Betting, Data Comparisons

UAH TLT March 2011: -0.099C

35 Comments

Roy Spencer revealed the UAH TLT anomaly for March: -0.099C. Talk about a tight contention on betting: Numerically, there was a 5 way tie! That means the big money goes to the bettors who entered their bets first. These are: plazaeme , Greg Meurer and J.Mens who took win, place and show positions. Last place was taking by yhivfbqu who bet the anomaly would be 8C. I would love yhivfbqu to reveal whether it is a math enabled bot, or whether that was a typo.

For those wondering how the TLT temperature compare to projections for surface temperatures, I created a graph comparing the two:

It is worth noting that the comparison in the graph is a bit apple-to-kumquats because the observations are TLT and the projections are surface temperatures.

Since the graph is new, I’ll elaborate a bit:

  1. Light Grey is observed monthly temperatures.
  2. Red is the least squares fit since 1980 with red-corrected uncertainties for the linear trend. This is an estimate of a mean or deterministic quantity.
  3. The solid black curve represents the multi-model mean for models forced using the A1B SRES. This is a projection for a mean of deterministic quantity.
  4. The dark dashed lines around the multi-model mean represents the standard error in the multi-model mean. (That is: It represents our estimate of the uncertainty associated with having a finite number of models. This uncertainty would approach zero if we had an infinite number of models.)
  5. The faint grey dashed lines represent the 1σ spread of the model-mean projections. In principle, this spread represents the structural uncertainty arising from l lack of understanding of the physics. In truth, it’s actually something of a bastard value because some models only have 1 run. So,the spread may include a large contribution from “weather”, which is a random sort of uncertainty”. (That said, this sort of 1σ spread was shown in figure 10.4 in the AR4.)
  6. The projections, observations and fits are all set to the 1980-1999 baseline. This means the mean temperature anomalies from Jan 1980-Dec 1999 for all three are mathematically forced to agree. However, the the baselining procedure does not force agreement for trends. You can see there happens to be quite good agreement between hindcast/projections and least squares fit in 1990.

So, make of this graph what you will!

Now for the detail on the winners:

