Network News

X My Profile
View More Activity
The new Washington
Post Weather website
Jump to CWG's
Latest Full Forecast
Outside now? Radar, temps
and more: Weather Wall
Follow us on Twitter (@capitalweather) and become a fan on Facebook
Posted at 11:30 AM ET, 05/24/2010

Hurricane chimp's cheap shot at climate science

By Andrew Freedman

* Drier weather develops: CWG's Full Forecast *

Are poor hurricane season forecasts a reason to doubt climate science? A conservative Washington think tank believes the answer is yes, and they are using a monkey -- pardon me, "Dr. James Hansimian" -- to prove it. (The monkey's name, by the way, is a dig at well known NASA climate scientist James Hansen, who incidentally does not forecast hurricanes).

According to the National Center for Public Policy Research (NCPPR), a monkey rolling a pair of dice can more accurately predict the number of hurricanes that will form in a given season than any of the more sophisticated human methods. Moreover, the think tank implied in a press release on May 18, all climate science projections are about as reliable as hurricane season forecasts.

Now I have nothing against monkeys, nor do I nor most climate scientists claim that climate change projections are 100 percent accurate, but NCPPR's comparison between seasonal forecasts of hurricane activity and long-term climate predictions makes no sense -- it's like comparing apples and giraffes -- and falls flat as an effort to attack the credibility of climate science.

The press release makes clear where the "monkeying around" is directed: "The video isn't intended to needle the National Oceanic and Atmospheric Administration (NOAA) for its erroneous forecasts, but to make a larger point about our current understanding of climate."

"NOAA's forecasts have been wrong not because of a lack of dedication or competence of its forecast team, but because climate science is really still in its infancy," NCPPR president Amy Ridenour said. "We should remember this as we consider whether to adopt economically-ruinous caps on energy. If we can't rely on 6-month forecasts, how can we rely on forecasts of what rising carbon concentrations will do to our climate 25, 50 or even 100 years out?"

The think tank's key mistake lies in conflating short-term prediction of small-scale weather events, such as hurricanes and tropical storms, with long-term forecasts of large-scale, global climate (e.g., average seasonal or annual temperature and precipitation for a region or the globe). The geographical scales and timeframes of the two are far different, as are the tools needed to project them with any accuracy.

NCPPR also fails to state exactly what part of climate science is in its "infancy." For example, it's been known since the 19th century that certain gases, such as carbon dioxide, warm Earth's climate and that pumping more of such gases into the air will cause temperatures to increase. That hardly qualifies as a newborn science.

A seasonal hurricane forecast is simply not analogous to a seasonal climate forecast for, say, the summer in Washington, or a multi-decadal climate forecast as cited in NCPPR's press release. Conflating the two is flawed in much the same way it is to argue that because weather forecasters can't predict the weather two days from now with perfect accuracy, that we can't make informed decisions based on climate projections for the more distant future. They are two different things.

Here's how: With hurricane season forecasts, meteorologists examine particular factors that are known to encourage or inhibit tropical storm and hurricane formation. Such factors include the presence or lack of El Nino (warmer-than-average sea-surface temperatures in the equatorial Pacific Ocean), sea-surface temperatures in the area of the Atlantic where most hurricanes form, and other variables. The end forecast is determined in part by looking back to the nature of hurricane activity in previous years when conditions were similar.

This year, forecasters are concerned that the disappearance of El Nino, combined with unusually warm sea-surface temperatures in the Atlantic, will lead to a rude awakening after last year's quiet season.

Although hurricanes are nature's largest and fiercest storms, they are weather phenomena that are significantly influenced by small-scale factors, ranging from ocean eddies to wind shear to the amount of dust in the atmosphere. Predicting how these influences will line up whenever a particular storm tries to develop is no easy task, for man or monkey.

To project climate change, on the other hand, scientists examine the larger factors that influence climate on long time scales. Such natural and manmade factors include variations in solar output, changes in the composition of Earth's atmosphere (including concentrations of greenhouse gases), and variations in Earth's rotational axis around the sun, among others. These natural and manmade factors are fed into long-term climate models, which are first calibrated by making sure they accurately simulate past climate.

