After Albert’s explanation of Afar’s volcanic and tectonic past, I will look into this region’s Holocene volcanism, and try to create the first comprehensive volcanic history of this very unstudied volcanic area.

I like Afar because the volcanoes are “naked”, what I mean is that due to a desertic environment, the lava is slow to develop a soil cover and doesn’t grow much vegetation. As such, using GIS data, like for example satellite surface reflectance imagery, digital elevation maps, elevation hillshade maps, or aerial imagery it’s possible to get a good understanding of the eruption history of Afar volcanoes. While absolute dates of eruptions are generally not known, it’s possible to use different techniques to figure out the relative order of events, including the superposition of lava flows, the interaction between lavas and water lakes that were probably formed during the African humid period, the color of the lavas as a function of weathering, silicic ash that has been found in lakes downwind of Afar, or the few lava flows whose absolute ages are known from historical accounts or surface-exposure dating. As such, despite being a region that remains very unknown, it has the potential to be understood far better than volcanic areas that are vegetated, or covered in snow, also because surface-exposure dating can be used everywhere in Afar. And it’s also a remarkably powerful volcanic province. My goal is to advance knowledge into the unexplored past of many Afar volcanoes, and hopefully motivate future research to look into it. So let’s commence!

 

The Stratoid Series.

The Stratoid Series is fundamental to understanding current Afar volcanism given that the area is still within this phase. This volcanic stage has erupted massive amounts of basaltic lavas and also some silicic material which has filled the Afar depression to a substantial thickness. In Central Afar, vertical cliffs of up to 1 km tall expose piles of basalt lava flows measuring hundreds of meters in thickness and often erupted within a few hundred thousand years, if not less. The bottom of the lava pile doesn’t outcrop anywhere in central Afar, only near the sides where it’s thinner, and the layer covers a circular area that is 300-400 km wide, some 100,000 square kilometers. Many maps put the Afar Stratoid Series as lasting between 4 and 1 million years ago, but this is based on dating from the 1970s that had very limited accuracy. The newest data is from a 2003 article titled, “New age constraints on the timing of volcanism in central Afar, in the presence of propagating rifts”, and puts the Stratoid Series across central Afar around 2-1 Ma ago. It should also be considered as ongoing given that vigorous volcanism continues to rapidly grow very thick basaltic lava piles in some northern areas like Alayta, the Erta Ale Range, or the Nabro Range.

It’s generally thought that the Afar Stratoid Series was caused by propagating rifts that swept west and north across the Afar depression starting from the Gulf of Aden and eventually reaching the current volcanism in the Erta Ale Range and other northern volcanoes. I believe, however, that another idea may be more appropriate. I think the Stratoid Series may have formed as magma, that was being produced under the Main Ethiopian Rift, was diverted laterally through the asthenosphere or lithosphere towards lower elevations. Based on data from Lahitte et al. 2003, the Stratoid Series would have started 3 million years ago just north of the “bottleneck” of Afar, in the area now occupied by the Yangudi and Gabillema volcanoes, this is where the Ethiopian Rift meets the Afar Depression. From here volcanism gradually swept in many directions. By 2 Ma vigorous volcanism had started in Central Afar over a broad area around present-day Manda Gargori volcano and reaching also to the Gulf Aden.

I’m not also not convinced that this was associated with propagating rifts given that the Danakil Depression where Erta Ale is located has a number of very old volcanoes, Alid and company in the north, which are located along the same rift of the Erta Ale Range, but are much older and likely predate the Stratoid Series in Central Afar, showing that rifting in this area likely has existed for quite some time. Instead, it appears to have been more of a volcanic wave spreading in the form of a series of volcanic chains, with volcanism occurring behind and ahead, but mostly within the wave. Like tentacles, the propagating chains of volcanoes leave rows of extinct volcanoes behind and concentrate most activity near the tip, the Erta Ale and Nabro ranges being prime examples of active tips, but also Dabbayra, Ma Alalta or Alayta. This leads to a dynamic environment of shifting volcanism, with volcanic systems quickly emerging and going extinct all over the region, which is a very distinctive characteristic of Afar, although is also seen over a more limited scale in Kenya and Tanzania.

