Series of inflation-deflation cycles have occurred during 2020-2024 in the center of the Svartsengi volcanic system, SW-Iceland. Since 27 October 2023, continuous inflation has been interrupted by deflation periods when nine dike injections and seven eruptions have occurred from 10 November 2023 to 8 December 2024 at the Sundhn & uacute;kur crater row and its extension. Extensive observations of ground deformation using GNSS (Global Navigation Satellite System) geodesy and interferometric analysis of synthetic aperture satellite (InSAR) images is here used to improve understanding of the dynamics of magma accumulation and transfer, both prior to and during repeated rifting events. Joint inversions of the GNSS and InSAR data, considering a deformation source within a uniform elastic half-space, infer pressure changes at about 4-5 km depth near the regional brittle-ductile boundary, with inflow causing volume increase rates of 2.4-9 m(3)/s. Geodetic modelling using GNSS has been undertaken in near real-time throughout the events, using deformation sources in fixed locations inferred in earlier joint inversions. The deflation periods began rapidly when a dike propagated from the eastern edge of the magma accumulation area. The estimated volume of dikes is in the range (1-133) x 10(6) m(3), with the first event being by far the largest and longest (similar to 15 km). Geodetic observations have contributed to success in forecasting diking/eruption onset in the medium and short term, using the expectation that a correlation exists between volume loss in the magma domain during a deflation event and subsequent volume recharge to the system before the next event is triggered.

The Surtsey Island was built from 130 m water depth during the period 1963–1967. Two tephra cones formed above sea level reaching a height of 170 m a.s.l. The final surface area of Surtsey reached 2.65 km2 with a volume of 0.8 km3 (dense rock equivalent). Repeated levelling and Global Position System (GPS) campaigns have been carried out every 5–10 years to monitor the internal deformation and subsidence of the island. The first levelling measurements were made in the summer of 1967. Surtsey follows an exponentially decaying subsidence curve, with total subsidence reaching 1 m since 1967. The 2000–2023 GPS surveys confirmed that Surtsey moves horizontally with the Eurasian plate. The average subsidence rate for the three GPS monitoring sites during 2000-2023 was 3.8 mm/yr, a total of 9 cm. About 2/3 of the total subsidence is contributed by compaction of sedimentary and volcaniclastic material, thermal contraction, and palagonization. Up to 1/3 of the total subsidence can be explained by loading of the island on the asthenosphere.

Following an increase in seismic activity in December 2019, a pressure increase began in the center of the Svartsengi volcanic system in January 2020, as inferred from geodetic observations. The first diking event occurred, however, in the nearby Fagradalsfjall volcanic system, about 10 km east of Svartsengi, 24 February – 19 March 2021, when an ~9 km long dike gradually formed with geodetically inferred initial volume increase rates up to 35 m3/s, during the first week of diking. The total dike volume was ~34 Mm3, based on joint inversions of InSAR and GNSS observations that have been extensively used to study this and later events in the area. This dike intrusion culminated in an eruption on 19 March 2021. The initial dike had minor incremental volume increase in association with opening of additional vents above the dike during the 6-month-long 2021 eruption, with near-surface opening in the top few hundred meters. Three additional dike intrusions occurred in the Fagradalsfjall area between December 2021 to July 2023, with initial magma flow rates between 22 to 70 m3/s. The Fagradalsfjall dikes were fed through a channel with an inferred cross-sectional area of about ~2-4 m2, passing through the lower crust from a source near the crust-mantle boundary, with a geodetically imaged deflation source at about ~12-13 km depth. Since late 2023, activity has been focused at the Svartsengi system, with 9 diking events and 7 eruptions in 2023-24. Initial diking there occurred on 10-11 November 2023 with inferred peak flow rates of ~7400 m3/s when an ~15 km long dike formed, following magma accumulation near the brittle-ductile boundary at about 4-5 km depth. The inferred cross-sectional area of the limiting part of the channel from the Svartsengi magma domain feeding the zone where dikes have formed in 2023-24 is on the order of ~2000 m2 or about 2-3 orders of magnitude larger than that inferred at Fagradalsfjall. This and the different depth of magma storage in the plumbing systems at Fagradalsfjall and Svartsengi explains their different behaviour in recent years, that are coupled in such a manner that only one of of the systems has been primarily magmatically active at each time since 2020.

