Guest essay by Dr. Tim Ball
I have no data yet. It is a capital mistake to theorize before one has data. Insensibly one begins to twist facts to suit theories, instead of theories to suit facts. Arthur Conan Doyle. (Sherlock Holmes)
There is no more common error than to assume that, because prolonged and accurate mathematical calculations have been made, the application of the result to some fact of nature is absolutely certain. A.N.Whitehead
The recent article by Nancy Green at WUWT is an interesting esoteric discussion about models. Realities about climate models are much more prosaic. They don’t and can’t work because data, knowledge of atmospheric, oceanographic, and extraterrestrial mechanisms, and computer capacity are all totally inadequate. Computer climate models are a waste of time and money.
Inadequacies are confirmed by the complete failure of all forecasts, predictions, projections, prognostications, or whatever they call them. It is one thing to waste time and money playing with climate models in a laboratory, where they don’t meet minimum scientific standards, it is another to use their results as the basis for public policies where the economic and social ramifications are devastating. Equally disturbing and unconscionable is the silence of scientists involved in the IPCC who know the vast difference between the scientific limitations and uncertainties and the certainties produced in the Summary for Policymakers (SPM).
IPCC scientists knew of the inadequacies from the start. Kevin Trenberth’s response to a report on inadequacies of weather data by the US National Research Council said
“It’s very clear we do not have a climate observing system…” “This may come as a shock to many people who assume that we do know adequately what’s going on with the climate, but we don’t.”
This was in response to the February 3, 1999 Report that said,
“Deficiencies in the accuracy, quality and continuity of the records place serious limitations on the confidence that can be placed in the research results.
Remember this is 11 years after Hansen’s comments of certainty to the Senate and five years after the 1995 IPCC Report. It is worse now with fewer weather stations and less data than in 1990.
Before leaked emails exposed its climate science manipulations, the Climatic Research Unit (CRU) issued a statement that said,
“GCMs are complex, three dimensional computer-based models of the atmospheric circulation. Uncertainties in our understanding of climate processes, the natural variability of the climate, and limitations of the GCMs mean that their results are not definite predictions of climate.”
Phil Jones, Director of the CRU at the time of the leaked emails and former director Tom Wigley, both IPCC members, said,
“Many of the uncertainties surrounding the causes of climate change will never be resolved because the necessary data are lacking.“
Stephen Schneider, prominent part of the IPCC from the start said,
“Uncertainty about feedback mechanisms is one reason why the ultimate goal of climate modeling – forecasting reliably the future of key variables such as temperature and rainfall patterns – is not realizable.”
Schneider also set the tone and raised eyebrows when he said in Discover magazine.
Scientists need to get some broader based support, to capture the public’s imagination…that, of course, entails getting loads of media coverage. So we have to offer up scary scenarios, make simplified dramatic statements, and make little mention of any doubts we may have…each of us has to decide what the right balance is between being effective and being honest.
The IPCC achieved his objective with devastating effect, because they chose effective over honest.
A major piece of evidence is the disparity between the Working Group I (WGI) (Physical Science Basis) Report, particularly the Chapter on computer models and the claims in the Summary for Policymakers (SPM) Report. Why did the scientists who participated in the WGI Report remain so silent about the disparity?
Here is the IPCC procedure:
Changes (other than grammatical or minor editorial changes) made after acceptance by the Working Group or the Panel shall be those necessary to ensure consistency with the Summary for Policymakers (SPM) or the Overview Chapter.
The Summary is written then the WGI is adjusted. It is like an executive publishing findings then asking employees to produce material to justify them. The purpose is to present a completely different reality to the press and the public.
This is to ensure people, especially the media, read the SPM first. It is released well before the WGI Report, which they knew few would ever read. There is only one explanation for producing it first. David Wojick, an IPCC expert reviewer, explained:
Glaring omissions are only glaring to experts, so the “policymakers”—including the press and the public—who read the SPM will not realize they are being told only one side of a story. But the scientists who drafted the SPM know the truth, as revealed by the sometimes artful way they conceal it
What is systematically omitted from the SPM are precisely the uncertainties and positive counter evidence that might negate the human interference theory. Instead of assessing these objections, the Summary confidently asserts just those findings that support its case. In short, this is advocacy, not assessment.
The Physical Basis of the Models
Here is a simple diagram of how the atmosphere is divided to create climate models.
Figure 1: Schematic of General Circulation Model (GCM).
The surface is covered with a grid and the atmosphere divided into layers. Computer models vary in the size of the grids and the number of layers. They claim a smaller grid provides better results. It doesn’t! If there is no data a finer grid adds nothing. The model needs more real data for each cube and it simply isn’t available. There are no weather stations for at least 70% of the surface and virtually no data above the surface. There are few records of any length anywhere; the models are built on virtually nothing. The grid is so large and crude they can’t include major weather features like thunderstorms, tornados, or even small cyclonic storm systems. The IPCC 2007 Report notes,
Despite the many improvements, numerous issues remain. Many of the important processes that determine a model’s response to changes in radiative forcing are not resolved by the model’s grid. Instead, sub-grid scale parameterizations are used to parametrize the unresolved processes, such as cloud formation and the mixing due to oceanic eddies.
