date: Wed, 07 Jul 2004 08:29:44 -0600
from: Tom Wigley
subject: sea level
to: Sarah Raper
Sarah,
Just to clarify a few things ...
OK about de-drift. I presumed you did Def. 2, but I just wanted to be
doubly sure.
By non-melt I mean the 'I', 'p' and 's' terms (see TAR p. 682,3)--- NOT
expansion. These terms are *defined* to be linear in time.
So are you saying that total GSIC ice of 19cm may be OK if Antarctic
(mainly Ant. Peninsula) (AGSIC) and Greenland GSIC (GGSIC) are not
included? You also imply that the latter could be as big or bigger than
the 'normal' GSIC.
I'm sure you have read the literature much more thoroughly than I have,
but this doesn't seem to be right. For instance, in the report by
Mark Dyurgerov that I mentioned before ("Glacier Mass Balance and
Regime: Data of Measurements and Analysis", INSTAAR Occ. Paper
No. 55, 2002) -- which I have only skimmed through -- he gives the
following area data: total (incl. AGSIC and GGSIC) = 680,000 km2;
AGSIC = 77,000 km2; GGSIC = 70,000 km2. So based on area it
seems that AGSIC and GGSIC add about only 28% to the 'normal' total.
The 680,000 value, by the way is as in TAR Table 11.3, which, according
to the Table footnote, does include A+GGSIC. This Table gives 140,000
km2 as the A+GGSIC term, a bit less than Dyurgerov.
If your 15cm for A+GGSIC is OK, and 19cm is the normal GSIC, then
you would have A+GGSIC adding 79%. I realize that area and volume
are different, but 25% vs 79% seems a big difference. The implied TAR
result is even more odd -- they give 50cm as total volume (see below)
so, if 19cm is correct, then the A+GGSIC part is 31cm!
Of course, this is all moot because the 19cm comes from the TAR
GSIC formula, and this applies to*all* GSIC including A+GGSIC.
I can see no way to justify the TAR formula.
In my last point, as in the text preceding it, I was still talking about
the total
GSIC volume. The TAR gives a formula for GSIC uncertainty, and, if
the central value implies a total volume of 19cm, then it is easy to get
the implied bounds for this total volume: viz. low = 5cm, high = 32cm.
(All the numbers here are relative to 1880.)
In any simple global formula, there should be at least two clearly
identifiable
sources of uncertainty. One is the sensitivity (d(melt)/dT) and the other
is the total available ice. In the TAR, the latter never comes into it
in their
analysis (i.e., the 'derivation' of the GSIC formula) -- but my point is
that it
*does* come in by accident due to the quadratic fudge factor. The total
volume range is 5-32cm, which is, at the very least, inconsistent with
other material in the chapter (see below). 5cm is clearly utterly
ridiculous.
Of course the TAR is doubly inconsistent here -- in the 'derivation' of the
GSIC formula, ice volume uncertainty is not directly considered. However,
Table 11.3 gives the total volume as 50cm +/- 10cm. (I am not sure what
the +/- refers to --- is it 1 sigma or 2 sigma? Perhaps this information is
in the text, but it should be given in the Table captions. Do you know?)
You appear to defend Jonathan in two ways. You seem to be saying
that the TAR formula, with its 19cm bound, might only apply to the GSIC
part that excludes A+GGSIC. I do not this this defence can be justified.
You also say that the GSIC formula (and other TAR formulae?) are only
meant to be applied out to 2100. This is more justifiable, but there seems
to be no clear statement that the GSIC (or any other) formula applies only
to 2100 -- perhaps you can tell me where this is? On p. 677 it says
that 'we cannot use the 21st century rate to deduce that there is a time by
which all glacier mass will have disappeared'. A generous person might
interpret this as saying that the formula used to get the 21st century rates
cannot be used beyond the 21st century. Do you know if there is a clearer
statement than this?
I'm sorry to keep on about this, but I am really searching for a way
to get more sensible long term results. The best I can do, I think, is
replace the TAR GSIC formula by one that agrees up to 2100 and then
has an asymptotic (very large warming) melt of 50+/-10cm. This would
at least be consistent with the TAR, and avoid the TAR's internal
inconsistencies. This is very ad hoc, but I need to get this done soon
and I can't wait for an AR4 result because that would defeat the whole
purpose of what I am doing.
Your comments are very helpful since they help me clarify my own thinking.
Tom.