E-Mail Interaction With Leonard Ornstein On My Post “Are The IPCC Global Climate Model Multi-Decadal Global Model?”

UPDATE: December 20 2010 pm  Len has sent us the following:

The URL for the “Irrigated Afforestation….” paper is:

http://www.springerlink.com/openurl.asp?genre=article&id=doi:10.1007/s10584-009-9626-y

The URL for the “Science in the Spectrum of Belief” is

http://www.pipeline.com/~lenornst/ScienceInTheSpectrumOfBelief.pdf

**********************************************

Leonard Ornstein and I have exchanged a set of e-mails related to my post

Are The IPCC Global Climate Model Multi-Decadal Global Model Predictions Accurate On Regional And Local Spatial Scales?

I have reproduced the e-mail exchange below, with Len’s permission [with some minor editing] .

Date: Thu, 16 Dec 2010 13:58:47 -0500
From: Leonard Ornstein <xxxxxx>
To: Roger A Pielke Sr <xxxxxx>
Subject: Are The IPCC Global Climate Model Multi-Decadal Global Model Predictions Accurate On Regional And Local Spatial Scales?

Hi Roger:

Setting aside the interpretation of “accurate”, I’d like to remind you of the attached paper that you kindly invited me to discuss on your blog on October 29, 2009;  “How To Quickly Lower Climate Risks, At ‘Tolerable? Costs’“.

I’d argue that the GISS ModelE provided a remarkably ‘believable’ prediction of what roughly might be expected by a rather enormous (and interesting) change in regional forcings.

We are, at present, reexamining “with ‘high expectations'” the performance with the GISS regional model RM3 (0.5º x 0.5º resolution), particularly with respect to the failure of ModelE, at (4º x 5º resolution) to demonstrate significant induced precipitation during the ‘winter half’ of the year .

To expect useful predictions with ‘small’ differences in forcings, is not too reasonable because of the high levels of ‘noise’ in the models, their parametrizations and in the observational data used to ‘tune’ them.

But, with very large differences in regional forcings, the ‘better’ designed GCMs should be expected to be able to tell us surprising things that are quite probable regional consequences.

Len Ornstein

Hi Len

 Thank you for your e-mail. In term of “a remarkably ‘believable’ prediction” this is not a scientifically robust approach. The regional model predictions must be compared against real world data [which is, of course, only possible in the hindcast mode in 2010, but can be accomplished in a truly predictive mode from 2011 and forward].

 However, there are two requirements for demonstrating regional skill. First, the current climatological statistics of societally and environmentally important weather events (e.g. droughts, floods, heat waves, ect) must be skillfully predicted.

Then, the even more difficult task of predicting the changes in these statistics due to the human intervention in the climate system. Large regional effects that are simulated by the GCMs is not a robust test unless corresponding observations can be shown to fit.

On your e-mail, can I post, along with my above reply?

On your paper, one concern I have with the irrigation idea is that while sensible temperature could fall, the moist enthalpy (i.e. heat) would be larger due to the added water vapor.

Best Wishes for the Holidays!

Roger

Date: Thu, 16 Dec 2010 21:03:52 -0500
From: Leonard Ornstein <xxxxxx>
To: Roger A Pielke Sr <xxxxxx>

Subject: Re: Are The IPCC Global Climate Model Multi-Decadal Global Model Predictions Accurate On Regional And Local Spatial Scales?

Hi Roger:
Of course, we agree that the “sensible temperature (does) fall [some 8º over the Sahara -see Figures 3f and 5e] – the moist enthalpy (i.e. heat) (IS) larger due to the added water vapor.  We didn’t claim this as a source of cooling. It’s the enormous sustainable sequestration of CO2 in the new ‘forests on the deserts’ that would ultimately lead to net global cooling; ~ 8 GtC/yr.

With this ‘clarification’, you may post.

Len

Date: Thu, 16 Dec 2010 22:29:55 -0700 (MST)
From: Roger A Pielke Sr <xxxxxx>
To: Leonard Ornstein <xxxxxx>
Subject: Re: Are The IPCC Global Climate Model Multi-Decadal Global Model Predictions Accurate On Regional And Local Spatial Scales?

Hi Len

 Before I post, please also discuss the albedo change issue. The darker surface will have a higher surface absorption of solar radiation.

Best Regards

Date: Sat, 18 Dec 2010 01:06:20 -0500
From: Leonard Ornstein <xxxxxx>
To: Roger A Pielke Sr <xxxxxx>
Subject: Re: Succinct Summary Of The Scientific Process

Hi Roger:
There is usually nothing distinctive about the low and high tails of an unbiased probability distribution of pertinent observations. “Truth’ and “falsity” are therefore usually symmetrical. It’s not either tail that tests the model  but the entire empirical record to date. For consistent  and ‘natural’ models, the empirical distributions converge rapidly, the distributions are tight and are they stable (as in much of natural science). We ‘like’ to look at these as foundational science  as ‘factual’; but they remain, at best, only the current leading models. However it’s just because of that, that strong forcing of such models either more easily reveals inconsistencies in model construction  or unexpected, but robust, elements of ‘probable’ predictive value. My current prose may yet fail to clearly convey this argument, but that’s part of what it’s about.

Len

On 12/17/10 8:01 PM, Roger A Pielke Sr wrote:
 Hi Len

 Thank you for the article.

 I do not see how any model is scientific if it is not testable (i.e.  shown to be false). Otherwise, it is another type of “proposition”). A  model is just a “hypothesis”.

 On your paper, when you are ready, why not present a guest weblog that  summarizes your paper and provides a link to it. This would be a very good topic to present and discuss further.

 Roger

 On Fri, 17 Dec 2010, Leonard Ornstein wrote:

 Hi Roger:

Sorry to be such a pest!

But no model can “at least potentially, be shown to be false” any more than it can be shown to be “true” (despite Karl Popper  or anyone else!). Please read my attached draft of  “The Sceptical Scientific Mind-Set in the Spectrum of Belief: It’s about models of  “reality” and the unavoidable incompleteness of evidence, for  “or against” any model.”

 Len

Date: Sat, 18 Dec 2010 01:37:27 -0500
From: Leonard Ornstein <xxxxx>
To: Roger A Pielke Sr <xxxxx>
Subject: Re:: Re: Are The IPCC Global Climate Model Multi-Decadal Global Model Predictions Accurate On Regional And Local Spatial Scales?

Hi Roger:

Attached are typical albedo anomalies. {NOTE from R. Pielke Sr: These figures are reproduced at the end of this post]

Len

Hi Len

 Examples would be useful. However, I remain unconvinced if the resultant albedo (averaged over the region and in time) would be higher than the already high albedo.

Roger

Comments Off on E-Mail Interaction With Leonard Ornstein On My Post “Are The IPCC Global Climate Model Multi-Decadal Global Model?”

Filed under Climate Change Forcings & Feedbacks, Climate Models, Guest Weblogs

Comments are closed.