New geoengineering article submitted

7 views
Skip to first unread message

Alan Robock

unread,
Mar 2, 2008, 11:58:52 AM3/2/08
to geoengi...@googlegroups.com
Dear Colleagues,

Luke Oman, Georgiy Stenchikov and I have completed a paper describing
our first climate model simulations of sulfate aerosol geoengineering,
and you can download it from my web site. The paper describes
the results I have already presented at AGU, AMS, and NCSE conferences.
Comments would be welcome.

Robock, Alan, Luke Oman, and Georgiy Stenchikov, 2008: Regional climate
responses to geoengineering with tropical and arctic SO2 injections.
Submitted to J. Geophys. Res.

http://climate.envsci.rutgers.edu/pdf/GeoengineeringJGR7.pdf

Alan

Alan Robock, Professor II
Director, Meteorology Undergraduate Program
Associate Director, Center for Environmental Prediction
Department of Environmental Sciences Phone: +1-732-932-9800 x6222
Rutgers University Fax: +1-732-932-8644
14 College Farm Road E-mail: rob...@envsci.rutgers.edu
New Brunswick, NJ 08901-8551 USA http://envsci.rutgers.edu/~robock

Alvia Gaskill

unread,
Mar 2, 2008, 4:25:00 PM3/2/08
to geoengineering
1. Line 92. The Mt. Laki eruption is probably irrelevant to the
analysis performed. It mostly involved the troposphere, not the
stratosphere. A better example of an Arctic eruption would be Katmai.

2. Lines 96-101. The contribution of Pinatubo to a reduction in
global precipitation is somewhat controversial. Shouldn't you respond
to the findings that a significant proportion was due to a concurrent
El Nino?

3. Line 100 "unbalanced."

4. Lines 102-108. What would be the impacts of the increased ozone
depletion? The reader is left thinking it will be catastrophic.

5. Lines 130-143. True that the Brewer Dobson circulation moves the
aerosol towards the poles, but that doesn't mean that releasing 3,000
tons of precursor per day at 30N and 3000 tons at 30S will be caught
by and diluted to a common level with 1000 tons per day released at
the equator, all releases occurring at the same time. The fact this
hasn't been observed with previous volcanic eruptions is because there
hasn't been a case where you had lots of volcanoes erupting at
different latitudes at the same time. So Teller's idea and mine also
that a non uniform release latitudinally, longitudinally, vertically
and temporally has not been voided by this work. (Lines 130-136).

6. Lines 187-193. As noted, the models do not consider the
possibility of an equilibration of the climate with a steady increase
in aerosol as would actually occur. Thus, the modeling is deficient
in addressing this important reality.

7. Lines 200-201. What was the basis for selecting 3Mt of SO2 (1.5Mt
S) for the Arctic model? Was that arrived at by running the model to
see what level would be required to produce a given reduction in solar
radiation? For just the area from 68N to the pole, that's a lot of
aerosol. I thought I read where Caldeira and Wood had proposed
300,000 tons of SO2.

8. Line 210. Although estimates of the Pinatubo eruption vary, I
think there is now a consensus that while the eruption injected 20MT
of SO2 into the atmosphere, not all of that made it into the
stratosphere or stayed there very long and only about 12MT actually
contributed to the global aerosol clouds.

9. Line 215. "lowermost stratosphere."

10. Lines 216-221. The Arctic injection could be easily accomplished
using 747's or our soon? to be retired fleet of KC-135's (assuming no
last minute bidding scandals emerge). Stratospheric balloons and
military fighters can easily reach the midpoint of the altitudes
chosen for the global releases (65,000 ft). So the technology to do
this exists in 2008.

11. 251-257. I am quite surprised by the hemispheric spread of the
Arctic generated aerosol, given the low altitudes of its creation. Is
there evidence from Katmai or other northern eruptions that aerosol at
33,000-50,000 ft would flow south so rapidly?