In this first article, I will focus on the area where everything started, around the triple junction, which is fundamental since the main age constraints of Holocene Afar volcanism are from here, which can be used to understand volcanism elsewhere.

 

The African humid period

While looking at Dama Ali volcano, which is roughly over the Afar triple junction, in GIS data, I noticed some unusual platforms of lava that are located around the volcano, which at first I didn’t give much importance, but then I realized they were lava deltas formed along the coast of a water body, and this was super-interesting, since it provided stratigraphic constraints on the lava. Why is this?

Once I found the Dama Ali lava deltas, I started seeing them all over the place. This is the basin where the Awash river/rivers go to die. It’s an endorheic basin, there are a few small lakes nowadays that are located at 240-340 meters in elevation, the lowest outlet from the basin is at 390 meters where a dry river channel can be seen to cut through a horst. I noticed that many young lava flows that reached down to the bottom of the Awash Basin formed deltas at elevations of 380 meters, in the Gabillema, Dama Ali, or Manda Hararo volcanoes, which means the (now almost entirely dry) basin must have been filled to the brim with water, making a 5,000 km2 lake that constantly or semi-constantly overflowed north and must have discharged into the Gulf of Aden (I think the outlet being perhaps 10 meters lower than it is today due to sedimentation that has taken place since). And this is almost certain to have been during the African humid period.

The African humid period, was an interval between 14,500 and 6,000-5,000 years ago when northern Africa was much wetter than today. The Sahara was riddled with large lakes, and covered in grass and trees. Before the onset of this period, the climate had been even more arid than it is today, and there hadn’t been another humid period for over 50,000 years.

The African humid period must have affected Afar too, given that rivers from the Ethiopian Highlands pour into the depression, and it’s likely that it was at this time that the Awash basin filled to the brim with water. It was likely not just the basin of the Awash River, Erta Ale volcano also has some lava deltas that lie at around -85 meters below sea level, and show that a lake about half the extent as the one in the Awash basin at some point covered the floor of the Danakil depression. Some other smaller salt pans in the Afar region, located in areas that are nowadays always dry may have contained salty lakes at the time. Four volcanoes are located along the shorelines of this ancient lake, Gabillema, Dama Ali, Manda Gargori, and Manda Hararo:

 

Gabillema

Gabillema volcano towers near the southeast corner of what used to be the lake inside the Awash River basin. It’s a long-lived volcanic system that lacks the dynamism of regions further north. A steep silicic stratovolcano crowns the mountain, surrounded by smooth basaltic slopes and the eroded ruins of older silicic stratovolcanoes of the same system. The summit hasn’t erupted during the Holocene, but the volcano has an extremely rare three-rift configuration, maybe being the triple junction itself, and along each of the rifts young lava flows and craters can be seen.

Two rifts, northeast and northwest, sourced basaltic lavas during the Holocene. An eruption from each rift formed a lava flow that reached down to the basin at the time when it contained a lake, making lava deltas. In the western lava delta, some lavas managed to flow beyond the shoreline and into the lake, where they formed a bizarre pattern of lava with small curved wrinkles and shaped into a network of winding tumuli and white sediments that kind of looks like a brain or a coral. Manda Gargori has a much larger and more spectacular example of these bizarre lavas that we will see later.

These lava flows are useful cause they can date the lavas to the African Humid Period, and also can be compared to other undated flows to know their rough age. Given the lava deltas show the presence of a lake then they must have formed somewhere in the 5,000-14,000 years old interval. The two flows of the lava deltas have a’a lava that doesn’t have soil (it’s black colored), and older a’a lava in the area is soil-covered or almost completely soil-covered, so a rough rule that soil-free a’a in Afar is of Holocene age can be established. A’a lava also changes in color as it weathers, where reflectance in the near-infrared part of the electromagnetic spectrum increases, which can be used to infer age. In this variable, both the lava delta flows show being highly weathered (for a soil-free flow). With pahoehoe lava, it’s much more difficult to tell, since pahoehoe comes in a myriad of colors depending on how it’s emplaced and it’s also hard to tell when soil starts to form.