O’Keefe and Kueter explain how a model works: “
The climate model is run, using standard numerical modeling techniques, by calculating the changes indicated by the model’s equations over a short increment of time—20 minutes in the most advanced GCMs—for one cell, then using the output of that cell as inputs for its neighboring cells. The process is repeated until the change in each cell around the globe has been calculated.”
Interconnections mean errors are spread and amplified. Imagine the number of calculations necessary that even at computer speed take a long time. The run time is a major limitation.
All of this takes huge amounts of computer capacity; running a full-scale GCM for a 100-year projection of future climate requires many months of time on the most advanced supercomputer. As a result, very few full-scale GCM projections are made.
A comment at Steve McIntyre’s site, Climateaudit, illustrates the problem.
Caspar Ammann said that GCMs (General Circulation Models) took about 1 day of machine time to cover 25 years. On this basis, it is obviously impossible to model the Pliocene-Pleistocene transition (say the last 2 million years) using a GCM as this would take about 219 years of computer time.
So you can only run the models if you reduce the number of variables. O’Keefe and Kueter explain.
As a result, very few full-scale GCM projections are made. Modelers have developed a variety of short cut techniques to allow them to generate more results. Since the accuracy of full GCM runs is unknown, it is not possible to estimate what impact the use of these short cuts has on the quality of model outputs.
Omission of variables allows short runs, but allows manipulation and moves the model further from reality. Which variables do you include? For the IPCC only those that create the results they want. Besides, because climate is constantly and widely varying so a variable may become more or less important over time as thresholds change.
By selectively leaving out important components of the climate system, likelihood of a human signal being the cause of change is guaranteed. As William Kinninmonth, meteorologist and former head of Australia’s National Climate Centre explains,
… current climate modeling is essentially to answer one question: how will increased atmospheric concentrations of CO2 (generated from human activity) change earth’s temperature and other climatological statistics? Neither cosmology nor vulcanology enter the equations. It should also be noted that observations related to sub-surface ocean circulation (oceanology), the prime source of internal variability, have only recently commenced on a consistent global scale. The bottom line is that IPCC’s view of climate has been through a narrow prism. It is heroic to assume that such a view is sufficient basis on which to predict future ‘climate’.
Static Climate Models In A Virtually Unknown Dynamic Atmosphere.
“Heroic” is polite. I suggest it is deliberately wrong. Lack of data alone justifies that position, lack of knowledge about atmospheric circulation is another. The atmosphere is three-dimensional and dynamic, so to build a computer model that even approximates reality requires far more data than exists, much greater understanding of an extremely turbulent and complex system, and computer capacity that is unavailable for the foreseeable future. As the IPCC note,
Consequently, for models to predict future climatic conditions reliably, they must simulate the current climatic state with some as yet unknown degree of fidelity. Poor model skill in simulating present climate could indicate that certain physical or dynamical processes have been misrepresented.
The history of understanding the atmosphere leaps 2000 years from Aristotle who knew there were three distinct climate zones to George Hadley in the 18th century. The word climate comes from the Greek word klima for slope referring to the angle of the sun and the climate zones it creates. Aristotle’s views dominated western science until the 16th century, but it wasn’t until the 18th century wider, but still narrow, understanding began.
In 1735 George Hadley used the wind patterns, recorded by English sailing ships, to create the first 3D diagram of circulation.
Figure 1. Hadley Cell (Northern Hemisphere)
Restricted only to the tropics, it became known as the Hadley Cell. Sadly, today we know little more than Hadley although Willis Eschenbach has worked hard to identify its role in transfer of heat energy. The Intergovernmental Panel on Climate Change (IPCC) illustrates the point in Chapter 8 of the 2007 Report.
The spatial resolution of the coupled ocean-atmosphere models used in the IPCC assessment is generally not high enough to resolve tropical cyclones, and especially to simulate their intensity.
The problem for climate science and modelers is the Earth is spherical and it rotates. Rotation around the sun creates the seasons, but the rotation around the axes creates even bigger geophysical dynamic problems. Because of it, a simple single cell system (Figure 2) with heated air rising at the Equator moving to the Poles, sinking and returning to the Equator, breaks up. The Coriolis Effect is the single biggest influence on the atmosphere caused by rotation. It dictates that anything moving across the surface appears to be deflected to the right in the Northern Hemisphere and to the left in the Southern Hemisphere. It appears that a force is pushing from the side so people incorrectly refer to the Coriolis Force. There is no Force.
Figure 2: A Simple Single Cell.
Figure 3 shows a more recent attempt to approximate what is going on.
Figure 3: A more recent model of a cross-section through the Northern Hemisphere.
Now it is the “Indirect Ferrell Cell”. Notice the discontinuities in the Tropopause and the “Stratospheric – Tropospheric Mixing”. This is important, because the IPCC doesn’t deal with the critical interface between the stratosphere and a major mechanism in the upper Troposphere in their models.
Due to the computational cost associated with the requirement of a well-resolved stratosphere, the models employed for the current assessment do not generally include the QBO.
This is just one example of model inadequacies provided by the IPCC.
What the IPCC Working Group I, (The Physical Science Basis Report) Says About the Models.
The following quotes (Italic and inset) are under their original headlines from Chapter 8 of the 2007 IPCC AR4 Report. Comments are in regular type.
