Bishop Hill The validity of climate models: a bibliography
Aug 3, 2013 Climate: Models A reader sent me some correspondence he had received from the Met Office. He had been enquiring about what evidence Prof Slingo et al had of the validity of the output of GCMs, and received in return the following bibliography:
Airey MJ, Hulme M, Johns TC (1996) Evaluation of simulations of terrestrial precipitation in UK Met Office/Hadley Centre climate change experiments. Geophysical Research Letters 23:1657-1660
Allan RP, Ramaswamy V, Slingo A (2002) Diagnostic analysis of atmospheric moisture and clear-sky radiative feedback in the Hadley Centre and Geophysical Fluid Dynamics Laboratory (GFDL) climate models. Journal of Geophysical Research-Atmospheres 107:7
Allan RP, Ringer MA, Slingo A (2003) Evaluation of moisture in the Hadley Centre climate model using simulations of HIRS water-vapour channel radiances. Quarterly Journal of the Royal Meteorological Society 129:3371-3389
Arritt RW, Goering DC, Anderson CJ (2000) The North American monsoon system in the Hadley Centre coupled ocean-atmosphere GCM. Geophysical Research Letters 27:565-568
Bellouin N, Jones A, Haywood J, Christopher SA (2008) Updated estimate of aerosol direct radiative forcing from satellite observations and comparison against the Hadley Centre climate model. Journal of Geophysical Research-Atmospheres 113:15
Bodas-Salcedo A, Ringer MA, Jones A (2008) Evaluation of the surface radiation budget in the atmospheric component of the Hadley Centre Global Environmental Model (HadGEM1). Journal of Climate 21:4723-4748
Collins M (2000) The El Nino-Southern Oscillation in the second Hadley Centre coupled model and its response to greenhouse warming. Journal of Climate 13:1299-1312
Collins M, Tett SFB, Cooper C (2001) The internal climate variability of HadCM3, a version of the Hadley Centre coupled model without flux adjustments. Climate Dynamics 17:61-81
Collins WJ, Bellouin N, Doutriaux-Boucher M, Gedney N, Halloran P, Hinton T, Hughes J, Jones CD, Joshi M, Liddicoat S, Martin G, O'Connor F, Rae J, Senior C, Sitch S, Totterdell I, Wiltshire A, Woodward S (2011) Development and evaluation of an Earth-System model-HadGEM2. Geoscientific Model Development 4:1051-1075
Cooper C, Gordon C (2002) North Atlantic oceanic decadal variability in the Hadley Centre coupled model. Journal of Climate 15:45-72
Corte-Real J, Qian B, Xu H (1999) Circulation patterns, daily precipitation in Portugal and implications for climate change simulated by the second Hadley Centre GCM. Climate Dynamics 15:921-935
Cusack S, Slingo A, Edwards JM, Wild M (1998) The radiative impact of a simple aerosol climatology on the Hadley Centre atmospheric GCM. Quarterly Journal of the Royal Meteorological Society 124:2517-2526
Gordon C, Cooper C, Senior CA, Banks H, Gregory JM, Johns TC, Mitchell JFB, Wood RA (2000) The simulation of SST, sea ice extents and ocean heat transports in a version of the Hadley Centre coupled model without flux adjustments. Climate Dynamics 16:147-168
Hardiman SC, Butchart N, Osprey SM, Gray LJ, Bushell AC, Hinton TJ (2010) The Climatology of the Middle Atmosphere in a Vertically Extended Version of the Met Office's Climate Model. Part I: Mean State. Journal of the Atmospheric Sciences 67:1509-1525
Hewitt HT, Copsey D, Culverwell ID, Harris CM, Hill RSR, Keen AB, McLaren AJ, Hunke EC (2010) Design and implementation of the infrastructure of HadGEM3: the next-generation Met Office climate modelling system. Geoscientific Model Development 3:1861-1937
Hewitt HT, Copsey D, Culverwell ID, Harris CM, Hill RSR, Keen AB, McLaren AJ, Hunke EC (2011) Design and implementation of the infrastructure of HadGEM3: the next-generation Met Office climate modelling system. Geoscientific Model Development 4:223-253
Inness PM, Gregory D (1997) Aspects of the intraseasonal oscillation simulated by the Hadley Centre Atmosphere Model. Climate Dynamics 13:441-458
James PM (2006) An assessment of European synoptic variability in Hadley Centre Global Environmental models based on an objective classification of weather regimes. Climate Dynamics 27:215-231
