Science by Press Release: The Story About Washington, DC's Heat
Update 9am Saturday: I'd like to direct your attention to the comments as there is an important discussion as to whether Scenario A or B is the correct interpretation of Dr. Hansen's paper.
Original Posting:
WASHINGTON — The relentless, weather-gone-crazy type of heat that has blistered the United States and other parts of the world in recent years is so rare that it can't be anything but man-made global warming, says a new statistical analysis from a top government scientist.
In a landmark 1988 study, Hansen predicted that if greenhouse gas emissions continue, which they have, Washington, D.C., would have about nine days each year of 95 degrees or warmer in the decade of the 2010s. So far this year, with about four more weeks of summer, the city has had 23 days with 95 degrees or hotter temperatures.
Hansen says now he underestimated how bad things would get.
--- Story by the Associated Press' Seth Borenstein
If Google News is correct, 430 news outlets ran with this story. Not only did I read it in yesterday's Wichita Eagle, I saw it on ABC News. But, does it convey an accurate picture of the science? I -- strongly -- do not believe so.
Let's break it down.
Below is Dr. Hansen's (fyi: not an atmospheric scientist) study. I have the entire paper and I have read it several times.
The part of the paper that, for a quarter of a century, has received the most attention is his forecast of future world temperatures. He keeps a copy of the forecast and a running tally of actual temperatures on his web site and it can be viewed here (scroll down to middle of page). I've placed a screen capture below.
In his paper, he posited three "scenarios." Scenario A was the worst case where CO2 levels continued to rise and the corresponding temperatures. "B" was the middle case where CO2 rises at a lesser rate (the blue line). Scenario C was the forecast of temperatures if CO2 levels were restrained beginning in the 1990's and continued to fall into the 21st Century. That is the purple forecast temperatures. The actual temperatures are the black line.
Turns out that Scenario A turned out to be correct. Dr. Hansen's forecast of future CO2 levels turned out to be excellent. That being the case, temperatures should followed the green (upper) line. Instead the actual temperatures have been colder than in any of his three scenarios, including C, which assumed falling CO2 levels!
So, while I cannot guess what Dr. Hansen had in mind when he said [he] "underestimated how bad things would get" the fact is the reverse is true. World temperatures are far, far below Hansen's forecast. FACT: In the real world, things are much cooler than Hansen forecast.
Dr. Hansen did make specific forecasts for Washington, D.C. in his 1988 paper. Here they are:
That said, Dr. Hansen is quoted in the article linking this summer's 23 days with 95° or hotter temperatures to global warming. Let's see how well this correlation actually works.
Below is Dr. Hansen's monthly world temperature data. I have highlighted the summers of 1980, 1988, 2008, 2009, and 2012.
In 1980, with significantly cooler world temperatures than today's, Washington, D.C. experienced 28 days of temperatures of 95° or higher -- five more than this summer. About the same number, 24, occurred in 1988 again with significantly cooler world temperatures.
In 2008, only 5 days (fewer than average) of 95 degrees temperatures occurred with world temperatures warmer than today's. In 2009, there were 8 days with temperatures about the same as today's
There is little or no correlation between Washington's summer temperatures and world temperatures.
If editors had known this, 430 of them might not have run with the story. This is the purpose of the the Friday press released timed for the Sunday papers with little opportunity for rebuttal.
This is "science by press release." Scientists with robust results are happy to go through the regular channels to release information.
Many times I have commented that the behavior of the pro-catastrophic global warming fringe is not consistent with people who are confident of their position. This is more of the same.
Addition: 9:40am Thursday August 9,
Two commenters (see below) quote excerpts of blog articles that are rationalizations of why Hansen's Scenario A forecast was so far off. While blogs can provide useful information, they misrepresent what is in Hansen's paper.