There is currently no consensus on the optimal way to divide computer resources among finer numerical grids, which allow for better simulations; greater numbers of ensemble members, which allow for better statistical estimates of uncertainty; and inclusion of a more complete set of processes (e.g., carbon feedbacks, atmospheric chemistry interactions).
Most don’t understand models or the mathematics on which they are built, a fact exploited by promoters of human caused climate change. They are also a major part of the IPCC work not yet investigated by people who work outside climate science. Whenever outsiders investigate, as with statistics and the hockey stick, the gross and inappropriate misuses are exposed. The Wegman Report investigated the Hockey Stick fiasco, but also concluded,
We believe that there has not been a serious investigation to model the underlying process structures nor to model the present instrumented temperature record with sophisticated process models.
FAQ 8.1: How Reliable Are the Models Used to Make Projections of Future Climate Change?
Nevertheless, models still show significant errors. Although these are generally greater at smaller scales, important large-scale problems also remain. For example, deficiencies remain in the simulation of tropical precipitation, the El Niño- Southern Oscillation and the Madden-Julian Oscillation (an observed variation in tropical winds and rainfall with a time scale of 30 to 90 days).
Models continue to have significant limitations, such as in their representation of clouds, which lead to uncertainties in the magnitude and timing, as well as regional details, of predicted climate change. Nevertheless, over several decades of model development, they have consistently provided a robust and unambiguous picture of significant climate warming in response to increasing greenhouse gases.
Of course they do, because that is how they are programmed.
8.2.1.1 Numerics
In this report, various models use spectral, semi-Lagrangian, and Eulerian finite-volume and finite-difference advection schemes, although there is still no consensus on which type of scheme is best.
But how different are the results and why don’t they know which is best?
8.2.1.3 Parameterizations
The climate system includes a variety of physical processes, such as cloud processes, radiative processes and boundary-layer processes, which interact with each other on many temporal and spatial scales. Due to the limited resolutions of the models, many of these processes are not resolved adequately by the model grid and must therefore be parametrized. The differences between parametrizations are an important reason why climate model results differ.
How can parameterizations vary? The variance is evidence they are simply guessing at the conditions in each grid and likely choosing the one that accentuates their bias.
8.2.2.1 Numerics
Issues remain over the proper treatment of thermobaricity (nonlinear relationship of temperature, salinity and pressure to density), which means that in some isopycnic coordinate models the relative densities of, for example, Mediterranean and Antarctic Bottom Water masses are distorted. The merits of these vertical coordinate systems are still being established.
8.2.3.2 Soil Moisture Feedbacks in Climate Models
Since the TAR, there have been few assessments of the capacity of climate models to simulate observed soil moisture. Despite the tremendous effort to collect and homogenise soil moisture measurements at global scales (Robock et al., 2000), discrepancies between large-scale estimates of observed soil moisture remain. The challenge of modelling soil moisture, which naturally varies at small scales, linked to landscape characteristics, soil processes, groundwater recharge, vegetation type, etc., within climate models in a way that facilitates comparison with observed data is considerable. It is not clear how to compare climate-model simulated soil moisture with point-based or remotely sensed soil moisture. This makes assessing how well climate models simulate soil moisture, or the change in soil moisture, difficult.
Evaporation is a major transfer of long-wave energy from the surface to the atmosphere. This inadequacy alone likely more than equals the change created by human addition of CO2.
8.2.4.1 Terrestrial Cryosphere
Glaciers and ice caps, due to their relatively small scales and low likelihood of significant climate feedback at large scales, are not currently included interactively in any AOGCMs.
How big does an ice cap have to be to influence the parameterization in a grid? Greenland is an ice cap.
8.2.5 Aerosol Modelling and Atmospheric Chemistry
The global Aerosol Model Intercomparison project, AeroCom, has also been initiated in order to improve understanding of uncertainties of model estimates, and to reduce them (Kinne et al., 2003).
Interactive atmospheric chemistry components are not generally included in the models used in this report.
8.3 Evaluation of Contemporary Climate as Simulated by Coupled Global Models
Due to nonlinearities in the processes governing climate, the climate system response to perturbations depends to some extent on its basic state (Spelman and Manabe, 1984). Consequently, for models to predict future climatic conditions reliably, they must simulate the current climatic state with some as yet unknown degree of fidelity. Poor model skill in simulating present climate could indicate that certain physical or dynamical processes have been misrepresented.
They don’t even know which ones are misrepresented?
8.3.1.2 Moisture and Precipitation
For models to simulate accurately the seasonally varying pattern of precipitation, they must correctly simulate a number of processes (e.g., evapotranspiration, condensation, transport) that are difficult to evaluate at a global scale.
Precipitation forecasts (projections?) are worse than their temperature projections (forecasts).
8.3.1.3 Extratropical Storms
Our assessment is that although problems remain, climate models are improving in their simulation of extratropical cyclones.
This is their self-serving assessment. How much are they improving and from what baseline?
8.3.2 Ocean
Comparisons of the type performed here need to be made with an appreciation of the uncertainties in the historical estimates of radiative forcing and various sampling issues in the observations.
8.3.2.1 Simulation of Mean Temperature and Salinity Structure
Unfortunately, the total surface heat and water fluxes (see Supplementary Material, Figure S8.14) are not well observed.
