UPDATE: this meeting is canceled, I will not be attending – Anthony
I’m pleased to announce that I and the entire WUWT community have been invited to a meeting and demonstration of computer modeling skills with Dr. Kevin Trenberth on November 10th in Boulder, CO. at NCAR. This meeting has been a behind the scenes negotiation with WUWT regular “R. Gates”, who has direct contact with Dr. Trenberth.
While some might question the wisdom of attending such a meeting, especially given some of the history, I’ll point out that a trademark of skeptics, illustrated here daily, is to listen to all available evidence and ask questions about it. This forum on how computer modeling works in climate science will provide just such an opportunity. I have tentatively agreed to attend.
One of the caveats I put forward is that Dr. Trenberth will not refer to me nor anyone in attendance as a “denier” such as he did with his AMS address. He has agreed to this. He has also agreed to allow me a short introduction and to have the event videotaped in entirety with it placed on the web unedited at some future date.
The Nov. 10th tentative agenda is:
====================================
Thursday November 10, 2011 9AM-1:30PM
9:00 arrival and greet in Damon Room
9:15 Dr Trenberth talk w/ Q&A
10:30 computer modeling demonstration in the visualization lab
11:15 short tour of the building-optional
11:45 lunch, on their own, in our cafeteria-optional ( we could reserve tables for the group)
1:00 explore climate exhibit floor and weather trail-optional
1:30 depart??
====================================
This meeting is free and open to any WUWT readers that can get there, but this is strictly a pay your own way event. I’m paying my own way as well.
Unfortunately, Dr. Roger Pielke Sr. will be in Florida at the time, and other scientists that I have invited have declined due to schedule conflicts and/or inability to justify travel for a half day event.
I can have up to 20 attendees, so attendance is strictly via RSVP.
If you can attend please use this contact form, providing your name and a valid address and email. This is required in order to get a visitor badge at the security gate.
Registration will be open until Tuesday and is on a first come first served basis. I hope you’ll be able to join me in person to help ask some serious questions. Thank you for your consideration.
Discover more from Watts Up With That?
Subscribe to get the latest posts sent to your email.
I would question that computational fluid dynamics via partial differential equations is at all valid when:
1. Grid cells are larger than hurricanes.
2. Insolation and IR irradiance are used as long-time averages.
3. Thermal equilibrium is assumed (how else could the Earth have a temperature?).
4. A multitude of important processes are either ‘parameterized’ or ignored.
The entire AGW-modeling schtick slyly confuses itself with short-term weather -modeling,
which has actual empirical success and shows some understanding of storms.
The entire climate system, however, is an entirely different animal that is many orders of magnitude more difficult to even understand, let alone computerize.
Just ask them this:
When you look back at your computer models 20 years ago, they must seem very limited and so weak as to be nearly useless by today’s standards, but this is how today’s models will seem in twenty years, so why pay them any heed today, especially considering the above-listed weaknesses?
Suggested question:
What historic data of Aerosol forcing and what assumptions about historic Aerosol forcing do they use to make the hindcasting fit?
A question I would like to see answered.
What steps were taken in the design and implementation of the models to prevent the “experimenter expectation” effect from influencing the results?
Specifically, what steps were taken in the model design to isolate the model from the experimenters, such as double blind techniques used in other fields of science?
For example, were the model parameters fully determined before the experiment, of did the experimenters adjust the parameters after observing the results of the experiment?
If the experimenters adjusted the parameters after seeing the results of the experiment, then is it not possible that the models have been adjusted to reflect the experimenter expectations. Otherwise, why adjust the model parameters except that it did not give the expected results?
In other words, have the models been cherry picked? Have those model parameters that did not fit experimenter expectations discarded, and those parameters that gave expected results retained.
Here is an example. Say for instance that cloud albedo is directly measured to be 0.73, but when that parameter is used in the model the results do not predict that past as well as using a figure of 0.77 in the model. Is the figure of 0.73 let untouched in the model, or is it replaced with 0.77?