Using these techniques, then the only Holocene basaltic eruption from the NE rift is the one that formed the eastern delta, which has a volume of 0.18 km3, given it’s the only one with black-colored soil-free aa lava. But the NW rift has a lot of lava flows that are Holocene, erupted from tens of vents that roughly line up, one of which overlies the western delta flow. I think many of the flows formed during a rifting event where the rift was intruded repeatedly in a matter of years producing closely spaced eruptions along fissures over a length of 10 kilometers, and perhaps some of them being complex long-lived eruptions with shifting vents. The weathering of other NW rift Holocene flows, when looking at surface reflectance data, shows to be about the same as the lava delta flows, so they are all likely older than 5,000 years. The western delta eruption had a volume of 0.09 km3, while the other NW rift flows have a total of 0.53 km3.

The south rift doesn’t have young basaltic lavas but contains a rhyolite lava dome and three ash layers, which I referred to as Gabillema-A, Gabillema-B, and Gabillema-C in order of likely increasing age given their preservation. I think that because these layers have not been removed by erosion they are of likely Holocene age, and I think Gabillema-A is probably the youngest eruption of the volcano, maybe occurring in the last few thousand years. The oldest tephra layer, Gabillema-C issued from a 700-meter wide explosion crater and covers a small area of 4 km2, which is merely the part that has endured erosion due to tephra being thick or coarse, the original eruption likely having showered a far larger area in tephra. Next was Gabillema-B which was issued from two small craters that line up and presently covers 12 km2 in a dark gray ash.

The last eruption of the south rift, and probably the volcano as a whole, was Gabillema-A which started as a subplinian eruption, that gave rise to a white pumice deposit on top of Gabillema B that currently spans an area of 20 km2. The eruption transitioned into an effusive eruption that formed a 0.09 km3 rhyolite coulee of pristine appearance and deep brown color. These three explosive eruptions were likely VEI-3 events, though I wouldn’t rule out Gabillema-A having reached a small VEI-4 size.

Other than having an unusual triple rift configuration, Gabillema is relatively similar to other volcanoes in the area, bimodal volcanoes with central alkaline rhyolites and medium-sized eruptions of alkali basalts, and overall a low level of volcanic activity that is mainly a side-consequence of rifting. The total volume of Holocene lavas erupted from Gabillema is 0.9 km3.

 

Dama Ali

To the north of Gabillema lies the Dama Ali volcano, this volcano has many distinctive aspects. One of them is the circular shape of the volcano on a horizontal section, as opposed to most volcanoes of the area which are elongated due to the presence of rifts. This is probably because Dama Ali doesn’t seem to erupt from the rifts, most of the lava comes from the summit or from a ring of vents that surrounds the outer caldera and is probably fed from cone sheet intrusions. The caldera itself is a wonderful complex of craters nested within each other, an outer 4-km wide caldera is mostly filled and overflowed and probably dates from before the Holocene. Another 2 km wide caldera lies inside which is filled with resurgence or lava almost to the brim, and has another deep 1 km wide caldera inside. The outer caldera is pre-Holocene, but the innermost two are probably from the last 3000 years as we will see later.

Another spectacular feature of Dama Ali is a series of lava deltas at 380 meters above sea level which run for a distance of nearly 60 kilometers around Dama Ali, encircling the volcano almost entirely. The volcano must have been an island at some point in time.  There are two types of lava deltas found here. The first type formed where pahoehoe flows reached the ancient lake, these solidified at the shoreline so formed platforms enclosed by cliffs, and very little or no lava flowed underwater. The second type was formed where a’a lavas entered the ocean, these lavas flowed underwater making some reddish-colored underwater lavas, which did pile up vertically building deltas, but smoother than those formed by the pahoehoe lava. This episode of lava flows must have taken somewhere between 14,000 and 5,000 years ago, during the African Humid Period, so I referred to them as AHP overflows. The lavas originated both from the summit and a ring of vents that surrounds the caldera complex, they are almost everywhere covered in ash, but on one side of the mountain, beyond the ash, the a’a lava is soil-free. The AHP overflows resurfaced most of the volcano with multiple sheets of lava likely emplaced one over another. Assuming mean thicknesses of 20-30 meters the volume of this phase might be around 4-6 km3.