While some sources of climate variability are not so predictable, such as volcanic eruptions that can cool the climate for several years, many of the factors are actually measured quite well, and there is a high degree of confidence in how they will change in the future. For example, in 2007 the U.N. Intergovernmental Panel on Climate Change concluded that there is a 90 percent chance or greater that recent warming is mainly due to human activities, and a high likelihood that more significant climate change is on tap for coming decades.

Determining the number of hurricanes that will form in a season is actually more complex than projecting that the global climate will be warmer in 20 years than it is now, based on the likelihood that climate-warming gases will continue to increase due to industrial activities.

Interestingly, NCPPR is not far off in its contention that hurricane season outlooks have little accuracy, at least when issued far in advance of the season's June 1 start. Last year, for example, NOAA's initial hurricane season forecast called for between nine and 14 named storms to form in the Atlantic Basin, with four to seven hurricanes and two major hurricanes (Category 3 or greater). In reality the season underperformed, due in large part to rapidly strengthening El Nino conditions in the Pacific.

However, the think tank's effort -- presumably politically motivated -- to spread the message that climate scientists know as much about Earth's future climate as a monkey rolling a pair of dice is laughable. Not just because monkeys are funny, but because predicting that the climate will warm in the next several decades, due largely to increased concentrations of greenhouse gases such as carbon dioxide, is fundamentally different than forecasting a particular hurricane season, regardless of whether a human or monkey issues the outlook.

The views expressed here are the author's alone and do not represent any position of the Washington Post, its news staff or the Capital Weather Gang.

By Andrew Freedman  | May 24, 2010; 11:30 AM ET
Categories:  Climate Change, Freedman, Science  
Save & Share:  Send E-mail   Facebook   Twitter   Digg   Yahoo Buzz   Del.icio.us   StumbleUpon   Technorati   Google Buzz   Previous: Forecast: Showers to dissolve, dry weather nears
Next: PM Update: Clouds linger a little longer

Comments

Well, considering the poor accuracy (and, in some cases, VERY poor) North Atlantic Hurricane-season forecasts made in the several years after Katrina's wake, from 2006 to 2009, I can understand, at least to some extent, them using a chimp to compare them to. We were told, in effect, that Katrina and Rita would be only the start of regular massive, destructive seasons that would occur regularly with global warming.......and, since 2005 and Katrina, nothing could have actually been further from what actually happened. 2006, in fact, turned out to be a almost total wimp of a season, and 2007, 2008, and 2009 were not very impressive either.

So, I'm sorry, no offense, Mr. Andrew, but in this case, I have to believe the chimp.

Posted by: MMCarhelp | May 24, 2010 11:45 AM | Report abuse

MMCarhelp, I would urge you to speak to someone in Texas, Louisiana, or especially Cuba before claiming 2008 a "wimpy" season.

Posted by: Brian-CapitalWeatherGang | May 24, 2010 12:04 PM | Report abuse

Andrew, that is not a monkey. That is a chimpanzee, who is no closer to being a monkey than you and I are.

The IPCC AR4 said little about tropical storms, only noting a potential for an increase in the numbers of the most intense tropical cyclones but with a decrease in the overall number of storms, but not putting a lot of confidence in that projection.

Like the seasonal prediction of hurricane frequency, the decadal prediction of hurricane frequency has a long way to go. There are not enough of these storms observed (thank goodness) to give statistical confidence, and while the energetics of already developed storms is fairly well understood, cyclogenesis of new storms remains a physically chaotic process.

Posted by: imback | May 24, 2010 12:16 PM | Report abuse

Even if the earth is warming, variability of day-to-day weather shows little impact in relation to long-term trends. This past winter we were largely below normal with record snowfall, but apparently the majority of the earth was warmer than normal...indicating we MAY be undergoing global warming.

The important factor in relation to climate change is determination whether the current "warming" is a long-term effect or whether it is part of a natural climatic cycle. The scary factor here is the data regarding disappearance of snow and ice at tropical latitudes [e.g., Andes, Kilimanjaro]; receding and disappearance of glaciers; shrinking of polar icecaps; warming in the tropics, etc. It may still need to be determined whether this apparent "warming" is due to increased greenhouse gases from artificial sources, or is part of a natural long-term cycle.

