are also trapped along fractures, but before the crystal has ceased growing at its periphery. To most igneous petrologists, both primary and pseudosecondary MI are "primary" in the sense that they are trapped within growing crystals and yield information on the composition of the silicate liquid during its evolution. Under some circumstances (such as magma mixing), there may be more than one primary melt from which the host crystal grew. It is difficult to envision how secondary MI may be trapped in volcanic phenocrysts (given the high viscosity of silicate melts), though Pasteris et al. (accepted) convincingly showed how non-silicate fluids can be trapped as secondary inclusions, particularly during fracturing events associated with magma ascent. Metasomatic secondary MI are commonly described in xenoliths found in volcanic ejecta (e.g., Schiano et al. 1994) and can potentially form in plutonic environments if a silicate melt infiltrates a previously crystallized rock.

This review concerns itself with primary silicate MI, inasmuch as the term "primary" indicates that the crystals were still growing from silicate melt at the time the MI were trapped. Though most MI contain only one phase at the time of entrapment (silicate liquid), during cooling, that phase may unmix to form a vapor bubble and daughter crystals (Sorby 1858; Roedder 1984). Another class of inclusions that will be discussed are called mixed MI, which form by entrapment of more than one phase besides the silicate melt (Roedder 1984). The other phases may include microphenocrysts or fluids, (used here to denote any volatile phase such as a low-density vapor or hypersaline liquid).

 

Entrapment Mechanisms

Roedder (1984: Chapter 2) and Sobolev & Kostyuk (1975) list a variety of mechanisms by which primary inclusions may form, including; 1) non-uniform supply of nutrients to the crystal face, resulting in skeletal growth, 2) undercooling, also resulting in skeletal growth, 3) formation of reentrants in the crystal during resorption events, followed by subsequent crystal growth, and 4) wetting by a separate immiscible phase (e.g., molten sulfide or a vapor bubble) that creates irregularities in crystal growth, resulting in entrapment of that phase as well as the melt.

Though primary MI are found in nearly all volcanic rocks, they are not present within every individual crystal. Some crystals may contain scores of inclusions, implying that irregularities of phenocryst growth may control inclusion distribution. Figure 2a-c shows secondary-electron images of b-form quartz crystals from the high-silica rhyolite erupted in 1912 at the Valley of Ten Thousand Smokes, Alaska (Clocchiatti 1975). The images illustrate the common occurrence of 'gaps' or 'hoppers' in the bipyramidal faces. Comparison with a thin-section photomicrograph from Lemmlein (1930; Fig. 2d) illustrates how this common semi-skeletal growth phenomenon (MacLellan & Trembath 1991) could result in the formation of reentrants and hence form MI in volcanic quartz.

Figure 2. Photographs of quartz and feldspar crystals from silicic volcanic rocks demonstrate typical growth forms that trap MI.

Secondary electron image of a well-formed bipyramidal b-form quartz phenocryst from the 1912 eruption at the Valley of Ten Thousand Smokes, Alaska (from Lowenstern 1993)

Secondary electron image of quartz phenocrysts from the the 1912 eruption at the VTTS, displaying semi-skeletal growth in which interiors of crystal faces grow more slowly than edges. This results in gaps filled by silicate melt that may later be trapped as inclusions. This results in gaps filled by silicate melt that may later be trapped as inclusions. Field is 1 mm across. From Clocchiatti (1975).

Secondary electron image of quartz phenocryst from the the 1912 eruption at the Valley of 10,000 Smokes, Alaska, displaying semi-skeletal growth in which interiors of crystal faces grow more slowly than edges. This results in gaps filled by silicate melt that may later be trapped as inclusions. This results in gaps filled by silicate melt that may later be trapped as inclusions. Field is 1 mm across. From Clocchiatti (1975).

Transmitted-light photomicrograph of a section parallel to the c -axis through the center of a quartz phenocryst from a liparite, showing skeletal growth pattern (adapted from Lemmlein 1930), as might be expected for cystals in previous two photos.

Transmitted-light photomicrograph of plagioclase from pumice fallout deposited during the climactic eruption of Mt. Mazama/ Crater Lake (from Bacon et al.; 1992). The dark MI are along zonal arrays, just outside of patchy zones that may be formed during resorption events. The elongate MI is 70 x 130 µm.

Transmitted-light photomicrograph of plagioclase from pumice fallout deposited during the climactic eruption of Mt. Mazama/ Crater Lake (from Bacon et al.; 1992). The dark MI are along zonal arrays, just outside of patchy zones that may be formed during resorption events. The MI is 90 x 120 µm.