There was also another eruption event during this time. Next to Dama Ali, perhaps being a flank vent, lies a volcanic structure that, according to Google, is called Āhīlē Isate Gemora YeFelek’ebet Bota. But Āhīlē Crater can also be used as an alternative. This volcano forms a complex of over 10 craters ranging from spatter cones to tuff cones that erupted pahoehoe lavas and formed an island in the lake during the AHP, now perched above the dry lakebed and on top of 35-meter tall cliffs the top of which are at 380 meters asl. The volume of the Ahile Crater eruption is 0.43 km3, it likely started as an intrusion into the lake sediments, followed by the opening of multiple vents above the intrusion and surtseyan activity, and finally sustained lava effusion for months or up to a few years.

Dama Ali must have been a prodigious explosive volcano too. The area around the caldera is covered in white pumice, that extends east, and almost the entire volcano is covered in a thick layer of gray ash that seems to follow topography, avoiding tall obstacles, so I think was probably emplaced as pyroclastic density currents. Parts of the flank that were probably protected from PDCs by the rim of the old 4-km wide caldera are instead covered in thick white pumice. There are no Holocene silicic lavas visible on Dama Ali, but some pre-Holocene eruptions did form rhyolite domes, and the Holocene pumice from the summit is probably rhyolitic. Within 3 kilometers to the east of the caldera, I think the ash is several meters thick or more since it obscures lava flow edges and levees. At first, I considered it as a single explosive event associated with the formation of the 2-km wide caldera, but ash layers recorded in nearby Lake Hayk may tell a different tale.

A 2018 article called “Glass compositions and tempo of post-17 ka eruptions from the Afar Triangle recorded in sediments from lakes Ashenge and Hayk, Ethiopia”, studies the age and composition of tephra layers found downwind of Afar volcanoes in Ethiopian lakes. I took a look into this data and compared the chemistry of these silicic tephras to other studied silicic products of Afar, and I was surprised to find that although the ash layers in Lake Ashenge had compositions that fit Ma Alalta and Dabbahu volcanoes, most of the compositions in Lake Hayk did not fit any volcano. The problem is that Lake Hayk ash has very low Na2O and K2O to fit potential sources of Afar explosive eruptions, the Nabro Range, the Erta Ale Range, Ma Alalta, Dabbahu, Yangudi, or Adwa, are all more alkaline than Lake Hayk ash. Among the layers with chemical data available, only the oldest dated ash (HT-12) has a composition like other sampled volcanoes and probably is from Dabbahu. Some volcanoes don’t have any chemical data like Tat Ali, Gabillema, or Dama Ali. Tat Ali activity I think is too old and I don’t see much reason why it would have an anomalous low alkalinity chemistry, particularly considering that sampled mafic lavas from this volcano are alkali basalts. Gabillema I expect has a similar chemistry as volcanoes further south like Yangudi or Adwa which erupt in a way that is very similar to Gabillema.

Dama Ali volcano has the best chance of producing the ash layers in Lake Hayk. First, it’s similarly close (200 km) to Lake Hayk as other major candidates like Dabbahu or Ma Alalta, which however do not have the chemistry to be responsible. Added to this the main dispersal of the ash seems to have been E or ENE, directly towards Lake Hayk, which would make it the likeliest candidate based on position alone. While the chemistry is not known, there is some possibility that the chemistry is unique. Dama Ali’s rift lines up with Manda Hararo to the NE which is the volcano in Afar that resembles the most a mature mid-ocean ridge in the way it erupts (it’s like a fragment of Iceland that’s embedded into Afar) so possibly the rift shared by Manda Hararo and Dama Ali has seen a greater transition towards an oceanic environment than elsewhere in Afar and could be producing lavas with unusually low alkalinity. This is why I think that Dama Ali is the most likely volcano to have sourced the ash layers in Lake Hayk, particularly it would have been responsible for at least six explosive eruptions that took place 2876−2460, 3635−2792, 3914−2913, 3981−3197, 4804−3835, 10514−8432 and 11045−8903 years ago (cal. a BP). Most explosions would have taken place after the African Humid Period, <5000 years ago, which is consistent with the thick ash that covers the AHP overflows. The layers in Lake Hayk are very thin, only up to 5 mm.