Posted by: Bombo47jea | May 24, 2010 12:16 PM | Report abuse

"In reality the season underperformed, due in large part to rapidly strengthening El Nino conditions in the Pacific."
So, the models are right, what actually happened is wrong!

Posted by: das75428 | May 24, 2010 12:54 PM | Report abuse

imback: Technically, you are correct. "Dr. Hansimian" is a chimp, not a monkey. However, when used informally (as in this blog), the distinction between monkeys and chimpanzees, as well as orangutans actually, is often blurred. But thanks for pointing it out, I appreciate it (no sarcasm intended!).

MMCarhelp: I clearly stated in the blog post that the hurricane season forecasts don't have a solid track record, so your point is in agreement with mine, rather than against it, as far as I can tell. Also, remember that it only takes one storm to make a season memorable, such as 1992's Andrew, which occurred in a less active year.

Posted by: Andrew-CapitalWeatherGang | May 24, 2010 1:42 PM | Report abuse

"The video isn't intended to needle the National Oceanic and Atmospheric Administration (NOAA) for its erroneous forecasts, but to make a larger point about our current understanding of climate."

ah, yes. the object is to sow seeds of doubt about climate change.... it's NOT even about hurricanes or hurricane predictions - it's about making mainstream scientists look stupid. the image of a chimp in a scientist costume will stick. damn, those bastages* are clever.....

(*see "johnny dangerously")

Posted by: walter-in-fallschurch | May 24, 2010 3:36 PM | Report abuse

How is the ad hom attack against Hanson different from his attack on specific people: "James Hansen, one of the world's leading climate scientists, will today call for the chief executives of large fossil fuel companies to be put on trial for high crimes against humanity and nature, accusing them of actively spreading doubt about global warming in the same way that tobacco companies blurred the links between smoking and cancer."

http://www.guardian.co.uk/environment/2008/jun/23/fossilfuels.climatechange

Or how is different from calling Bush a chimp (not that anyone did that here)? To me this ad hom attack is light hearted, meant to be humorous. There is no doubt that the stronger hurricanes in 2005 were used to make general predictions of more strong hurricanes which didn't turn out that way.

As for the models, climate models MUST model ENSO and all of the other factors that impact hurricanes if they have a shred of hope for predicting climate. They don't necessarily need to predict ENSO to do that, but they have in fact been used to predict ENSO. So the argument that they are not relevant for seasonal hurricane predictions is a little weak.

Posted by: eric654 | May 24, 2010 5:05 PM | Report abuse

so, when CWG offers a long-range forecast, say for the 2011 winter snowfall totals, would you call that a "weather" or a "climate" prediction/projection?
================================

eric, did you get "satisfaction" from john cook on the ice modeling issue at that skepticalscience website?

Posted by: walter-in-fallschurch | May 24, 2010 5:16 PM | Report abuse

So are you saying that the predictions for climate change are, in fact, more accurate than for hurricane developement? If so, where is the proof of that. Your statement on the 2007 IPCC climate forecast does not tell us anything about its accuracy nor do you mention the accuracy of any previous climate change models (except to say that many of the factors are actually measured quite well). What you fail to note is just how accurate the models from 20-50 years ago for today were/are. Telling me that the presumably politically motivated IPCC is 90% confident in future forecasts tells me nothing about the record of how well models have done in the past when measured against actual results.

While the method of these forecasts are fundementally different, I think the purpose was to point out the overall accuracy of weather data models and predictions of complex phenomena, be they short term or long term. If you don't provide evidence that climate change models are more accurate than hurricane forecasts, your arguement falls short. Essentially you are trying to answer the following question: "If we can't rely on 6-month forecasts, how can we rely on forecasts of what rising carbon concentrations will do to our climate 25, 50 or even 100 years out?" You can't do that by simply describing how the two predictions are different (apples and giraffes).

All that said, it is inherently more difficult to determine a long range forecast's accuracy since baselining the data is difficult for any number of reasons including having different and better measurements and technology than we did decades ago. By the end of the year, we will be able to count the number of named storms with close to, if not, 100% accuracy (something we could not do in 1900) and have the answer to how accurate the forecast was. By the end of the year we will have less than three hundreths of the answer for the 2007 IPCC forecast. Between now and the rest of the century, measurement points will change, techniques will change, technology will change making it much more difficult to determine how well the model actually worked.