Johns TC, Durman CF, Banks HT, Roberts MJ, McLaren AJ, Ridley JK, Senior CA, Williams KD, Jones A, Rickard GJ, Cusack S, Ingram WJ, Crucifix M, Sexton DMH, Joshi MM, Dong BW, Spencer H, Hill RSR, Gregory JM, Keen AB, Pardaens AK, Lowe JA, Bodas-Salcedo A, Stark S, Searl Y (2006) The new Hadley Centre Climate Model (HadGEM1): Evaluation of coupled simulations. Journal of Climate 19:1327-1353
Jones PD, Hulme M, Briffa KR, Jones CG, Mitchell JFB, Murphy JM (1996) Summer moisture availability over Europe in the Hadley Centre general circulation model based on the Palmer Drought Severity Index. International Journal of Climatology 16:155-172
Joshi MM, Webb MJ, Maycock AC, Collins M (2010) Stratospheric water vapour and high climate sensitivity in a version of the HadSM3 climate model. Atmospheric Chemistry and Physics 10:7161-7167
Martin GM, Bellouin N, Collins WJ, Culverwell ID, Halloran PR, Hardiman SC, Hinton TJ, Jones CD, McDonald RE, McLaren AJ, O'Connor FM, Roberts MJ, Rodriguez JM, Woodward S, Best MJ, Brooks ME, Brown AR, Butchart N, Dearden C, Derbyshire SH, Dharssi I, Doutriaux-Boucher M, Edwards JM, Falloon PD, Gedney N, Gray LJ, Hewitt HT, Hobson M, Huddleston MR, Hughes J, Ineson S, Ingram WJ, James PM, Johns TC, Johnson CE, Jones A, Jones CP, Joshi MM, Keen AB, Liddicoat S, Lock AP, Maidens AV, Manners JC, Milton SF, Rae JGL, Ridley JK, Sellar A, Senior CA, Totterdell IJ, Verhoef A, Vidale PL, Wiltshire A, Had GEMDT (2011) The HadGEM2 family of Met Office Unified Model climate configurations. Geoscientific Model Development 4:723-757
Martin GM, Ringer MA, Pope VD, Jones A, Dearden C, Hinton TJ (2006) The physical properties of the atmosphere in the new Hadley Centre Global Environmental Model (HadGEM1). Part I: Model description and global climatology. Journal of Climate 19:1274-1301
Osprey SM, Gray LJ, Hardiman SC, Butchart N, Bushell AC, Hinton TJ (2010) The Climatology of the Middle Atmosphere in a Vertically Extended Version of the Met Office's Climate Model Part II: Variability. Journal of the Atmospheric Sciences 67:3637-3651
Pope VD, Gallani ML, Rowntree PR, Stratton RA (2000) The impact of new physical parametrizations in the Hadley Centre climate model: HadAM3. Climate Dynamics 16:123-146
Pope VD, Pamment JA, Jackson DR, Slingo A (2001) The representation of water vapor and its dependence on vertical resolution in the Hadley Centre Climate Model. Journal of Climate 14:3065-3085
Ringer MA, Martin GM, Greeves CZ, Hinton TJ, James PM, Pope VD, Scaife AA, Stratton RA, Inness PM, Slingo JM, Yang GY (2006) The physical properties of the atmosphere in the new Hadley Centre Global Environmental Model (HadGEM1). Part II: Aspects of variability and regional climate. Journal of Climate 19:1302-1326
Slingo A, Pamment JA, Allan RP, Wilson PS (2000) Water vapor feedbacks in the ECMWF reanalyses and Hadley Centre climate model. Journal of Climate 13:3080-3098
Spencer H, Sutton RT, Slingo JM, Roberts M, Black E (2005) Indian Ocean climate and dipole variability in Hadley Centre coupled GCMs. Journal of Climate 18:2286-2307
Stratton RA (1999) A high resolution AMIP integration using the Hadley Centre model HadAM2b. Climate Dynamics 15:9-28
Turner J, Connolley WM, Lachlan-Cope TA, Marshall GJ (2006) The performance of the Hadley Centre climate model (HADCM3) in high southern latitudes. International Journal of Climatology 26:91-112
Walters DN, Best MJ, Bushell AC, Copsey D, Edwards JM, Falloon PD, Harris CM, Lock AP, Manners JC, Morcrette CJ, Roberts MJ, Stratton RA, Webster S, Wilkinson JM, Willett MR, Boutle IA, Earnshaw PD, Hill PG, MacLachlan C, Martin GM, Moufouma-Okia W, Palmer MD, Petch JC, Rooney GG, Scaife AA, Williams KD (2011) The Met Office Unified Model Global Atmosphere 3.0/3.1 and JULES Global Land 3.0/3.1 configurations. Geoscientific Model Development 4:919-941
Wang KY, Shallcross DE (2005) Simulation of the Taiwan climate using the Hadley Centre PRECIS regional climate modeling system: The 1979-1981 results. Terrestrial Atmospheric and Oceanic Sciences 16:1017-1043
Webb M, Senior C, Bony S, Morcrette JJ (2001) Combining ERBE and ISCCP data to assess clouds in the Hadley Centre, ECMWF and LMD atmospheric climate models. Climate Dynamics 17:905-922
Woodward S (2001) Modeling the atmospheric life cycle and radiative impact of mineral dust in the Hadley Centre climate model. Journal of Geophysical Research-Atmospheres 106:18155-18166
Perhaps readers would like to pick a title at random and see how much comfort it gives.
