Re: [asa] The Climate Science Isn't Settled

From: Rich Blinne <rich.blinne@gmail.com>
Date: Tue Dec 01 2009 - 13:48:55 EST

On Tue, Dec 1, 2009 at 8:23 AM, Randy Isaac <randyisaac@comcast.net> wrote:

> This article came out the night before I was debating a friend of mine on
> global warming as guest lecturers in a classroom last week. He surprised me
> with it since I had missed it. Lindzen is a well known figure in this
> debate. He has a reasoned tone, at least in this piece, and raises specific
> points that can be discussed. I think that is very helpful and healthy for
> the debate. Here are some initial reactions I would have to this article:
>
> "It is generally accepted that a doubling of CO2 will only produce a change
> of about two degrees Fahrenheit if all else is held constant." I don't have
> the citation handy right now but I think the value I've seen in the
> literature is running closer to 3 degrees Centigrade. That's a big
> difference. I wonder where he gets that.
>

Out of the air it appears. The lower bound of climate sensitivity is
well-bounded with a > 90% confidence that the real climate sensitivity of >
1.5 degrees Celsius which is greater than Lindzen's "accepted" value. Knutt
and Hegerl 2008 did a review article on amongst other things climate
sensitivity.

http://climatechange.pbworks.com/f/The+equilibrium+sensitivity+of+the+Earth's+temperature,+2008.pdf

Some early estimates of climate sensitivity drew on palaeoclimate
information. For example, the climate of the Last Glacial Maximum (LGM) is a
quasi-equilibrium response to substantially altered boundary conditions
(such as large ice sheets over landmasses of the Northern Hemisphere, and
different vegetation) and diff erent atmospheric CO2 levels. Simple
calculations relating the peak cooling to changes in radiative forcing
yielded estimates mostly between 1 and 6 C, which turned out to be close to
Arrheniuss estimates9,5456. Simulations of the LGM are still an important
testbed for the response of climate models to radiative forcing57.

In some recent studies, parameters in climate models have been perturbed
systematically to estimate *S *(refs 14, 58, 59). The idea is to estimate
the sensitivity of a perturbed model by running it to equilibrium with
doubled CO2 and then evaluate whether the same model yields realistic
simulations of the LGM conditions. This method avoids directly estimating
the relationship between forcing and response, and thus avoids the
assumption that the feedback factor is invariant for this very different
climatic state.Instead, the assumption is that the change in feedbacks with
climate state is simulated well in a climate model. The resulting estimates
of climate sensitivity are quite different for two such attempts58,59,
illustrating the crucial importance of the assumptions in forcings (dust,
vegetation or ice sheets) and of differences in the structure of the models
used60.

A few people have used palaeoclimate reconstructions from the past
millennium to gain insight into climate sensitivity on the basis of a large
sample of decadal climate variations that were influenced by natural
forcing, and particularly volcanic eruptions61,62. Because of a weak signal
and large uncertainties in reconstructions and forcing data (particularly
solar and volcanic forcing), the long time horizon yielded a weak constraint
on *S *(ref. 62) (see Fig. 3), arising mainly from low-frequency temperature
variations associated with changes in the frequency and intensity of
volcanism. Direct estimates of the equilibrium sensitivity from forcing
between the Maunder Minimum period of low solar forcing and the present are
also broadly consistent with other estimates63.

Some studies of other, more distant palaeoclimate periods 64,65 seem to be
consistent with the estimates from the more recent past. For example, the
relationship between temperature over the past 420 million years64 supports
sensitivities that are larger than 1.5 C, but the upper tail is poorly
constrained and uncertainties in the models that are used are signifi cant
and diffi cult to quantify.

*There are few studies that yield estimates of S **that deviate
substantially from the consensus range, mostly towards very low values.
These results can usually be attributed to erroneous forcing assumptions
(for example hypothesized external processes such as cosmic rays driving
climate66), neglect of internal climate variability67, overly simplified
assumptions, neglected uncertainties, errors in the analysis or dataset, or
a combination of these6871. These results are typically inconsistent with
comprehensive models. In some cases they were refuted by testing the method
of estimation with a climate model with known sensitivity50,7274**.*

Several studies and assessments have discussed the available estimates for
climate sensitivity in greater detail4,5,17,18,23,24,75. In summary, most
studies find a lower 5% limit between 1 and 2 C (Fig. 3). The combined
evidence indicates that the net feedbacks *f *to radiative forcing (equation
(2)) are significantly positive and emphasizes that the greenhouse warming
problem will not be small.* Figure 3 further shows that studies that use
information in a relatively complete manner generally find a most likely
value between 2 and 3.5 C and that there is no credible line of evidence
that yields very high or very low climate sensitivity as a best estimate.
However, the figure also quite dramatically illustrates that the upper limit
for **S **is uncertain and exceeds 6 C or more in many studies. The reasons
for this, and the caveats and limitations, are discussed below.*

*On the basis of the available evidence, the IPCC Fourth Assessment Report
concluded that constraints from observed recent climate change**18 support
the overall assessment that climate sensitivity is very likely (more than
90% probability) to be larger than 1.5 C and likely (more than 66%
probability) to be between 2 and 4.5 C, with a most likely value of about 3
C (ref. 24). More recent studies support these conclusions8,45,51,64, with
the exception of estimates based on problematic assumptions discussed above
67,69,71.*

**

So, Randy, your recollection is absolutely correct. That's the most likely
value. Lindzen picked a value that that is completely outside of the 90%
probability range. How that's "generally accepted" is to torture the English
language. Figure 3 in the paper above also underscores your consilience
point. The real "generatly accepted" climate sensitivity is generated around
multiple lines of evidence" the instrumental record, climate modelling,
proxy data over the last millenium, and proxy data from millions of years
ago. Dismissing the HADCRUT database (which is not the only repository of
the instrumental record, BTW) would do nothing to dismiss this conclusion as
the other lines of evidence that came up with the SAME ANSWER.

Again from Knutt and Hegerl:

"Although uncertainties remain large, it would be presumptuous to say that
science has made no progress, given the improvements in our ability to
understand and simulate past climate variability and change18 as well as in
our understanding of key feedbacks4,5. *Support for the current consensus
range on **S **now comes from many different lines of evidence, the ranges
of which are consistent within the uncertainties, relatively robust towards
methodological assumptions (except for the assumed prior distributions; see
below) and similar for different types and generations of models.* *The
processes contributing to the uncertainty are now better understood."*

Another thing to look at is to consider is the "boundedness" of the value.
The lower bound is tightly bounded while the upper bound is not. If we are
wrong in our estimates of the climate sensitivity more likely than not we
are *underestimating* the problem. This makes it very difficult to determine
a "safe" level of CO2 for public policy consumption.

Finally, the following bears repeating:

*"There are few studies that yield estimates of S that deviate substantially
from the consensus range, mostly towards very low values. These results can
usually be attributed to erroneous forcing assumptions (for example
hypothesized external processes such as cosmic rays driving climate66),
neglect of internal climate variability67, overly simplified assumptions,
neglected uncertainties, errors in the analysis or dataset, or a combination
of these6871. These results are typically inconsistent with comprehensive
models. In some cases they were refuted by testing the method of estimation
with a climate model with known sensitivity50,7274."*

Rich Blinne

Member ASA

To unsubscribe, send a message to majordomo@calvin.edu with
"unsubscribe asa" (no quotes) as the body of the message.
Received on Tue Dec 1 13:49:06 2009

This archive was generated by hypermail 2.1.8 : Tue Dec 01 2009 - 13:49:06 EST