8.3.2.2 Simulation of Circulation Features Important for Climate Response
The MOC (meridional overturning circulation) is an important component of present-day climate and many models indicate that it will change in the future (Chapter 10). Unfortunately, many aspects of this circulation are not well observed.
8.3.2.3 Summary of Oceanic Component Simulation
The temperature and salinity errors in the thermocline, while still large, have been reduced in many models.
How much reduction and why in only some models?
8.3.3 Sea Ice
The magnitude and spatial distribution of the high-latitude climate changes can be strongly affected by sea ice characteristics, but evaluation of sea ice in models is hampered by insufficient observations of some key variables (e.g., ice thickness) (see Section 4.4). Even when sea ice errors can be quantified, it is difficult to isolate their causes, which might arise from deficiencies in the representation of sea ice itself, but could also be due to flawed simulation of the atmospheric and oceanic fields at high latitudes that drive ice movement (see Sections 8.3.1, 8.3.2 and 11.3.8).
8.3.4 Land Surface
Vast areas of the land surface have little or no current data and even less historic data. These include 19 percent deserts, 20 percent mountains, 20 percent grasslands, 33 percent combined tropical and boreal forests and almost the entire Arctic and Antarctic regions.
8.3.4.1 Snow Cover
Evaluation of the land surface component in coupled models is severely limited by the lack of suitable observations.
Why? In 1971-2 George Kukla was producing estimates of varying snow cover as a factor in climate change. Satellite data is readily available for simple assessment of the changes through time.
8.3.4.2 Land Hydrology
The evaluation of the hydrological component of climate models has mainly been conducted uncoupled from AOGCMs (Bowling et al., 2003; Nijssen et al., 2003; Boone et al., 2004). This is due in part to the difficulties of evaluating runoff simulations across a range of climate models due to variations in rainfall, snowmelt and net radiation.
8.3.4.4 Carbon
Despite considerable effort since the TAR, uncertainties remain in the representation of solar radiation in climate models (Potter and Cess, 2004).
8.4.5 Atmospheric Regimes and Blocking
Blocking events are an important class of sectoral weather regimes (see Chapter 3), associated with local reversals of the mid-latitude westerlies.
There is also evidence of connections between North and South Pacific blocking and ENSO variability (e.g., Renwick, 1998; Chen and Yoon, 2002), and between North Atlantic blocks and sudden stratospheric warmings (e.g., Kodera and Chiba, 1995; Monahan et al., 2003) but these connections have not been systematically explored in AOGCMs.
Blocking was a significant phenomenon in the weather patterns as the Circumpolar flow changed from Zonal to Meridional in 2013-14.
8.4.6 Atlantic Multi-decadal Variability
The mechanisms, however, that control the variations in the MOC are fairly different across the ensemble of AOGCMs. In most AOGCMs, the variability can be understood as a damped oceanic eigenmode that is stochastically excited by the atmosphere. In a few other AOGCMs, however, coupled interactions between the ocean and the atmosphere appear to be more important.
Translation; We don’t know.
8.4.7 El Niño-Southern Oscillation
Despite this progress, serious systematic errors in both the simulated mean climate and the natural variability persist. For example, the so-called ‘double ITCZ’ problem noted by Mechoso et al. (1995; see Section 8.3.1) remains a major source of error in simulating the annual cycle in the tropics in most AOGCMs, which ultimately affects the fidelity of the simulated ENSO.
8.4.8 Madden-Julian Oscillation
The MJO (Madden and Julian, 1971) refers to the dominant mode of intra-seasonal variability in the tropical troposphere. Thus, while a model may simulate some gross characteristics of the MJO, the simulation may be deemed unsuccessful when the detailed structure of the surface fluxes is examined (e.g., Hendon, 2000).
8.4.9 Quasi-Biennial Oscillation
The Quasi-Biennial Oscillation (QBO; see Chapter 3) is a quasi-periodic wave-driven zonal mean wind reversal that dominates the low-frequency variability of the lower equatorial stratosphere (3 to 100 hPa) and affects a variety of extratropical phenomena including the strength and stability of the winter polar vortex (e.g., Baldwin et al., 2001).. Due to the computational cost associated with the requirement of a well-resolved stratosphere, the models employed for the current assessment do not generally include the QBO.
8.4.10 Monsoon Variability
In short, most AOGCMs do not simulate the spatial or intra-seasonal variation of monsoon precipitation accurately.
Monsoons are defined by extreme seasonality of rainfall. They occur in many regions around the word, though most only associate them with Southern Asia. It is not clear what the IPCC mean. Regardless, these are massive systems of energy transfer from the region of energy surplus to the deficit region.
8.4.11 Shorter-Term Predictions Using Climate Models
This suggests that ongoing improvements in model formulation driven primarily by the needs of weather forecasting may lead also to more reliable climate predictions.
This appears to contradict the claim that weather and climate forecasts are different. As Norm Kalmonavitch notes,
The GCM models referred to as climate models are actually weather models only capable of predicting weather about two weeks into the future and as we are aware from our weather forecasts temperature predictions…
In 2008 Tim Palmer, a leading climate modeller at the European Centre for Medium-Range Weather Forecasts in Reading England said in the New Scientist.
I don’t want to undermine the IPCC, but the forecasts, especially for regional climate change, are immensely uncertain.
