Dear Gerald,

I assume that many accretionary lapilli may have formed also in pdfs and may not have been frozen. The largest accretionary lapilli I saw occurred on the inner slope of the Surtsey crater Surtur I, where they reached 3 cm in diameter and contained vesicles in the outer layer. Similar ones, up to 2 cm in diameter I discovered in the palagonite beds on the eastern wall of the Askja caldera. I attach the publication on vesiculated tuffs which describes also these accretionary lapilli with vesicles in the outer layer.

Best wishes,

Volker

Dear Volker,

Thank you for the reply and paper.

I was actually referring to acc-laps that fall out from volcanic clouds.

Actually Jim Moore also showed me photos from one of his papers, years ago, of acc-laps that he colected at Surtsey. I do not recall all the details but I believe they were from fall deposits that Jim had sampled.

Besides, Sigurdur Thorarinsson, if I recall this correctly, directly observed and reported about "small" frozen acc-laps falling from the Surtsey volcanic clouds onto them. This was November and the ground temperature was close to zero degrees. Sadly this was the early 1960s and the samples were not described in detail or studied in the lab afterwards as far as I know.

Nonetheless of course many acc-laps are found in pdfs, as you know. The Laacher See deposits are a classic case in hand. Rolf Schumacher and Hans Schmincke did a fine study documenting acc-laps in tephra deposits including fallout, flow and surge deposits.  

In dense pdfs, streaming steam through hot fluidized pyroclastic deposits  can also accrete ash around a coarser pumice nuclei or a crystal and lead to very large acc-laps (a wet accretion growth process).

In locations such as those you just mentioned, the inner slope of the first Surtsey crater and the hydrated glass beds (palagonite) in the Askja crater, the context is also one where one can expect extreme fragmentation is due both to a phreatomagmatic fragmentation and to crater wall collapse, milling and recycling again and again into the exploding crater. I suspect one mechanism by which large acc-laps could be made up in this context is by fall and rotation of an acc-lap rolling down the wet (vesiculated ash) slope of the inner crater, even though one could envisage various other possible mechanisms not requiring any rolling on the wet ash slope (and rather up and down motion into the turbulent particle-laden water drop laden steamy mix in the crater with one explosion (or base-surge cloud) ultimately depositing the ash balls onto the inner crater wall slope..

In Surtseyan-style eruptions, George (GPL Walker) used to speak of the washing machine effect that goes on in the crater and involves crater wall collapse, refragmentation, increased mass loading in the jet, overloading of the crater wall etc. I studied this in the lab and it is just as George described. I presented the results at conferences (eg. CEV IAVCEI Workshop at Clermont Ferrand in 2009 or so) but there aren't in any papers. Similar processes where mass loading pulses due to crater wall collapse, recycling, milling etc.. also occur in violent strombolian eruptions (eg. Rowland et al paper focusing on the Jorullo eruption, in the GPLW special publication after the George Memorial Meeting in Iceland in 2005 or so; the last paper we worked on and finished with George).

I shall check your paper and get back to you (probably in a few days).

Many thanks again,

Best wishes and kindest regards,

Gerald

Hello Volker,

Thank you for reminding me of your classic study of vesiculated tuffs.

I am very grateful and I think it is crucial when exploring various mechanisms to go back to the original studies where diagnostic features were first described and interpreted.

In your milestone paper, the features of vesiculated tuffs and all associated features are described in very great detail and quite systematically. Sadly this is not often enough the case in many field studies.

The accretionary lapilli you described, for example those from Surtsey (inner crater of Surtur I crater) have a large basalt lapilli fragment as core, up to about 1.5cm across and a somewhat less extensive coat of vesiculated ash (5-10mm or so).

I cannot recall what exact terms have been used in the literature by some of our colleagues but I guess these have been described also as some form of "coated lapilli" or "armoured lapilli", rather than accretionary lapilli sensu stricto.

The terms in themselves are not the most important aspect to me. However I believe they likely correspond to different formation mechanisms.

Personally I recall having observed "armoured lapilli" closely similar to the ones you have reported from Surtsey (and other sites) in vesiculated tuffs from the inner crater of Irazú volcano in Costa Rica (see PhD by Guillermo Alvarado). These were also emplaced by Surtseyan style eruptions that occurred there in alternance with more dry magmatic phases during the 1965-1967 eruptions (occurred on and off for 2 years, quite like in Surtsey).

I have reread your interpretation of those "armoured lapilli" and I agree that it fits all the observations and measurements you have made so carefully and systematically.

