Rgbatduke (aka Robert G Brown of Duke University aka the DuKE), Monckton, Spencer & Christy and now Willis Eschenbach and WM Briggs having a right old ding-dong battle. It's another battle of the DuKEs.
Here is where it started, here is a continuation, and then WM Briggs "statistician to the stars" dealt a blow to deniers by saying what Professor Brown wrote was complete and utter nonsense.
This is the latest (h/t Nick Stokes).
"Wondering" Willis Eschenbach told WM Briggs what he thought of his tearing down of Professor Brown's rant. This is part of what Willis wrote:
I know you’re a great statistician, and you’re one of my heroes … but with all respect, you’ve left out a couple of important priors in your rant ….
1. You assume that the results of the climate model are better than random chance.
2. You assume that the mean of the climate models is better than the individual models.
3. You assume that the climate models are “physics-based”.
As far as I know, none of these has ever been shown to be true for climate models. If they have, please provide citations.
As a result, taking an average of climate models is much like taking an average of gypsy fortunetellers … and surely you would not argue that the average and standard deviation of their forecasts is meaningful.(Odd that Wondering Willis doesn't know that "climate models" are based on physics - or at least the ones I believe he's referring to are.)
Robert G Brown is embarrassingly wrong...wronger than televised wrestling
Here are some excerpts from WM Briggs' response to Willis. This is about as big a slap down as one climate science denier (WM Briggs) can give to two others (Willis and the DuKE) and probably bigger than a science denier would give to a scientist (my bold):
1. Do ensemble models make statistical sense in theory?
Yes. Brown said no and wanted to slap somebody, God knows who, for believing they did and for creating a version of an ensemble forecast. He called such practice “horrendous.” Brown is wrong. What he said was false. As in not right. As is not even close to being right. As is severely, embarrassingly wrong. As in wrong in such a way that not one of his statistical statements could be repaired. As in just plain wrong. In other words, and to be precise, Brown is wrong. He has no idea of which he speaks. The passage you quote from him is wronger than Joe Biden’s hair plugs. It is wronger than Napoleon marching on Moscow. It is wronger than televised wrestling....
...3. A model does not have to explain the physics to be good.
Stop and re-read that before continuing.
The rigorous statistics of "it looks like"...
2. Are the ensemble climate models good? As I said originally, not for long-range predictions, but yes for very short-range ones. If Brown wants to claim long-range models are poor, even useless, then I am his brother. But if he wants to say that they do not make statistical sense, then I am his enemy. Being “good” and making “statistical sense” are different and no power in Heaven or on Earth can make them the same.
I tried to find where WM Briggs said anything about "long range predictions vs short range predictions" in his original article and its updates. About the closest I could find was this:
TWO Are the ensemble models used in climate forecasts any good? They don’t seem to be; not for longer-range predictions (and don’t forget that ensembles can have just one member). Some climate model forecasts—those for a few months ahead—seem to have skill, i.e. they are good. Why deny the obvious? The multi-year ones look like they’re too hot.
Well, that's weird. WM Briggs argues at length based on "statistics" but then dismisses climate models because "it looks like they're too hot". That doesn't sound to rigourous an assessment, does it. And he talks about climate model forecasts of "a few months ahead". I don't know of any such climate model forecasts. Climate models by definition model climate, not a few months of weather.
The "short-termism" of climate science denial
This is how WM Briggs is wrong. He said that ensemble climate models are good for short range predictions. But climate models model climate, not "short range predictions". Climate is long range, not short range. Climate models do not make any claim of being "right" in the short term. That is, the noise of weather dominates in the short term. In the short term, the noise masks the signal. What the models are designed for is to help understand what the various elements of the earth system will be like over coming decades to centuries, not over the coming days, weeks or months. As noted on realclimate.org:
Short term (15 years or less) trends in global temperature are not usefully predictable as a function of current forcings.Fifteen years or less! Not "a few months', not even "a few years". It would be fair to say there is more difference between models and model runs on a daily or weekly basis than there is on a longer term basis. This is an extract from the full comment from the article on RealClimate.org (my bold):
In interpreting this information, please note the following (mostly repeated from previous years):
- Short term (15 years or less) trends in global temperature are not usefully predictable as a function of current forcings. This means you can’t use such short periods to ‘prove’ that global warming has or hasn’t stopped, or that we are really cooling despite this being the warmest decade in centuries. We discussed this more extensively here.
