E-mail from C.L. Perason [2009-10-29]

Dear Mr Keenan,

Thank you very much for your email.

I will address each of your specific comments below, but first I'd like
to make a more general response.

This paper reports preliminary findings based on a small sample of trees
from Porsuk and is clear about this from the outset. Although
potentially exciting and sufficient to build a case for future work,
data from such a small sample of trees would never have been sufficient
to unequivocally prove a direct link between the growth anomaly and a
volcanic eruption, nor to specify the Minoan eruption as an exact source
- as we stated ourselves in the concluding sentence of our abstract:

"Although a volcanic association is implied, the new data are not
sufficient to prove a link to the exact eruption source."

In representing just four trees (all our budget would permit) we could
not begin to conduct the level of detailed analysis you correctly
advocate as required to draw any scientifically quantifiable conclusions
— hence in our text we do not make claims, rather we put forward
hypotheses. The data were however sufficient to provide a firm basis to
offer alternate hypotheses as to the origin of the Porsuk growth anomaly
and worthy of publication as they demonstrate the potential of
dendrochemistry to contribute to evidence for or against specific years
put forward by various scholars for the date of the Thera eruption. I
stand by our conclusion based on these data - exactly as stated in the
paper- that a direct volcanic association is implied and that this
hypotheses now requires rigorous testing.

As proposed in the paper - what is now required is new analyses of a
continuous sequence of tree-rings from the Aegean covering the entire
time period over which the Thera eruption may have occurred (i.e. dates
argued by proponents of both the high and low chronologies). This should
be at annual resolution, from as many trees as are available for
dissection in any given period. In this way we could test whether
elemental change is evident in just the rings of the Porsuk growth
anomaly or in other years which have also been put forward as possible
dates for Thera and other eruptions. Using an improved analytical
strategy we would aim to produce a full suite of elements (including the
rare earth elements and hard to get elements such as Cl) with which we
could then undertake further comprehensive objective analysis.

In this way we would hope to identify years of chemical change in tree
rings from the Aegean Bronze Age and then attempt to determine the
origin of that change based on what these more detailed well replicated
chemical analyses show us. We are currently preparing a funding
application to do exactly this plus isotopic analysis
(oxygen and carbon) of the same growth rings to establish a
separate high resolution record of climatic trends for this period. It
is however often difficult to obtain funding for this type of detailed
high resolution study due to the high costs of such analysis (c.$50 per
sample for full ICP-MS analysis of a single year and $20 per sample for
isotopic analysis of a single year). If we can successfully secure
funding I would very much like to incorporate the points you raise
(in particular re. Aniakchak as an alternate eruption source) as
alternate hypotheses to test.

I will now attempt to address each of your comments more directly:

> A central claim of your paper is that there is an increase in S at
> the time of the growth spike. Data for two trees is presented
> (fig.4): one tree shows a definite, but not huge, increase; the other
> tree shows almost nothing. Hence the S deposit at Porsuk does not
> seem to have been extremely large.

The amount of S actually present in the wood cannot be used to measure
the amount of S deposited by a particular pollution event. If a whole
range of growth environment variables (soil chemistry in particular)
were known, along with key physiological requirements of the species of
tree to be investigated, then it might be possible to model uptake of S
and other elements and thereby retrospectively attempt to quantify the
actual amount of acidic deposition required for the type of response
observed. The fact that for one of the trees there was an observable
increase of S and that for others there were increases of other
essential elements fits with a hypothesis of element mobilization
relating to increased environmental acidity. (Based on the findings of
50 or more studies on the impact of anthropologically induced acid rain
on terrestrial ecosystems).

This hypothesis is put forward in addition to evidence for relatively
major S deposition during the Bronze Age from at least three independent
paleoenvironmental records (e.g. Frisia et al. 2008, Siklosy et al. 2009,
Eastwood et al. 2002) - which all conclude the Thera eruption to be the
source.

All of this could be more fully explored by analyses of a wider sample
of trees from Porsuk. A model could be developed to indicate the likely
concentration of acidity required to mobilize particular elements and
how this would equate to proximal and distal volcanic acidity sources.