8.5.2 Extreme Precipitation
Sun et al. (2006) investigated the intensity of daily precipitation simulated by 18 AOGCMs, including several used in this report. They found that most of the models produce light precipitation (<10 mm day–1) more often than observed, too few heavy precipitation events and too little precipitation in heavy events (>10 mm day–1). The errors tend to cancel, so that the seasonal mean precipitation is fairly realistic (see Section 8.3).
Incredible, the errors cancel and since the results appear to match reality they must be correctly derived.
8.5.3 Tropical Cyclones
The spatial resolution of the coupled ocean-atmosphere models used in the IPCC assessment is generally not high enough to resolve tropical cyclones, and especially to simulate their intensity.
8.6.2 Interpreting the Range of Climate Sensitivity Estimates Among General Circulation Models
The climate sensitivity depends on the type of forcing agents applied to the climate system and on their geographical and vertical distributions (Allen and Ingram, 2002; Sausen et al., 2002; Joshi et al., 2003). As it is influenced by the nature and the magnitude of the feedbacks at work in the climate response, it also depends on the mean climate state (Boer and Yu, 2003). Some differences in climate sensitivity will also result simply from differences in the particular radiative forcing calculated by different radiation codes (see Sections 10.2.1 and 8.6.2.3).
Climate sensitivity has consistently declined and did so further in IPCC AR5. In fact, in the SPM for AR5 the sensitivity declined in the few weeks from the first draft to the final report.
8.6.2.2 Why Have the Model Estimates Changed Since the TAR?
The current generation of GCMs[5] covers a range of equilibrium climate sensitivity from 2.1°C to 4.4°C (with a mean value of 3.2°C; see Table 8.2 and Box 10.2), which is quite similar to the TAR. Yet most climate models have undergone substantial developments since the TAR (probably more than between the Second Assessment Report and the TAR) that generally involve improved parametrizations of specific processes such as clouds, boundary layer or convection (see Section 8.2). In some cases, developments have also concerned numerics, dynamical cores or the coupling to new components (ocean, carbon cycle, etc.). Developing new versions of a model to improve the physical basis of parametrizations or the simulation of the current climate is at the heart of modelling group activities. The rationale for these changes is generally based upon a combination of process-level tests against observations or against cloud-resolving or large-eddy simulation models (see Section 8.2), and on the overall quality of the model simulation (see Sections 8.3 and 8.4). These developments can, and do, affect the climate sensitivity of models.
All this says is that climate models are a work in progress. However, it also acknowledges that they can only hope to improve parameterization. In reality they need more and better data, but that is not possible for current or historic data. Even if they started an adequate data collection system today it would be thirty years before it would be statistically significant.
8.6.2.3 What Explains the Current Spread in Models’ Climate Sensitivity Estimates?
The large spread in cloud radiative feedbacks leads to the conclusion that differences in cloud response are the primary source of inter-model differences in climate sensitivity (see discussion in Section 8.6.3.2.2). However, the contributions of water vapour/lapse rate and surface albedo feedbacks to sensitivity spread are non-negligible, particularly since their impact is reinforced by the mean model cloud feedback being positive and quite strong.
What does “non-negligible “ mean? Is it a double negative? Apparently. Why don’t they use the term significant? They assume their inability to produce accurate results is because of clouds and water vapor. As this review shows there are countless other factors and especially those they ignore like the Sun. The 2001 TAR Report included a table of the forcings with a column labeled Level of Scientific Understanding (LOSU). Of the nine forcings only two have a ”high” rating, although that is their assessment, one is medium and the other six are “low”. The only difference in the 2007 FAR Report is the LOSU column is gone.
8.6.3.2 Clouds
Despite some advances in the understanding of the physical processes that control the cloud response to climate change and in the evaluation of some components of cloud feedbacks in current models, it is not yet possible to assess which of the model estimates of cloud feedback is the most reliable.
The cloud problem is far more complicated than this summary implies. For example, clouds function differently depending on type, thickness, percentage of water vapor, water droplets, ice crystals or snowflakes and altitude.
8.6.3.3 Cryosphere Feedbacks
A number of processes, other than surface albedo feedback, have been shown to also contribute to the polar amplification of warming in models (Alexeev, 2003, 2005; Holland and Bitz, 2003; Vavrus, 2004; Cai, 2005; Winton, 2006b). An important one is additional poleward energy transport, but contributions from local high-latitude water vapour, cloud and temperature feedbacks have also been found. The processes and their interactions are complex, however, with substantial variation between models (Winton, 2006b), and their relative importance contributing to or dampening high-latitude amplification has not yet been properly resolved.
You can’t know how much energy is transported to polar regions if you can’t determine how much is moving out of the tropics. The complete lack of data for the entire Arctic Ocean and most of the surrounding land is a major limitation.
8.6.4 How to Assess Our Relative Confidence in Feedback to controls Simulated by Different Models?
A number of diagnostic tests have been proposed since the TAR (see Section 8.6.3), but few of them have been applied to a majority of the models currently in use. Moreover, it is not yet clear which tests are critical for constraining future projections. Consequently, a set of model metrics that might be used to narrow the range of plausible climate change feedbacks and climate sensitivity has yet to be developed.