I prefer to consult Hansen's paper to learn what Hansen's paper says. It states (p. 9343),
"We define three trace gas scenarios to provide an indication of how the predicted climate trend depends on upon trace gas growth rates. Scenario A assumes that growth rates of trace gas emissions typical of the 1970's and 1980's will continue indefinitely; the assumed annual growth averages about 1.5% of current emissions, so the net greenhouse forcing increases exponentially. Scenario B has decreasing trace gas growth rates, such that the annual increase of the greenhouse climate forcing remains approximately constant at the present level. Scenario C drastically reduces trace gas growth between 1990 and 2000..."
Below is a graph of carbon dioxide concentrations since 1958. The rate of growth is the same now as it was in the 1970's and 1980s. Therefore, Scenario A applies.
Mike, am wondering why you think the rebuttal linked below is wrong? If you think it was a bad forecast you should submit it to a journal. From what I have read his original paper now provides pretty convincing evidence that climate sensitivity is around 3-3.4C.
ReplyDeleteSee this: http://www.skepticalscience.com/Hansen-1988-prediction-advanced.htm
Hi Dan,
ReplyDeleteIt isn't that I think he is wrong, the data clearly shows he is wrong.
With regard to the skeptical science piece, it ties Hansen's forecast to Scenario B. That is incorrect. Both qualitatively and quantitatively Scenario A's forecast is the one that matches current CO2 concentrations.
Thanks for the comment.
The CO2 is not the only forcing in the Hansen 1988 paper - you need to judge the scenarios based on net forcings, not just CO2. And by that measure scenario B was closest though somewhat higher than observed:
ReplyDeletehttp://www.realclimate.org/index.php/archives/2007/05/hansens-1988-projections/
Similarly, your picking of 4 individual years for the DC numbers are not a fair comparison. No-one is claiming that there is no variability from year to year - rather the Hansen claim was for a shift in the distribution - and your analysis doesn't address that at all.
I have made an addition to the essay addressing the Scenario A versus Scenario B comments.
ReplyDeleteWith regard to the DC numbers, it is Hansen who made a forecast of the DC numbers (p. 9356) and then, in his press release, states that things are worse than that forecast. I'm simply examining the relationship, if any, between world temperatures (which are substantially below Hansen's forecast) and DC temperatures. No relationship appears to exist.
Mike,
ReplyDeleteI noticed that you first read about Hansen's press release on the Wichita Eagle. Do you live in the area.
My hometown is Wichita.
I do live in Wichita.
ReplyDeleteMike, this is very simple. Hansen 88 projected 2011 CO2 concentrations of 393.7, 391, and 367.8 ppmv for Scenarios A, B and C respectively. The actual CO2 concentration was 390.5. The 2011 concentration for NO2 (322.5 ppbv) was also less than the Scenario B projection of 329.94. All other projected gas species (CH4, CFC11 and CFC12) were less than the scenario C projections. Your refusal to look at that data does not make your claim about Scenario A being "what happened" true. Rather, it shows that your article represents indoctrination rather than science.
ReplyDeleteTom,
ReplyDeleteWhere are you seeing those in the paper?
Mike
Mike, they have been available on the net for some time at Real Climate:
ReplyDeletehttp://www.realclimate.org/data/H88_scenarios.dat
I have also independently reproduced a close approximation of those figures for CO2 using the method described in the paper. My replication of the method described shows a projected CO2 concentration of 397.4, 394, and 367.2 ppmv for scenarios A, B and C respectively.
Note that the concentrations reported were linked in both the Real Climate article whose URL is provided by Admin above, and in the Skeptical Science article whose URL is provided by Dan Satterfield.
Tom, thanks for the comment.
ReplyDeleteThe paper should speak for itself as opposed to articles written about the paper two decades later.
The problem is that your estimation does not agree with the qualitative explanation of those figures as stated in the paper itself. Please see the excerpt from page 9343 above.
But, let us assume you are correct. For 2012 the forecast temperature under Scenario A is +1.2. Under Scenario B it is +1.1. Hansen's (GISTEMP) actual temperature is +0.5 which is cooler than even Scenario C!
