volcanology references

Voight, 1988

likelihood of eruption can be evaluated by the idea that volcanic eruptions are always accompanied by rock failure

Voight, 1988

assumes volcanology obeys the laws of rock mechanics

Pyle et al., 2013

we can use remote sensing to monitor volcanoes, predict paths of lahars and PDCs and monitor thermal disturbances

Pyle et al., 2013

dense vegetation cover and water vapor makes remote sensing of volcanoes at the tropics difficult. remote sensing required ground truthing

Pritchard et al., 2018

unrest can be evaluated as magnetic or hydrothermal based on gravity change, ground deformation, temperature, seismicity, gas flux and hydrothermal activity

Pritchard et al., 2018

it is often not possible to distinguish between magnetic and hydrothermal unrest

Girona et al., 2021

statistically determined that heat flux before an eruption could not be created by random variation

Girona et al., 2021

they only use 5 volcanoes and assume they are representative of all different types of volcanoes. only able to analyse volcanoes that have erupted only once or twice in the last decade

Biggs and pritchard, 2017

classic volcano deformation suggests a magma chamber gradually inflates prior to eruption, but this is an over simplification. different deformation shapes show different processes

Biggs and pritchard, 2017

we cannot detect deformation if inflation is too deep or very small. we need to consider mass balance rather than volume

Parks et al., 2012

The volume of lava erupted from nea kameni is directly related to the time elapsed since the last eruption. Seismicity and unrest in 2011 was caused by the influx of ¼ of the magma required for an eruption.

Kilburn et al., 2023

the last eruption of campi flegrei occurred in 1538 after 100 years of inflation. there has been new inflation occurring since 1950. in 1984 the extension required for tensile failure was 100cm, now it is only 37cm meaning rupture can happen at a lower strain threshold

Kilburn et al., 2023

they assume youngs modulus remains unchanged through time they also assume the corrected uplift and crustal strain rate correlate and that this is a consistent relationship through time

Hickey et al., 2022

the last eruption of Soufriere hills, Montserrat was in 2010. we have had 14 years of unrest without an eruption. 94% of dome building eruptions are less than 20 years long

Dempsey et al., 2020

Used machine learning to recognise patterns in tremor data that were seen before an eruption at Whakaari to create early warnings

Dempsey et al., 2020

false alarms and missed eruptions, not a perfect system so public might not trust. not good at recognising unusual patterns that havent been seen before. need training data

Camejo-Harry, 2024

used zoning of olivine crystals to determine how long magma takes to accumulate. when a zoned crystal sits in a magma system, diffusion evens out the zoning which is stopped on eruption. the state of diffusion can tell us how long the crystal sat in the magma reservoir

Camejo-Harry, 2024

the 2 models used to model the diffusion had different estimates for the diffusion time, many differences were small but some were on the order of hundreds of days. we can only do this in reterospect

Yeo et al., 2022

used satellite imagery to view the growth of temporary islands and water disturbances caused by small submarine eruptions of lake’iki, tonga. used sea floor sampling to show the magma chemistry hasnt changed in 50 years but the lava dome is different

Santana-Casiano et al., 2023

during submarine eruption of El Hierro the ocean pH decreased to <6 due to emissions of CO2, SO2, and H2S.volcanoes release lots of dissolved species which act like fertiliser allowing for recovery

Preine et al., 2024

took 4 cores in the santorini caldera to find an additional 2km3 of the 726CE pumice when previously we could only see <1m layer, despite reports of explosions and a pumice raft seen 400km away. bulk volume estimate of 3.1km3 making it VEI5

Preine et al., 2024

can only give minimum volume estimates because cant trace the deposit under kameni volcano or see the size of the pumice raft

Yeo et al., 2024

VEI3 eruption of volcano F with no warning creating a pumice raft that inundated fiji and reached austrailia. navigation disruption, destruction of boats and coastal infrastructure

Yeo et al., 2024

cloud cover limited the back tracking of pumice rat to volcano F

Carn et al., 2022

VEI 5.7 eruption of Hunga tonga after 2 months of surtseyan eruptions. the plume produced was larger than pinatubo but lower in SO2 because of submarine scrubbing

Carn et al., 2022

~35% error across all sensors. unknown how much water-magna interaction there was. didnt look at the pH changes in the water