Posted by: amaranthpa | May 24, 2010 5:27 PM | Report abuse

The chimp version is somewhat rigged because two dice are biased toward a result around 7. (The odds of getting a 6, 7, or 8 is only a little short of 50%.) Since (I think) that's about the average number of hurricanes there are in a season, the game's been rigged from the outset.

Posted by: kevinwparker | May 24, 2010 6:16 PM | Report abuse

NCPPR president Amy Ridenour: "If we can't rely on 6-month forecasts, how can we rely on forecasts of what rising carbon concentrations will do to our climate 25, 50 or even 100 years out?"

hahahaha good one, amy...

bad skeptic argument #45....(the difference between weather and climate) http://www.skepticalscience.com/weather-forecasts-vs-climate-models-predictions.htm

i mention this because it relates to my earlier question about whether seasonal hurricane predictions are "weather" or "climate". i'd say they're more like weather...

Posted by: walter-in-fallschurch | May 24, 2010 7:01 PM | Report abuse

amaranthpa: You seem to still be conflating weather prediction and short-term climate predictions with projecting long range changes in the earth's climate system. As for the reliability of climate models to date, Skeptical Science http://www.skepticalscience.com/climate-models.htm has some good info on that, and the folks at realclimate.org do as well.

I would be much more supportive of NCPPR's chimp tactics if they made an apples-to-apples comparison. Relating the number of hurricanes to 50-year climate projections just doesn't work from a scientific standpoint.

However, it is a good PR stunt, I'll give 'em that.

Posted by: Andrew-CapitalWeatherGang | May 24, 2010 8:08 PM | Report abuse

amaranthpa, for climate model evaluation, start with IPCC AR4 WG1 chapter 8. Follow the references for more detail. Then go on to chapters 9 and 10.

Browsing IPCC AR4 just now, I came across FAQ 9.1 Can Individual Extreme Events be Explained by Greenhouse Warming? which seems relevant to the current topic.

Excerpt:
Extreme events usually result from a combination of factors...The formation of a hurricane requires warm sea surface temperatures and specific atmospheric circulation conditions. Because some factors may be strongly affected by human activities, such as sea surface temperatures, but others may not, it is not simple to detect a human influence on a single, specific extreme event...[An example of a probabilistic approach for heatwaves is then given.]...The value of such a probability-based approach – 'Does human influence change the likelihood of an event?' – is that it can be used to estimate the influence of external factors, such as increases in greenhouse gases, on the frequency of specific types of events.

Posted by: imback | May 24, 2010 8:12 PM | Report abuse

Walter, seasonal prediction lurks in the netherworld between weather and climate, and it seems to inherit issues from both disciplines. Particularly difficult is predicting tropical storms, which are quite small in scale, both in space and in time, compared to the global climate scale.

Probabilistic prediction is the only way to go. The CSU seasonal storm forecast is a single number, which I find pretty misleading. NOAA forecasts a probabilistic range which is more faithful to the forecast uncertainty. Here was the 2009 outlook from last May; for the record, the 2009 Atlantic season had 9 named storms, 3 hurricanes, and 2 major hurricanes, on the low edge of the forecasted range.

Posted by: imback | May 24, 2010 8:18 PM | Report abuse

kevinwparker: The way I see it with a pair of dice, a single throw has 11 possible outcomes (2-12). Hence, the odds of getting 6, 7, or 8 are only 27% (3/11) - not that this nitpicking has anything to do with anything.

More relevant is that this is to be the first year that the CPC will be using the NCEP climate model as an additional input to the seasonal hurricane prediction. In testing, the model demonstrated a "fair" level of skill in predicting interannual variability of seasonal tropical storm activity. I'll have more to say on this pending what's including in the official release later this week

Posted by: SteveT-CapitalWeatherGang | May 24, 2010 8:30 PM | Report abuse

imback,
great points. the netherworld, huh.... and interesting to see the actual numbers (9,3,2) versus predicted (9-14,4,1) and see that it WAS at the low end of last year's prediction.

steve,
there are many more ways of rolling a 7 than there are of rolling a 2.

