Sunday, April 8, 2012

3176.txt

date: Mon, 11 Aug 1997 21:23:03
from: Sirje Keevallik <sirjeatXYZxyzbus.emhi.ee>
subject: unep
to: m.hulmeatXYZxyz.ac.uk

Mike,
I try to send you by Attachment a very first draft of my piece.
Please feel free to correct my mistakes as well as to tell me if
something is missing or something should be removed.

Additionally, I have got one question, two comments and two problems.

1. My version of SCENGEN contains 14 GCMs. In your text you show that
the number is 11. Which GCMs have been left out?

2. You should point out in your text more clearly that CHANGES can be
estimated only on a grid of 5 degrees and that a resolution of 0.5
degrees can be obtained only when these changes are added to
basic climatology.

3. At first I tried to get scenarios that correspond to extreme
parameter combinations (MAGICC + emissions scenarios). This ended up
with rather unbelievable results: According to ECHAM3TR, the maximal
warming in January will be 17 degrees by 2100. To my mind, this
refers to the possibility that regional patterns are not constant
when scaling with large values of global warming projections.

4. What to do with aerosol? At the present version of my text there
is a discrepancy between MAGICC (that takes aerosol inti account) and
regional patterns (that do not).
Is this a very stupid idea, when I find an earlier year (instead of
2100) that corresponds to a global warming WITH aerosol and ascribe
this pattern to 2100? More clearly: High emission/high climate
sensitivity gives by 2100 3.5 K with aerosol and 4.7 K without.
Without aerosol the warming of 3.5 K will be reached by 2077. Can I
regard this as year 2100 WITH aerosol? I repeat that in Estonia the
influence of aerosol should be significant.

5. People here want to have one central scenario instead of two. Can
I recommend to take precipitation estimates from HadCM2 and
temperature estimates as averages between HadCM2 and ECHAM3TR?
If you look at the histogram we constructed at Norwich you see that
HadCM2 predicts central estimate of precipitation among all models.
This feature is also characteristic of similar histograms of
different seasons.

Best regards,
Sirje

No comments:

Post a Comment