Winnings in Quatloos for UAH TTL March, 2011 Predictions.
Rank Name Prediction (C) Bet Won
Gross Net
— Observed -0.099 (C)
1 plazaeme -0.1 1 14.681 13.681
2 Greg Meurer -0.1 3 35.235 32.235
3 J.Mens -0.1 5 46.98 41.98
4 Paul Ostergaard -0.1 5 37.584 32.584
5 AFPhys -0.095 2 12.027 10.027
6 Bob Koss -0.104 2 9.622 7.622
7 Don B -0.105 4 15.394 11.394
8 Zer0th -0.112 5 15.394 10.394
9 Cassanders -0.113 5 12.316 7.316
10 Pavel Panenka -0.084 3 5.911 2.911
11 EddieO -0.114 5 7.882 2.882
12 Owen -0.115 4 5.044 1.044
13 Gareth -0.08 5 5.044 0.044
14 ErnieP -0.119 5 4.036 -0.964
15 B Buckner -0.12 5 3.228 -1.772
16 steve T -0.12 3 1.55 -1.45
17 MikeP -0.121 5 2.066 -2.934
18 Pieter -0.123 4 1.322 -2.678
19 MarcH -0.123 5 1.322 -3.678
20 intrepid_wanders -0.07 5 1.058 -3.942
21 Myrna -0.068 3 0.508 -2.492
22 George Tobin -0.13 3 0.406 -2.594
23 John Norris -0.131 5 0.542 -4.458
24 Boris -0.064 5 0.433 -4.567
25 Joe Prins -0.06 5 0.347 -4.653
26 Bill Illis -0.14 1 0.055 -0.945
27 cagw_skeptic99 -0.14 5 0.222 -4.778
28 HaroldW -0.142 5 0.177 -4.823
29 Troy_CA -0.143 3 0.085 -2.915
30 BobN -0.05 5 0.114 -4.886
31 Rob R -0.05 3 0.055 -2.945
32 SamuelJ -0.15 2 0.029 -1.971
33 David Frith -0.046 5 0.058 -4.942
34 Adam Gallon -0.04 5 0.047 -4.953
35 nzgsw -0.163 3 0.022 -2.978
36 SteveF -0.163 4 0.024 -3.976
37 Golfgeek -0.165 5 0.024 -4.976
38 Guy Fardell -0.032 3 0.011 -2.989
39 Jennie -0.032 3 0.009 -2.991
40 mccall -0.167 4 0.01 -3.99
41 Eds -0.17 5 0.01 -4.99
42 christopherf -0.17 2 0.003 -1.997
43 David Smith -0.17 3 0.004 -2.996
44 moschops -0.026 3 0.003 -2.997
45 Chris in Ga -0.022 2 0.002 -1.998
46 Dudley Robertson -0.02 3 0.002 -2.998
47 KÃ¥re Kristiansen -0.18 4 0.002 -3.998
48 Anton -0.183 4 0.002 -3.998
49 Don -0.011 1 0 -1
50 harrywr2 -0.01 5 0.001 -4.999
51 Tamara -0.01 5 0.001 -4.999
52 Freezedried -0.005 3 0.001 -2.999
53 ivp0 -0.002 5 0.001 -4.999
54 Diego Cruz -0.197 5 0.001 -4.999
55 Anamoi 0 1 0 -1
56 Ed Forbes 0 5 0 -5
57 Vernon 0 1 0 -1
58 Les Johnson 0 5 0 -5
59 enSKog 0.001 3 0 -3
60 Nyq Only -0.2 5 0 -5
61 Geckko -0.2 5 0 -5
62 RobB -0.2 2 0 -2
63 Hal -0.2 5 0 -5
64 KAP 0.006 5 0 -5
65 Andrew Kennett 0.01 4 0 -4
66 Bob B -0.21 3 0 -3
67 Spellbound 0.02 4 0 -4
68 Arfur Bryant 0.03 2 0 -2
69 dallas 0.03 3 0 -3
70 Jarmo 0.04 5 0 -5
71 stephan -0.25 5 0 -5
72 hempster -0.25 5 0 -5
73 Jon P -0.261 5 0 -5
74 Harold McCard -0.265 3 0 -3
75 pdjakow 0.07 5 0 -5
76 Jim Thomason 0.08 2 0 -2
77 John F. Pittman 0.08 3 0 -3
78 George Tobin -0.28 3 0 -3
79 AMac 0.09 2 0 -2
80 Paul Butler 0.09 2 0 -2
81 colin aldridge -0.29 5 0 -5
82 YFNWG -0.32 3 0 -3
83 pete m 0.123 5 0 -5
84 DeNihilist -0.33 3 0 -3
85 TimTheToolMan 0.135 5 0 -5
86 douggerel 0.145 2 0 -2
87 Robert Leyland 0.146 4 0 -4
88 David -0.35 5 0 -5
89 Larry Goldberg -0.4 4 0 -4
90 Steve Brown -0.43 3 0 -3
91 Cameron Rose 0.24 2 0 -2
92 Simon Smith 0.328 5 0 -5
93 denny 0.35 3 0 -3
94 Paul Demmert 0.42 4 0 -4
95 Tim W 0.501 5 0 -5
96 yhivfbqu 8 1 0 -1

The net winnings for each member of the ensemble will be added to their accounts.

35 thoughts on “UAH TLT March 2011: -0.099C”

  1. I suspect if this keeps happening, the comments here will dry up, except from skeptics like me who will be shouting from rooftops NO WARMING! hahahaha

  2. Lucia

    Your friend Stephan [redacted]

    … aka Stephan, continues to misread the La Nina induced decrease as a sign of no warming. The same thing happened in 1998, yet temperatures rebounded

    To try to help [redacted], I would like to point him to this chart that shows the significant but temporary nature of El nine – La Nina cycles on global temperatures using UAH data.

    http://processtrends.com/images/UAH_El_nino_La_Nina_status.png

    Here is a multiple regression model that shows the role of El Nino – La Nina as well as volcanoes (SATO index).

    http://chartsgraphs.files.wordpress.com/2011/03/uah_nino34_sato_regression.png

    It is time for …., aka Stephan, to stop the silly one liners and show his analysis for his claims.