Dr. Hansen's temperatures are warmer than some of the other indices. For example, the current value of the UAH index (rounded up) is +0.3.
So, even if you take Scenario B, Hansen's forecast is far too warm. There is no legitimate way Dr. Hansen can claim that, compared to his '88 forecast, current global temperatures are "worse."
Since you cite blogs, even people who worry about the eventual effects of global warming more than I do believe Dr. Hansen's linking of the heatwave to global warming is "distortion." See, for example, http://cliffmass.blogspot.com/2012/08/climate-distortion.html
Mike
Mike, the data I linked to from real climate is not part of the blog post. It is data supplied by Sato (a collaborator with Hansen in reviews of Hansen 88) to Schmidt. In the event that that provenance is not sufficient for you, as mentioned I calculated the concentration projections for CO2 using the methods described in Appendix B of the paper. Using those methods, the CO2 concentration projected for Scenario B is even higher above the actual CO2 concentrations for 2011 than from the data made public by Schmidt.
ReplyDeleteYour continued reference to the quote in your update puzzles me. It is very plain from the CO2 concentration graph that the growth in CO2 concentration did not increase by 1.5% annually after 1988. On the contrary, it had lower growth rates in the 1990s, causing a noticeable dip in the graph. It then recovered, and grew rapidly in the 2000s. The net effect is that the end point, even on visual inspection, matches a linear increase since 1987 better than an exponential increase over that period - ie, scenario B better than Scenario A.
More importantly, Hanen 88 states that:
"The abundance of the trace gases in these three scenarios are specified in detail in Apendix B."
(Page 9345)
In turn, Appendix B states that:
"In scenario B the growth of the annual increment of CO2 is reduced from 1.5% today to 1% in 1990, 0.5% in 2000, and 0% in 2010; thus after 2010 the annual increment in CO2 is constant, 1.9 ppmv yr^-1."
(Page 9361)
As previously mentioned, I calculated the projected atmospheric concentrations using this formula, and the result is a CO2 concentration greater than atmospheric concentrations in 2011.
Tom,
ReplyDeleteI accept you at your word with regard to your calculations. This is not where my disagreement lies.
If I read the rest of your posting correctly, you believe Scenario B is correct. Fine. In my 8/9/12 9:45am posting, I showed there is a difference of 0.1°C. That is about what you would expect given the small difference in CO2 concentrations.
Scenario B is still far too warm. Scenario C is too warm.
The whole posting is about the bad science behind Hansen's claim that things are "worse" than predicted. I see nothing in your comments that indicates things are "worse" than predicted.
Mike, I have split this comment of from the previous due to length. In this comment, I will make two very simple points, and one not so simple. First, with regard to Cliff Mass, I was not impressed:
ReplyDeletehttp://cliffmass.blogspot.com.au/2012/08/climate-distortion.html?showComment=1344555832877#c6809688225340088070
Second, Hansen 88 represents single model runs, not multi-run means. As such, they are a single "realization" of a possible future climate for each scenario. We expect that the realization will remain withing error (0.13 C) of the multi-run mean 95% of the time. (See discussion on page 9346) We also expect the actual temperatures to do the same. Consequently, we expect the two to remain within 2 times error (because one may be above, and one below the multi-run mean at any given time) significantly more than 95% of the time. Given that actual forcings have been less than Scenario B over the entire period, and indeed, less than Scenario C throughout the 1990s, we would expect temperatures to have tracked within 0.26 C (2 * error) of a middle point between the Scenario B and Scenario C projections most of the time. As you can see from your graph, they certainly have done so, only leaving that range in the last couple of years. Because of the stochastic nature of the predictions (and weather) such a short excursion is to short to have falsified Hansen 88.
