Note: two events at AGU13 this morning dovetail in with this essay. The first, a slide from Dr. Judith Lean which says: “There are no operational forecasts of global climate change”.
The second was a tweet by Dr. Gavin Schmidt, attending Lenny Smith’s lecture (which I couldn’t due to needing to file a radio news report from the AGU press room) that said:
Smith: usefulness of climate models for mitigation 'as good as it gets', usefulness for adaptation? Not so much #AGU13
— Gavin Schmidt (@ClimateOfGavin) December 10, 2013
With those events in mind, this essay from Dr. William Gray (of hurricane forecasting fame) is prescient.
Guest essay by Dr. William M. Gray
My 60-year experience in meteorology has led me to develop a profound disrespect for the philosophy and science behind numerical climate modeling. The simulations that have been directed at determining the influence of a doubling of CO2 on Earth’s temperature have been made with flawed and oversimplified internal physical assumptions. These modeling scenarios have shown a near uniformity in CO2 doubling causing a warming of 2-5oC (4-9oF). There is no physical way, however, that an atmospheric doubling of the very small amount of background CO2 gas would ever be able to bring about such large global temperature increases.
It is no surprise that the global temperature in recent decades has not been rising as the climate models have predicted. Reliable long-range climate modeling is not possible and may never be possible. It is in our nation’s best interest that this mode of prophecy be exposed for its inherent futility. Belief in these climate model predictions has had a profound deleterious influence on our country’s (and foreign) governmental policies on the environment and energy.
The still-strong—but false—belief that skillful long-range climate prediction is possible is thus a dangerous idea. The results of the climate models have helped foster the current political clamor for greatly reducing fossil fuel use even though electricity generation costs from wind and solar are typically three to five times higher than generation from fossil fuels. The excuse for this clamor for renewable energy is to a large extent the strongly expressed views of the five Intergovernmental Panel on Climate Change (IPCC) reports, which are based on the large (and unrealistic) catastrophic global warming projections from climate models.
The pervasive influence of these IPCC reports (from 1990 to 2013) derives from the near-universal lack of climate knowledge among the general population. Overly biased and sensational media reports have been able to brainwash a high percentage of the public. A very similar lack of sophisticated climate knowledge exists among our top government officials, environmentalists, and most of the world’s prestigious scientists. Holding a high government position or having excelled in a non-climate scientific specialty does not automatically confer a superior understanding of climate.
Lack of climate understanding, however, has not prevented our government leaders and others from using the public’s fear of detrimental climate change as a political or social tool to further some of their other desired goals. Climate modeling output lends an air of authority that is not warranted by the unrealistic model input physics and the overly simplified and inadequate numerical techniques. (Model grids cannot resolve cumulus convective elements, for example.) It is impossible for climate models to predict the globe’s future climate for at least three basic reasons.
One, decadal and century-scale deep-ocean circulation changes (likely related to long time-scale ocean salinity variations), such as the global Meridional Overturning Circulation (MOC) and Atlantic Thermohaline Circulation (THC), are very difficult to measure and are not yet well-enough understood to be included realistically in the climate models. The last century-and-a-half global warming of ~0.6oC appears to be a result of the general slowdown of the oceans’ MOC over this period. The number of multidecadal up-and-down global mean temperature changes appears also to have been driven by the multidecadal MOC. Models do not yet incorporate this fundamental physical component.
Two, the very large climate modeling overestimates of global warming are primarily a result of the assumed positive water-vapor feedback processes (about 2oC extra global warming with a CO2 doubling in most models). Models assume any upper tropospheric warming also brings about upper tropospheric water-vapor increase as well, because they assume atmospheric relative humidity (RH) remains quasi-constant. But measurements and theoretical considerations of deep cumulonimbus (Cb) convective clouds indicate any increase of CO2 and its associated increase in global rainfall would lead to a reduction of upper tropospheric RH and a consequent enhancement (not curtailment) of Outgoing Longwave Radiation (OLR) to space.
The water-vapor feedback loop, in reality, is weakly negative, not strongly positive as nearly all the model CO2 doubling simulations indicate. The climate models are not able to resolve or correctly parameterize the fundamentally important climate forcing influences of the deep penetrating cumulonimbus (Cb) convection elements. This is a fundamental deficiency.