Reader Comments (106)
I volunteer to read the two papers by R McKitrick. :)
That was it. I worked there one summer myself as a vacation student. I visited the site some years back but, if I remember, the main buildings had gone and what remained was pretty much abandoned and in ruin.
.....yield a trend pattern negatively correlated with observations.
Aug 3, 2013 at 7:48 PM Ross McKitrick
Haha
KnR says at Aug 3, 2013 at 12:55 PM
I disagree. The author knows her own work well. She will be more confident in defending it. It is the obvious first place to go to when asked for evidence.
Furthermore J Slingo is a top scientist. That is why she is the one being asked the question. Why shouldn’t her papers be the papers that fit the bill?
Well, actually I do think her papers don’t fit the bill because the two of hers I looked at just compare models to models and ignore the real world or at best acknowledge that the world doesn’t agree.
The papers I read:
Spencer H, Sutton RT, Slingo JM, Roberts M, Black E (2005) Indian Ocean climate and dipole variability in Hadley Centre coupled GCMs. Journal of Climate 18:2286-2307
And (I read the abstract only)
Slingo A, Pamment JA, Allan RP, Wilson PS (2000) Water vapor feedbacks in the ECMWF reanalyses and Hadley Centre climate model. Journal of Climate 13:3080-3098
But that wouldn’t be any better if she hadn’t co-written the papers.
I prefer empirical data.
Aug 3, 2013 at 11:38 AM | John Marshall
An excellent idea. Please give me links to your empirical data.
Trolls are a bit like buses.
You wait for ages and the two turn up at once.
Happily the buses don't always have 'Out of Service' signs on them!
Jones PD, Hulme M, Briffa KR, Jones CG, Mitchell JFB, Murphy JM (1996) Summer moisture availability over Europe in the Hadley Centre general circulation model based on the Palmer Drought Severity Index. International Journal of Climatology 16:155-172
Didn't anyone tell them the PDSI was designed only for _dry_ lands, like Texas?
That ammounts to another "nature trick".
RoyFOMR
Interesting that you call us trolls, but have no answers to our arguments.
With regards to
It appears to be available in full in pdf form at this link
The abstract ends with
The full paper includes the following comment:
Comparing radiative observations with models the paper explains:
and goes on to detail the models reproducing broad distributions but all three having significant and different errors of detail where contrasting boundary layers exist, including indicating results opposite to the observations.The paper itself is not so much about the models but about finding a method for comparing observations with model output in a way that would allow people to identify why and where the models are going wrong. Even so, it is hard to ignore the comments throughout the paper concerning the inability of the three models to replicate observations.
[Errors and Omissions excepted - I had to type the quotes rather than copy and paste]
EM
First, I recommend that you speak for youself.
Second, as the aim of blog is to promote discussion and debate.
Maybe those who have a need to argue should seek solace elswhere?
is available at
http://centaur.reading.ac.uk/30593/7/jgrd14409.pdf
----------------------------------------------------------
On a brief scan, their previously published algorithm produced a clear-sky aerosol direct radiative forcing of -1.89 W/m2 which is updated to yield the less negative value of -1.3 W/m2. This is compared to the -0.63 W/m2 simulated by the Hadley Centre climate model HadGEM2-A.
I hope you'll excuse me for not going over-the-moon about that.
It apparently claims some skill in that
If "remarkable agreement" can be construed as skill, that just leaves measurements over cloudy oceans, measurements over cloudy land, and measurements over clear land, as not being skillfully modelled.
That was published in 2008. So how well were they doing before that?
Not much comfort for me there, Bish.
.
Must have been somewhere else. Google maps shows the main building is still there.