If 0.73 is replaced with 0.77 then the model has been invalidated. The adjustment was made to meet experimenter expectations that the model should be able to predict the past. The correct figure to use is 0.73 because that matches observations. The conclusion that should have been made if the model doesn’t predict the past is that there are still unknowns and further observations are required to find them.
It is my understanding that the computer models parameters have been adjusted to made the models predict the past, similar to what is done with curve fitting. In which case the models are not predicting the future, they are predicting what the experimenters expect the future to look like.
Hmm, so R Gates, who started out as an eyes wide open fence sitter allegedly, has now morphed into a mate of someone who calls anyone who asks questions a d*****.
Run Luke, its a trap…..
Interstellar Bill says:
October 16, 2011 at 10:33 am
“I would question that computational fluid dynamics via partial differential equations is at all valid when:
1. Grid cells are larger than hurricanes.”
Bill, the entire approach of the GCM’s rests on the fact that grid cells are larger than one cloud, or one hurricane, so that they can use statistics. So, instead of simulating one cloud, they say, we know how a thousand clouds behave on average, and that’s why our approach works. Similar to how you can’t predict the exact trajectory of one atom in a gas but you can make valid statements about the behaviour of a volume of gas.
It’s a dilemma for them: Convective fronts come in all sizes, they can get larger than the grid cell size and cannot be described statistically in that case. It also stops them from improving the accuracy by making the grid cells smaller – the statistical approach breaks down in that case.
Some of the reasons why GCM’s don’t work as a predictive tool…
Anthony
If you have to shake hands with Trenberth I suggest you count your fingers afterwards.
Stevo
Looking in the mirror too much isn’t healthy.
Just a few questions I’d like to be answered as I can’t be there.
Which is the most accurate (best) model and how does he know?
Why take the average of models instead of just using the best when predicting?
If all CO2 was slowly removed from the current atmosphere what would the resultant average global temperature be according to the best model? And how much water vapour would remain in the atmosphere?
Would this zero CO2 climate be the same as snowball earth? If not why not?
Is the best model sophisticated enough to resonably carry out this exercise? If not what is the missing physics and what are the unreasonable/missing assumptions in the model? If it is how does he know?
The answers will be of great interest to me in assessing how believable the models are.
If you can get some good people to go such as Willis, maybe even one of the M & M’s, I’d be glad to contribute to travel expenses. Ric Werme is pretty good with computers.
REPLY: …maybe even one of the M & M’s
Regular or peanut? ;p – Anthony
I will not be able to attend, but would love to hear the answer to this question:
“If one runs the ‘best’ climate models forward, how long is it until the next ice age if we remove the effects of human-induced increase in CO2? ”
And the obvious followup,
“How does human-induced increase in CO2 change the timing of the next ice age?”
Anthony,
Surely it is possible for this to be set up for video conferencing so that all those who are best placed to evaluate and question this presentation can participate. Otherwise, it is extremely unlikely that this will do anything other than obfuscate.
REPLY: We are asking for that, but the time may not allow for extended Q&A – Anthony
[snip -great idea – I don’t want to give it away just yet – Anthony]
Timing on Nov 10 is a little short notice, unfortunately.
A good question to courteously ask is about the error from a certain fundamental circularity of the GCM argument; a circularity which the IPCC has supported.
For help framing the question, see Chapter 28 – “Spinning Straw into Gold”, [Laframboise, Donna (2011-10-09). The Delinquent Teenager Who Was Mistaken for the World’s Top Climate Expert (Kindle Location 1598). Ivy Avenue Press. Kindle Edition. ]
John
I would expect that the quid pro quo was that you were not allowed to mention the “travesty” nor its implications?
If unedited, it would certainly be a showpiece for ummmm, a lessening of tensions? (As long as their are no fisticuffs and only verbal fustigations… 😉 )
Are there any other denizens of the modelling world that will be in attendance?
Anthony,
Maybe get a short list of pivotal questions from Spencer, Pielke Snr, et all to be asked.
I signed up!