Three, the CO2 global warming question has so far been treated from a “radiation only” perspective. Disregarding water-vapor feedback changes, it has been assumed a doubling of CO2 will cause a blockage of Outgoing Long-wave Radiation (OLR) of 3.7 Wm-2. To compensate for this blockage without feedback, it has been assumed an enhanced global warming of about 1oC would be required for counterbalance. But global energy budget considerations indicate only about half (0.5oC, not 1oC) of the 3.7 Wm-2 OLR blockage of CO2 should be expected to be expended for temperature compensation. The other half of the compensation for the 3.7 Wm-2 OLR blockage will come from the extra energy that must be utilized for surface evaporation (~1.85 Wm-2) to sustain the needed increase of the global hydrologic cycle by about 2 percent.
Earth experiences a unique climate because of its 70 percent water surface and its continuously functioning hydrologic cycle. The stronger the globe’s hydrologic cycle, the greater the globe’s cooling potential. All the global energy used for surface evaporation and tropospheric condensation warming is lost to space through OLR flux.
Thus, with zero water-vapor feedback we should expect a doubling of CO2 to cause no more than about 0.5oC (not 1oC) of global warming and the rest of the compensation to come from enhanced surface evaporation, atmospheric condensation warming, and enhanced OLR to space. If there is a small negative water-vapor feedback of only -0.1 to -0.3oC (as I believe to be the case), then a doubling of CO2 should be expected to cause a global warming of no more than about 0.2-0.4oC. Such a small temperature change should be of little societal concern during the remainder of this century.
It is the height of foolishness for the United States or any foreign government to base any energy or environmental policy decisions on the results of long-range numerical climate model predictions, or of the recommendations emanating from the biased, politically driven reports of the IPCC.
###
William M. Gray, Ph.D. (gray@atmos.colostate.edu) is professor emeritus of atmospheric science at Colorado State University and head of the Tropical Meteorology Project at CSU’s Department of Atmospheric Sciences.
The idea that although you cannot predict weather more than 72 hours ahead worth a dam you can be throwing a enough numbers at it predict climate for years or decades ahead to two decimal places of accuracy , when all climate really is weather over a long period . Always seemed odd to me .
In fact in the past this was a none issues , not because we got better at it since , but because everyone accepted that was the nature of weather dam hard to predict . But now with the ‘settled science’ and all the money and personal ego ridding on it the ‘need’ to pull this trick off has become much stronger.
Excellent summary! In statistical model building terms, what he’s saying is not only are the main effects not known, but the interaction terms aren’t known either.
Take the equation C=f(w, x, y, z….) + e as shown in the first graphic. Assuming only linear relationships in only w, x, y, and z the full empirical model is thus: f(w,x,y,z)=a1w+a2x+a3y+a4z+a5wx+a6wy+a7wz+a8xy+a9xz+a10yz+a11wxy+a12wxz+a13wyz+a14xyz+a15wxyz +e.
Now add more terms and add non linearity (like a simple Taylor expansion) and the complete number of terms you have to understand is astronomical.
Anytime you hear the terms “feedback” “feedforward”, “dependence on” you are hearing about an interaction term.
Terry Oldberg says:December 10, 2013 at 6:17 pm
From online Dictionaries
Predict – To state, tell about, or make known in advance, especially on the basis of special knowledge.
Project – To calculate, estimate, or PREDICT (something in the future), based on present data or trends
Your bullshit exposed.
markx says: December 11, 2013 at 1:48 am
Ice core strongly suggests that CO2 increases lag Temperature Increases by approximately 600-800 years. The increase in CO2 we are seeing now could easily be related to past increases in temperature as we come out of the last Ice Age.
Richard M says: December 10, 2013 at 7:31 pm “Whether the MOC is behind the PDO/AMO cycles is another interesting question”
But what drives the changes in the MOC and PDO/AMO cycles is the really BIG question.
wrecktafire says:
December 10, 2013 at 9:46 pm
Forecasting is just guessing with computers.
The slide at the top says it all: error effects saturate weather forecasting models when run out to the 10-14 day mark. The same thing happens with climate models that get the clouds and the water vapor wrong.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
Forecasting (weather) is integrating forward in time the whole Earth’s weather (a snapshot of chaos). Now because we necessarily cannot quantify every last detail of that snapshot (incomplete/imperfect observation) AND because each chaotic moment has it’s own sensitivity (may spin more/less from the current chaotic state) …. NWP models are constrained (though considerable progress has been made in the last 40 years – a 5 day forecast verifies as well now as did a 2 day one then).
GCM’s however do NOT model weather (your clouds/water vapour), because of the above reasons and because the computational expense constraints. Parametrization is used to model such processes instead.