Perhaps it was just the security cabin at the entrance that was in ruin.
AUWE main building in 2008
Richard Drake:
Very amusing, but I still reckon reading the IPCC source papers is the best way to understand how very weak the non-97% arguments are. Problem is, there's so many papers and so little spare time. Good thing is, they are probably all making very similar arguments.
Thanks Martin - it brings it all back.
Now, back to the CGCMs......
The more that I look for GCM validation in these papers (and fail to find any), the more my mind keeps drifting back to a question that has really been bugging me: do the models simulate insolation properly (using a rorating planet and taking account of Holder's Inequality), or do they simulate the flawed IPCC concept of a static planet with uniform, quarter power insolation, 24 hours a day?
Does anybody know the answer?
It would be a good question if the models were not flawed in far deeper ways. As it is, that would just be the cherry on the unvalidated and erroneous cake.
Since programming the time dependence of incoming radiation cannot be very complicated, it would be a bit surprising if it had not been done. And it would be very surprising if outgoing radiation were assumed uniform over the Earth's surface (Holder's).
Why not pose the question on My Climate and Me?
"Why not pose the question on My Climate and Me?"
I tried to, but was only allowed 100 characters. I might email the Met Office, but I fear that their patience with me must be wearing a bit thin, and the question would be regarded as vexatious.
The Met Office is always banging on about "communicating the science of climate change," yet when they're given an opportunity to do some communicating without exception they simply list the IPCC, Met Office on Climate Change, or in this case 35 papers to answer the question. There never seems to be any dialogue, or any attempt to explain in lay terms what they're doing. Just to set them straight on this as a communication effort, outside the halls of academia listing 35 papers without a lay persons explanation is a GFY response.
If you believe it's too complicated to explain in lay person's terms then you clearly don't understand your topic enough.
It is worth remembering that, if there is no evidence to support GCM validation in the literature, it can only be concluded that Baroness Verma misled Parliament with her response to Parliamentary Question HL1080:
“The Parliamentary Under-Secretary of State, Department of Energy and Climate Change (Baroness Verma): General circulation models developed by the Met Office are continually reassessed against observations and compared against international climate models through workshops and peer reviewed publications. The validity of general circulation modelling has been established for over four decades, as evident in the peer-reviewed literature. Such models are further developed in light of improvements in scientific understanding of the climate system and technical advances in computing capability.”
Gareth beat me to it (Aug 4, 2013 at 1:32 AM). Last night I started at the far end of the list and worked back to find a free paper, and that was the Webb et al. one. Once again, as one finds so often in the papers cited by those who wish to alarm us about the climate and CO2, it does not provide much support for what they claim.
Thanks to building-in (through the convenience of the 'external forcing' wheeze much beloved by programmers) their conclusions about CO2, the models can be run to produce rising global mean temperatures and that is what their role in life has been - at least their money-making role. One day, we might be able to actually model CO2 in GCMs but that seems a long way off, as does the modelling of other sub-grid phenomena, but I hate to think of the further harm such models and their masters will be able to do before then.
I think any match to other phenomena, such as regional effects on temperatures or precipitation or ocean temperatures etc is a bonus, a secondary one but useful nevertheless for the politically-motivated in search of some credibility for these models. But even papers from institutions 'on-side' such as this one by Webb et al. give little succour. It more or less admits that clouds are really important and that the models they evaluate do them so poorly they can have little confidence in them. Gareth gives some telling quotes from the abstract. Here are some quotes from later in the paper:
My comments are in square brackets [ ] within the above extracts.
I look forward to the NIPCC report later this year. I expect there will be a section on computer models, and it will be interesting to see what kind of a catalogue of failures they come up with.
In the meantime, this list of references is proving to be a valuable resource. I just wish more of us could find the time to check some more out. Some no doubt will be highly favourable to the models. Surely? But then there will be a slightly tougher problem of finding what hoops they had to jump through to get there. Just quoting from the text of the papers reported on so far provides grounds for not wanting to bet the farm on model outputs.
With regards to
It is available here in text and pdf format.
The paper demonstrates some ability of the models to reproduce climate patterns (in 'broad agreement'...) but points out that the models fail to reproduce what is described as "a strong annual-cycle modulation". It also describes the models as having a top of the atmosphere warm bias of up to 5K.
From the conclusion:
Martin,
As you say, "...that would just be the cherry on the unvalidated and erroneous cake". But it is the cherry that says "game over, do not pass go and do not collect £200"...
Therefore, I am having one final attempt to get a straight answer from the Met Office - I just sent this to their website:
"Dear Sir / Madam,
Thank you for answering my recent enquiries.
