After a wonderful series on Krakatau by Albert, a mysterious island by Carl and a seismic “Intermezzo” by Lurking, I have decided to take things back to Iceland for a while. Most of us know Greip by now, and we are going to take another look at it, from a more seismic perspective with some fresh plots.
Here are links to last two articles on Greip, by Gaz and Carl. They contain important information and theory on Greip, which is a must read.
After all these posts, what is there even left to say? Greip has not erupted (for now) and it just does what it does best: transporting magma upwards in a noisy fashion. I like to follow all the data I can get on Greip, which is for now mainly seismic. Carl has talked about GPS signals in his post linked above, but it’s not easy, since Vatnajokull is rather crowded with volcanic features in that part.
This year so far, was quite productive for Greip. In January-May period, we recorded the highest number of earthquakes for any 5 month period in Greip so far. That is also due to the better monitoring equipment in last few years, but generally in the last decade (or prior), no period comes close. For this purpose I have produced a new type of plot for Greip, where we can see earthquake numbers and the seismic moment release. Both parameters are part of the story. Data used for the plot was filtered by depth and magnitude. Holuhraun eruption period was also removed due to obvious reasons. It produced noise since it was in the same region as the plot. Some of the earthquakes in 2015 are also cooling quakes of the Holuhraun dyke, but it would be too hard and time-consuming to manually filter and remove all that data. So I decided to use cut-off filtering at 9km, which is very effective, since that is the upper limit of Greip core and lower end of the Holuhraun dyke. In the 2001-2012 period, there were M2s recorded and a lot of higher-end M1s, which over those years pushed up the energy total, while keeping the earthquake numbers at a relatively low count for such a long period.
It is obvious right away that the earthquake numbers are increasing. Energy on the other hand, is stagnant or even reducing. It takes more and more earthquakes to release the same amount of energy. What this obviously tells us, is that the magnitudes are reducing. We see weaker earthquakes, but more in numbers. The most obvious and “practical” reasoning behind it, is that the crust in the core area of Greip is getting cracked to the point where not much more strain can be build up for stronger magnitudes. Total moment energy release so far from Greip core (9km+ depth), based on the data recorded in the catalogs, is equal to an M3.4 earthquake. That is earthquake energy only, not including the accompanying tremor energy. That is a decent amount, considering being in the hotter/warmer part of the crust.
On 31st. May 2019 we have seen one of the strongest dyke intrusions at Greip in the past few years, with tremor being recorded/detected on SIL stations from Askja to Katla. There were 17 earthquakes recorded in the core with that dyke intrusion, but the strongest earthquake detected was just M0.9 in magnitude. Crust in/around the core area of Greip is likely getting hot and cracked enough to start slowly loosing its “structural” integrity, which slowly reduces/limits the amount of solid/cold material to crack in higher magnitudes. The crust around it is likely still strong enough that higher (M2+) magnitudes can happen, if there were dyke intrusions outside the current main core area. But that is currently low probability, since the dyke intrusions seem to be focused very well on the specific area for now. I usually compare such “dyke swarms” like Greip, to lightning. Just like lightning, magma also pushes in the path of least resistance, with each individual dyke intrusion being like an individual lightning bolt.
The drumplot from the DJK station (the closest station to Greip), shows the dyke intrusion at Greip from 8:39am UTC to 8:51am UTC.
Looking at the magnitude vs depth distribution, we can see that the main area of energy release in Greip is between 10-20km deep. Below that, it is possible that the crust is getting too hot/ductile, or the main cracking has happened well before monitoring began on a high enough level. I used 2011-2019 data for this plot, which is the period where monitoring was getting a lot better across the region. Data was filtered by depth, showing only earthquakes below 8km depth, to remove some of the noise from the cooling Holuhraun dyke (LD2). This plot does not show any actual shape of Greip, but only presents earthquake magnitude distribution by depth.
Since this plot includes long-term data, I asked myself how the Holuhraun seismicity fits into this model. I have made a few cross sections that include Greip and the LD2. What is notable right away, is that the Holuhraun dyke is not completely linear in vertical shape. It made a turn at around 10km depth, tilting NW, away from Greip core area. It is also slightly tilted away above 9km.