@TB: your reply to me hints that weather modeling and climate modeling are somehow different, due to parameterization as opposed to integration. I don’t believe that is a proper description. A climate model integrates physical processes over time in exactly the same fashion as a weather model. The parameterization is a shorthand for stuff that is too resource intensive to compute, so we put in something simpler that we *hope* is an adequate substitute. Performing this procedure billions of times results in significant errors.
KNR says:
December 11, 2013 at 1:59 am
The idea that although you cannot predict weather more than 72 hours ahead worth a dam you can be throwing a enough numbers at it predict climate for years or decades ahead to two decimal places of accuracy , when all climate really is weather over a long period . Always seemed odd to me .
In fact in the past this was a none issues , not because we got better at it since , but because everyone accepted that was the nature of weather dam hard to predict . But now with the ‘settled science’ and all the money and personal ego ridding on it the ‘need’ to pull this trick off has become much stronger.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
This seems to be a common conception – but is a fallacy.
See my post above.
Climate is NOT weather projected into the future. Climate creates weather – hang on – in the sense that the Earth acts as a heat engine. It receives energy from the Sun and it is reflected/absorbed/re-radiated back to space. Only two factors essentially govern it’s working – energy in vs energy out, this is constrained by albedo and radiative forcing (and for past epochs by orbital eccentricity). Weather is the chaos in the system – the noise on the general climate trend (up or down). To suppose that climate is weather projected forward is missing seeing the wood by only seeing the trees.
Let’s say there was no global warming scare, no IPCC, no fear mongering etc. People would adapt and mitigate as they have always done since the end of the last glaciation. No need for global action. What an utter waste of time, can’t these people find something else better to do?
MAN-MADE
✔Thames flood barrier
✔Dutch Polders
✔Flood defences
✔Clothing, hats, gloves, UV sunglasses
✔Air conditioning, fans, central heating
✔Double glazing, triple glazing (Scandinavia)
✔Homes on stilts around the world, igloos, tree houses
✔Bad weather warnings
✔New farming methods, livestock breeding for desired traits
✔Water wells, boreholes and desal plants
✔Use of fire
✔Concrete
✔Sand replenishment on beaches
✔White painted homes and walls in the tropics
✔Tornado hatches
etc., etc., etc.
NATURAL
✔Bangladesh gaining landmass over the last few decades
✔Coral island atolls rise with sea level rise
✔Marshlands change with sea level rise
✔Plants move uphill / plants move downhill
✔ Animals move north / south
✔Birds bigger / birds smaller 😉
✔Greening of the biosphere in recent decades
✔Gaia hypothesis at work?
✔Natural climate oscillations
✔More clouds / less clouds
✔More NH winter snow, bigger albedo
✔Less NH spring snow, lower albedo
✔Antarctica sea ice extent up (allegedly caused by global warming)
✔Antarctica recent cold record shattered! (This may be soon blamed on AGW)
✔2009-2011 extreme snowfalls in East Antarctica (maybe due to AGW)
etc., etc., etc.
Yet we feel the need to call special global conferences to deal with things we have been dealing with at a local and regional level since we discovered fire.
“Smith: usefulness of climate models for mitigation ‘as good as it gets’, usefulness for adaptation? Not so much #AGU13—
Gavin Schmidt (@ClimateOfGavin) December 10, 2013”
The models predict warming, but it isn’t warming.
Smith and Schmidt are both liars.
Spot on!
Terry Oldberg says:
December 10, 2013 at 7:08 pm
Perhaps you could train the UKMO to distinguish between projections and predictions:
Reliable long-range climate modeling is well possible if you can forecast the solar signal at a monthly noise level and successfully model all the regional atmospheric and oceanic responses. Long range regional forecasting at the scale of weather for say UK/Euro based on short term solar effects on the AO/NAO is far simpler than modelling global averages or trends.
“CO2 global warming question has so far been treated from a “radiation only” perspective”
Dr. Gray is close to “getting it”, but not quite.
The net effect of radiative gases in our atmosphere is cooling at all concentrations above 0.0ppm.
To understand this, you only have to understand 4 simple points –
1. Radiative cooling at altitude is critical to continued strong vertical tropospheric circulation in the Hadley, Ferrel and Polar tropospheric convection cells.
2. The observed lapse rate below the topopause is a result of strong vertical circulation of gases across a gravity induced pressure gradient.
3. Incident LWIR cannot heat nor slow the cooling rate of liquid water that is free to evaporatively cool. (sorry Willis, but empirical experiment proves you wrong)
4. Without radiative cooling at altitude and the resultant strong vertical tropospheric circulation, our atmosphere would trend isothermal, with its temperature driven by surface Tmax, not surface Tav. (sorry Dr. Spencer, but empirical experiment proves you wrong)
That’s it. That’s all you need to know. Adding radiative gases to the atmosphere will not reduce the atmospheres radiative cooling ability. Global warming is a physical impossibility.