Regarding your seminal interpretation of vesiculated tuffs, citing from your article so that anyone willing can follow the discussion:

"The process of fall-out of ash from a base surge can be looked at in terms of defluidization of such a fluidized system. In order for the gas phase of the cavities to be trapped cohesion of the wet ash particles must increase rapidly up to a level where massing or packing of the particles is sufficiently dense to prevent escape of gas of the cavities, ie. the cavities must be enclosed completely by ash particles and liquid that binds the particles together". (p. 287, in: Lorenz V (1974) Vesiculated tuff and associated features. Sedimentology 21: 273-291

In addition, larger armoured lapilli are at the lower points and smaller armoured lapilli found further up and this, as you also interpreted, is consistent with the rolling down of the basalt fragment in the wet vesiculated tuff and accretion of this wet and vesiculated ash to form a coat of vesiculated ash around the basalt fragment core for each of the armoured lapilli.

All this is consistent with the armoured lapilli having a large fragment of basalt as the core and a rather irregular and thinner coat of vesicular ash....so that there are not particularly spherical in shape...to me they are of a different shape to accretionary lapilli sensu stricto that I have studied (and equally important and interesting).

Amongst other deposits where vesiculated tuffs were subsequently described, there is an interesting study by Mauro Rosi where, in that particular case, Mauro could make a convincing case that vesiculated tuffs can also form where accretionary lapilli fell out accompanied by too much water, so that the acclaps got all smooshy and trapped air where space had originally been present between the acclaps. I have personally observed similar vesiculated tuffs with acclaps in the US1 deposits from Santorini, where this may also have worked.

In your paper you have also documented several types of mudflow or slurry flow erosional channels up to 50cm wide. In the US1 deposits of Santorini, similar erosional channels occur at the top of the US1 sequence. There these are also close to vent and up to 6m or so wide and 1-2m deep or so. George (GPL Walker) was able to document such channels systematically cutting through classic phreatomagmatic phase deposits at Taupo. There he documented that the width and depth of gullying decreased exponentially with distance from vent. Even though this seems to be the case also for the US1 deposits, the George study at Taupo is the only one I am aware of where the exponential decrease in gullying was ever quantified.

A key point is that at Taupo and at US1 (and I think also in your reported observations) gullying occurs at the end of a phreatomagmatic phase of deposition (which may resume some time later).

In your paper you speak of en masse deposition from the inflated base surges. The mass loading rapidly increases during deflation so that single ash particles do not settle individually. Rather a convective instability is expected to occur and downward plumes of ash move towards the ground, resulting in hindered settling, which as you know is equivalent to fluidization in the features it produces.

Regarding some of the gullies you describe, something similar to what I have considered (with Adam Durant) for the US1  deposit gullies can be invoked. When successive base surges are generated, a co-surge plume rises above the vent. Often it adds to a plume generated by explosions also responsible for the base surges. When a main sequence of such activity sudddenly stops, all the fine ash that has recycled for a while in the plume, often aggregated and frozen (so that in some cases there can be a huge amount of recycling water on ash) is no longer supported from below. A convective instability occurs with downgoing plumes of ash and water crashing down  from high up within minutes. The result is the US1 or Taupo gullies as well as potentially very hazardous mud flows. I believe this was the case during the Rabaul 1937 phreato-subplinian eruption where mud flows dug 100 feet deep gullies according to reports.

Could it be that some of your larger gullies and mud flows at Surtsey also formed in this way ?

Thank you again for all the crucial and systematic measurements you made at Surtsey and elsewhere. They certainly inspired my own work.

Clearly there are ash balls and ash balls....and yet more types of ash balls !

There is more to ash balls than what first meets the eye.

I am still curious as to whether anyone has collected large acc-laps and studied them immediately after they fell. Thank you for sharing any insights you may have.

Kind regards to all,

Gerald

Hi Gerald and Volker,

very interesting and in-depth discussion.

Let me point out one paper by Richard Brown and co-workers (https://www.researchgate.net/publication/257603831_A_review_of_volcanic_ash_aggregation) and my AGU abstract (https://www.researchgate.net/publication/258467915_Experimentally_constraining_the_boundary_conditions_for_volcanic_ash_aggregation).

One of my PhD students has succesfully reproduced accretionary lapilli in a lab environment, using fluidised bed technology. This paper is currently in review.

Furthermore, as a general comment, I disagree with using accretionary lapilli as a proxy for phreatomagmatic fragmentation. I have seen them in related deposits, but also in deposits from dry/magmatic activity. I have seen and samples accretionary lapilli from a magmatic eruption of Tungurahua, Ecuador. I don't believe that freezing had a significant influence on their generation.

Can I propose an ash aggregation related meeting?

Two possible opportunities come to my mind:

- EGU 2016

- 3rd workshop of VERTIGO

Article A review of volcanic ash aggregation

Conference Paper Experimentally constraining the boundary conditions for volc...