To explain the graphic, we are looking at the data within the yellow lines (cross-section). Blue is the Holuhraun dyke, red is the pre-eruption seismicity, and white is post-eruption seismicity. White dots in LD2 are cooling earthquakes. We can see a tilting of LD2 starting around the area where the Greip core starts, and going down in a similar angle as Greip goes up.
Looking specifically at pre-eruption data we can see how Greip core was well established. Some older data includes shallower earthquakes, which are not really seen since the monitoring grid upgrades, so it is hard to judge how legitimate those locations really are. But the Holuhraun eruption and its related seismicity did affect the stress fields in the area, so it could be related to that aspect.
And comparing with post-eruption seismicity till the present time, we see the cooling quakes in the LD2, and a very good-looking Greip core area, which by this time has a more circular and cylindrical shape. It does seem expanded a bit towards E/SE, while the border between LD2 and Greip to the NW remains fairly stable. This data is also of a higher quality/accuracy, thanks to the much more sensitive monitoring grid.
Looking at the longitude components from Greip (9km+ depth data), we can see the slow movement/expansion towards east in recent years.
Looking specifically at the seismicity in the past year, we see that Greip is tilted/leaned slightly towards east with height That was evident also on pre-eruption and historical data.
It is evident that there was some force present that managed to “deflect” the Holuhraun dyke away from Greip below 10km depth. It is quite likely that there is possible magma accumulation in that area in Greip, which was present before the Holuhraun eruption. That should at least in theory provide some alteration to the surrounding stress fields to possibly deflect the Holuhraun dyke only at that specific area and depth. Looking closely at the plot above, we can see that Greip has two more features. A shallow seismic “cap”, and a gap in seismicity between 5-10km depth. The Holuhraun dyke is also tilted a bit away above 9km depth.
Looking at the SW-NE profiling of Greip, the shallow cap is nicely seen. It is a rather new feature, seen more in the recent year. It is possible that it is a stress field response to the LD2 formation and cooling, or it could be a shallow response to the potential inflation stress of Greip, being more prone to seismicity as a potential extension of the Grimsvotn fissure swarm. Usually it is a combination of many things. On the cross-section, we also see the large seismic gap between 5km and 10km depth. There is currently no valid explanation for that. It is possible that the crust there is cold and dense, so Greip intrusions were not yet able to crack that part of the crust. If this is true, than this is a potential area for higher magnitude earthquakes, perhaps even a star or two, if pressure from below is high enough. The second option is that there is perhaps magma accumulation there, like a magma chamber, which usually does not have much or any earthquakes. If that were the case, that would explain the connected shallow seismic cap. But such a large accumulating body of magma, should also have seismicity surrounding it on the horizontal level as well, as it expands horizontally, not just vertically above and below it. So this is not a favorite option for now, but it remains plausible. There was some seismicity recorded in this layer, but has gone quiet in recent years. Time will eventually show what this area really is.
Also of note on this graphic, is potential sill formation between 18-26km depth.
Speaking of time, I have to include a time plot of Greip, showing its evolution. Some data is of course from the cooling Holuhraun dyke, especially in 2015, but generally it tells the tale. Holuhraun eruption period is removed for obvious reasons. We can see the seismic gap between 5-10km in the recent times. Some of the data in that depth are cooling quakes from LD2.
To put Greip into perspective, I have to finish off with a size comparison between Greip and some central volcanoes. By size, Greip is no pushover, and is in itself strong enough, and likely transporting enough magma, to kick-start many Icelandic volcanoes, were it “used” in that way.
Greip is a noisy bugger, which has entered a new type of seismic pattern in the past year, where we see more tremor accompanying the earthquakes, and the elevated shallow sesimicity above its core in the 0-5km depth layer. It will only get more interesting with time, since it is hard to believe that its momentum would stop, given that we are entering (or have entered already) a new plume pulse. Being a 3D plotter, I naturally have to include a 3D spin-around video of Greip, showing 2012-2019 seismicity. You can see the Holuhraun dyke (LD2) passing by it. I recommend watching in HD and full-screen.
Down Under (@Recretos)