A 15 km long dike formed rapidly in the Reykjanes Peninsula oblique rift on 10 November 2023 and propagated under the town of Grindavík. From just before noon on 10 November until midnight, around 25 MW≥4 earthquakes occurred, two of which were of MW~5.2. Three-dimensional ground deformation is well resolved both temporally and spatially with dense Global Navigation Satellite System (GNSS) geodetic observations, which record cumulative displacements up to about 80 cm occurring mostly over 6 hours in the evening of 10 November and continuing at much reduced rates in the following days. Interferometric analysis of synthetic aperture radar images using Sentinel-1, COSMO-SkyMed, and ICEYE satellites records also well the dike deformation, which occurred simultaneously with deflation over the nearby central part of the Svartsengi volcanic system. Geodetic modelling, assuming uniform elastic host rock behavior, infers a dike volume of (130-139)×106 m3, with up to ~8 m dike opening, as well as some strike-slip shear motion. Deflation at Svartsengi in our model is best fit using a spherical point source with a volume decrease of (76-82)×106 m3up until 12 November. The temporal evolution of the dike opening was further modelled using hourly GNSS displacements, allowing better derivation of the temporal evolution of the flow rate into the dike and the contraction volume of the subsidence source. The maximum flow rate into the dike is inferred to be ~9500 m3/s, between 18:00 and 19:00 on November 10. We infer that the massive magma flow into the dike was established with only modest overpressure in the feeding magma body, a sufficiently large pathway opening at the boundary of the magma body, and pre-failure lowering of pressure along the pathway that had occurred through gradual build-up of high tensile stress over the previous eight centuries. This explains the unprecedented fast maximum magma flow rates that we infer. Such high flow rates provide insight into the formation of giant dike swarms under conditions of high tensile stress, and imply a high hazard potential for dike intrusions, considering their potential to transition into eruptions.

Neighboring volcanic systems sometimes show evidence of some form of interconnection, for example by inflating or deflating either in phase or in anti-phase. We review here the course of events on the Reykjanes Peninsula (RP) in the ongoing unrest since approximately 2020, using volcano geodesy. There are several volcanic systems on the RP, from west to east: Reykjanes, Svartsengi, Fagradalsfjall, Krýsuvík, Brennisteinsfjöll, and Hengill, with Fagradalsfjall being the least developed. All these volcanic systems, except Brennisteinsfjöll, have shown signs of unrest in the past years and decades: Three uplift episodes occurred at Svartsengi during 2020; one or two subtle deformation events further west on Reykjanes in 2020, and further uplift episodes at Svartsengi in May 2022, October 2023, November 2023, and December 2023 - January 2024. Inflation was observed at Krýsuvík during the summer of 2020; and a M5.6 earthquake occurred in Krýsuvík in October 2020. As of beginning of 2024, three eruptions have occurred at Fagradalsfjall (in 2021, 2022, and 2023) and one eruption at Svartsengi in December 2023. Each eruption has been preceded by a dike intrusion, often intertwined with complex patterns of faulting, near-surface fracturing over wide areas, and creep along segments of the plate boundary. Additional dike intrusions in December 2021 in Fagradalsfjall and in Svartsengi in November 2023 did not breach the surface. The dike growth has spanned timescales of just over an hour to several weeks; furthermore, small dikelets accompanied new vent openings during the 2021 eruption. The dikes were emplaced in the brittle crust, above ~8 km depth, spanned several decimeters to meters in thickness, and released locally a great amount of plate-tectonic stresses. Re-inflation following each eruption or dike intrusion is usually observed, however, the temporal style of uplift rates varies considerably from time to time. Co-eruptive deflation was observed during the 2021 Fagradalsfjall eruption and the 2023 Svartsengi eruption. The detailed deformation observations and modeling for the unrest periods reveal complex interactions of tectonics and magmatism across several volcanic systems on the RP. During 2020-2024, localized deformation and seismicity have alternated between different volcanic systems on the RP, such that only one system is inflating or erupting at a time. This observation may be interpreted in terms of deep pressure coupling between the systems. Furthermore, the deformation events cause significant stress changes at neighboring volcanic systems, affecting the probability of dike propagation and seismicity as well as conditions for magma accumulation.