The IPCC chapter on climate models appears to justify use of the models by saying they show an increase in temperature when CO2 is increased. Of course they do, that is how they’re programmed. Almost every individual component of the model has, by their admission, problems ranging from lack of data, lack of understanding of the mechanisms, and important ones are omitted because of inadequate computer capacity or priorities. The only possible conclusion is that the models were designed to prove the political position that human CO2 was a problem.
Scientists involved with producing this result knew the limitations were so severe they precluded the possibility of proving the result. This is clearly set out in the their earlier comments and the IPCC Science Report they produced. They remained silent when the SPM claimed, with high certainty, they knew what was going on with the climate. They had to know this was wrong. They may not have known about the political agenda when they were inveigled into participating, but they had to know when the 1995 SPM was published because Benjamin Santer exploited the SPM bias by rewriting Chapter 8 of the 1995 Report in contradiction to what the members of his chapter team had agreed. The gap widened in subsequent SPMs but they remained silent and therefore complicit.
To the old guys with their PDPs and VAXs. My first machine experience was the Bendix G15D running the Intercom interpreter. We used it on undergraduate engineering classroom problems. Fun. Even more so was operating the analog computers of the day. The Bendix had the neat feature of ringing a bell when it encountered a programming error at which point you received “the smile” from your associates in the room.
Government supported science is directed to supporting government goals, good and not so good. It can be corrupted when the goals are regulations and taxes ( by any other name). Climate “science” ?
Having worked in industry, a government lab and academia, I encountered the strictest accountability for my work in industry, reasonable accountability in the lab and observed academic accountability to be a function of the requirements of the funding body. Many (most?) of my industry associates are cAGW doubters. Many of my academic associates are believers.
Steven Mosher says:
March 21, 2014 at 10:46 am
Whom are you trying to kid?
As of 2010, GISS relied on just 3846 stations. Many were lost in the 1980s & early ’90s:
http://www.appinsys.com/globalwarming/GW_Part2_GlobalTempMeasure.htm#historic
Coverage, duration, siting & reporting are all terrible. Taking a planetary average temperature based upon such shoddy to nonexistent “data” is ludicrous, if not criminal, even before the shameless adjustments & interpolations.
From the above link:
It is important to note that the HadCRU station data used by the IPCC is not publicly available – neither the raw data nor the adjusted data – only the adjusted gridded data (i.e. after adjustments are made and station anomalies are averaged for the 5×5 degree grid).
@KevinK
>>If you can’t get it done in 128k, it’s not worth doing.”
>You had 128k ?, boy back in my day all we had was ones and zeros, and sometimes we ran out of zeros and we had to use o’s.
Reg Barrow, an up and coming programmer based himself in a company that sold data sharing in the 60’s using an IBM 360. He was delighted to hear the bosses were doubling the RAM to 512 K for the princely sum of $250,000 and hopped to the door on Saturday as the technician arrived to perform the upgrade.
“Where is the equipment?” he asked astonished. The technician carried only a briefcase.
“It is here in my case,” was the reply.
They proceeded to the main computer room. The technician pulled out a cable and plugged it into two empty connectors next to an existing cable that provided access to the first 256k of RAM.
“You mean the memory is already in there? asked an astonished Reg.
“Yup. Just have to put in the cable.”
“And for this you charge $250,000?”
“Yup.”
The computer was sold to timeshare companies who created contracts with dumb terminal users. Sixty clients per IMB machine was the mantra. Break even for the purchase cost and financing was 40 users. 66%. In fact as the number of users logged on approached the high 30’s the time to log on increased exponentially, with the 40th user taking literally forever. In practice it was impossible to purchase the machine and make money from selling timeshare. Everything all those timeshare companies made was handed to IBM so they sold programming/consulting services to make actual profits.
The morality in the world of big computing hasn’t change much in 50 years.
My Father was a technician working on one of the first computers in Philadelphia.
Total approximate memory about 2000 bytes, though not as we understand memory today. His job was to find and replace burnt tubes and circuits in between runs and fix them.
It was strictly a serial computer. One step led to another single step, card after card till the last card was read and the final card punched.
I started first with a PDP and then later a little more seriously (student wise) with FORTRAN at one of Penn State’s branch facilities; both using punch cards, submitted runs during specific open times and collecting output much later at the Computer Operations mail center. From both I would receive total CPU time used out of my allotted CPU budget.
Penn State closed that ‘computer branch’ for refurbishment and I migrated to a community college sporting a brand new massively outfitted 3270 IBM mainframe that even had a pack of 3270PCs that were left alone by the staff and students to my glee.
No punch cards. No constant reminders of ‘allocated CPU budget’. No restrictions on times for submitting jobs and output was immediate, especially if prints were kept in a spool library instead of printed.
After running out of CS courses to take along with business and Finance I moved on. But not before work discovered a blue collar worker who could maintain PCs and interface with a mainframe.
In the finance world I ran into the issue of trying to replicate the multiple dimensions of reality in code.
128K, 512K, 1M, 2M of memory coupled with stepwise single function code takes forever.
Multidimensional arrays hanging in memory coupled with threads and multiple processors allows for some mighty fast and fancy computing.
I get a kick out of seeing historical relics and even looking them over. I am not interested in keeping them for even one day as ‘reminders’ of the good old days. They were anchors then and they’re not even decent for anchors today.
Back about 1990 the Federal government sought to ‘recover’ some of the half million dollars it cost to buy a fully fleshed Wang 100 mini Computer and terminals. They published their intent to sell at auction with ads and flyers throughout town.