Seabrook et al., 2023

partial collapse of eruption column during hunga tonga eruption led to large submarine PDCs and bathymetric changes, damage to sea floor cables and sea floor life. created a simple model to predict where PDCs would go which turned out to be very accurate. tsunamis throughout pacific

Seabrook et al., 2023

model is very simple and doesnt take into account material addition and eruption magnitude, meaning it may not be applicable to other eruptions

joseph et al., 2022

predicting and responding to the eruptive transition of 2021-2022 eruption of la soufriere. made a model looking at seismic, deformation, gas monitoring and dome growth. used a combination of communication methods to reach people

Joseph et al., 2022

hard to confidently attribute signals to changing behaviour

Sigmundsson et al., 2024

massive flow of magma into dikes can be established with only modest overpressure in extensional regimes with reference to grindavik dike

Sigmundsson et al., 2024

dike flow rate is 2 orders of magnitude higher than fagradalsfjall despite same tectonic setting as grindavik dike. assumptions are made to simplify the model so crustal and magma complexities are not considered.

anderson et al., 2024

basaltic fissure eruption at kilauea 2018 led to caldera collapse because of crater lake drainage. stepwise caldera collapse led to pressurisation of magma chamber, sustaining the fissure eruptions

anderson et al., 2024

gaps in data at the start of the eruptions so some interpretations are based off limited data. simplifications in the modelling of the magma system

Galetto et al., 2022

we can predict magma inflow rate from deformation. if we know reservoir size, we can predict whether a dike will propagate. fast inflow and it will propagate in a year, slow and it either wont propagate or will take longer

galetto et al., 2022

only looked at very few volcanic systems

crozier et al., 2023

explosive eruption at kilauea 2018 was due to abrupt subsidence of reservoir roof, pressurising pocket of gas at the top of the reservoir

crozier et al., 2023

the magma chamber and gas pocket size are not well constrained and they are not sure the gas pocket is even there, but the mechanism doesnt work if there is no gas pocket

caricchi et al., 2021

eruptions are triggered by internal and external processes. External: landslides, earthquakes, tides, stress changes, loading/unloading and rainfall. internal: magma recharge, volatile exsolution. magma is transported through fluid fracturing

caricchi et al., 2021

dike propagation is sensitive to regional and local scale heterogeneities

Cassidy et al., 2018

intra-eruption transitions are driven by conduit processes: changes from open to closed degassing, conduit geometry. inter-eruption transitions are driven by magma reservoir processes (injection of new magma, mixing, cooling, crystalisation, second boiling)

Cassidy et al., 2018

complex interactions between deep crustal and shallow processes makes predicting eruptive transitions challenging

Sparks, 1981

at kilauea, the probability of an eruption in a week centred on a tidal maximum is twice of this for a week on a tidal minimum

Sparks, 1981

weak correlation. correlation doesnt hold when looking at mauna loa. some volcanoes dont have a long enough record to see whether tidal forces have any impact

Sparks, 2003

precursory signals indicating magma movement do not always lead to an eruption. timescales of eruption build up can very from hours to years which makes eruptions difficult to predict

Sparks, 2003

external triggers further complicate predictions

Weber and Sheldrake, 2022

volcanoes that have produced caldera forming eruptions have larger geochemical variability and silica content. this is because they have large magma bodies that leads to more fractionation. large chemical variability can be used as an indicator of likelihood of caldera forming eruption

Weber and Sheldrake, 2022

the volcanoes they look at are not evenly distributed globally, focusing on the americas. not backed up any of these findings with geophysics to show if there are large magma bodies

Popa et al., 2021

Magma stored at <4km is more likely to be explosive. Magmas with very low or very high water contents will be effusive. Deeper storage allows for more water. magmas with 4-5.5wt% water with <30vol% crystals will be explosive. >40% crystals will be effusive in dome forming.

Popa et al., 2021

focus on andesites and rhyolites. the analysis is restricted to volcanoes with storage regions around 2kbar to avoid variations in water saturation due to pressure, but this limits the applicability of their depth predictions. they classify all dome forming eruptions as effusive even if an explosive phase comes first

Wandsworth et al., 2022

2011-2012 eruption of Cord’on Caulle produced effusive and explosive eruptions simultaneously which cannot occur under the classic ascent rate model. Rocks fragment at depth, as they rise to low pressures they fuse together to form a plug that erupts effusively. Cracks form in the plug, through which explosive eruptions occur.