“considerations of deep cumulonimbus (Cb) convective clouds indicate any increase of CO2 and its associated increase in global rainfall would lead to a reduction of upper tropospheric RH and a consequent enhancement (not curtailment) of Outgoing Longwave Radiation (OLR) to space.”
Willis’s Thermostat anyone?
TB says:
December 11, 2013 at 3:27 am
Climate is NOT weather projected into the future. Climate creates weather – hang on – in the sense that the Earth acts as a heat engine. It receives energy from the Sun and it is reflected/absorbed/re-radiated back to space. Only two factors essentially govern it’s working – energy in vs energy out, this is constrained by albedo and radiative forcing (and for past epochs by orbital eccentricity).
I believe this is partly wrong according to Dr. Gray. Dr. Gray is telling us that the energy already in the system is very important. The changes in the speed of the MOC change the amount of energy released to the atmosphere by the oceans. While big changes in solar energy would make a difference, they don’t appear to happen for the most part. The earth itself through various feedbacks (like clouds) tends to balance it out. This leaves the oceans as the primary driver of climate *change*.
—————-
On a separate note, Dr. Gray’s description fits with so many others in Nature. It’s always quite amazing to see how many systems work synchronously. This appears to be another one. As CO2 increases you get both a small increase in temperature and a small increase in rain. All three of these are key ingredients for plant growth (which then supports a growth in fauna). It fits perfectly that they would work in such a synchronous fashion. From what I can tell the ideal amount of CO2 would be 1200 ppm. That would raise the global temperature another .5C and increase rainfall about 3%. That might be just enough to ward off the next glaciation event. It would also help feed a growing global population. What’s not to like about it.
Whenever my hero speaks, I listen. And while not one of the 4 Gospels, Dr William Grays
http://typhoon.atmos.colostate.edu/Includes/Documents/Publications/gray2012.pdf
is as close as you can get without being in scripture… at least to me.
TB says:
“Weather is the chaos in the system – the noise on the general climate trend (up or down).”
Indices of the North Atlantic Oscillation and the Arctic Oscillation show correlations on the day-to-day timescale with the solar wind speed…
http://www.sciencedirect.com/science/article/pii/S0273117713005802
Now if we transformed that four-syllable abbreviation into a two-syllable acronym, we get something much smoother (and sillier!):
DirkH: ‘The models predict warming, but it isn’t warming. Smith and Schmidt are both liars ‘.
If I let water flowing into my bath tub, my common sense model predicts that the water level will rise unless something unexpected happens in between, for example that someone removed the stop. My prediction is conditional and its failure does not mean that I am a liar (how dare you). Because the climate models failed in their prediction, they can be saved by the unexpected event that all heat disappeared into the deep ocean. Unexpected means that this heat flow occurs perhaps once in every million years. If the stop is put back, surface temperatures will rise to unprecedented levels, unless something unexpected happens in between.
Climate is not weather.. true, yet the patterns of weather over prolonged periods of time become climate. In the end, you are asked to believe that the design of the system along with the sun, the oceans and stochastic events, will be changed in an way that it will change where it would go naturally, but a trace gas that is needed for life on the planet and has been far higher with COLDER CONDITIONS, and lower with warm.
It is fantasy, and nature, not man, controls the climate of the planet
So like a weather forecast.. one must weigh the parameters given and then make the forecast. Given the flip of the pdo and soon the amo, the planet will ( has started ) cool. Gray opined we would warm in the 1970s, was spot on about the leveling and begin of the turn around. Far better than the screams of tipping points by people who have never made a forecast in their life that they have been held accountable for. And that is the big difference also. In what I do in the private sector, against many super talented private and gvt mets, if I do not bring value to the table, I get fired. Well I am 58 and I am still standing. But if you never learn that, because you simply make statements that this will happen and that will happen in some far off time, or when the opposite happens say its all part of what you thought, you can again see what the real difference between weather and climate is… THE BOTTOM LINE. In forecasting, you have to be judged by the result of your idea. Apparently the AGW crowd thinks that isnt an issue, that eventually they will be right.
Anyone who says they can forecast the climate 50 or 100 years ahead is either a liar or a fool. Or, most likely, both.
I’m sure the three technical reasons given are valid. But even if the models had a perfect understanding of how the climate works, they would still almost certainly fail, because climate systems are fundamentally chaotic and the initial conditions can never be known with sufficient precision.