Posted by: walter-in-fallschurch | May 24, 2010 8:54 PM | Report abuse

oops,just demonstrated I've never played the dice

Posted by: SteveT-CapitalWeatherGang | May 24, 2010 9:19 PM | Report abuse

Steve, I'm looking forward to your coming post. But stay away from the craps table. There are 6x6=36 different pair of dice rolls, with 5 of them totaling 6, 6 of them totaling 7, and 5 of them totaling 8, so 16/36=44% of the rolls give a 6, 7, or 8. That said, I don't see why this means it's rigged. Essentially predicting an average number of storms would have zero skill value.

Posted by: imback | May 24, 2010 9:36 PM | Report abuse

Walter, I guess I'll have to spend some time on the "skeptic argument #45" thread now. It leaves out the single most important fact about weather and climate models which is that the climate models must model "short term (e.g. ENSO) fluctuations" not ignore them. Not only that, but climate models must model weather, not parameterize it. Otherwise they are GIGO.

The simplest example I can give is convection. If convection becomes more concentrated in a world with higher CO2, then that world will warm less than if it becomes less concentrated. The cooling effect of concentrated convection is well established. In order to model convection, the climate models must either have a much finer grid (which they don't due to high computation cost) or must incorporate the results of finely gridded weather models with initial conditions fed by the climate model. I honestly don't know if they do that and I am interested if anyone has the answer.

On SS.com in general, it's a good format to debate the talking points for dismissal (easy in some cases) or clarification or rewording (needed in some cases). I would say that some of the talking points are only red herrings which might have emanated from a "skeptic" or "denier" but have no value. Posting the red herrings is a rather transparent ploy to make "deniers" look like "idiots".

Specific answer? Not yet, but a productive discussion.

Posted by: eric654 | May 24, 2010 10:20 PM | Report abuse

imback and walter: Coincidentally, I just returned from a weekend in Atlantic City as a guest of friends. I was prepared to drop a quarter in a slot machine, but the minimum was inserting a dollar bill, which was over my limit. I also commented to friends when strolling by various table games that I don't have the slightest idea how to play any of them (really). Guess I know too much about probabilities, even after forgetting about the details of possibilities with a throw of the dice.

I'll stick to probabilities when dealing with uncertainties in weather and climate predictions.

Posted by: SteveT-CapitalWeatherGang | May 24, 2010 10:21 PM | Report abuse

steve,
i guess you were imagining rolling one 12-sided die.

we used to go to atlantic city annually for family vacations - something there for kids and adults. like you, i've never enjoyed "the tables" much. it was probably the time i lost $100 in blackjack in about 5 mins that soured me. i thought of all the other better ways i could have spent that money... i am always amused by people who think they're going to be the one to "beat the house" - as if the whole thing isn't created for the house to win....

Posted by: walter-in-fallschurch | May 24, 2010 11:30 PM | Report abuse

Eric, almost all weather and climate models simulate cumulus convection through parameterization. Cloud parameterizations are based on the known physics of clouds and are separately tested against field observations. They operate on the model resolution and compute a statistical description of the clouds within each model grid cell. From the highest resolution operational hurricane model to the lowest resolution millenial climate model, they all use similar cloud parameterization toolboxes. The parameterizations are getting good enough now that forecasting hurricane intensity with some skill is becoming a reality. The future probably will involve resolving individual clouds explicitly, but we don't quite have the required computer power yet. See section 8.2.1.3 in the IPCC chapter I linked to above for some more information.

I never heard of "concentrated convection" but I assume you're referring to a Lindzen hypothesis? Could you point to a peer-reviewed source?

Deniers often need no help to look like idiots. The same old red herrings, no matter how often debunked, are continually brought back to stink up the joint.

Posted by: imback | May 24, 2010 11:50 PM | Report abuse

imback, the response of models to changes in parameterizations have been thoroughly studied. I believe Lindzen's hypothesis is somewhat related (not too sure). Eschenbach had a complicated explanation on WUWT. The best source I found was here: http://www.ccsm.ucar.edu/publications/PhD%20and%20Masters%20Theses.htm for all kinds of people and papers on climate model runs. In this case (#3088067) the researcher varied the convection from concentrated to diffuse.