The stratigraphy and weathering of Dama Ali flows show that in the last few thousand years, a series of small fissure eruptions took place from the summit and upper flanks of the volcano. I numbered these flows from 1 to 8 in order of increasing age. The first eruption was Dama-8, a thin basalt or basaltic andesite flow, with a volume of 0.03 km3, that weathering shows happened a long time after the AHP overflows. Dama-8 overlies most of the ash from the earlier explosive events but is covered by some pumice and ash from what I think was the youngest explosion recorded in Lake Hayk. What I think happened is that this lava flow erupted shortly before the explosive eruption of ~2600 years ago, which may have accompanied the formation of the 2-km wide caldera at the summit. Other older explosive events may have had this mechanism when rifting events caused the summit to collapse and accidentally triggered rhyolite explosive eruptions from a shallow chamber, older generations of calderas that were shallowed by the 2-km wide collapse.

The last eruption period comprises a series of basaltic or basaltic-andesitic lava flows which overlie all the silicic tephras from Dama Ali, and probably took place within the last 2000 years. This period includes 4-6 fissure eruptions from the upper flanks of the volcano which I numbered as Dama-2 to Dama-7, which had negligible volumes. At the same time, lava lakes erupted inside the 2-km wide caldera and filled it up, and huge lava fountains issued from inside the crater which formed black scoria deposits that extend up to 1 km beyond the rim. A second caldera collapse followed, 1-km wide, with no explosions associated given that the lava lake filling the ~2600 years old caldera doesn’t have any ash on top. The final eruption of Dama Ali was a small flow that erupted from the bottom of the 1-km wide caldera, Dama-1, ending the productive eruption history of Dama Ali.

In total, Dama Ali erupted over 4.5 km3 of lava outside the caldera, plus lava lakes that filled calderas, plus probably six or more sizable Holocene silicic explosive eruptions, at least some of which must have reached VEI 4.

 

Manda Gargori

Manda Gargori is a small volcano north of Dama Ali, unlike its neighbors it does not have a central volcano but instead consists of a field of lava flows, shields, and domes. There are three Holocene eruptions, including two basaltic flows (referred to as Gargori 1 and Gargori-2), and a rhyolite lava dome. The rhyolite lava is brown-colored and has a volume of 0.11 km3 with a well-preserved surface.

The Gargori-2 eruption is very interesting, half the lava flow is normal a’a lava, but the other half which occurs below 380-370 masl has a bizarre and unique morphology. The surface of lava is covered in curved wrinkles and the flow itself forms a structure of winding tumuli that in places resemble the texture of a brain and elsewhere resemble corals. The white sediment or salt deposited in concave areas makes beautiful patterns with the lava. If I had seen this landform with no context I don’t think I would have guessed it to be lava, but due to being the continuation of a normal lava flow below the level of the ancient lake, and because it resembles similar structures below the western lava delta of Gabillema volcano, it’s clear that these patterns were emplaced as voluminous underwater lava flows over the flat lakebed of the Awash basin somewhere 5,000-14,000 years ago. The Gargori-2 lava flow was a large long-lived eruption with a volume of 0.45 km3.

Gargori-2 was partly covered by the later Gargori-1 flow, in which I haven’t seen any signs of water interaction, but is almost entirely above the AHP water level, and the AHP wasn’t continuous either. The weathering is about the same as Gargori-2 so I think it’s older than 5,000 years, and has a volume of 0.06 km3.