Never-the-less I expect Hansen 88 to be falsified shortly. That is because it has been known on theoretical grounds since at least Myhre 98 that the model used in Hansen 88 overstated the effects of radiative forcing.
http://folk.uio.no/gunnarmy/paper/myhre_grl98.pdf
That brings up the third point. If Hansen in the quote above is saying that climate sensitivity is worse than he thought in 88, then he is certainly wrong. But he does not discuss climate sensitivity in his article, so that would be a perverse interpretation in that it completely ignores context. Rather, he is more naturally interpreted as saying that the frequency of extreme heat events has increased faster than he projected. This can occur with a lower climate sensitivity if the variability of temperatures increases, as Hansen argues it does. In that regard, your test of his claim using Washington DC climate is inadequate. His claim was a statistical claim about average frequency of high heat days in given decades relative to the average frequency in the period 1951-1980. Pointing out single years with more high temperature days than in 2012 todate is not germaine. You need to find the average number of high temperature days per year for each of the decades 1980-90, 1991-2000, 2001-2010. If the later is significantly below 9, his claim about Washington DC is falsified. If the later is not above 9, his claim that things are worse than he projected is false, at least with regard to Washington. You would still need to show that it was false with regard to New York, Memphis and Omaha, and indeed the world. If one of the cities did not have the increase Hansen expected, but the others (and the world) did, that would not show Hansen was wrong. It is, after all, just a natural consequence of the variability of weather.
Correction, I treated the 0.13 C figure as being 2 Standard deviations, whereas it is actually 1 standard deviation. In my second point you need to double the numerical values accordingly.
ReplyDeleteTom,
ReplyDeleteThank you for the effort you have put into these comments. They are interesting and useful.
I agree that this period of time, by itself, is too short to falsify Hansen. But, that is not my primary point in writing this point.
My point is that a couple of hot summers in D.C. is also too short to validate Hansen. But, two hot summers do not support it, either. As I point out in my graph, they had two cooler than average summers recently.
I strongly disagree with his statement ( http://www.kansascity.com/2012/08/04/3743401/new-study-ties-global-warming.html ) that this faster than he thought warming is "scientific fact"-- that is in addition to his statement about it occurring faster than he predicted.
I dislike the context. Yes, it has been hot in the half of the U.S. But, much of the world has been cooler than average this summer. He never points that out.
Finally, I dislike "science by press release" with things timed to go out on a Friday so it can be in the Sunday papers without time for rebuttal. This is manipulative advocacy, not science.
Thank you again for the dialog.
Mike
Mike @5:26 PM, you now have sufficient information to know that actual forcings rose more in line with Scenario B than with Scenario A, and indeed, slightly below Scenario B. See the fourth figure at the following SkS post, and related discussion:
ReplyDeletehttp://www.skepticalscience.com/Hansen-1988-prediction-advanced.htm
That is not a matter of conjecture. It is easily established by one of two methods - either consulting the data of gas concentrations used in Hansen 88 as provided by Gavin Schmidt, or by recreating the method used as I did. It has also been established in a couple of peer reviewed papers by Hansen that looked back explicitly at the 88 predictions. Given this, you have more than enough information to correct your Update of August 9th.
It is of the nature of honest discussion that it corrects errors when made. You have made an error, however, inadvertently, and from our discussion I hold a reasonable hope that you will correct it.
I will return to discuss further issues when I have more time.
Tom,
ReplyDeleteI have added a note at the top of the posting directing readers to this discussion.
Thank you again for the dialog.
Mike
So after all that, actual the actual increase in temps is STILL below what Hansen predicted, yet he claims that they are above what he predicted.
ReplyDeleteIt is a typical climate troll tactic to avoid the point. When Hansen says that 'it's worse than I thought,' he's lying. He certainly can't be mistaken - he's too intelligent for that.
Mark B, Hansen does not claim temperatures are above what he predicted. He claims that the frequency of extreme heat events is greater than he predicted. It is possible for the frequency of extreme heat events to be greater while having the same or lower mean increase provided variability in climate increases, which is what Hansen claims has happened.
ReplyDeleteTom,
ReplyDeleteYour interpretation is correct. However, three seasons -- in one city --do not a trend make.
Mike