I have one, final question:
Do the Met Office general circulation models that are used to predict long term (decadal) average planetary surface temperatures simulate insolation with a rotating planet, and taking account of Holder's Inequality, or do they simulate the IPCC concept of a static planet with uniform, quarter power insolation, 24 hours a day?
In the case of the former, please could you let me know the typical time step in the numerical integration over decadal simulated timeframes?
Yours sincerely,
Roger Longstaff"
The Met Office has always answered my email enquires in a timely and polite manner, so I expect that they will do so again.
If you believe it's too complicated to explain in lay person's terms then you clearly don't understand your topic enough.
Aug 4, 2013 at 11:07 AM | Unregistered Commentergeronimo
Absolutely. The endless talk about the "need to communicate more effectively" is nothing but an evasion. For years, skeptics have asked for engagement on myriad topics regarding models and the evidence for them. The IPCC, the Met Office, and all fellow travelers have refused such engagement as Dr. Slingo does here.
michael hart
I reckon that gives the game away! They are trying to model climate and yet find data apparently derived from observation is in "remarkable agreement". In other words "we didn't expect out model accurately to reflect reality".Yet again note the crap grammar: "depths" and "appears"!
Roger, I stand by my cynical observation. You weren't supposed to read the papers. In fact, if it were me I'd regard it as an insult that the papers listed just don't support the contention at all.
I'm inclined to go back to my one square metre question. Can you model one square metre of the surface, sea or land, and measure the reality to check the result? Too difficult, I suppose? Like the lab experiment and the real-life observation of AGW in progress.
rhoda, buried within the hundreds of thousands of words of referenced literature you can be sure that there will be a few statements that the Met Office maintain will validate GCM methodology. They will be hidden in plain sight, but the MO will not point us to them, rather claiming that the body of work as a whole, which is peer-reviewed, proves their case. Of course, this is obfuscation, but the battle is impossible to win on these terms...
But I am not insulted, I expected it. It was, however, necessary to show others what we are dealing with.
But please, ask your one square metre question. Email it to "enquiries@metoffice.gov.uk" and you will get a polite and timely reply. And then please tell us what the answer is!
If the Bishop will permit me, there is an interesting essay at McIntyre's blog that makes some interesting points about models and evidence. http://climateaudit.org/2013/07/26/guy-callendar-vs-the-gcms/
"..are continually reassessed against observations and compared against international climate models through workshops and peer reviewed publications.
Murry Salby in his Hamburg talk made the point that the General Circulation Models cited by the IPCC had excellent mutual agreement in predicting the climate (ie global temp) of the last 15 years - but very poor agreement with actual temp.
I find it hard to understand that they can say such things (as the quote above) with a straight face.
John Shade, Gareth
As model aircraft flyer and ex-PPL I have spent a lot of my life looking at the sky. Even to a layman it is obvious that a large proportion of the physics generating low and medium level cloud takes place on a much more local scale than the 50km grid usually used in models. I doubt that improved simulation of cloud formation and behaviour will be achieved until a smaller grid size becomes practicable.
Entropic Man -
You wrote, "a large proportion of the physics generating low and medium level cloud takes place on a much more local scale than the 50km grid usually used in models." The question is, then, why do you put so much credence into the results of those models? They seem to be inaccurate even in the prediction of the most basic metric (global average temperature), and are worse at regional temperatures, or precipitation.
I do not doubt that as time progresses, and computing power increases, and a *lot* of work is put into refining the models, GCMs will become more faithful to reality. But at the moment, I don't see how one can believe that long-term trends from models have any reliability. All indications I have seen are that models over-estimate warming at the decadal time scale, to a greater or lesser extent, and can only be described as speculative (that is, untested) for longer periods. It is distressing to see extrapolations to 2100, 2200 or even 3000 based on such flimsy material, treated with the same respect as, say, eclipse predictions.
EM - Do you have any reason to believe that the Met Office's current GCM's do simulate the mechanism of cloud formation?
Just asking.
(I'd assumed it was one of the things they "parameterised".)
Murry Salby in his Hamburg talk made the point that the General Circulation Models cited by the IPCC had excellent mutual agreement in predicting the climate (ie global temp) of the last 15 years - but very poor agreement with actual temp. Aug 4, 2013 at 9:16 PM | Martin A
Martin, the Parliamentary and Met Office statements are also impossible to reconclie with the IPCC's own analysis - for example AR5 (draft) figure 1.4:
http://wattsupwiththat.com/2012/12/14/the-real-ipcc-ar5-draft-bombshell-plus-a-poll/
To a layman it looks like GCM methodolgy is on the verge of falsification. In fact, you could probably say that it has already been falsified.