> Your paper claims that "Given the approximate date of the anomaly,
> the most logical source of sulfuric deposition would be volcanic, and
> given its proximity, Thera would seem the most likely candidate".
> The contrary seems indicated: Thera seems an unlikely candidate
> because it would be expected to deposit much more S, as well as
> substantial Cl--and no Cl is mentioned. (The issue with Cl is that
> HCl does not survive stratospheric transport: the presence of HCl is
> a near-conclusive sign of proximity to an eruption.) Too, HF is
> similar to HCl in this.

With regard your argument for the quantity of S evidenced from the tree
rings not being enough to be from Thera I hope I have answered this
above — with the limited dataset presented in the paper we cannot begin
to estimate 'how much' additional S was present at the onset of the
growth anomaly — just that there was a change in environmental acidity.
As to Cl — this element is particularly mobile in the terrestrial
environment and sparsely studied in terms of wood physiology, so has not
so far been a key focus. It is problematic to detect with ICP-MS due to
interferences from other elements, and as yet we have not been able to
gain sufficiently good detection via SXFM to find it. If we are able to
move forward with some further analyses this would be a particular
priority — as you rightly suggest — to establish or not the likelihood
of a 'local' eruption source.

> Given the approximate date of the anomaly, Aniakchak is a highly
> plausible candidate for a volcanic source. Yet your paper makes no
> mention of Aniakchak. Here is a scenario: Aniakchak erupted, injected
> lots of S into the stratosphere, some of which was later deposited
> at Porsuk; the volcanogenic aerosol (in the stratosphere) greatly
> increased precipitation and cloud cover in parts of the eastern
> Mediterranean (such increases are attested by other eruptions), and
> that, together with the acidic deposition, induced the tree-growth
> spike at Porsuk.

This is a valid alternate scenario that could be further explored,
however in the case of the existing paper please note that the primary
objective was to test whether any elemental change could be detected at
the onset of the Porsuk growth ring anomaly.

A change was observed so secondary hypotheses were put forward to
explain this observation. A volcano is a logical source of S in the
Bronze Age and a source in close proximity to the growth environment of
the wood is surely more likely than one on another continent - although
I take your point regarding the need to test this against the scale of
particular volcanic events. Future work should aim first to
unequivocally establish a volcanic cause for the observed change and
then go on to attempt more in-depth analysis to identify further
evidence to suggest which exact eruption. In other words, more complex
discussion of a volcanic source is best left until we have enough data
to do it justice.

On the basis of evidence at the time of writing, the paper suggests (not
claims) that Thera would be the most likely candidate as a source of S
deposition based on the following reasoning: According to field
observations of tephra deposition and associated isochrone maps, the
Porsuk trees were growing downwind of Thera, and so are likely to have
received some fallout when the eruption occurred. The fact that they are
only 400km away — the same distance (though not quite the same
direction) as a speleothem record in which a large increase in S
(concluded to be the result of the Thera eruption) has been
independently identified during the Bronze Age (Frisia et al. 2008)
makes this a logical assertion. The fact that other researchers have
linked acidic deposition directly with the Minoan tephra in lake
sediments (e.g. Eastwood et al. 2002 — albeit in that case not to
detrimental effect) can also be taken to support a logical argument for
Thera being the most probable source for a change in environmental
chemistry at Porsuk.

> Your paper makes little mention of volcano-climatic effects. The
> closest it comes to doing so seems to be this: "Rather than the
> indirect, albeit well-established effect of sulfur dioxide on climate
> (Robock and Mao, 1995), it seems plausible that the growth anomaly
> was caused by more direct volcanogenic impact, a type of
> fertilization effect, either from tephra deposition, or due to the
> soil/chemical impact of volcanogenic acid." Your paper presents no
> good reason to accept this. It presents data that seems to
> counter-indicate this. And it ignores that volcanoes have been known
> to induce precipitation increases in parts of the eastern
> Mediterranean region.