Abstract Unrest began in July 2021 at Askja volcano in the Northern Volcanic Zone (NVZ) of Iceland. Its most recent eruption, in 1961, was predominantly effusive and produced ∼0.1 km3 lava field. The last plinian eruption at Askja occurred in 1875. Geodetic measurements between 1983 and 2021 detail subsidence of Askja, decaying in an exponential manner. At the end of July 2021, inflation was detected at Askja volcano, from GNSS observations and Sentinel‐1 interferograms. The inflationary episode can be divided into two periods from the onset of inflation until September 2023. An initial period until 20 September 2021 when geodetic models suggest transfer of magma (or magmatic fluids) from within the shallowest part of the magmatic system (comprising an inflating and deflating source), potentially involving silicic magma. A following period when one source of pressure increase at shallow depth can explain the observations.

Tectonic controls on dyke emplacements, eruption dynamics and locations have been observed in multiple volcanic areas worldwide. Mapping of active structures is therefore key for assessing potential tectonic and volcanic hazards in active regions. We used wrapped interferograms from the TerraSAR-X satellite to map active fracture movements over a 2-year period of a volcano-tectonic unrest at the onshore Reykjanes Peninsula plate boundary in SW Iceland. As of 1 December 2023, the unrest has included at least six inflation events and five dyke injections resulting in three eruptions of the Fagradalsfjall volcanic segment. In addition to the deformation associated with the 2019–2021 inflation events and intrusions, the interferograms reveal fracture movements over a wide area surrounding the active plate boundary segment. This first-order mapping of active fractures complements previously mapped structures, as InSAR allows for the detection of subtle ground movements, even in areas where young lava flows cover older structures. Our fracture data therefore fill in some of the apparent voids in previous fracture and fault maps of SW Iceland. Furthermore, our investigation reveals aseismic movement on previously unknown fractures directly beneath the town of Grindavík, as well as a N45 ^∘ E striking fracture co-located with the longest lasting volcanic vent of the subsequent 2021 eruption. The mapping method we present in this study is relevant for active volcano-tectonic regions where InSAR can be applied to detect small-scale fracture movements to advance understanding of ongoing unrest and volcano-tectonic hazards.

Many examples of exposed giant dike swarms can be found where lateral magma flow has exceeded hundreds of kilometers. We show that massive magma flow into dikes can be established with only modest overpressure in a magma body if a large enough pathway opens at its boundary and gradual buildup of high tensile stress has occurred along the dike pathway prior to the onset of diking. This explains rapid initial magma flow rates, modeled up to about 7400 cubic meters per second into a dike ~15-kilometers long, which propagated under the town of Grindavík, Southwest Iceland, in November 2023. Such high flow rates provide insight into the formation of major dikes and imply a serious hazard potential for high–flow rate intrusions that propagate to the surface and transition into eruptions.

<p>Precursors to volcanic eruptions vary widely between volcanic systems and their individual eruptions. Volcanic systems in Iceland undergoing unrest include the Reykjanes, Svartsengi, Fagradalsfjall, and Krísuvík systems on the obliquely spreading Reykjanes Peninsula. Main precursors prior to the Fagradalsfjall eruptions in 2021 and 2022 were signals associated with the formation of dikes releasing stored tectonic stress over weeks and days, respectively. If volcanic activity occurs at Fagradalsfjall in coming years it may be associated with shorter warning time, as less stored tectonic stress remains. In contrast, the nearby Svartsengi system experienced cumulative uplift of about 15 cm in multiple inflation episodes during 2020 to 2022, modeled as repeating sill intrusions. Prior to, in-between, and following the intrusive events, the surface subsided. We find that the onset of diking accompanied by a sudden increase in seismicity and deformation rates is a likely scenario prior to future eruptions on the Reykjanes Peninsula. A decline in seismicity and/or deformation may occur as unrest activity progresses, as experienced prior to the 2021 and 2022 eruptions. In other areas of Iceland, since 2020 magma storage areas with increasing pressure have been identified at the Askja, Grímsvötn, Krafla, and Bárðarbunga calderas, as well as at Hekla volcano. Increasing pressure buildup in the roots of these volcanoes, is expected to a varying degree prior to next eruption, with different amounts of inflation and seismicity. Tectonic stress release as observed during the 2014/15 Bárðarbunga rifting event may occur or not. The largest capacity for pressure increase is expected at the Askja caldera, where the surface over the magma chamber subsided by more than 1 m from 1983 to 2021, but since August 2021 over 45 cm of uplift has occurred and deformation continues. The amount of subsidence prior to present uplift may indicate the scale of further inflation needed to reach critical conditions, assuming that the current inflation is sourced in a similar crustal volume as the deflation, and the strength of the surrounding material remains similar (e.g., no new faulting/fracturing). Examples of intermittent flow of magma to shallow depth, or pressure increase beneath calderas, occurred during 2017-2018 at Öræfajökull, where a slight increase in seismicity has been detected in recent months, and inflation 2018-2019 at Torfajökull caldera. It remains a challenge to promptly identify seismic swarms that may be indicative of formation of magma feeding conduits versus those indicating intermittent increases in seismic activity due to high stress levels, e.g., caused magma recharging, changes in geothermal activity, or glacial retreat. Experience from the Northern Volcanic Zone and the Reykjanes Peninsula oblique rift, suggest precursory activity may take place simultaneously over wide parts of plate boundary areas, indicating to some extent coupled activity of nearby volcanic systems.</p>

