EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

Pre-existing fault-controlled eruptions from the lateral tips of a laccolith in SE Iceland

Vincent Twomey1, William McCarthy1, Craig Magee2, and Michael Petronis3
Vincent Twomey et al.
  • 1University of St. Andrews, St. Andrews, United Kingdom of Great Britain and Northern Ireland (
  • 2School of Earth and Environment, University of Leeds, Leeds, United Kingdom of Great Britain and Northern Ireland (
  • 3Environmental Geology, Natural Resource Management Department, New Mexico Highlands University, Las Vegas, USA (

Volcano eruption forecasting relies on models of sub-volcanic magmatic plumbing systems that link ground deformation to sub-surface magma movement. However, many of these models typically assume that eruption sites occur directly above laccolithic reservoirs. Furthermore, many of these models assume deformation of the host rock is exclusively elastic with few studies highlighting the role inelastic deformation (e.g., faulting/fracturing). Whilst the dynamics of magma flow have previously been well constrained in ancient in sub-volcanic systems, its geometrical and kinematic relationship with the corresponding host rock deformation remains poorly understood which, is critical to volcanic hazard assessment.

Here, we examine the structure of the shallow-level (i.e. intruded <1 km below the palaeosurface), silicic Reyðarártindur laccolith in SE Iceland, and demonstrate how the underlying mechanisms of lateral magma flow coupled with pre-existing host rock structures influenced the localisation of volcanic activity. In particular, we use anisotropy of magnetic susceptibility (AMS) fabric analysis and show that the intrusion contains several laterally emplaced magma lobes, with magma flowing along a SW-NE axis, parallel to the strike of pre-existing, steeply dipping fault arrays in the host basalt lavas. Lateral magma flow and inflation of the lobes promoted upward intrusion along these pre-existing faults, which we posit acted as preferential pathways for magma to reach eruption sites that were laterally offset by tens to hundreds of metres from the underlying main intrusion.

Our interpretation provides field evidence for the reactivation of pre-existing structures as inclined magma conduits to eruptive vent sites on the outer margins of subjacent lateral magma bodies. This supports seismic observations where (i) Volcanoes overlie the lateral tips of subjacent intrusions in subvolcanic systems; (ii) ground, and host rock deformation preceding eruptions can be most prominent in areas adjacent to the volcano site; and (iii) volcanoes overlie and are aligned along fault traces suggesting that pre-existing normal faults influence the localisation of volcanic activity.


How to cite: Twomey, V., McCarthy, W., Magee, C., and Petronis, M.: Pre-existing fault-controlled eruptions from the lateral tips of a laccolith in SE Iceland, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-17969,, 2020

Comments on the presentation

AC: Author Comment | CC: Community Comment | Report abuse

Presentation version 1 – uploaded on 07 May 2020
  • CC1: Comment on EGU2020-17969, Emma Rhodes, 08 May 2020

    Hi Vince,

    What an incredible data set you have here. I spent an enjoyable few hours going through it. And thanks for doing all this AMS! The NE-SW direction of the AMS, parallel to the regional tectonics is certainly interesting. The contrast of the long axis of the laccolith pluton to the regional tectonics has always been of interest to me. 

    My first question is, why do you say the pluton is <1km keep in your abstract? Have you managed to constrain this somehow?

    My second question is, what constaints do you have on the boundary between the two main lobes? 

    Thirdly, your model has eruption sites only from the inclined sheets at the toe of your lobes, are you also considering that other large dykes could have erupted?

    That will be enough for now,

    Looking forward to your answers and discussing further.


    • AC2: Reply to CC1, Vincent Twomey, 08 May 2020

      Cheers Emma, 

      Many thanks for the feedback. It probably is a good idea to discuss the area in further detail soon.

      To answers your quetions for now I will keep it brief

      Q1... I say the pluton was emplaced at <1km depth, whic I wrote down with a grain of salt. I made that call for now based on the rhyolitic texture of theupper section of the intrusion. I don't know how deep intrusves keep a rhyolitic texture. I note that Sandfell is called a Rhyolite and is inferred to be emplaced at 800m depth? (off the top of my head, Tobias and Steffi will correct me) as this is the structural thickness of the forced fold (the gently dipping host rock beyond that is inferred previously as an 'unconfromity' in the lave piles. So I assumed the same regarding the rhyolitic texture in Reydara. But, I do note previous literature regarding the level of erosion in SE Iceland based on extrapolation of the level of zero dip of the lava, zeolite zones and extrapolation of intrusion density which I need to look into more and I'm open to correction on it. But Ultimately it does not change the model of these inclined sheets acting as ascent conduits be it immediately below the surface or >1km down.

      Q2... The Constraints on the two main lobes, is based on the particular concentration of the gently inclined K1 fabric wth a steep Mag fol. plane, also it is where the river valley with in the intrusion occurs so that could have been a site of weakness between the lobes. 

      Q3... This is a fair point, My schematics here are simplified with respect to how the AMS can be fully interpreted. I agree that looking at the roof there are likey to be more vents or conduits up there. So yes, magma ascent may not be  be fully exclusive to the lateral tips of Reydara. The process however could be simuilar as these vents could be linked with smaller magma fingers within the larger scale lobes. Furhter more, the NE-SW fractures appear persistant across the entire structure aureole. This can be tested by looking athte distribution of the dip of the magnetic foliation plane, coupled with the Tj of the AMS with respect to structures in the host rock 

      Look forward to discussing this further




      • CC5: Reply to AC2, Emma Rhodes, 11 May 2020

        Thanks for your response Vince!