Hmm, please ask him: Why does he imput shortwave, a.k.a. light, from the Sun as heating land and oceans, which is physically impossible, and misses out all the thermal infrared, a.k.a. heat, thermal energy direct from the Sun to the Earth’s surface (which we feel as heat) which actually does heat land and oceans? And, Who suggested this gi? I suggest this is the missing heat he’s been looking for..
If you can’t ask all the questions posted here at the time, seems there’s not much of it allotted for Q&A, would you give them to him as a written form to fill in and send back?
The main problems with the current models are clouds and aerosols. The first are parametrised as a positive feedback on warming atmospheres (by CO2), while it is more and more clear that clouds provide a negative feedback (see the contribution of Dr. Spencer of lately).
The second is an important one too: If aerosols have a huge cooling effect (as several models imply to explain the 1945-1975 cooling period), then CO2 must have a huge effect and vv., but that is very questionable. See the discussion I had in the early days of RC, before they censored about halve my comments:
http://www.realclimate.org/index.php/archives/2006/02/an-aerosol-tour-de-forcing/ see my comment at #6.
The effect of aerosols can be reduced by 3/4, which makes that the effect of CO2 can be halved, without changing the tracking of the temperature record of the past century:
http://www.ferdinand-engelbeen.be/klimaat/oxford.html
The interdependence of the sensitivity of the climate for aerosols and CO2 can be seen in the first graph at:
http://www.realclimate.org/index.php/archives/2005/07/climate-sensitivity-and-aerosol-forcings/
And the comment of Kaufmann and Stern, that a simple linear model (on a spreadsheet!), only taking the forcings as base, outperforms the multi-million dollar models for hindcasting global temperatures is of interest too:
http://climateaudit.org/2005/12/23/gavin-vs-kaufmann/
Anthony,
Perhaps the time allocated for the tour of the building could instead be allocated to an extended Q & A. A building is a building.
Agree with lot of comments along the line – “Simply no time for any serious discussion”. If he wants to organise trully scientific meeting, then it should be 2-3h around-the-table open debate.
DB – UK
Simple questions:
1. How models handles thermodynamic laws?.
2. How there can be some backradiation from cooler atmosphere/gases to the warmer surface that warms it and transfers energy.
3. Do the models include heatstoring capasity from different gases?
9:15 Dr Trenberth talk w/ Q&A
I would be extremely concerned about this format – it appears that the may well be NO chance for questions, or questions cut short due to lack of time.
I would also urge someone to have a quality audio recording of the session – test the equipment before the gathering in the room – leave nothing to chance.
posters here are suggesting some very good questions – i fear the tragedy is that there will be no opportunity to ask them AND get answers – not change of subject OR the excuse of scientific ignorance of attendees justifies simplistic answers.
I fear you may be being set up AND tricked into having your ‘appearance’ morphed into a sort of ‘see, we have convinced the doubters of our correctness’ !
remember, there IS a snake in the grass.
Kudos to Kevin for taking this step. Do everything to make him glad that he did. Civility and comity are an absolute.
Dr. Dave says:
October 16, 2011 at 10:57 am
[snip -great idea – I don’t want to give it away just yet – Anthony]
—————
Dr. Dave & Anthony,
It is not nice for both of you to be such teases with your hints of spicy things to come.
John
Anthony – A simple question:
The reports of atmospheric warming are always couched as average global atmospheric temperatures. However, the effect of CO2 is to scatter heat energy. The actual atmospheric temperature rise caused by that heat will be dependent on the enthalpy of the atmosphere which is largely driven by its water vapor content. Why report atmospheric heat content as a temperature – when temperature is not a measure of heat content? And why average that inappropriate measure globally despite the huge variances in humidity?
Given the amount of time available I don’t think you will be able to get many questions asked and answered. I think the most important thing is, and there should be time for this, to get public admission that these are only process models. Research vehicles really. And as such they have not gone through any official validation and verification by software professionals. Which makes them useless for making predictions for public consumption. The “scenarios” are really just what-if questions and the “projections” merely “experiments”. While useful for the researchers, they can’t be used to form public policy in any way.