The total volume of lava erupted by Manda Gargori during the Holocene is 0.6 km3, so not as much as the other volcanoes but highly interesting due to the bizarre underwater lavas it has, making this area in general a prime location to study the interaction between lava and water, or it would be if it wasn’t so hostile.

Manda Gargori is as far as I go in this post, in the future I hope to analyze more Afar volcanoes, but I’m not sure when the next Afar will be.

 

Fentale (and other volcanoes) update

There are a lot of volcanic situations ongoing. Kilauea has entered a series of fountain episodes that are getting gradually more intense as the conduit widens and matures, the eruption could easily go on for months and keep growing the lava fountains in height or maybe their height will max out at some point, some lava spurts are likely approaching 200 meters in height. Etna has erupted after several quiet months and is erupting a lava flow from a fissure, as always with Etna it’s hard to know if this will be a months-long effusion or a short event that will die out rapidly. Sundhnukur in Reykjanes should erupt soon too. There is a volcanic swarm ongoing near Santorini and Kolumbo volcanoes in Greece which is unusual as it’s happening away from the central volcanoes and could be a large dike intrusion, regarding volcanic outcomes, other than nothing happening, the most likely is probably a monogenetic volcano eruption, but at this point it’s unclear what the relation to the central volcanoes is and whether it could somehow cause them to erupt.

But the most important situation regarding this article is the Fentale dike intrusion. I was wrong to bring Dofen into the mix when I talked about the Fentale situation in my last article, although the phreatic explosion and mud volcanoes were next to Dofen it appears they were related to the Fentale dike, which is the only volcano involved. The dike itself did not directly erupt, which in hindsight should have been expected, but there have been developments at the caldera of Fentale volcano. Clouds of smog started to appear inside the caldera following the intrusion and seemed to be caused by degassing from a series of locations within the caldera floor, possibly from an incipient caldera fault. Yesterday another development took place, a M-6 earthquake occurred at the summit of Fentale carrying the same rare focal mechanism as the caldera collapse events of Kilauea and Bardarbunga. So it seems almost certain that there has been a small collapse of Fentale’s summit caldera (the ground sinking some meters). The dike drained magma from Fentale’s shallow reservoir which seems to have triggered yesterday’s collapse. A possibility is that collapse might eventually cause a rhyolite explosive eruption, similar to what happened to Askja volcano in 1875, but we haven’t seen any other such collapses happen which makes it hard to know if the Askja 1875 eruption is a usual outcome of the current situation, or a rare case, in which case an eruption would be unlikely. I think more dike intrusions are likely to take place from Fentale or even other volcanoes in the area in the future, which might eventually lead to eruptive activity.

*Edit: Investigating into the past of Afar volcanoes didn’t leave much time to be on top of the ongoing volcanic situations around the world, but thankfully a lot of people have been posting interesting details and data on Volcanocafé, that I don’t think would have been easy to find otherwise. Yesterday Mike Ross brought up the focal mechanism of the M 6 at Fentale, Squonk has been sharing seismograms of tremor events and other signals, jbean45 brought up the Fentale smoke, and Eolienne also shared seismic station data on Kolumbo, among a lot of other great comments for which I am very grateful, and would have otherwise been hard to stay up to date. 

References

Lahitte, P., P.-Y. Gillot, T. Kidane, V. Courtillot, and A. Bekele (2003), New age constraints on the timing of volcanism in central Afar, in the presence of propagating rifts, J. Geophys. Res., 108, 2123, doi:10.1029/2001JB001689, B2.

C.M. Martin-Jones, C.S. Lane, N.J.G. Pearce, V.C. Smith, H.F. Lamb, C. Oppenheimer, A. Asrat, F. Schaebitz (2017), Reprint of Glass compositions and tempo of post-17 ka eruptions from the Afar Triangle recorded in sediments from lakes Ashenge and Hayk, Ethiopia Quaternary Geochronology, Volume 40, May 2017, Pages 92-108, doi: https://doi.org/10.1016/j.quageo.2016.10.001