Plagioclase and pyroxene tend to trap MI in zonal arrays, permitting some studies on the sequence of inclusion formation. Bacon et al. (1992) studied MI in plagioclase and noted that large inclusions were commonly found adjacent to (outside of) areas of patchy zoning; these were interpreted to have formed during resorption (Figs. 2e, f). This arrangement seems to imply that, at least for plagioclase, resorbed crystal surfaces are favorable substrates for inclusion entrapment during subsequent growth events. Watson (1976) found that inclusion-rich zones correspond to sharp compositional gradients in the host plagioclase, implying that a considerable variation in either temperature or melt composition (e.g., H2O concentration), or both, enhanced crystal growth and inclusion entrapment.

Leaked or 'hourglass ' Inclusions

Occasionally, the host crystal does not completely enclose the included melt, and the system remains open during cooling and decompression. As shown in Figure 3, such inclusions remain connected to the outside of the host by a narrow capillary filled with either silicate melt or a vapor phase. Hourglass inclusions fill the continuum between completely enclosed MI and reentrants in phenocrysts (Anderson 1991). By comparison of the various parts to Figure 3, it becomes apparent that petrographic examination of a sample in 3-dimensions is required to assess whether an MI was (or still is) an hourglass inclusion (e.g., Fig. 3d, e). In addition, one can see that growth processes resulting in hourglass inclusions and reentrants can be confused with resorption textures.

Hourglass inclusions seem to form by the same sort of crystal-growth process that results in the gaps seen in Figure 2c and 2d. The capillary remains open if a pressure gradient keeps melt or vapor flowing through it during ascent of the magma (Anderson 1991). Because populations of these "leaked" inclusions may be variably degassed, they present a unique opportunity to constrain the timing and kinetics of gas exsolution during eruptive ascent. This was first recognized by Anderson (1991) who gave a detailed discussion of 'hourglass' inclusions and showed that they allow one to characterize the degassing and ascent history of quenched volcanic rocks. For other types of studies, however, hourglass inclusions should be avoided, as they do not retain volatile concentrations representative of the melt phase at the time of entrapment.

FIG. 3. Transmitted-light photomicrographs of hourglass inclusions and reentrants in quartz.

 

 

(A) Crystallized hourglass inclusion from the Pine Grove porphyry, Utah. Small capillary connects the bulk of the inclusion with the outside of the host crystal. (B) Hourglass inclusions form a continuum with simple reentrants, as illustrated by three hourglass/reentrants (arrows) in quartz grain from Pine Grove porphyry. (C) Glassy hourglass inclusion (i) in quartz (q) contains a large bubble (v) grown at inclusion end of capillary (from 1912 rhyolite erupted at Valley of 10,000 Smokes, Alaska; Lowenstern 1993).

 

(D) Hourglass MI from the Valenza unit at Pantelleria (Lowenstern & Mahood 1991). Large bubble within remelted MI sits at end of long, thin capillary (arrows) that extends to crystal surface. Bright spots below bubble are other MI, out of focus beneath the inclusion of interest. Bright spot northeast of bubble is a refractory quartz bleb (Lowenstern 1994a).

 

 

(E) Closeup of inclusion, bubble and capillary described above in (D).

Does the Melt Inclusion Trap Representative Liquid?

One process of concern to researchers is the formation of boundary layers adjacent to growing crystals (Watson et al. 1982; Bacon 1989). Such layers can contain non-representative melt compositions because of the slow diffusion rates of incompatible melt components away from the crystal interface. If such a melt boundary-layer were trapped as an inclusion, it would provide misleading information about melt compositions.

Many studies have tried to assess the affect of boundary-layer phenomena. Most have noted the similar major-element compositions of MI and matrix glass and have concluded that the inclusions represent melt from which the crystals grew (Lowenstern & Mahood 1991; Dunbar & Hervig 1992b; Bacon et al. 1992). Sometimes, trace-element concentrations vary significantly between inclusions and associated matrix glass. For example, both Cl and Ti concentrations in melt inclusions from the Lower Bandelier tuff vary by a factor of three (Dunbar & Hervig 1992a). The matrix glass contains lower amounts of Cl, yet higher Ti than MI, a relationship inconsistent with entrapment of enriched boundary layers, which should contain higher concentrations of both elements. Lu et al. (1992) found that U and La were negatively correlated within a group of quartz-hosted MI from the Bishop tuff, consistent with crystal fractionation, but not enrichment in boundary layers (see also Anderson 1991). If immobile trace elements are not enriched in MI, it is unlikely that faster-diffusing species such as H2O and CO2 are greatly affected by boundary-layer phenomena. In summary, most workers have concluded that diffusion-related gradients such as boundary layers have little effect on analyzed populations of MI, and that the effects of boundary layers can be avoided by by studying inclusions larger than about 25 µm (Anderson 1974a).