Wandsworth et al., 2022

new model is based on one eruption of one volcano. where multiple explanations can be made for the data, they chose the one that fits their model best and dont really explain why

Jorgensen et al., 2024

looked at clinopyroxene zoning to understand magma storage and evolution. low viscosity melt can accumulate at depth because of high amounts of exsolved volatiles. explosive eruptions can be driven by the expansion of these volatiles even though the magma is low viscosity

Jorgensen et al., 2024

a lot of the high temperature data comes from a single apatite crystal. this style of eruption with these conditions hasnt been seen at other volcanoes

Cassidy et al., 2019

explosive eruption of Kelud, Indonesia in 2014 after a 7 year period of repose after dome building in 2007. explosive eruptions = cooler water saturated magma. dome building = hotter, water undersaturated magmas. little warning because triggered by top-down volatile overpressure

Burton et al., 2013

CO2 emitted from volcanoes comes from 3 different sources: CO2 dissolved in the mantle, recycled co2 from subducted crust and decarbonation of shallow crustal material. isotopes help us distinguish between them

Burton et al., 2013

only consider co2 for carbon emissions despite other carbon species being released. extrapolation from a limited set of volcanoes.

Gerlach, 2011

anthropogenic co2 emissions are 135 times higher than volcanic emissions. The belief that volcanic CO2 exceeds anthropogenic co2 implies either unbelievable volumes of magma production or unbelievable concentrations of co2 in the magma.

Gerlach, 2011

they assume constant volcanic co2 emissions over the last 100 years and that they are able to accurately constrain submarine degassing

Suarez et al., 2019

studies from past atmospheric co2 content from LIPs and bolide impacts help us to identify tipping points that will mean anthropogenic emissions lead to irreversible damage

Suarez et al., 2019

paper focuses on catastrophic events and overlooks the cumulative effects of smaller scale perturbations on the climate cycle. earths early history has been erased by plate tectonics

Aiuppa et al., 2021

CO2/S ratios can be used to differentiate between hydrothermal and magmatic degassing. higher carbon content in magma = deeper storage depth

Aiuppa et al., 2021

only used 12 volcanoes with a very specific set of characteristics. in their plot, the depth trend is actually quite unclear and many of the volcanoes don’t fit

Hunt et al., 2017

re-evaluation of the passive degassing from the main Ethiopian rift by measuring degassing from background flux, fault zones and geothermal fields. this was then extrapolated to the whole rift to find that passive degassing was significantly lower than previous estimates

Hunt et al., 2017

implies a constant, stable flux of co2. biogenic co2 degassing was deemed negligible, so ignored. lots of extrapolation

Black et al., 2019

reevaluation of CO2 flux from LIPs because the CO2 exsolved before melt inclusion entrapment cannot reliably be reconstructed. can also look at trace element ratios for elements that are similarly incompatible as carbon

Black et al., 2019

cannot say for certain that modern analogues are good for LIPs. assumes degassing from LIPs was constant. dont give any chemical comparison to see if these lips are similar to modern analogues

cashman et al., 2017

melt is heterogeneously distributed in transcrustal magma systems. TCMS evidenced by low-velocity electrically conductive zones. crystal zoning records different crystallisation conditions and temperature and pressure/depth

cashman et al., 2017

inferred that their data is representative of all magmatic systems even though they utilise only a few samples from sporadic locations. dont explore the role of tectonic setting on magma system

Lundstrom and Glazner, 2016

challenges the typical magma chamber concept because of lots of mismatch between different geophysical observations and chemical measurements

Lundstrom and Glazner, 2016

geophysical monitoring only captures a snapshot of the subsurface

Sparks et al., 2019

melt accumulates in low pressure settings, rheological and density boundaries and in weaknesses or discontinuities. zoned crystals and disequilibrium textures are evidence of magma recharge

Sparks et al., 2019

do not consider the role of plate tectonic setting. they don’t state which volcanic systems they are getting any of their information from.

Latypov et al., 2022

from the bushveld complex we can see that large magma chambers have existed in the past. the thickness of pyroxene layer shows the magma chamber would have been atleast 5km thick

Latypov et al., 2022

the model of magma accumulation in a single pulse seems unlikely. the model they produce has non-unique solutions. hard to date layers because happen too quickly for radiogenic dating.

