Betting

Bet on June UAH

42 Comments

Want to win some quatloos? It’s time to bet on the June UAH. Bets will close June 12.

[sockulator(../musings/wp-content/uploads/2011/UAHBets5.php?Metric=UAH TTL?Units=C?cutOffMonth=6?cutOffDay=12?cutOffYear=2012?DateMetric=June, 2012?)sockulator]
Cut off date: June 12!

42 thoughts on “Bet on June UAH”

  1. Hopefully we get to see the May result first. I have to appropriately calibrate my WAG-o-meter.

  5. It is. I don’t have time right now to parsomatize (?) this month vs last month but it would be easy.

  6. I don’t know about the first, without checking, but last months was 3rd order.
    In reality the 3rd and 4th order are very similar.
    I wonder why he has changed.

  7. Ray,

    I did have the time! Well, to do the quickie version I plotted the data in Excel. The two fits are indistinguishable on a chart and the r^2 is identical to 4 decimal places. Carrick’s friend Akaike might have something to say about this.

  8. That’s the only way I can do it!
    Anyway, it’s also the way Roy Spencer does it.
    It’s funny, but 6 months ago, the 4th order poly seemed to show a slight upward trend at the end.

  11. BillC –
    I’ve noticed and been pondering the same thing (for obvious reasons…)
    It’s quite striking isn’t it?
    If the correlation between Ch5 and the monthly anomaly holds up reasonably well, then the June figure looks likely to be in the ballpark of 2010. Oddly, even though that looks considerably higher than 2011, it was only 0.07C.

    It also looks inconceivable that this June won’t be warmer than June 2011 – which narrows down the range of last minute quatloo-netting.

  12. Anteros, you mean only 0.07c higher?
    The daily AQUA CH5 temp. may be slightly higher than the equivalent day in 2010, but the cumulative figure isn’t.
    Cumulative AQUA CH5 for June 2010 = 254.00k
    Cumulative AQUA CH5 for June 2011 = 253.80k
    2011 0.2k lower.
    UAH figure for June 2010 = 0.39c
    UAH figure for June 2011 = 0.32c
    2011 0.07c lower.
    Cumulative AQUA CH5 for June 2012 (so far) = 253.726k
    But even if the temp stays the same for the rest of the month, which seems unlikely, the cumulative figure will be 253.83k, slightly higher than 2011.

  14. This year’s anomalies

    Jan -0.08
    Feb -0.11
    Mar 0.11
    Apr 0.3
    May 0.29

    This year’s anomalies as predicted by
    Avg(AquaCh5 Daily 2012) – [Avg (AquaCh5 Daily 2011) – Anomaly 2011]

    Jan -0.09
    Feb -0.13
    Mar 0.09
    Apr 0.33
    May 0.25

  15. Billc –
    Closer than my eye-balling intuition suspected.

    Ray – I was wondering the same. Also why no Hadcrut3 update since March? Is there an expected date for the last 18 months of Hadcrut4 values to be published?

  16. If you look at the first 9 days of June 2012, they average 0.214 degrees warmer that the first 9 days of June 2011 according to the daily UAH plot. If that average were to persist (and that is a mighty big if), we could estimate the June 2012 anomaly by adding the 0.212 to the 0.375 anomaly reported for June, 2011 by UAH, giving us a predicted anomaly of +0.587 which is warmer than any month in 2010 on the UAH record. However, I don’t think that high average will hold, and I’ve found that direct addition of averages does not usually work too well.

  17. Owen –
    I’ve got 0.316 for June 2011, but I take your point. Something doesn’t add up very well…

  18. Anteros (Comment #97459)
    June 11th, 2012 at 3:02 pm
    Owen –
    I’ve got 0.316 for June 2011, but I take your point. Something doesn’t add up very well…
    ——————————————–
    You are probably right. I used Wood for Trees raw data link, and I picked a value for 2012.5, thinking that was June (it may well be July).