Following two periods of dike intrusion in 2021 at Fagradalsfjall, Iceland, one of which led to an eruption, a third dike intrusion commenced on 30 July 2022. A sudden increase in seismicity occurred within the diking area, with approximately 1700 automatically detected earthquakes > M1 within 24 h. Strong earthquakes were felt over several days within a wider area (largest M W 5.3). The timeline and spatial distribution of seismicity suggested it resulted from diking, together with triggered seismicity in nearby areas releasing stored tectonic stress. Geodetic observations revealed displacements consistent with a dike intrusion, and geodetic modeling on 2 August revealed a best-fit model with a shallow top depth of the dike (~1 km), and high magma inflow rate (~49 m 3 /s). Also considering a decline in seismicity, a warning was issued that the likelihood of a new eruption in the coming days was high. An effusive eruption started the next day (3 August) on a ~375-m-long fissure, with an initial extrusion rate of 32 m 3 /s. The projected surface location of the dike (from the optimal model) was within 49–110 m of the eruptive fissure. We present a timeline of the activity and monitoring response in the days both preceding and following the eruption onset. We compare the details of the activity that occurred prior to this diking and eruption to the previous events at Fagradalsfjall to improve understanding of unrest preceding eruptions.

- The Gr & iacute;msv & ouml;tn volcano, one of the most active volcanoes in Iceland, is covered to a large extent by the Vatnaj & ouml;kull glacier. High geothermal activity within its caldera maintains an ice-covered caldera lake with variable water level. Large floods from the lake (j & ouml;kulhlaups) are initiated when the water breaks through an ice dam and flows out of the caldera. In several cases the falling lake level is known to have triggered eruptions of the volcano, e.g, in 1922, 1934, and 2004. The eruptions of 1983, 1998, and 2011, however, were not triggered by j & ouml;kulhlaups, and most j & ouml;kulhlaups have not triggered eruptions, including those of 2008 and 2010. All these processes, i.e. volcanic activity, water floods, and geothermal activity, are accompanied by seismic tremor that is detectable by the surrounding network of seismic stations. By comparing tremorplots of the j & ouml;kulhlaups of 2008 and 2010, and the eruptions of 2004 and 2011, we can identify three types of tremor: Water flood tremor. J & ouml;kulhlaups from the caldera are always accompanied by high-frequency tremor (2-9 Hz), recorded on the seismic stations near the caldera. It starts when the lake level begins to drop and increases gradually with increasing water discharge from the lake. This tremor is usually detected a few days before the subglacial flood reaches the glacier edge. Geothermal tremor. The second type of tremor appears to be switched on when the drop in water level reaches 10-30 m. It remains after all water has been drained from the lake. The tremor is characterized by relatively high frequency (2-6 Hz) and sudden changes in amplitude. The distance range of this tremor is short, it is seldom recorded beyond the edge of the glacier. We suggest that it is generated by flash-boiling of the geothermal system within the caldera, triggered by the pressure drop of the lake level. Eruption tremor. Eruptions of Gr & iacute;msv & ouml;tn are accompanied by tremor that begins simultaneously with the eruption and is distinctly different from the other two types of tremor. It contains lower frequencies (0.5-4 Hz) and has a wider distance range. It is recorded beyond the edge of the glacier, possibly because of its frequency content, but other effects such as crustal structure and depth of the tremor sources may also play a part.