12. Line 296, delete "the" at end of line.

Alan Robock

unread,
Mar 2, 2008, 5:24:23 PM3/2/08
to Alvia Gaskill, Georgiy Stenchikov, Luke Oman, geoengineering
Dear Alvia,

Thanks for the quick reading of our paper. Please see my responses
below.

Alan

Alan Robock, Professor II
Director, Meteorology Undergraduate Program
Associate Director, Center for Environmental Prediction
Department of Environmental Sciences Phone: +1-732-932-9800 x6222
Rutgers University Fax: +1-732-932-8644
14 College Farm Road E-mail: rob...@envsci.rutgers.edu
New Brunswick, NJ 08901-8551 USA http://envsci.rutgers.edu/~robock


On Sun, 2 Mar 2008, Alvia Gaskill wrote:

>
> 1. Line 92. The Mt. Laki eruption is probably irrelevant to the
> analysis performed. It mostly involved the troposphere, not the
> stratosphere. A better example of an Arctic eruption would be Katmai.
>

Please read our papers on Katmai and Laki, referenced in this paper and
all available at

http://www.envsci.rutgers.edu/~robock/#Publications:

Actually, 85% of the Laki emissions went into the stratosphere.

> 2. Lines 96-101. The contribution of Pinatubo to a reduction in
> global precipitation is somewhat controversial. Shouldn't you respond
> to the findings that a significant proportion was due to a concurrent
> El Nino?
>

There are indeed multiple causes of climate change, and just looking at
one case cannot unravel the various contributions. I would not say this
is controversial, however. That is why we do climate model simulations
to be able to control the different factors. In any case, Pinatubo was
a large eruption and produced a large forcing, while the El Niño was
rather ordinary and only produced large effects on precipitation around
the Pacific.

> 3. Line 100 "unbalanced."
>

Thanks for catching this.

> 4. Lines 102-108. What would be the impacts of the increased ozone
> depletion? The reader is left thinking it will be catastrophic.
>

Simone Tilmes and colleagues at NCAR presented model simulations at the
AGU session I organized in December 2007 that were indeed catastrophic.
I understand they are working on a paper on this now. We have not done
ozone calculations.

> 5. Lines 130-143. True that the Brewer Dobson circulation moves the
> aerosol towards the poles, but that doesn't mean that releasing 3,000
> tons of precursor per day at 30N and 3000 tons at 30S will be caught
> by and diluted to a common level with 1000 tons per day released at
> the equator, all releases occurring at the same time. The fact this
> hasn't been observed with previous volcanic eruptions is because there
> hasn't been a case where you had lots of volcanoes erupting at
> different latitudes at the same time. So Teller's idea and mine also
> that a non uniform release latitudinally, longitudinally, vertically
> and temporally has not been voided by this work. (Lines 130-136).
>

I am not familiar with this idea of yours and Teller's. Please send me
the references to your journal articles.

> 6. Lines 187-193. As noted, the models do not consider the
> possibility of an equilibration of the climate with a steady increase
> in aerosol as would actually occur. Thus, the modeling is deficient
> in addressing this important reality.
>

I am not sure what you mean by "important reality." As we point out in
the paper, such scenarios could be investigated, but first you have to
decide on how you want to control the climate. Do you want to keep it
constant? Cool it back to pre-industrial times? Tom Wigley already
used a gradually ramped up aerosol loading in his 2006 Science paper,
and other presented such scenarios at AGU. What we present is the
climate equilibration to a constant SO2 emission and gradually changing
anthropogenic forcing. The results are as interesting as any other
hypothetical forcing scenario.

> 7. Lines 200-201. What was the basis for selecting 3Mt of SO2 (1.5Mt
> S) for the Arctic model? Was that arrived at by running the model to
> see what level would be required to produce a given reduction in solar
> radiation? For just the area from 68N to the pole, that's a lot of
> aerosol. I thought I read where Caldeira and Wood had proposed
> 300,000 tons of SO2.

It was an estimate of what would keep the Arctic from warming,
considering our experience with modeling high latitude eruptions. If
you could confine the sulfate aerosols to just a polar cap you would
need less, but that is not how the atmosphere works. There are no walls
to confine the aerosols over the pole.