Ooops, I misused the keyboard....

Anyway, I was nearly done, we will be hosting the third workshop of Vertigo in May next year in the Eifel in Germany (see below and click on EVENTS, more info will follow soon).

I would be very happy if we could meet directly and discuss this issue "live". I would bring my PhD student (Sebastian B. Mueller) along.

Best regards, Ulli

http://vertigo-itn.eu/

https://www.researchgate.net/profile/Sebastian_Mueller11

Dear Ulrich,

Thank you for the helpful reply and diverse infos. I shall look into all of it.

I just want to clarify in which case freezing appears necessary to me.

Ash aggregation can occur in a variety of context:

1. In hot dense fluidized pyroclastic flows. Clearly no ice here.

2. In Surtseyan deposits such as those described by Volker, armoured lapilli form (acc-laps can also form...). As discussed in the 2 previous exchanges, here no need for freezing of any kind to account for the armoured lapilli.

3. In clouds ascending above base surges or above pyroclastic flows generally (co-ignimbrite clouds) or in the case of ascending vertical clouds above vent, ash aggregation also takes place and the ash aggregates then fall out.

If the atmosphere is pretty dry, low density electrostatically-bound ash clusters are produced (work by Mike James, etc..).

If the cloud is water vapour rich or if the atmosphere is moist, dense ash aggregates are observed.

By wet accretion, ash aggregates or water drops (to produce rain) cannot grow to a size exceeding 5-6mm diameter, from the classic cloud physics literature.

The reason is a hydrodynamic instability. Shear drag forces tear these large drops apart. The break-up probability increases exponentially as size increases and no stable drop can exist above 6mm. So if ash-filled drops are not frozen a little before reaching the drop break-up limit, dense ash aggregates exceeding 6mm across are not expected to be able to form. They cannot form by wet accretion above 6mm across.

It is in that case that I believe freezing and growth by riming as in hailstone formation takes over at that point. Besides I presented numerous observations and measurements consistent with this growth model during the IAVCEI 2004 international conference in Chile. This is already very long ago.

I am of course curious about your observations at Tungurahua and shall look them up.

Great many thanks.

Best wishes and kindest regards,

Gerald

Hi Gerald, will you be attending AGU in San Francisco? Sebastian will be presenting our results. I believe there is a way to allow ash to aggregate (with water/moisture/humidity present) without thinking of water drops collecting ash and allow for significant growth. Ulli

Hi Ulli,

No I shall not attend AGU. If you can invite me to Vertigo (at Laacher See in May ?) I may be able to attend and participate actively.

Clearly it would be good to invite the likes of  Volker Lorenz, Mauro Rosi, Rafaello Cioni, Jim Moore if he is still about and active, Jenni Gilbert and Steve Lane, Rolf Schumacher and Hans Schmincke, Adam Durant, Bill Rose, Steve Self, Mike James, Costanza Bonadonna, Michael Herzog (ATHAM)...(and anyone I have forgotten to mention), as well as experts on hailstone formation and precipitation dynamics in thunderstorms.

I still have to look up all the info you sent. I'll get back to you after this.

Best,

Gerald

Hi Gerald and Ulli, sorry for not having answered earlier or in between. I am trying to finish 2 manuscripts - the deadline is approaching fast.   A Vertigo workshop in the Eifel next year sounds highly attractive.

Best wishes

Volker

Hello Gerald, Ulli, Volker,

There quite a few nice examples I'm aware of where accretionary lapilli were sampled soon after fallout. One is the 2002 study by Bonadonna et al. on the 1995-1999 eruption at Montserrat (Geol. Soc. London Memoirs). Another is the 2009 eruption of Redoubt, Alaska, described by Kristi Wallace et al. in JVGR, 2013. The latter samples provided the basis of my study of 'volcanic hail' using radar observations and plume modeling with ATHAM in Nature Communications. As suggested by Gerald, these samples were indeed frozen upon landing, though that's not too surprising for an Alaskan eruption. It is clear from lab experiments that neither ice nor salt precipitates are required to grow aggregates >1cm diameter. I do not think aggregates need to undergo freezing, however, it does seem to be very common and does affect their growth rates and internal structure. 

There are also some gorgeous, layered accretionary lapilli from the 18 May 1980 eruption of Mount St. Helens, which have been archived for the past 35 years. I've recently thin-sectioned these samples and am currently working on comparing their internal structures to the eruptive stratigraphy and photographic record of the eruption dynamics. Hopefully I'll be presenting this work at the Cities on Volcanoes meeting in Chile this year. Ulli and I were talking about organizing a workshop on ash aggregation there if there's enough interest. 

All the best

Alexa