The big day arrived, the auction room filled and people proceeded to watch as item after item went without bid. Some of the disk drives sold for a few bucks, mostly to get the removable drive components themselves.
The VS100 central CPU sold for $100 with a few bidders slowly adding their $5 bid increments.
All of the unsold items were put into a dumpster behind the building over the next week. We copped a few parts just in case our Wang VS100 needed them. Completely ignored were all of the Wang WP terminals (monitors and keyboards) and miles of double cable.
So also went the various PDPs and early DEC lans along with those early IBM PCs and clones.
I suppose someone could’ve started designing 1024 bit processor run computers without need for doubleword code or split addressing. But who would’ve bought them back then? $4500 for the original IBM 8086/8088 PC without hard drive was very expensive back then. Several thousand dollars more and one added a separate ‘box’ with a 10MB hard drive and massive wire connection back to the CPU.
So what’s the statement? Something like; “Get over it”. This is the 21st Century and we have little concept what the computer field will be like in twenty years.
Still; I have a lingering thought that the fancy new computers in climate modeling are not being fed state of the art code runs. That annoying one step at a time mainframe approach still seems prevalent in many places. Yes, even state of the science computers can not ‘perform’ old code any better than old computers; just a tad faster.
8.4.11 “The GCM models referred to as climate models are actually weather models only capable of predicting weather about two weeks into the future”.
Exactly. The whole basis of the GCMs is ridiculous for a climate model. You can’t tell the decade/century future of climate by dividing the globe into little 20-minute cubes, just like you can’t for example predict world food production a decade or century ahead by dividing the surface area into little 20-minute squares.
you don’t have to prove models are wrong and you can’t, THEY must prove how models are acurate, and, they can’t.
Thank you, Dr. Ball. yet again, this time for a devastating critique of Warmism’s faith in the “Models”. Your breadth and detail of knowledge on the subject should be rather startling for any Warmist Reader. The idea of a computer model sounds good, but in practice they are little more use than a large Lego Model for modelling the reality of climate.
VAXes SMAXes 🙂 you kiddies were spoiled, spoiled I tell you. Who here remembers sitting for hours in front of an 026 or 029 keypunch loading up a 1000 card box full of F2, Watfiv or Watfor, (or yikes COBOL!) for the sysop to run, only to have batch spit out your deck halfway through due to compile errors, never getting to run… Or the perverse joy of watching a modified IBM Selectric “flying ball”, hooked to a System 370 stream out pages of APL gibberish before you figured out your program was wrong, and oh btw, your booked time allotment on the 370 is up… If your input didn’t weigh at least 20lbs, you could hardly call yourself a programmer…
Then we all went out and bought TRS80 Model Is and ran North America’s middle class until IBM claimed to have invented the PC…
lemiere jacques:
you don’t have to prove models are wrong and you can’t…
I dunno about that. If n models say n contradictory things, I can reasonably conclude that n-1I/i> models are wrong, and maybe all n.
Well nuts. Should read n-1 models are wrong and maybe all n.
That GCMs are useless in making 100-year forecast of climate is known long ago. Dr. Patrick Frank published “The Climate of Belief” in 2008. He showed that GCM forecasts are no better than random guesses. To my knowledge, none of IPCC scientists have refuted Frank’s conclusion. Even the editor of Journal of Geophysical Research of AGU and IPCC scientist admitted to me the large uncertainty in modeling clouds.
Dr. Ball: Thank you for taking the time to copy these passages from AR5 and providing some context, especially when the IPCC’s words vastly out number you context.
Having adequate data to initialize models is essential for weather forecast model, but I’m not sure why initialization data is important for century scale climate projections. ENSO variability averages out over a century, but not one or two decades. Climate models don’t exhibit much variability on the decade+ time scale, so initializing PDO and AMO shouldn’t change much.
Until one computer modeler can predict next week’s ‘lotto’ (6 from 45) then they are propping up a gigantic tax fraud on people.
Stop taxing or seeking to tax our salt!
izen says:
March 21, 2014 at 10:05 am
==
Please read what you wrote, think about it, and report back. Your analogy doesn’t fly or even get off the ground. Think before you respond.
Reblogged this on The GOLDEN RULE and commented:
This is appropriate given that two posts later, WUWT reveal what Professor Mann claims is science to be believed, when its computer-modelled graphs and predictions right from the start use a 2013 global temperature level already incorrect. That’s how their ‘warmist’ science “works”.
From this post we learn the importance and influence of the agenda factor – “Here is the IPCC procedure:
Changes (other than grammatical or minor editorial changes) made after acceptance by the Working Group or the Panel shall be those necessary to ensure consistency with the Summary for Policymakers (SPM) or the Overview Chapter.”
Blatant admission from the IPCC that reports are amended to prove the premise that ‘man is causing catastrophic warming’. You can’r find anything clearer than that. Yet warmists persist in their false (demonstrably) claims and scaremongering lies.
Gamecock, I hope you realized my “running out of zero’s” joke was just in fun. I use it when our younger software professionals start to get “uptight” about “delays” in product deliveries.