Roger
"To a layman it looks like GCM methodolgy is on the verge of falsification. In fact, you could probably say that it has already been falsified."
It has never been validated (MO statements to the contrary notwithstanding). That's the only reason we are even talking about falsification. In a sane world, unvalidated models would be assumed false (for decision making) unless shown to be otherwise.
Had it been properly validated, falsification criteria would have been built in to the validation process.
With regards to:
It is available here
The paper is describing a newer model (HadGEM3-AO) coupling together several component models and how it performs compared to older ones.
From the introduction:
The science is settled!
Later:
The filtering is described as being necessary for the sea components because as the regions being modeled converge at the north pole the accuracy and stability of the models became questionable due to something called 'grid point noise'. What this comment says is that for the older model HadGEM1 they didn't solve it, they just eliminated the sea component problem by putting a pretend island at the top of the world.
To improve the models this new one uses a grid that still doesn't converge at the north pole. In order to avoid grid point noise it simulates the north pole as a normally spaced grid running from northern Canada to Russia. This is referred to as a tripolar grid. Illustration Fig. 2 on page 4 should help anyone here understand what this means - the southern hemisphere is a regular grid as there is no sea component convergence issue at the south pole and the northern hemisphere has in effect two regions where the grid points are close together but don't actually converge and a normally spaced strip inbetween them. I suppose this is an improvement...
Later on there is an interesting passage that provides an insight into how the scientists operate these models:
The need to have initial conditions is clearly going to be dictated by when the Met Office can catalogue an entire planet of conditions to start from. That is an understandable predicament as we didn't have satellite observations and wide station coverage until relatively recently. But is this significant shortcoming known to our representatives, policy makers and the activists?
The paper also seems to be suggesting that only ozone, aerosols and greenhouse gas emissions are the forcings. From that I can't see how you could ever use them to prove that greenhouse gases cause the observed warming - all you are showing is that the model is programmed to use greenhouse gas levels as a driver for climate change.
In addition to those issues they are testing the performance of a new model with initial sea ice conditions generated by an older model, not observations. Again this may be due to a lack of observational data but is such models on top of models on top of models widely known outside the modelling community?
One thing that is striking me reading these papers is that this science is up to date stuff but still displays many limitations due to complexity, processing power, a lack of observational data and other things, yet we have been subjected to model projections apparently proving CO2 induced warming for a good two decades now. They must have been much more limited in the past yet formed the basis of the consensus and policy decisions we are now living with.
Exactly.
As someone said, the outputs from a model are illustrations of a hypothesis - they are not evidence.
Model based climate science is one great big multi gigaflop tautology.
Nice find Gareth.
".......such in-consistencies are common place in climate modelling. For example, "pre-industrial control experiments" are often initialised with present day conditions but then forced with pre-industrial radiative forcings."
It makes you want to weep......
Quote from Gareth at Aug 5, 2013 at 1:12 PM: “the regions being modeled converge at the north pole”
Sorry, I haven't had a chance to go through these papers. Can anyone indicate where the computational grids are described?
It seems to me they might all be using the wrong sort of grid for these calculations. In the very few references I have found to actual grid details, I have only ever seen grids which look roughly like standard map grids, with surface lines corresponding approximately to latitude and longitude.
Such grids when applied to a nearly spherical object will always have geometrical singularities, where the distance between adjacent grid lines goes to zero. This implies numerical difficulties and loss of accuracy.
Instead, they should be using what are known as unstructured grids, an example of which would be the patches that go to make up a football. In this case the patches are pentagons and hexagons and they give a nearly uniform coverage, no region of the surface being much better or much worse than any other. To get more cells, which is very necessary, different surface elements would of course be used.
The earliest 3D calculations would naturally choose orthogonal map-type grids because the equations actually solved would be far simpler: it would not be necessary to include the changing angle between grid lines, for example, and all that implies. However, re-writes would be expected to cater for a more general geometry, which had the ability to increase local accuracy where spatial changes turned out to be greatest. The question is, how often are the programs re-written from scratch? Does program development simply mean updating tables of correlations perhaps – with no change to the underlying solver?
I hope they really are using unstructured surface grids but up to now I have not seen any reference to them.
I have come to this thread rather late, but I did pick up on Roger Longstaff's quote from the Met Office:
"Thank you for your email regarding references referred to in a response to a recent Parliamentary Question. Attached is a list of peer-reviewed literature since 1996 that have established the validity of general circulation modelling."