That damned butterfly in an Amazonion forest has a lot to answer for….
Chris
TB says:
December 11, 2013 at 3:27 am
“Climate is NOT weather projected into the future. Climate creates weather – hang on – in the sense that the Earth acts as a heat engine. It receives energy from the Sun and it is reflected/absorbed/re-radiated back to space. Only two factors essentially govern it’s working – energy in vs energy out, this is constrained by albedo and radiative forcing (and for past epochs by orbital eccentricity). Weather is the chaos in the system – the noise on the general climate trend (up or down). To suppose that climate is weather projected forward is missing seeing the wood by only seeing the trees.”
This wrong in so many ways, I don’t know what to point out first!
1. The Earth is not in equilibrium, it does not have ‘A Climate’ in the sense you use it:
“Moreover, it hardly needs stating that the Earth does not have just one temperature. It is not in global thermodynamic equilibrium – neither within itself nor with its surroundings.
It is not even approximately so for the climatological questions asked of the temperature
field. Even when viewed from space at such a distance that the Earth appears as a point
source, the radiation from it deviates from a black body distribution and so has no one
temperature [6]. There is also no unique “temperature at the top of the atmosphere”. The
temperature field of the Earth as a whole is not thermodynamically representable by a single
temperature.”
[6] Essex C., Kennedy D., Berry R. S., How hot is radiation?, Am. J. Phys., 71 (2003),
969–978.
2. To talk about the Earth’s ‘climate’ as if it was independent of its Geography, let alone its Geology is absurd. I’ll list some of the ‘essential factors’ below:
a. Oblate spheroid, rotating on axis creating uneven “energy in”! We know these as the seasons! Precession is in flux.
b. Diameter and hence, speed of rotation at equator faster, creating Coriolis effect, which dominates the global circulation patterns (The trade winds). Creating the major climatic zones (Desert/Jungle)
c. The shape and geographical distribution of land masses and bodies of water (All of which are in flux).
e. The Earth’s magnetic field, without which, there would be no atmosphere (It would have blown away in the solar wind)
f. The Moon and it’s effect on rotation and tides
g. The earth’s elemental composition. Carbon is the fourth most abundant element in the universe yet most of the Earths is locked in the core. “Energy in” and carbon-dioxide are utilised by all life on earth and much is stored as biological mass via photosynthesis.
I probably didn’t point out the most important but you get the idea.
On the subject of “The Futility of long range Climate Predictions”
Exhibit A is the IPCC Science Fiction from AR5
2.5 Projected changes in the climate system pages 30ff
Headlines (expect to see these in the media soon):
CO2 will drive warming for centuries to come
Surface air temperature will be up to 4.8C higher by 2100
More hot extremes and heat waves
Less rain in dry areas and more rain in wet regions
Stronger and wetter cyclones
Nearly ice-free Arctic by mid century
NH spring snow cover reduced up to 25%
Near surface permafrost reduced up to 81%
Global glacier volume by 2100 reduced up to 85%
Global mean sea level rises up to 1 m by 2100
Ocean surface ph 0.3 more acid by 2100
I think we are supposed to fearful enough to stop using fossil fuels.
TB says:
December 11, 2013 at 3:27 am
Weather is the chaos in the system – the noise on the general climate trend (up or down).
============
That is what the models believe, but it is an over simplification. If it was true then climate would be predictable, in the sense that it would be subject to the Law of Large Numbers. Over time you would expect to see a statistically predictable trend. ie: you could predict if climate was statistically more likely to warm or cool.
However, that is not what you see. At all time scales climate is a fractal distribution. It does not converge about an average, because it has no constant mean. As a result most statistical analysis of climate is misleading at best.
Weather is not the noise in the climate system. Weather and Climate are measurements of the same physical process at different time scales. As you expand the time scale, weather becomes climate and remains as unpredictable.
So what is a fractal distribution and how does it differ? When one graphs any physical process, typically you get some sort of a wavy line. If you expand the time scale, if the line becomes less wavy then the process is becoming more predictable over time.
If however the line does not become less wavy, if it maintains the same irregularities at different scales, then you likely have a fractal distribution. This sort of process does not become statistically more predictable as you increase the time scale.
Now look at a graph of earth’s average temperature over the past million years as compared to the past 1000 years or the past 1 year or the past day? Does climate show any less variability at longer scales? No. If anything climate over the past 1 million years shows greater variability, which shows that climate is no more predictable than weather. The farther you look into the future, the less reliable the prediction.