- The CMIP3 model simulations were an ‘ensemble of opportunity’ and vary substantially among themselves with the forcings imposed, the magnitude of the internal variability and of course, the sensitivity. Thus while they do span a large range of possible situations, the average of these simulations is not ‘truth’.
- The model simulations use observed forcings up until 2000 (or 2003 in a couple of cases) and use a business-as-usual scenario subsequently (A1B). The models are not tuned to temperature trends pre-2000.
- Differences between the temperature anomaly products is related to: different selections of input data, different methods for assessing urban heating effects, and (most important) different methodologies for estimating temperatures in data-poor regions like the Arctic. GISTEMP assumes that the Arctic is warming as fast as the stations around the Arctic, while HadCRUT4 and NCDC assume the Arctic is warming as fast as the global mean. The former assumption is more in line with the sea ice results and independent measures from buoys and the reanalysis products.
- Model-data comparisons are best when the metric being compared is calculated the same way in both the models and data. In the comparisons here, that isn’t quite true (mainly related to spatial coverage), and so this adds a little extra structural uncertainty to any conclusions one might draw.
There is a paper by Santer et al (often misquoted by deniers) in which it is found that to determine a change in climate based solely on trends in global surface temperature generally requires multiple decades, not months or years. In their analysis, a period of 32 years yielded a clear signal over the noise. In the abstract they maintain you need at least 17 years. This is from the paper's conclusion (my bold):
The clear message from our signal-to-noise analysis is that multi-decadal records are required for identifying human effects on tropospheric temperature. Minimal warming over a single decade does not disprove the existence of a slowly-evolving anthropogenic warming signal.
From what I have read, all current climate models project these (among other things) over the longer term to varying degrees depending on the emissions scenario:
- Average global surface temperature will rise
- Sea level will continue to increase
- Ice will continue to melt - including Arctic sea ice and the worlds glaciers and ice sheets.
|Source: IPCC AR4 WGI|
|Source: IPCC AR4 WGI|
(And just in case there is anyone who thinks the world has stopped warming, check this out before you go making as big a fool of yourself as the DuKE, WM Briggs and Wondering Willis.)
Denier models are very short term and embarrassingly wrong
Denier "models" on the other hand, tend to the short term and "project" weird and silly stuff like this, from David Archibald. His "model" projects that before seven years is out, by 2020, the average surface temperature will have dropped below the lowest temperature in the entire Holocene:
Why deniers deny science...
It's enlightening watching the deniers fight it out among themselves. They all bring their own "wrong" to the table. One thing many of them have in common is embodied in this comment from DAV on WM Brigg's blog (my bold):
Policymakers are relying on these models to represent the RANGE of possible future climates that are consistent with known physics and chemistry.
To do what, exactly? Are they making preparations or just looking for a revenue source?
PS We can add Judith Curry to the list of deniers promoting the DuKE from Duke's rubbish. She aligns herself with the denialiati every chance she gets. (No link from me this time. I can't usually be bothered with Curry tripe, with some exceptions like here and here and here and here.)
PPS (23 June 2013) Looks as if WM Briggs is trying to backtrack from his comment about short term climate models by saying he meant the models used for weather forecasting and seasonal outlooks. I know they are called "climate" models too, although they model weather, not climate. And they are constantly updated with real data through data assimilation unlike the climate models rgbatduke was writing about. So this "fervent, ill-educated activist" is sticking to her guns. WM Briggs was wrong, plain and simple in what he said. His article was in relation to IPCC models, which are a somewhat different beast, used for a different purpose and having different features. The climate models being discussed here are built to learn about climate not to forecast weather.