>
> 8. Line 210. Although estimates of the Pinatubo eruption vary, I
> think there is now a consensus that while the eruption injected 20MT
> of SO2 into the atmosphere, not all of that made it into the
> stratosphere or stayed there very long and only about 12MT actually
> contributed to the global aerosol clouds.
>

That is not correct. As Bluth et al. [1992] measured with TOMS, about
20 Mt is what the eruption put into the stratosphere.

> 9. Line 215. "lowermost stratosphere."
>
> 10. Lines 216-221. The Arctic injection could be easily accomplished
> using 747's or our soon? to be retired fleet of KC-135's (assuming no
> last minute bidding scandals emerge). Stratospheric balloons and
> military fighters can easily reach the midpoint of the altitudes
> chosen for the global releases (65,000 ft). So the technology to do
> this exists in 2008.

This is not correct. The ceiling for a 747 is only 45,000 ft.

http://www.boeing.com/history/boeing/747.html

The ceiling for the KC-135 is 40,000 ft.

http://www.dfrc.nasa.gov/gallery/photo/KC-135/HTML/index.html

Even if you flew all the KC-135s as high as you could as often as you
could, how much mass could you put into that highest layer per year?

How do you define "easily" for balloons? How many would you need and
how many would you need every day? How much would it cost.

Please explain what military fighters you mean that can get to 65,000
ft. Even if they can, they have a tiny payload. How many would you
need and how often would they have to fly? Furthermore, once they have
started an acid cloud, how could subsequent planes safely fly into this
cloud?

For all the airplanes, if they are fully loaded they cannot fly as high,
and the ceiling in the Arctic must be lower, because the air is less
dense at the same height as lower latitudes.

>
> 11. 251-257. I am quite surprised by the hemispheric spread of the
> Arctic generated aerosol, given the low altitudes of its creation. Is
> there evidence from Katmai or other northern eruptions that aerosol at
> 33,000-50,000 ft would flow south so rapidly?
>

We know of no observations of the distribution of the cloud, but would
like to find out. However, the climate response, the same as we
modeled, is clear. The lowest flow in the period 1900-1980 of both the
Niger and Nile Rivers was in 1913, the year after the Katmai eruption,
caused by the impact of the aerosol cloud on the African monsoon. After
that, the Sahel drought caused slightly lower flows, and the Aswan Dam
stopped the Nile record in 1967.

> 12. Line 296, delete "the" at end of line.
>

Thanks.

Alvia Gaskill

unread,
Mar 2, 2008, 8:58:07 PM3/2/08
to geoengineering
1. We need to get clear on what is the stratosphere and what is the
troposphere.

According to http://seismo.berkeley.edu/~manga/LIPS/thordarson03.pdf,
most of the Mt. Laki aerosol was formed and stayed between 9-13Km, a
region referred to in the paper as the upper troposphere/lower
stratosphere (30,000-43,000 ft). However, this is in the same
altitude range as your Arctic example.

4. I think it is irresponsible to toss around words like
"catastrophic" without giving some context. Catastrophic to whom or
what? Catastrophic to some of the ozone molecules, no doubt, but what
about the rest of us? The media will seize on words like catastrophic
even though your paper does not address ozone depletion.

5. The reference to Teller came from lines 130-134 of YOUR paper. I
have proposed this as an idea. It needs to be verified. I think it
is important to determine what works and what doesn't so we won't
waste lots of time and resources. I am somewhat bemused by scientists
who seem mainly concerned with pointing out what won't work. This is
important also, but so is finding a solution that will work if that is
possible. If all the research into geoengineering is confined to
modeling studies that limit the conditions to outcomes that are
unfavorable, then we will never know the truth.