The concentration of convection brings more heat to the upper atmosphere where it escapes to space and also creates more subsidence areas where the ocean or earth's surface radiates more effectively to space. The secret to climate sensitivity for warming due to CO2 is how concentrated the tropical convection is but also how much the tropics expand in a warmer world. Climate models are very good at predicting the latter. They cannot model the former which is why it is often parameterized.

You are right that the parameterizations are getting better, but it's because they run small scale weather models which are quite good to feed into the climate models. Also computing power is chipping away at the grid resolution getting the climate models closer to where the weather models used to be. Eventually my point will be moot; climate models will be adequate to model convection all by themselves (I would say no more than 10-20 years).

Posted by: eric654 | May 25, 2010 5:45 AM | Report abuse

imback, the response of models to changes in parameterizations have been thoroughly studied. I believe Lindzen's hypothesis is somewhat related (not too sure). Eschenbach had a complicated explanation on WUWT. The best source I found was here: http://www.ccsm.ucar.edu/publications/PhD%20and%20Masters%20Theses.htm for all kinds of people and papers on climate model runs. In this case (#3088067) the researcher varied the convection from concentrated to diffuse and found more upper atmospheric drying with concentrated convection.

The concentration of convection brings more heat to the upper atmosphere where it escapes to space and also creates more subsidence areas where the ocean or earth's surface radiates more effectively to space. Hurricanes are a perfect example, while alarmists point out they transfer heat to the poles, their far bigger role is to transport heat to space. More hurricanes = cooler planet.

The secret to climate sensitivity for warming due to CO2 is how concentrated the tropical convection is but also how much the tropics expand in a warmer world. Climate models are very good at predicting the latter. They cannot model the former which is why it is often parameterized.

You are right that the parameterizations are getting better, but it's because they run small scale weather models which are quite good to generate those parameters. Also computing power is chipping away at the grid resolution getting the climate models closer to where the weather models used to be. Eventually my point will be moot; climate models will be adequate to model convection all by themselves (I would say no more than 10-20 years).

Posted by: eric654 | May 25, 2010 5:53 AM | Report abuse

Upon further research I see that the poleward heat transport from tropical cyclones is within the ocean, not atmosphere. The Kerry paper "Contribution of tropical cyclones to meridional heat transport by the oceans" http://www.agu.org/pubs/crossref/2001/2000JD900641.shtml discusses that. The question after I read that paper is still open: what is the net effect in both the oceans and atmosphere of an increase in TCs.

A 2007 paper https://www.purdue.edu/climate/pdf/Sriver%20and%20Huber%20Nature05785.pdf by Sriver and Huber does a followup. Results still not integrated into GCMs so I can't see how they can say it warms the global mean climate. For that claim they reference another paper: http://www3.interscience.wiley.com/journal/118697338/abstract?CRETRY=1&SRETRY=0 "Why ocean heat transport warms the global mean climate" Interestingly, that paper does not have any notion of changes in convection patterns and in fact seems to assume diffuse convection. Unfortunately I only have the abstract.

Posted by: eric654 | May 25, 2010 6:26 AM | Report abuse

eric, you said,
"Walter, I guess I'll have to spend some time on the "skeptic argument #45" thread now."

excellent. i'd LOVE for you to post your questions at "skepticalscience" and even realclimate. i bet they'd answer your questions. some guys at realclimate can be pretty abrasive, though.... let me know if you do so i can follow up.

Posted by: walter-in-fallschurch | May 25, 2010 11:44 AM | Report abuse

The conclusions based on statistics are only as good as the model. A coin toss has a 50/50 chance of coming up heads or tails. But that's a *model* of a coin toss, not the reality. A coin toss is actually highly deterministic and can be described with classical physics. You can build a coin tossing machine that makes the same side come up each time.

http://comptop.stanford.edu/preprints/heads.pdf

It's just that a human-flipped coin has enough variability in the initial conditions to make it almost 50/50 for practical purposes. But for thousands of runs or under certain conditions the bias becomes apparent. It's one thing to explain the probability in a model, quite another to say the model fits reality. Just look at the "quants" whose models predicted that the failure of certain complex financial instruments was virtually impossible. I think a little humility is needed when making conclusions about complex models. When defending the conclusions of these models, people too often fall back on what the stats say, which is simply saying they did the math right. They need to spend more time arguing why they believe the model is a good representation for the reality they want to describe.