The Fagradalsfjall eruption commenced on 19 March 2021 on a 180-m-long eruptive fissure, following a 23-day dike intrusion. New eruptive fissures opened northeast of the initial eruption site on 5, 6–7, 10, and 13 April 2021. The northernmost eruption occurred on 5 April, approximately 1 km northeast of the initial fissure, with the other fissure openings between this and the initial eruptive vents. Still images from web cameras and time-lapse cameras are available for five of the fissure openings. These data show that the eruptions were preceded by steam emitted from cracks in the exact locations where the eruptions started. The time between the first steam observations and the visual appearance of glowing lava ranged between 15 s and 1.5 min during night observations and from 9 to 23 min during daytime observations. The difference in observation time is likely explained by the different lighting conditions. The eruptive vents are located where the north-easterly oriented dike intersected pre-existing north-south-oriented fractures, inferred to be strike-slip faults. These fractures could be identified on a high-resolution ICEYE interferogram as well as on pre-existing aerial photographs and digital elevation models. This interferogram spanned the first day of the eruption (19–20 March 2021). It not only displays deformation related to the pre-eruptive dike intrusion but also shows lineations in locations where eruptive vent openings occurred later in April 2021. These findings demonstrate how Interferometric Synthetic Aperture Radar Analysis (InSAR) can be used to forecast likely locations of subsequent eruptive vent openings, which is of great importance for hazard assessment and defining exclusion zones during fissure eruptions.

The two volcanic eruptions of 2021 and 2022 at Fagradalsfjall in SW Iceland occurred within the Reykjanes Peninsula Oblique Rift, a segment of the complex boundary in Iceland between the North America and Eurasia Plates. Two of the plate boundary segments are highly oblique to the overall plate velocity vector, i.e., the Reykjanes Peninsula and the Grímsey oblique rifts. They contain both volcanic systems and seismogenic strike-slip faults. Oblique spreading leads to extensive volcanism and large earthquakes, a combination that is otherwise uncommon in Iceland. The fissure swarms of individual volcanic systems contain normal faults and fissures, arranged en echelon along the plate boundary. The fissure swarms fade out as they extend into the plates on either side. These volcano-tectonic rift structures on the Reykjanes Peninsula are overprinted by sets of parallel, N-S striking transcurrent faults that generate the largest earthquakes in the zones, up to M 6.5. Their surface expressions are en echelon fracture arrays and push-up structures. The distance between them varies from 0.3 to 5 km. They are most prominent in the areas between the overlapping fissure swarms, and together they form a bookshelf-type fault system taking up the shear component of plate movements across the oblique rift zones. The Fagradalsfjall volcanic system is located between the fissure swarms of the Svartsengi and Krísuvík fissure swarms. It lacks its own fissure swarm, which is otherwise one of the characteristics of Icelandic volcanic systems. We present maps of surface fracturing structures of the area and interpret them as the result of strike-slip displacement on underlying N-S faults. About 20 faults are implied along a 8-km-long section of the plate boundary. In addition to these bookshelf-type faults, several areas have been identified where earthquakes appear to line up along ENE-WSW-striking, fault-like structures. These structures have so far only been seen at the surface in one place despite a thorough search. Taken together, the N-S and the ENE-WSW faults form a conjugate set of faults. The implied tectonic stress field has a horizontal maximum principal stress with a N45°E orientation, and a minimum principal stress with a N135°E orientation, perpendicular to the fissure swarms on the peninsula and the dike intrusion that preceded the Fagradalsfjall eruption in 2021. It is postulated that bookshelf faulting is one of the characteristics of unstable or immature plate boundaries.