10. Your Arctic scenario called for aerosol to be created between
33,000 and 49,000 ft. The 747 could easily achieve the midpoint of
that range. It would not be able to sustain flight at 49,000 ft. The
KC-135 can fly above 50,000 ft. The example you chose is incorrect.
Also, why does all the aerosol precursor have to be created in "the
highest layer?" Didn't seem to matter to Mt. Laki.

http://www.airforce-technology.com/projects/kc135/specs.html

I've run the numbers on the KC-135 and the B-52 and the F-15c and the
MIG-31. It is possible to carry sufficient payload (4-5 tons H2S per
flight for the fighters) to meet the demands of a global aerosol
program with the fighters and the Arctic program could be done with
the other planes. The fighters would fly 4-5 times per day. The
existing F-15s are in need of refurbishing and would be stripped of
unnecessary weight for this single use. The cost of the fighter
program would be on the order of $5 billion per year. I refer you to
the files section of this group for more details. Stratospheric
balloons could be used with the gas blended in with the lifting gas.
In your paper, you used SO2. I would use H2S and for the lifting gas
H2. Balloons are cheap. They are made of plastic sandwich bag
material. The costs would probably be less than that for the
airplanes.

We now routinely deliver payloads to 120,000 ft of up to 8000 lbs with
stratospheric balloons. If one balloon can deliver 4 tons and the
daily requirement is 10,000 tons, then you need 2500 balloons per
day. That's a lot of balloons, but since the balloons can reach
90,000 ft, the aerosol lifetime will be greater than at 75,000 ft, so
less will be needed. Whether it's 500 planes per day or 1000 balloons
or some combination, it will not be easy to run a stratospheric
aerosol program to completely offset all man-made GHG forcing. The
point is that it is possible, contrary to your conclusion.

As to flying in the acid cloud and in the thin air of the Arctic,
these concerns would need to be investigated. Although the impact of
volcanic ash is well known and understood, the effect of sulfuric acid
aerosol is not. Corrosion of acrylic windows, the air frame and
engine parts will have to be considered. See p. 5 of
http://www.flightsafety.org/fsd/fsd_may93.pdf

The fighters would probably not be up there longer than an hour nor
would the 747s. It is true that service ceilings for aircraft are
somewhat dependent on the weight of the plane. However, the 747 is a
very large airplane and the additional weight of the gas would not
significantly degrade its ceiling in the Arctic. The fighters combat
ceilings include the full weight of armaments that would be replaced
by the H2S tanks. So 65,000-70,000 ft fully loaded is possible.
> >> New Brunswick, NJ 08901-8551  USA      http://envsci.rutgers.edu/~robock- Hide quoted text -
>
> - Show quoted text -

Alan Robock

unread,
Mar 2, 2008, 9:17:08 PM3/2/08
to Alvia Gaskill, geoengineering
Dear Alvia,

Thanks for the additional information. We will certainly use these an
other comments to revise our paper before publication.

The ceilings were referring to your suggestion of 65,000 feet, but I now
see that you were referring to the tropics. Yes, the 747 and KC-135
could fly into the Arctic lower stratosphere.

We plan to evaluate many other potential scenarios. We are not out to
prove what can't be done. We are out to evaluate different suggestions
for emission scenarios and we are just beginning.

I was just responding to your use of "catastrophic" for the ozone
effects, and we do not use that word in the paper. We do use the word
for the effect of a large reduction in Asian monsoon rainfall, and will
have to consider whether or not to rephrase it.

Tom Wigley

unread,
Mar 2, 2008, 11:18:09 PM3/2/08
to rob...@envsci.rutgers.edu, Alvia Gaskill, geoengineering
Dear Alan,

I too would advise against the use of the word "catastrophic".

I do not think we know enough about the impacts of any change in
the monsoon (changes in interannual variability may be more
important than changes in the mean) to use any definitive adjective.

This is clearly an area where more research is needed. Peter Webster
has done relevant work.

Tom.