        I was enquiring about the depth because it is something I am curious about. I have been referring to the ca. 2km depth from Walker data, but otherwise your guess is as good as mine. I agree that this difference doesn't influence your model.

        I have been focussing on geochemistry recently, and learning how to plot and interpret geochemistry of course. So this is my first look into the structural data for now. I think the idea of the ring dykes at the tips of the lobes is super interesting. Especially in combination with the AMS data. In your model, on page 4, the lobes are being emplaced horizontally, is this the case here? Or are they more inclined?


        • AC4: Reply to CC5, Vincent Twomey, 15 May 2020

          Hi Emma, 

          The lobes are intepreted to have been emplaced  horizontalally/sub-horizontally. This is based off interpretations of the AMS fabrics. The lobes are defined based on the magnetic foliations paralleling and slighly curving towards magnetic lineation.  The magnetic lineation trend parallels the flowaxis but may splay down flow. But post emplacement, the laccolith was likely to have been gently tiltled towards the WNW with the lava piles during rifting/extension.

  • CC2: Comment on EGU2020-17969, Ingi Þorleifur Bjarnason, 08 May 2020

    Hello Vincent,

    You mention pre-existing ENE-WSW regional, extensional faults within the basaltic host rock. How do you explain this regional direction? In which direction do these faults dip?

    Thank you,  Ingi

    • AC1: Reply to CC2, Vincent Twomey, 08 May 2020

      Hi Ingi,

      Many thanks for you question. 

      The main trend of minor intrusions sheets/dikes in SE Iceland strike NE-SW (see Torfason 1979  PhD, & Burchardt et al., 2012) which, in map view is parallel to the main horizontal compressive stress direction (Sig3).  The direction of regional extension at the time can be therefore deduced (to the NW). Looking at a regional map of Iceland, many of the volcanc centres are alligned NE-SW. Also, the regional dip of the lava piles is 10deg to the WNW, i.e perpendicular to these regional faults. 

      Based on the observations of the inclined sheets in my area, I rekon there faults are composite, dipping steepy both to the NW and SE

      • CC3: Reply to AC1, Ingi Þorleifur Bjarnason, 08 May 2020

        Hello Vincent.

        Thank you for your answere and for studying the geology of Iceland.

        I got confused but also interested about you mentioning 'pre-existing ENE-WSW regional, extensional faults' .

        > The direction of regional extension at the time can be therefore deduced (to the NW).

        So you are saying that the volcanic zone had approx. the trend of the active EVZ, while the NVZ trends more NNE. Was this apparent extension at the time (NW) constant over much of the geological history of the East?

        Cheers, Ingi


        • AC3: Reply to CC3, Vincent Twomey, 10 May 2020

          Hi Igni,

          Yes, that would be my thought process. Akin to how the active fissure swarms today are liigned perpendiclar to the spreading direction. But the linking the laccolith here to other subvolcainic laccoliths and and volcanic centres to make regional constraints may be beyond the scope of my PhD.

          p.,... Iceland was a great place both geologically/scenically/socially to do fieldwork

          • CC6: Reply to AC3, Ingi Þorleifur Bjarnason, 15 May 2020

            Hi Vincent,

            Can you speculate on the stress state in the crust, needed to form the laccolith you are studying. Does maximum stress have to be horizontal? If so, why would it be horizontal in a rifting environment? Can you determine the azimuth of the maximum principal stress.

            Cheers, Ingi

  • CC4: Questions and answers from the live chat during EGU2020, Michael Heap, 11 May 2020

    Q: Is the pre-existing fault created by the intrusion, or a regional fault?

    A: We belive it to be a rgional as, they match the trend of the regional dykes and, these structures are also pervasive across the host rock in the area

    Q: You mention pre-existing ENE-WSW regional, extensional faults within the basaltic host rock. How do you explain this regional direction? In which direction do these faults dip?

    A: Regionally, the strike of regional fissure/dyke swarms

    Q: Do you have other evidence except for the structural to suggest the location of eruption sites?

    A: Well, you can visiblly see the inclinded sheets in the field, paralleling the exposed contact of the main laccolithic body

    Q: A similar technique was used to study the Sandfell laccolith/cryptodome, also in east Iceland. Did you compare your model with those developed for this intrusion?

    A: I have seen the AMS from Sand fell. It does not appear to have the same disinct, consistant fabric seen here in Reyfara. Additionally, the host rock deformation in Sandfell appears much differnt and may suggest a different emplacement process

    Q: Reydarartindur is a small pluton. Do you think that volcanoes in genera are located above the lateral ends of their source chambers? If so, why?

    A: You could argue we do not know the true size of the Reydara Laccolith, If yhe lobes are flowing along a NE-SW axis, the lonitudinal section of the intrusion could go much furher back into the subsurface along a NE-SW direction. Only the outer extent may currentlly be exposed. Also, volcanism associated may not be fully exclusive to the lateral tips of Reydara. There could be further vents along the roof. These vents could be linked with smaller magma finger within the larger scale lobes. This can be tested by looking athte distribution of the dip of the magnetic foliation plane, coupled with the Tj of the AMS with respect to these structures