Using ground deformation measurements of high spatial and temporal resolution SAR, the understanding of new vents created during volcanic eruptions can be improved with 3D mapping of the activated shallow magma plumbing system. Interferometric analysis of radar data from ICEYE X-band satellites with daily coherent ground track repeat (GTR) provides unprecedented time series of deformation in relation to the opening of 6 eruptive vents over 26 days in 2021, at Fagradalsfjall, Iceland. Unrest started in this location at the end of February and tens of thousands of earthquakes were recorded during the following four weeks. The seismicity was linked to gradual formation of a magma-filled dike in the crust and triggered seismicity along the plate boundary. On 19 March, an eruptive fissure opened near the center of the dyke. New vents and eruptive fissures opened on the 5th, 7th, 10th, and 13th April. The daily acquisition rate of the ICEYE satellite facilitated the observation of the ground openings associated with each new vents. Each event can be observed individually and with minimal loss of signal caused by new lava emplacement, which would occur if images were acquired at a slower rate. Being able to retrieve deformation near the edge of the fissure ensures that we have the optimal constraints needed for modelling the subsurface magma path. The ICEYE dataset consists of Stripmap acquisitions (30x50km) in the period 3-21 March, and Spotlight acquisitions (5x5 km) from 22 March and onward. Images have a resolution of about 2 m x 3 m, and 0.5 m x 0.25 m, respectively. The descending 1-day interferogram covering each individual event is used to invert for the distributed opening along the dike plane. We find that each fissure was associated with opening of up to 0.5 meters in the topmost 200 m of crust. The conduits propagated vertically at least 50–80 m/h. The new fissure locations were influenced by local conditions and induced stress changes within the shallow crust.

The basaltic effusive eruption at Mt. Fagradalsfjall began on March 19, 2021, ending a 781-year hiatus on Reykjanes Peninsula, Iceland. At the time of writing (January 7, 2022), no eruptive activity has been observed since September 18, 2021. To monitor key eruption parameters (i.e., effusion rate and volume), near-real time photogrammetric monitoring was performed using a combination of satellite and airborne stereo images. By late September 2021, 32 near real-time photogrammetric surveys were completed, usually processed within 3–6 hours. The results are a significant achievement in full-scale monitoring of a lava flow-field providing temporal data sets of lava volume, thickness, and effusion rate. This enabled rapid assessment of eruption evolution and hazards to populated areas, important infrastructure, and tourist centers. The lava pathways and lava advancement were very complex and changeable as the lava filled and spilled from one valley into another and short-term prediction of the timing of overflow from one valley to another proved challenging. Analysis of thickness maps and thickness change maps show that the lava transport into different valleys varied up to 10 m3/s between surveys as lava transport rapidly switched between one valley to another. By late September 2021, the mean lava thickness exceeded 30 m, covered 4.8 km2 and has a bulk volume of 150 ± 3 × 106 m3. Around the vent the thickness is up to 122 m. The March–September mean effusion rate is 9.5 ± 0.2 m3/s, ranging between 1–8 m3/s in March–April and increasing to 9–13 m3/s in May–September. This is uncommon for recent Icelandic eruptions, where the highest discharge usually occurs in the opening phase. This behavior may have been due to widening of the conduit by thermo-mechanical erosion with time, and not controlled by magma chamber pressure as is most common in the volcanic zones of Iceland.

The formation of the island of Surtsey over 3.5 years, remains one of the best-documented volcanic, island-forming eruptions to date. The basaltic submarine volcanic activity was detected on November 14, 1963, where ocean depth was 130 m prior to the eruption at the southern end of the Vestmannaeyjar archipelago. The eruptions occurred in several phases involving explosive and effusive activity, including the initial submarine phase on November 12–13, 1963. Separate phases of subaerial volcanic activity occurred during November 14, 1963–January 1964, January–April 1964, April 1964–May 1965, May–October 1965, December 1965–August 1966, and August 1966–June 1967. Seismic data quality from this period is inferior compared to that of modern monitoring systems. Four permanent seismic stations were operated in Iceland at the time, whereof only two, located at 115 and 140 km distance, had the sensitivity to record tremor from Surtsey. Nevertheless, the scanned analog seismograms (http://seismis.hi.is/) show that the eruptive activity was accompanied by considerable seismic activity, both earthquakes, and volcanic tremor. Earthquakes were primarily associated with changes in vent location. Both spasmodic and harmonic tremor was identified, both with low (\>3 Hz) and higher (3–5 Hz) characteristic frequencies. The results indicate a complicated relationship between tremor and magma flow rate or style of activity. During the explosive eruption, the highest magma flow rates occurred in the first 10–20 days, a period with little observed tremor. The highest tremor is observed in December 1963–March 1964, after the discharge rates had dropped substantially, and on a timescale of hours-to-days, no clear relationship between tremor and eruption style is observed. The same applies to the effusive activity, where no seismic tremor was observed during most of the effusive eruption of Surtungur, despite the fact that magma flow rates were ∼3 times higher than during later phases where some tremor was observed.