  4. Kelly, does CO2 cause El Nino or La Nina?

    Does CO2 cause the PDO?

    No, both are natrual variability.

    Claiming a positive anomaly when the PDO is positive is kind of silly.

    Why not do two trendlines, one to the end of the positive PDO, one after.

    PS Kelly, could you do a 5 year trend on sea level on your charts. It would be a shocker.

  6. Kelly, I haven’t been following the threads where Stephan has been active. Is there a reason for outing him?

  7. Layman Lurker

    Here is a log of Stephan’s comments and my responses’

    Stephan Trail:

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72216

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72242

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72253

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72254

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72258

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72267

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72392

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72397

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72400

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72401

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72402

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72429

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72432

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72444

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72445

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72447

    http://rankexploits.com/musings/2011/bet-on-march-uah/#comment-72684

  8. Lucia,

    I like your blog.

    I read it fairly regularly, although I post infrequently.

    I would say you are even handed, honest, forthright and friendly. (Just my opinion, but what other opinion would I have.)

    As to Stephan’s hit and run remark, if you look at the absolute temperature, and not the “trend” (well the linear trend anyway), the March 2011 temperature is about what it was back in 1980 and numerous times since then.

    Please don’t take this as criticism; I am honestly asking why we should be overly concerned about a “trend” when absolute temps (at least in the satellite record) seem to be oscillating around the “zero” anomaly mark (admittedly slightly more often above it than below it) for the last thirty some years?

    If my car’s temperature gauge read slightly warmer than usual and then slightly lower than usual (and continued this dance in both directions but never overheated) and seemed to return to the “normal” mark every so often I don’t think I would even notice let alone get worked up enough to open the hood.

    I know the “linear trend” is the most often cited number but why should it be? If there were some frightening phenomenon occurring wouldn’t that “zero” anomaly point be far behind in the review mirror?

    Also I keep reading that the satellite and surface temperature records are in close agreement but I don’t see GISS or CRU dipping below the zero anomaly point when I look at the last thirty years.

    Am I missing something here?

  9. Kelly, I don’t blame you for being annoyed with Stephan – especially when he borrows the fine work on display at your blog. As an anonymous commenter, I have mixed feelings about seeing people outed.

  10. Lance

    I am honestly asking why we should be overly concerned about a “trend” when absolute temps (at least in the satellite record) seem to be oscillating around the “zero” anomaly mark (admittedly slightly more often above it than below it) for the last thirty some years?

    Whether you should be concerned about the trend is separate from whether or not there is a trend. Also, the issue of the anomaly needs to be understood. So let’s separate that a bit. Stephan keeps arriving and suggesting that it’s not warming. But he does that by just ignoring the fact that there is a pretty clear warming trend over a short period of time. Overall, it been warming. It hasn’t been cooling.

    Do temperatures osscilate? Yes. But the trend is up and the fact that we are in La Nina right now doesn’t change that.

    On the anomalies: The anomalies are defined on a particular baseline. UAH recently changed the years used in their baseline. If this is done every 10 years, we could warm rapidly and forever, and the temperatures would always seem to hover around an anomaly of 0. This is because the “anomaly” would always be defined relative to the average temperature over the most recent 30 years.
    To see if it’s warming it’s best to ignore the location of the 0 anomaly and look at the trend. The trend is up.

    So, the trend is up. The ‘0’ on the anomaly is a mathematical artifice you should pay all that much attention to. You should pay attention to trend, over all changes etc.

    Is the positive trend something to be “overly concerned” about? That depends what “overly concerned” means. It is evidence that suggests persistent on going warming.