Volcano Hazards Program | U.S. Geological Survey Skip to main content
U.S. flag

An official website of the United States government

Filters

Learn More

The map displays volcanoes, earthquakes, monitoring instruments, and past lava flows.

Get Help

Earthquake Plots
To view an earthquake cross-sectional plot, click "Draw Plot" to start drawing on the map. Once the plot area is selected, click on "View Plot" to view chart or "Clear Plot" to start over.

Results
Filters

Learn More

The map displays volcanoes, earthquakes, monitoring instruments, and past lava flows.

Get Help

Earthquake Plots
To view an earthquake cross-sectional plot, click "Draw Plot" to start drawing on the map. Once the plot area is selected, click on "View Plot" to view chart or "Clear Plot" to start over.

Settings

x

Map Legend

x

Map Legend

Results

Volcano Hazards Program

Find U.S. Volcano

There are about 170 potentially active volcanoes in the U.S. The mission of the USGS Volcano Hazards Program is to enhance public safety and minimize social and economic disruption from volcanic unrest and eruption through our National Volcano Early Warning System. We deliver forecasts, warnings, and information about volcano hazards based on a scientific understanding of volcanic behavior.

Learn more about NVEWS

Volcano Activity Notifications

Ash-rich plume rises out of Halemaʻumaʻu Crater, Kilauea Volcano Hawaiʻi

USGS Volcano Observatories release regular updates and notifications to communicate increases or decreases in volcanic activity and explain unusual or hazardous circumstances. 

Get Latest Updates

Volcano Hazards Assessments

Mount Rainier seen from Puyallup, Washington

Long-term hazards assessments identify locations at risk from volcanic hazards. They are released as high-resolution hazards-zonation maps with accompanying explanations, which help guide eruption preparedness plans.

View Assessments

U.S. Volcano Observatories

Volcano Observatory staff monitor, research, and issue formal notices of activity for volcanoes in assigned geographic areas. Scientists also assess volcano hazards and work with communities to prepare for volcanic eruptions.

Learn More

News

Volcano Watch — The 1969 Maunaulu eruption: 12 lava fountaining episodes

Volcano Watch — The 1969 Maunaulu eruption: 12 lava fountaining episodes

Photo & Video Chronology — January 12, 2026 — Kīlauea episode 40

Photo & Video Chronology — January 12, 2026 — Kīlauea episode 40

Volcano Watch — Hau’oli Makahiki Hou: a round-up of fireworks from Kīlauea’s fountains

Volcano Watch — Hau’oli Makahiki Hou: a round-up of fireworks from Kīlauea’s fountains

View All

Publications

The magmatic-hydrothermal system of the Three Sisters volcanic cluster, Oregon, imaged from field gravity measurements The magmatic-hydrothermal system of the Three Sisters volcanic cluster, Oregon, imaged from field gravity measurements

From 2019 to 2024, gravity surveys were conducted at the Three Sisters volcanic cluster (TSVC), measuring 246 gravity sites using a spring relative gravimeter. We calculated the residual Bouguer anomaly and identified three main zones with negative anomalies, ranging from −4 to −8 mGal, located southwest and west of South Sister, within an area that has been uplifting for the past two...
Authors
Helene Le Mevel, Nathan Lee Andersen, Annika E. Dechert, Josef Dufek
By
Volcano Hazards Program, Volcano Science Center

The anatomy and lethality of the Siberian Traps large igneous province The anatomy and lethality of the Siberian Traps large igneous province

Emplacement of the Siberian Traps large igneous province (LIP) around 252 Ma coincided with the most profound environmental disruption of the past 500 million years. The enormous volume of the Siberian Traps, its ability to generate greenhouse gases and other volatiles, and a temporal coincidence with extinction all suggest a causal link. Patterns of marine and terrestrial extinction...
Authors
Seth D. Burgess, Benjamin A. Black
By
Volcano Hazards Program, Volcano Science Center

Mitigation of human cognitive bias in volcanic eruption forecasting Mitigation of human cognitive bias in volcanic eruption forecasting

Modern operational eruption forecasting methods rely heavily on human judgment in the face of uncertainty and are thus susceptible to myriad cognitive biases and errors by the scientist-forecasters. Recent developments in the behavioral sciences have elucidated cognitive biases across a wide spectrum of human behaviors and found ways to mitigate them. These insights have led to...
Authors
Heather M. Wright, J. D. Pesicek, Stephen A. Spiller
By
Volcano Hazards Program, Volcano Science Center
View All
Was this page helpful?