Increased rates of deformation and seismicity are well-established precursors to volcanic eruptions, and their interpretation forms the basis for eruption warnings worldwide. Rates of ground displacement and the number of earthquakes escalate before many eruptions 1–3 , as magma forces its way towards the surface. However, the pre-eruptive patterns of deformation and seismicity vary widely. Here we show how an eruption beginning on 19 March 2021 at Fagradalsfjall, Iceland, was preceded by a period of tectonic stress release ending with a decline in deformation and seismicity over several days preceding the eruption onset. High rates of deformation and seismicity occurred from 24 February to mid-March in relation to gradual emplacement of an approximately 9-km-long magma-filled dyke, between the surface and 8 km depth (volume approximately 34 × 10 6 m 3 ), as well as the triggering of strike-slip earthquakes up to magnitude M W 5.64. As stored tectonic stress was systematically released, there was less lateral migration of magma and a reduction in both the deformation rates and seismicity. Weaker crust near the surface may also have contributed to reduced seismicity, as the depth of active magma emplacement progressively shallowed. This demonstrates that the interaction between volcanoes and tectonic stress as well as crustal layering need to be fully considered when forecasting eruptions.

The 2021 effusive eruption at Mt. Fagradalsfjall, on the Reykjanes Peninsula oblique rift in Iceland, was preceded by a 14-month long period of volcano-tectonic unrest (comprising both significant ground deformation and intense seismicity). A seismic swarm was initially detected in the Fagradalsfjall region between the 15th to 20th December 2019. Following a short quiescence, activity re-commenced on the 21st January 2020, with a small cluster of earthquakes near Grindavík (~ 10 km west of Fagradalsfjall). Concurrent deformation was detected on two GNSS stations in this area and on Sentinel-1 interferograms. Geodetic modelling of these observations indicated the deformation most likely resulted from the intrusion of a magmatic sill, directly west of Mt. Thorbjörn, at a depth of about 4 km. This was followed by two additional sill-type intrusions in a similar location, between 6th March - 17th April and 15th May - 22nd July 2020 respectively. The three intrusions comprised a total volume change of about 9 million cubic meters. In mid-July 2020, inflation was again detected on the Reykjanes Peninsula, this time in the Kýsuvík volcanic system to the east of Fagradalsfjall. This episode of inflation lasted several weeks and geodetic inversions indicated the observed signal was produced by the combination of a deflating sill-like source at a depth of ~16 km and inflation of a body at a depth of ~6 km. The latter, corresponding to a volume change of about 5 million cubic meters. During this period of intrusive activity, seismicity shifted along various regions across the Peninsula, in relation to a combination of processes – magma migration, triggered seismicity and tectonic earthquakes. Intense seismic swarms commenced on the 24th February 2021, concentrated at both Fagradalsfjall and also extending across a 20 km segment along the plate boundary – including triggered strike-slip earthquakes up to Mw5.64. At the same time, deformation was detected on local GNSS stations, and subsequent Interferometric Sythethic Aperture Radar Analysis (InSAR) of Sentinel-1 data confirmed the observed deformation was primarily the result of a dike intrusion and slip along the plate boundary. Geodetic inversions indicated a ~9 km long dike with a total intruded volume of around 34 million cubic meters (Sigmundsson et al., in review). During this period, stored tectonic stress was systematically released, resulting in a decline in deformation and seismicity over several days preceding the eruption onset, on 19th March 2021 in Geldingadalir at Mt. Fagradalsfjall. The eruption continued until the 18th September 2021 and produced a lava field covering an area of 4.8 km2 with an extruded bulk volume of 150 ± 3 × 106 m3 (Pedersen et al., in review). References Sigmundsson et al. (in review). Deformation and seismicity decline preceding a rift zone eruption at Fagradalsfjall, Iceland. Pedersen et al. (in review). Volume, effusion rate, and lava transport during the 2021 Fagradalsfjall eruption: Results from near real-time photogrammetric monitoring. DOI:10.1002/essoar.10509177.1.