Keller et al., 2015

rift volcanics have similar chemistry to MORB reflecting anhydrous decompression melting. plutonic magmas are rich in MgO and are hydrated which is why they freeze in the subsurface

Keller et al., 2015

only looked at very young samples

Huber et al., 2012

identified 2 groups of ignimbrites: crystal rich and crystal poor. crystal rich magmas are highly viscous so require heating to unlock them prior to eruption, mixing them and making them homogenous. crystal poor magmas are not locked so dont require heating and mixing so are heterogenous.

Huber et al., 2012

dont really define how the rocks are homogenous. don’t discuss whether the crystals found in the ignimbrites formed on ascent or in the magma chamber.

Gualda et al., 2012

Rhyolite:MELTS thermodynamic model adaptation of MELTS to incorporate rhyolitic composition

Gualda et al., 2012

doesnt really consider volatiles. decreased accuracy at >2GPa. conditions are based on only one sample. works best when hydrous phases do not play a critical role in magmatic evolution

Riel et al., 2022

MAGEMin. gibbs energy minimisation to focus on the effects of water and other volatiles

Riel et al., 2022

computationally intensive. dependent on the quality of the thermodynamic database. assumes equilibrium conditions

Holland et al., 2018

KNCFMASHTOCr. used experimental and theoretical data to emphasise the interactions between different mineral phases

Holland et al., 2018

limited consideration for volatiles. simplified melt structure. below 10kbar, treats aqueous fluid as pure water which may not be accurate.

Hegde et al., 2025

Outlines how volcanic SO2 affects the atmosphere and climate systems and how it affects feedback loops

Hegde et al., 2025

Uses an idealised one dimensional model, so isn’t necessarily an accurate representation of real-world conditions.

Autumn et al., 2025

The findings suggest that the volcanic systems of Santorini and Kolumbo operate independently despite their geographical proximity

Autumn et al., 2025

They exclude the crust above 8km when looking for magma storage. The exact methods of magma recharge beneath Santorini still remains uncertain.

Romano et al., 2025

Machine learning techniques can enhance insights into volcanic activities and often surpass classical methods, machine learning automates tasks and uncovers hidden patterns in complex datasets.

Romano et al., 2025

There are challenges with working with noisy datasets and real-time decision making.

Kennedy et al., 2025

Emphasises the importance of indigenous perspectives into volcanology learning to enhance engagement and understanding of volcanic risks in new zealand.

Ogden et al., 2025

Columnar and insitu gas measurements during the 2021 eruptions of la palma using solar absorption Fourier Transform Infrared (FTIR) spectroscopy. It highlights the value of solar absorption measurements for volcanology, atmospheric research, and air-quality monitoring during eruptions. The good agreement between different measurement techniques and retrieval products

Ogden et al., 2025

Study focuses on clear-sky conditions for spectroscopic measurements which may not be representative. Retrieval of gas concentrations relies on mixing ratio profiles and so are not direct measurements.

Schiavon, 2024 (phd thesis)

Violent explosions at stromboli are driven by fast magma ascent and rapid water exsolution, as revealed by textural parameterization of plagioclase crystals.

Schiavon, 2024 (phd thesis)

focuses on petrological data without consideration for geophysics

Coppola et al., 2025

The study successfully monitored the Fernandina eruption using satellite infrared data, providing rapid estimations of key parameters . It demonstrates the value of combining different satellite sensor types with varying spatial and temporal resolutions for a comprehensive understanding of effusive eruptions

Coppola et al., 2025

The accuracy of the derived parameters depends on the quality and resolution of the satellite data and the algorithms used for processing. Focuses on the effusive phase.

Özsoy, 2025 (PhD thesis)

Detailed tephrostratigraphy of pyroclastic sequences over last 450kyr, outlining their distribution, characteristics, and ages. This has contributed to a more comprehensive and detailed record of explosive volcanism within the Ulukışla Caldera

Özsoy, 2025 (PhD thesis)

The study relies on the assumption that the identified pyroclastic deposits originated from the Ulukışla Caldera. The interpretation of eruption dynamics is based on the characteristics of the preserved deposits.