Until the 80’s and 90’s most of the signalling systems used on US railroads consisted of mechanical relay logic. Not “reprogrammable” as such, and slow but fast enough for the job at hand. The “computers” were in all those silver metal boxes you see along the railroad right of way and in the control towers. The towers contained “interlocking machines” which had an elegant system of levers/cams/bars that prevented a tower operator from putting two trains on the same track (usually leads to a “train wreck”). Dammed rugged stuff, probably survive an EMP blast just fine.
Personally, I prefer a Hammer to a computer, they are all Rev 1.0, and never need rebooting, ha ha ha …..
Dr. Ball, very nice essay, thanks for your time demonstrating the flaws in climate models.
Cheers, Kevin.
Paul Coppin says:
March 21, 2014 at 4:08 pm
Ahhh . . . the Trash80!
At the heart of a scientific scandal that will make Piltdown Man look like a practical joke, that’s where…
====================
When I learned about Piltdown Man, I thought “how could people back then have been so stupid?”. I was so obviously a fraud. Yet everyone believed because they wanted to believe.
Cars are a nuisance. We want to believe they are bad. So we can get rid of them and replace them with something better. Like the bus?? Or the bicycle?? Or walking?? Or the horse??
Frank says:
March 21, 2014 at 5:51 pm
Having adequate data to initialize models is essential for weather forecast model, but I’m not sure why initialization data is important for century scale climate projections. ENSO variability averages out over a century, but not one or two decades
===========
If initialization doesn’t matter, why keep temperature records? Why train models using past data? If ENSO average out over 100 years, where is the data to support this? Or are these simply assumptions?
NoFixedAddress says:
March 21, 2014 at 6:02 pm
Until one computer modeler can predict next week’s ‘lotto’ (6 from 45)
========
NASA GISS has just such a computer. It cost $10 billion. Every week it predicts the wrong answer for next week’s ‘lotto’. However, in hind casting it manages to predict last weeks’ number almost 1/2 the time.
My first work computer was a Royal McBee LGP-30, circa 1962. Input was via keyboard; programs were stored on 1″ mylar punched tape. No air conditioned sanctum sanctorum with priests in white lab coats, I could get in and use it anytime it was open. Later that year, one of my professors said, “Ach! Vun of you has done his homevork vit a computer. Vell, I didn’t zay dot you couldn’t. Bezides, vun day, all homevork vill be done on computers. Ve vill all have computers to do our calculations.” He was right. I’m still impressed.
I once worked near the ocean and was impressed on my morning commute by how many different ways the sun could reflect off the water. Ocean albedo is a function of solar zenith angle, wind direction and velocity, tides, salinity, seafoam, air and water temperature, currents, pollution, and (so help me!) plankton. There is no way that the GCM’s account for all of these continuously, rapidly changing variables. Oceans make up 71% of the Earth’s surface. If you can’t get the ocean albedo right, you can’t model the system.
Matthew R Marler says:
March 21, 2014 at 6:56 am
It isn’t clear which of the problems you address are simply problems with current models
==========
The problem is inherent in prediction from first principles. Chaos makes any such prediction mathematically impossible because of round off errors in digital computers.
There are other techniques that have proven to work. We can predict the tides by decomposition of the orbital frequencies of sun, moon and planets. However modern science rejects this because the underlying principle in tidal calculation is Astrology.
However, to forecast the tides according to first principles as is done with climate models? Please, it has never been done successfully, even after centuries of trying. For $15.95 the Old Farmers Almanac uses Astrological principles and routinely outperforms $$ Billion dollar climate models used by the IPCC.
izen: There is a common error in both the posted article and many of the responses that has to do with the difference between models intended to predict a specific state and models intended to simulate a physical process.
I think you need to direct that comment toward anyone who proposes or accepts that the current GCMs form a reasonable basis for informing policy decisions. I personally think that the current GCMs are amazing achievements, and incorporate a great deal of scientific knowledge in computable format. That is, a complex computer program is a summary and codification of knowledge in the same sense that the periodic table was, but translates all knowledge representations into explicit computational procedures and makes use of estimates of physical constants. However, the current GCMs also are wrong in their forecasts, showing that they are worthless for anticipating or planning for the future. Either the codified “knowledge” is incorrect, the models are inadequate, the computational techniques are inadequate, the parameter estimates are too inaccruate, too much is unknown and omitted from what may eventually be an adequate and accurate model, or (add to the list, which is not exhaustive) some of all of those.
We commentators are not those making the error that you accuse us of (or at least “many” of us, undifferentiated.). We are showing that the limits of accuracy imply the existence of important shortcomings in the knowledge of the physical processes that they simulate. And we claim that the demonstrated inaccuracy to date provides adequate(!) reason to doubt that the simulations have any practical utility beyond guiding the development of better models.
I agree with Matthew Marler on the uselessness of the current GCMs for the purpose of planning for the future but seem to disagree with him on the merits of these models. Firstly, the GCMs reference no underlying statistical population but it is the observed events in this population that tie the associated model to reality. Secondly, they lack means for updating the state of the climate to the observed state at beginning of each event in the population; as the GCMs lack these means, their errors are unbounded on the up side. Thirdly, the GCMs fail to extract from the observational data all of the information that is available in it but no more than this information. Had the planners of the global warming research program addressed each of these shortcomings from the outset, they would have realized that 30 year forecasts of climatological outcomes are not a possibility during the next 4500 years, for the observed events will be far too few.