  11. Does anyone know why the multi-model mean declines slightly the last 2 years?

  12. SteveReynolds–
    Some of hte models have very violent ‘weather’ and very few runs. It’s due to weather not being averaged out. It picks up again soon.

  13. Lucia,

    Thanks for that clear and concise answer, but I still have a question or two.

    So am I correct that the various “zero” anomaly points are the average temps over some arbitrary reference time? If so this still means that you can say that the temperature (relative to the arbitrary baseline temperature) at some given time in the past was higher, lower or the same as the temperature (relative to that same baseline temperature) at some other time correct?

    Despite the chosen “baseline” for the “zero” anomaly, shouldn’t the fact that the “anomaly” this month is the same as the “anomaly” back in 1980 (and indeed many times during the last thirty years), relative to whatever “baseline” is selected, be consistent across temperature sets?

    Wouldn’t the choice of anomaly “baseline” just compress or expand the vertical axes of the graphs relative to each other? A casual glance at the various (GISS, CRU, UAH, RSS) anomaly graphs seems to indicate that this is not the case.

    What gives?

    UAH and RSS dip and weave above and below their “baselines” while GISS and CRU haven’t dipped below their baselines in over two decades (assuming I’m looking at the correct graphs).

    Now I understand that they could still all give about the same least squares (or what ever) linear fit despite this discrepancy but this would mean that the data sets were describing very different radiative behavior.

    For example a car that slowly accelerates from rest to a given speed can have the same average rate of acceleration as one that jack rabbits to a high speed, stops, then reverses, then stops and then jack rabbits to a high speed.

    Would saying that they both had the same linear trend be meaningful in trying to understand the dynamics of both systems? Could both graphs, showing these very different characteristics, represent the physical behavior of a single system?

    If the satellite records are correctly representing the earth’s radiative behavior can the surface data sets also being doing so?

  14. Lance,

    So am I correct that the various “zero” anomaly points are the average temps over some arbitrary reference time? If so this still means that you can say that the temperature (relative to the arbitrary baseline temperature) at some given time in the past was higher, lower or the same as the temperature (relative to that same baseline temperature) at some other time correct?

    Yes and yes.

    Despite the chosen “baseline” for the “zero” anomaly, shouldn’t the fact that the “anomaly” this month is the same as the “anomaly” back in 1980 (and indeed many times during the last thirty years), relative to whatever “baseline” is selected, be consistent across temperature sets?

    Well, sort of yes and no. Think about this: Suppose it were january in Chicago and the temperature was 60F. You tell your friend it’s warm for January. He comes back and says you can’t say that and points out it was warmer in August. You then say but it’s warm for January. Then he says: But it was wamer in August. You go round and round in circles.

    Right now, if you compare the temperature right now this month to the temperature in 1980, you will be comparing a temperature during a La Nina to a temperature during an El Nino. So, noticing it was warmer in 1980 is a bit like noticing it was warmer in August than in January and then insisting that 60F is not warm even for January. (It’s not quite as bad because ENSO is not as regular as the annual cycle, but we do know it’s currently La Nina and it was El Nino in 1980.)

    Now, 2010 was a strong El Nino, so it’s somewhat fair to compare that to 1980. Then compare this year to 1982 or 1985. Or better yet, look at the trend. That tends to average out all the El Nino and La Nina’s.

    There are other ways to look at it, but you shouldn’t compare a La Nina year to an El Nino year and conclude there is no warming.

    Wouldn’t the choice of anomaly “baseline” just compress or expand the vertical axes of the graphs relative to each other? A casual glance at the various (GISS, CRU, UAH, RSS) anomaly graphs seems to indicate that this is not the case.

    It just shifts everything up or down a constant amount.

    UAH and RSS dip and weave above and below their “baselines” while GISS and CRU haven’t dipped below their baselines in over two decades (assuming I’m looking at the correct graphs).