Posted by: CM_in_Fairfax | May 25, 2010 2:03 PM | Report abuse

CM_in_Fairfax wrote, "It's one thing to explain the probability in a model, quite another to say the model fits reality. Just look at the "quants" whose models predicted that the failure of certain complex financial instruments was virtually impossible. I think a little humility is needed when making conclusions about complex models. When defending the conclusions of these models, people too often fall back on what the stats say, which is simply saying they did the math right. They need to spend more time arguing why they believe the model is a good representation for the reality they want to describe."

Well said. Very well said. I agree 100%.

Speaking of the wonderful models, what about all of those wonderful models that predicted more frequent and more powerful hurricanes because of global warming? Don't hear much about them any more.

Or how about the models that predicted that the Arctic would be completely ice free any day now. Funny how there is no more talk of that. They go right on preaching gloom and doom as if those failed predictions were never made. Now that the Arctic is back up to normal, you don't hear a peep about the Arctic from Mr. Freedman.

Likewise about hurricanes. I remember when Mr. Freedman was touting the studies that linked global warming and stronger and more frequent hurricanes. Now he wants to distance himself from anything linking hurricanes and global warming. Heck, these days he will write an entire column to repudiate the linkage that others have drawn between hurricanes and global warming (albeit their linkage was to mock the accuracy). Funny how times have changed.

We are supposed to believe that the models got all the details wrong, but they are still accurate.

Mr. Q.

Posted by: Mr_Q | May 25, 2010 3:28 PM | Report abuse

mr.q., you said,
"...Now that the Arctic is back up to normal, you don't hear a peep about the Arctic from Mr. Freedman."

that's argument #68
http://www.skepticalscience.com/Has-Arctic-sea-ice-recovered.htm

http://www.skepticalscience.com/Has-Arctic-sea-ice-returned-to-normal.html

http://nsidc.org/data/seaice_index/images/daily_images/N_stddev_timeseries.png

you're about a month late with that one. was GREAT for april, though. it's now BELOW 2007 levels..... hahaha

Posted by: walter-in-fallschurch | May 25, 2010 3:54 PM | Report abuse

Hey Walter, the pendulum swings both ways (hehe in advance)

Posted by: eric654 | May 25, 2010 4:21 PM | Report abuse

eric,
huh? if it were a "pendulum" ice levels would be swinging to the other side of the historical average.

instead skeptics were "pleased" when for about 6 weeks this winter the level soared to approach, but still remained below, the historical average.

Posted by: walter-in-fallschurch | May 25, 2010 4:47 PM | Report abuse

eric, re pendulum:
or is there some unknown-to-me "cooling" data you're about to spring?

Posted by: walter-in-fallschurch | May 25, 2010 5:07 PM | Report abuse

Walter, the "historical average" happens to be the satellite record which coincides with a peak in ice in 1979. Be that as it may, I have no prediction for sudden cooling other than what has already taken place (the end of El Nino). Also the ice will respond to other factors on a short term basis (like wind) with no particular relation to warming. Lots of things could happen, that's all I'm saying.

Posted by: eric654 | May 25, 2010 5:46 PM | Report abuse

eric,
oh, ok. when you said "hehe in advance", i thought you had something up your sleeve....

Posted by: walter-in-fallschurch | May 25, 2010 6:02 PM | Report abuse

Eric, you can often find a paper on the lead author's institutional site, at least after the journal's embargo is lifted. In this case, http://www.ldeo.columbia.edu/res/div/ocp/pub/herweijer/herweijer_swc_2005.pdf
They used two different atmospheric models, the GISS model and the NCAR CCM. Both had prognostic clouds, meaning clouds were produced where convection was favorable and compensating downdrafts were naturally produced elsewhere. The clouds were simulated implicitly by the grid-scale parameterizations, not explicitly, but the global statistics should have been more or less OK.

Anyway, I still don't see much evidence of any "concentrated convection" in the climate change literature.