+++++++++++++++++

Alan Robock

unread,
Mar 3, 2008, 9:39:41 AM3/3/08
to Tom Wigley, Alvia Gaskill, geoengineering
Dear Tom and Alvia,

Thanks for your comments. Because of a bug in the JGR system, I am not
completing the paper submission until today, so I changed the manuscript
to fix the typos Alvia found and change the word "catastrophic" to
"serious." The revised manuscript is now posted on my home page at

http://climate.envsci.rutgers.edu/pdf/GeoengineeringJGR7.pdf

Alan

Alan Robock, Professor II
Director, Meteorology Undergraduate Program
Associate Director, Center for Environmental Prediction
Department of Environmental Sciences Phone: +1-732-932-9800 x6222
Rutgers University Fax: +1-732-932-8644
14 College Farm Road E-mail: rob...@envsci.rutgers.edu

New Brunswick, NJ 08901-8551 USA http://envsci.rutgers.edu/~robock

Alvia Gaskill

unread,
Mar 4, 2008, 7:56:04 PM3/4/08
to geoengineering
Some additional comments and questions about the impacts on
precipitation cited in the paper and the presentation format in
general.

1. I know it is the convention to use these color coded Mercator
projections of the globe to show temperature and precipitation shifts,
but they are very hard to read and reconcile with conclusions made in
papers, not just this one. I had to blow the figures up 400-600% to
clearly identify the areas of note. Perhaps papers would benefit from
a table that augments these figures when changes are found. Another
possibility is to "subtract" the control from the one with the altered
forcing so that the differences can be more clearly discerned.

2. In the case of this paper, the precipitation changes were also
identified by marking them with the deltas, -0.25, -0.50, etc.
However, it seemed to me that the colors didn't always correspond to
the numbers shown. In my following comments, I refer to the numbers
and if they are not what the conclusions are based on, please correct
me. All the comments refer to Figure 8.

3. I've only looked at India and the authors may also need to review
Bangladesh (not mentioned in the paper, but in a zone with what
appears to be a severe reduction for June, July, August (JJA) for the
5Mt case, Northeastern China and the U.S. Midwest (not mentioned in
the paper), and central Africa.

4. The reduction for India is stated to be 0.5mm/day for JJA, which is
45mm. There is essentially no reduction for December, January and
February (dry months anyway) and no data for the rest of the year.
Assuming that the -0.5 applies to 6 months, that's 91mm for the year.
Actually, the total delta shown for the year is almost 0, so 91mm may
be too high.

According to the Indian Meteorological Dept. (IMD)http://
www.imd.ernet.in/section/nhac/dynamic/endofmonsoon.htm, the average
precipitation for the swath of land identified in the paper as at risk
from the aerosol strategy is between 700-1000mm/year. The 91mm
reduction then corresponds to between a 9-13% reduction in
precipitation, a deficit that the IMD classifies as within the normal
range (+/-19%).

If all this is true, then how can one conclude that the aerosol at 5Mt
SO2 per year poses a threat to the Indian food supply? Note also that
not all Indian agriculture is based on precipitation generated
rainfall. A lot of crops are irrigated using river water fed by snow
melt in the Himalayas that would either benefit from or not be harmed
by the aerosol.

5. If the 10Mt case were examined, the reductions might be more
severe. A 5Mt aerosol loading would be enough to return forcing to
pre-industrial if applied today or enough to offset forcing added
between 2000 and 2030. However, most of the predictions of severe
climate change driven droughts in Asia are based on the time period
2050 and later, so perhaps the conditions and time period used in this
study were not adequate to model the consequences.

6. I note also the reference to the famine due to Mt. Laki in the
discussion of the Arctic aerosol strategy. This eruption was nearly
30 times larger (in terms of aerosol) than the 3Mt of SO2 used in the
model and would still be 8 times larger than a Pinatubo size aerosol
program if compared to a global strategy. Comparisons with Mt. Laki
are therefore irrelevant and the authors should seek out information
on the famine if any produced by Mt. Novarupta, a comparable sized
eruption to the proposed Arctic strategy which injected most of its
precursor SO2 into the Overworld and not Upper Troposphere/Lowermost
Stratosphere.
> > +++++++++++++++++- Hide quoted text -
Reply all
Reply to author
Forward
0 new messages