    UAH and RSS have baselines starting after 1980. GISS and CRU’s baselines start earlier before it warmed so much. I rebaseline all the time so my graphs with GISS and CRU will show lower anomalies. If set to the same baseline you’d see the data sets all show similar– but not identical– amounts of warming. The satellite instruments measure slightly different things so some of the differences arise from that. But the graphs are largely similar. I can create some for you tomorrow.

  16. Lucia: have you ever done “The Wisdom of Crowds” calculation, to see how the mean guess at temperature is?

  17. Lucia I will not be posting here or visiting this site anymore could you please remove my personal information posted by Kelly above thank you

  19. Les–
    No. I plan to.

    Stephan.
    Thanks for bringing that to my attention. I think I got all mentions of your real name removed now.

  22. Kim– I see that comment. What’s your point? That Layman asked Kelly his reasons for outing Stephan and Kelly answered Layman’s questions?

  23. Kim– I thought to search the archives for Kelly’s references to stephan’s real name. I found one more and moderated that one so that google searches by people researching either stephan or the real-name person purported to be stephan don’t find the connection at my blog.

    I don’t really think this is going to protect stephan’s identity because it appears he left comments under both his real name and pseudonymn at Kelly’s blog.

  24. Heh, lucia, I’m so late to the party I thought ‘Trail’ was his last name. Sorry ’bout dat.
    =========

  25. Kelly, posting Stephan’s real name here was a Bush League move, imo.

    Edit: I guess he posted under that name at Kelly’s? Okay, i take it back.

  26. Kelly O’day,
    If you subtract the regression based best estimates for SATO and Nino 3.4 effects from the UAH data, then what is left is a trend showing only the ‘unexplained’ part of the warming in the UAH data (GHG’s and everything else). I think this would be interesting to look at for two reasons:
    1. It would show clearly that the “adjusted” post 1979 trend in the lower troposphere, taking into account the effects of ENSO and volcanoes, is much lower than the unadjusted data would suggest (that is, 0.0107 C per year, not 0.0141 C per year).
    2. The remaining pattern of deviation of the data from the trend line might help identify one of more of the factors that make up the ‘everything else’ I mention above.

  27. Lucia — Instead of apples to kumquats, shouldn’t it be apples to slightly larger apples? The climate models predict that the troposphere will warm up more rapidly than the surface. Any divergence between the trends would be an under-representation of the actual divergence between the models and observations.

  28. Ken–
    My understanding is the troposphere is supposed warm more quickly. So… yes. You can interpret it the way you do. But it’s still comparing two different things. I want to be sure people know that.

  29. Bruce (Comment#73018)
    April 6th, 2011 at 9:02 am
    “5 years trend lines anyone?
    UAH? Sea Level?”
    ———————————————–
    As Master of the Short Term, you might appreciate several other examples:

    11 year period fro 1987-97 (inclusive): (http://www.woodfortrees.org/plot/wti/from:1987/to:1997/trend/plot/wti/from:1987/to:1997)

    Or perhaps from 1978 to 1987: (http://www.woodfortrees.org/plot/wti/from:1978/to:1987/trend/plot/wti/from:1978/to:1987)

    There – proof of no warming from 1978 to 1997.

    Need more help? Let me know.

  30. “As Master of the Short Term”

    I prefer “Noticer of Cycles Ending”

    Or “Noticer of HADCRU January anomaly being identical to 1944 and 1942 … ”

    Its sure is cold … considering.

    http://sealevel.colorado.edu/current/sl_ib_ns_global.txt

    2010.7415 28.119
    2009.7642 31.028
    2008.7868 23.752
    2007.7552 25.546
    2006.7507 26.391

    1.7mm in 4 years … terrifying.

  31. Was just going to post that I’d love to see Lucia do a similar graph comparing sea level predictions (perhaps Rahmstorf’s?) with satellite observations over the past 20 years. I guess the issue is that I’m not sure Rahmstorf backdates his models?

  32. Throwing out that extreme value of 8 degrees here is how the rest of the crowd did matching the -0.099C.

    46 people guessed higher.
    49 people guessed lower.
    crowd average = -0.072
    crowd median = -0.100

Comments are closed.