Pea under the thimble, as McKintyre would say, and builds on a suspicion I have had for some time about these models. Building a general circulation model is a reasonable objective. It is an emulation of the flow patterns of the atmosphere eg Hadley cell, coriolis effect, jet stream evolution etc. But you can build a reasonable emulation of the global circulation which incorporates all the observable features and therefore looks quite good, without including any climate change parameters at all. Such a model might be trendless, but nonetheless it would pick out major wind patterns and the temperature distribution of the globe. That is what a GCM does - it shows how the temperature would be distributed across the globe.
To make a climate predcition you don't really need a 3D GCM model, a 0-D model will suffice (see the recent ClimateAudit article on the Callendar model which outperforms all the GCM's.)
The pea under the thimble is this: by dressing up the climate change model inside a GCM (possibly ocean-coupled) you can misleadingly present the success of the GCM at representing general circulation in the hear and now as being indicative of the quality of climate change predictions in the future. That is the fallacy that is being perpetuated by these Met Office statements.
A further point about the GCM's and climate models that many are not aware is that they are incapable of predciting absolute temperature with any accruacy, so theuy tend to get baselined or presented as anomalies. The paper Tuning the climate of a global model by Mauritsen et al , 2012 notes:
"[6] Yet, the span between the coldest and the warmest
model is almost 3 K, distributed equally far above and
below the best observational estimates, while the majority
of models are cold-biased. Although the inter-model
span is only one percent relative to absolute zero, that
argument fails to be reassuring."
Interestingly I don't see the following publication from the Met Office on that list:
Improved Surface Temperature Prediction for the Coming Decade from a Global Climate Model
Doug M. Smith,* Stephen Cusack, Andrew W. Colman, Chris K. Folland, Glen R. Harris, James M. Murphy
"Previous climate model projections of climate change accounted for external forcing from natural and anthropogenic sources but did not attempt to predict internally generated natural variability. We present a new modeling system that predicts both internal variability and externally forced changes and hence forecasts surface temperature with substantially improved skill throughout a decade, both globally and in many regions. Our system predicts that internal variability will partially offset the anthropogenic global warming signal for the next few years. However, climate will continue to warm, with at least half of the years after 2009 predicted to exceed the warmest year currently on record."
Recall that the TAR in 2001 stated that all natural causes were accounted for and so warming could only be attributed to anthropogenic causes. Then in 2007, at the time of AR4, the Met Office publishes the above paper. Just in case you didn't read the first senetence of the abstract, I'll repeat it:
"Previous climate model projections of climate change accounted for external forcing from natural and anthropogenic sources but did not attempt to predict internally generated natural variability."
and the last sentence:
"However, climate will continue to warm, with at least half of the years after 2009 predicted to exceed the warmest year currently on record."
Has any thought whether the TLA "GCM" means the same thing for the Met Office in all situations?
To me GCM = Global Circulation Model but in the paper above we have a "Global Climate Model".
Mark Well (Aug 5, 2013 at 2:37 PM) asks for information on model grids.
Some information is given in the 26-author paper Walters et al (2011) in the above list. It can be obtained here: http://www.geosci-model-dev.net/4/919/2011/gmd-4-919-2011.pdf
See paragraphs at ‘3.1 Dynamical formulation and discretization’.
The paper itself is not very helpful for the stated purpose of the list provided for the Met Office. It mostly consists of describing recent efforts at the Met Office to produce a ‘seamless’ universal, portable, model that will handle weather forecasting over hours ahead, seasonal forecasting over months ahead, decadal forecasting, and climate forecasting a hundred years or more ahead. It provides a useful summary of some of their parameterisations, and hints at possible conflicts of interest between those intent on numerical weather forecasting (a noble, worthwhile, and successful pursuit that contributes a lot to society) and those intent on climate forecasting using GCMs (an ignoble – thanks to the IPCC, foolish, and unsuccessful pursuit that has been nothing but a nuisance, to say the least, and a burden on society). The climate folks have vested interest in associating closely with the weather ones because that can give the climate modelling a hint of verisimilitude – see for example how easily Paul Nurse was duped in his own tv documentary as he gazed like a schoolboy at weather simulations when he was meant to be addressing climate issues – that they could not hope to get on their own merits. The weather folks on the other hand may be grateful for all the money pouring in for new computers and possibly new research on the weather modelling front. They have not achieved this ‘seamlessness’ because there is ‘a continued small set of long-standing differences between our NWP [that’s Numerical Weather Prediction] and climate models rather than deficiences in any new developments. Although we will not be using GA3.0 [their latest climate model of the atmosphere] in operational NWP systems, we are assessing is performance in NWP trialling.’ That last bit refers, I guess, to looking out for the likes of BBQ summers and snowless winters.