Posted by: imback | May 25, 2010 10:15 PM | Report abuse

CM_in_Fairfax, climate models *are* based on classical physics. See the IPCC chapter I linked to above.

Posted by: imback | May 25, 2010 10:17 PM | Report abuse

imback: That was never in doubt. What else would they be based on?

Posted by: CM_in_Fairfax | May 26, 2010 9:16 AM | Report abuse

By "classical" I meant that a coin toss is not purely random, but deterministic based on initial conditions and forces applied. The 50/50 ideal coin toss model handles the uncertainties in the practical world, but has limits. I was making a broader comment about how debate gets muddled when people confuse arguments about how a model works with whether it's a good model. Just think of the classic question: If you toss a coin 100 times and it comes up heads each time, what's the chances of it coming heads the 101st? I sure hope your answer would be "nearly certain" since observation should show that the ideal coin model (0.5 probability every toss, independent trials) just doesn't hold in that case.

I'd be interesting in seeing an analysis explaining why scientists believe a certain model is good, not just that they did the math and this is what it shows.

Posted by: CM_in_Fairfax | May 26, 2010 9:40 AM | Report abuse

imback, thanks very much for the paper. The conclusion is that ocean heat transport causes global warming all other things being equal. As they admit in the paper, all other things aren't equal even in the limited context of changing OHT (e.g. OHT changes wind patterns which changes the overall warming or cooling). In the broader context of tropical cyclones, their conclusion is very much irrelevant and should not be used to suggest that tropical cyclones cause global warming.

In fact there are basic axioms of weather and global warming that have been confirmed by models. Those axioms however can't be used to say there will be global warming or not especially when they are overgeneralized. For example, lots of people on both sides will say things like "more clouds causes X or there will be fewer clouds, etc" But the axioms are that high clouds tops are generally warming and low cloud tops are generally cooling. This is easy to prove from a look at an IR satellite picture which shows high (cold) cloud tops throwing less heat into space. Can't make any general conclusions about mid-level cloud tops.

Diurnal clouds are almost always cooling, clouds that cover land, water or ice cool or warm depending mostly on the temperature of what they cover (they cool over ice, warm over warm waters, etc). Concentrated convection creates large areas of tropical clear sky (cooling) due to subsidence (note the much larger subsidence area around a hurricane than the cold cloud tops of the hurricane itself). Further concentrated convection elevates heat where it can escape to space. These effects dwarf the small global warming effect from OHT, and also they alter the main OHT GW variable: water vapor. Another axiom of GW is that variations in WV are cooling whereas uniformity of WV is warming

Posted by: eric654 | May 26, 2010 10:02 AM | Report abuse

CM_in_Fairfax,
argument #5?
http://www.skepticalscience.com/climate-models.htm

or is that just more of just "trusting the math"? i think it's a combination of trusting the math and checking the results. (then improving the models.) sure, we're limited by certain unknowns (and even unknown unknowns...), but who's to say the models aren't underestimating the effects?

Posted by: walter-in-fallschurch | May 26, 2010 10:42 AM | Report abuse

Walter, I no longer know what CM_in_Fairfax's point was. A coin toss model is constructed from classical physics, like initial vector, Newton's laws, gravity, etc. It is validated against a series of real coin tosses. If it gets over 90% right it is pretty good. A climate model is constructed from classical physics, like constituent molecules, conservation of mass, Newton's laws, gravity, laws of thermodynamics, radiative transfer, changes of state, etc. It is validated against a series of past climate observations. If it gets over 90% right it is pretty good.

Eric, you're making unjustified statements without supporting evidence. For instance, "Further concentrated convection elevates heat where it can escape to space." Elevating hot air will expand and thus cool, and cool air radiates less to space.

Posted by: imback | May 27, 2010 1:32 AM | Report abuse

imback,
'preciate all your comments. are you a "scientist"? "engineer"? interested citizen?

i think CM_in_Fairfax wonders whether scientists just create these models and assume they're accurate, or if they test them empirically - or something like that. basically, i thought he was questioning how "good" the models are, and how do we know if they're working.

Posted by: walter-in-fallschurch | May 27, 2010 8:49 AM | Report abuse

The comments to this entry are closed.

 
 
RSS Feed
Subscribe to The Post

© 2012 The Washington Post Company