Discussions of the Porsuk growth anomaly in relation to climate change
resulting from volcanic impact have already been published and these
publications were cited in the paper before going on to suggest an
alternate hypothesis pertinent directly to the new findings being
reported. If we were able to undertake the further analysis I propose we
test the climate change hypothesis against the Porsuk growth rings by
looking at the oxygen and carbon isotope record for the same period — if
there was an increase in precipitation it would be logical to expect a
change in the isotopic record with the onset of the anomaly. Such
results would provide an opportunity to put together a new paper
regarding the potential climatic impact of various Bronze Age eruptions
on Aegean tree rings.

> Your paper additionally says that "Hf has been measured in the Minoan
> eruption deposits by numerous research teams": yes, but at less than
> one part per 100000; and tephra deposits at Porsuk were likely
> extremely tiny, as your paper states. Also, why would Hf become any
> more abundant than any of the other numerous elements? There was some
> Hf present at Porsuk prior to the eruption (fig.5), and biological
> processes that led to a rapid mobilization (as your paper states) of
> Hf seems to be a plausible a cause for the Hf abundance increase.
> Your paper appears to treat such mobilization cavalierly: Whilst it
> could be argued that the observed increase reflects the impact of
> increased environmental acidity on REE availability in the
> soil/regolith, fresh deposition of acidic volcanic ash seems more
> probable. For example, Hall et al. (1990) found increased REE
> indicative of known deposition from an eruption of Mt St Helens. How
> does the example of Hall et al. make fresh deposition of volcanic ash
> more probable?

The reason we suggest that fresh deposition
from volcanic ash might be more probable is the nature of the increase
in Hf which is abrupt and tied to the onset of the anomaly. If it were
the result of mobilization from pre-existing Hf in the growth
environment one would expect the increase to be gradual as the
acidity level increased and then lag as it slowly leached out over time.
The singular nature of the Hf increase is rather more indicative of a
sudden introduction of new material.

The answer to settle all of this is really to undertake further analysis
— I could find no references on the uptake and distribution of Hf in
tree species, thus a model explaining how and why it might specifically
become more abundant in tree rings as a result of acidification could be
developed. The fact that Hall et al. found rare earth uptake as a direct
result of known tephra deposition simply demonstrates that others have
found that this can occur.

> A related quote is this. “A recent study (Frisia et al., 2008)
> provides speleothem evidence for a major sulfate peak in the early
> Late Bronze Age (which the authors attribute to the Thera eruption),
> from a cave in northwestern Anatolia. The cave is located further to
> the north than Porsuk, but at approximately the same distance from
> Thera (see Fig. 1) and provides independent evidence of widespread S
> deposition across the region. Given this evidence, it seems
> reasonable to conclude that the root cause of the Porsuk growth-ring
> anomaly is not the indirect impact of an eruption on climate, but
> rather a direct impact from volcanogenic deposition.” This quote
> purports to argue against a distant eruption, but provides no actual
> reasoning.

I think I have covered the reasoning behind this in my previous comments
- It is agreed by most scholars that the Thera eruption occurred
somewhere between 1450 and 1700 BC. It is also agreed that it was a
relatively big eruption which dispersed S and tephra in a broadly
easterly direction. Sofular cave and Porsuk are both approximately
downwind of Thera. The Sofular cave record shows a big deposition of S
in the Bronze Age which they conclude to be from the Thera eruption. If
there was sufficient deposition at Sofular for this to be incorporated
in the speleothem record then it is not unreasonable to argue that there
would have been a significant increase at Porsuk at the time of the
eruption. Dating of the actual event aside — the Sofular record is
convincing evidence for me that the Minoan eruption resulted in
sulphuric deposition that would have been wide ranging enough to affect
trees at Porsuk. It does not rule out a more distant eruption but
logically, in my opinion, supports a local source.

> I also have a request: please publish all your raw data, e.g. on the
> web.

I completely agree with the idea of publishing the raw data but this is
not practicable. The complete raw datasets amount to many gigabytes of
data and make little sense until the raw spectra are processed into
elemental abundance.

Thank you for raising some interesting issues — I hope I have gone
some way towards addressing them.

Kind regards,

Charlotte.

--
Dr Charlotte L Pearson
Cornell Tree Ring Laboratory
Cornell University
Ithaca, NY 14853
USA
Tel: +1 607 255-8650
Email: c.pearson@cornell.edu
Web: http://dendro.cornell.edu/personal/charlotte.php