A small part of the paper compares their newest models with the previous generation. I am not sure there is the detail there to enable a passer-by like me to evaluate what they have done. One first impression is that they have computed mean and rms errors for each of these generations of models over a 30 year simulation versus observations, and then subtracted one set from the other for various parameters such as ‘mean precipitation rate’, ‘mean outgoing sw radiation flux’, and so on. In other words, an earlier generation output could be off target by a mile, and the current one by a mile plus or minus a few inches. Do the subtraction and the mile disappears leaving only the inches behind for the graphical displays. I despair of finding time to look at this section more adequately, so readers of my comments here beware!
The model evaluation is not a primary purpose of the paper, and is merely noted in the ‘Summary and conclusions’ as ‘some initial assessments showing encouraging performance’. I do not think this paper provides convincing evidence that their GCMs are fit for any policy-making purpose with regard to future climates. Nor does it make any claim to. It is more like a technical document of the kind useful for sales-forces and marketeers promoting their shiny new model as being slicker and a bit better than their previous one. The Met Office products are, after all, being used by research institutes and meteorological services across the globe. It is perfectly reasonable for the Met Office, as it would be for a company, to publish such a document to provide technical details of their new product and assurance to customers that it is on a dynamic development path.
Johm Shade "Although we will not be using GA3.0 [their latest climate model of the atmosphere] in operational NWP systems, we are assessing is performance in NWP trialling."
But didn't Julia Slingo, at the ClimateGate hearings, state to the science & technology committee that the weather models and climate models are the same code?
Aug 5, 2013 at 6:20 PM | John Shade
Thanks for that lead.
By the way, there seem to be two authors called Slingo in the above reference list, not only ‘JM’ (twice) but also ‘A’ (5 times).
Yes, she did. And she explicitly said that their climate models are therefore validated each weather forecast.
By the way, there seem to be two authors called Slingo in the above reference list, not only ‘JM’ (twice) but also ‘A’ (5 times).
Aug 5, 2013 at 8:29 PM Mark Well
If I have it right, they were husband and wife. I believe that Mr Slingo, sadly, died some years ago.
Rapidly skimmimg: http://www.geosci-model-dev.net/4/919/2011/gmd-4-919-2011.pdf
I found on page 5: "Full radiation calculations are made every 3h using the instantaneous cloud fields and a mean solar zenith angle" I asked a question about this yesterday, and perhaps this is the answer.
Now, because of the fourth power relationship between power and temperature, I simply cannot see how raditaion calculations every 3 hours, using a mean solar zenith angle, can possibly be used to calculate a sensible planetary temperature distribution. Perhaps someone could explain it to me?
Unfortunately I need to go away for the rest of the week, but I will look forward to getting back to this at the weekend.
Roger Longstaff
GCMs are multivariate versions of the Game of Life, also known as cellular automata.
Your comment suggests that the generation time for the simulation corresponds to three hours in the real world.
I doubt that "Full radiation calculations are made every 3h using the instantaneous cloud fields and a mean solar zenith angle" refers to averaging over the whole planet. More likely it refers to the radiation summed over three hours for each 50km grid square.
Released by the American Geophysical Union (AGU) today. Climate sensitivity of around 0.57. No need for political action of any kind. Game over.
**
Human activities are changing Earth’s climate. At the global level, atmospheric
concentrations of carbon dioxide and other heat-trapping greenhouse gases have increased
sharply since the Industrial Revolution. Fossil fuel burning dominates this increase.
Human-caused increases in greenhouse gases are responsible for most of the observed
global average surface warming of roughly 0.8°C (1.5°F) over the past 140 years. Because
natural processes cannot quickly remove some of these gases (notably carbon dioxide)
from the atmosphere, our past, present, and future emissions will influence the climate
system for millennia.
http://www.agu.org/sci_pol/pdf/position_statements/AGU_Climate_Statement_new.pdf
EM, a quick thought before I head off -
Yes, I assumed that 3 hours means simulated elapsed time, and a gridded structure. But, insolation is the primary driver of temperature, and temperature controls conduction, convection and phase changes. 3 hours could span the rising or setting of the sun, so how can you model thermodynamic mechanisms within this period that correspond to the fourth power of insolation, or lack of it?