  19. Anteros, regarding HadCrut3, I guess you are looking at the CRU website for the figure, but that site tends to be late.
    In fact, the April figure was published late May on the UKMO/Hadley Centre website and was 0.482c.
    You can find it here:
    http://www.metoffice.gov.uk/hadobs/index.html
    Click on HadCRUT3 in the COMBINED LAND/MARINE section, then click on “various diagnostics” at the bottom left of the page.
    Then click on, for example the mean of northern and southern hemisphere link and on the next page, scroll down to the data file link and click on that and you’re there. Sometimes the pages may not load but just keep clicking.
    I don’t have any info regarding HadCRUT4 but the HadCRUT3 figures seem to be getting later each month.

  20. Billc,
    Those figures are interesting and it should be easy to estimate the UAH anomaly fairly accurately, given the final AQUA CH5 figure for the month, but estimating the CH5 figure is the problem!

  21. Owen:

    If you look at the first 9 days of June 2012, they average 0.214 degrees warmer that the first 9 days of June 2011 according to the daily UAH

    9 days is too short a period to be making any predictions a month ahead of time. See e.g. the temperature spikes in this figure. (Related to inter-seasonal oscillations.)

    If you could predict where the next spike hits that could improve your accuracy. I haven’t tried any sort of forecasting on that so far. But notice the amplitude of the spikes ranges from about 0.5 to a full 1K. That’s a big range to have to predict from if the occurrence of the temperature spikes is entirely random.

    But notice the amplitude of the the

  22. Ray
    Thanks. You’re right – I was looking at the CRU website assuming they’d update the data when it was available. How foolish of me!
    Also the data hasn’t arrived at Woodfortrees which I aslo assumed would get it soon after it was published.

  23. Carrick –
    Excuse my eyes (brain?) seeing patterns where none might exist, but those look like monthly spikes.. I can’t offhand think of any remotely convincing monthly reason for such a thing, but they don’t cry out ‘random’ to me. Any thoughts?

  25. Anteros, Carrick linked you to a graph, he can be forgiven for not linking you to the discussion thread, here is a good comment to start with.

  27. Anteros, Here is Roy’s discussion.

    The spikes in temperature are real, they are not periodic with a 1-month period, they are not entrained to the annual cycle, and as far as I can tell, do not follow any particular logic when they occur.

    Hence you might consider them part of the chaotic (yet organized) part of short-period climate.

  28. A spike corresponds to broad-band frequency response.

    Hence the flat spectrum. This spectrum was computed for 1980 through 2011 inclusive.

    The high-frequency roll-off at 20 year^–1 may or may not be “real”… that is it may be associated with UAH, or it could be real and a result of destructive interference of high-frequency components once you average over the globe. EIther way, we can set the minimum sampling rate needed to capture this short period fluctuations to ≥ 40 year^-1.

    (Recall the Nyquist theorem says you have to sample at ≥ 2x the maximum frequency for your signal, which we can “safely” take at the 6-dB = factor-of-two-in-amplitude roll-off.)

    I’ve suggested to Steven Mosher (and anybody else who will listen involved in temperature reconstructions) that they have their global temperature products produce outputs at e.g. 52 samples-per-year instead of the arbitrarily chosen 12-samples-per-yera.

    I’ve gone to the day-resolution for arctic ice extent for the same reason (you can see short-period fluctuations in ice level on roughly a 7 day period).

    Let me know if this makes sense.

  29. The basic stats this month:
    MAX 0.565
    MIN -0.300
    MEAN 0.270
    MEDIAN 0.298
    STD DEV 0.168
    MEAN 1-18 0.181
    MEAN 19-36 0.360
    MEAN PLUS 1 SD 0.438
    MEAN MINUS 1 SD 0.102
    The average appears to have doubled between the first and second half of the betting, but that is mainly due to the figure of
    -0.3c at #5.
    The average for the last 9 bets was quite consistent at 0.368c, and an sd of 0.017c.

  30. Owen –
    Lucia’s system uses what we call ‘normal’ time here in the UK aka Greenwich Mean Time.

    I know – it catches out many people Stateside.

    Perhaps, to avoid disappointment for those who haven’t followed the dictum of “Bet early, Bet often!” Lucia could add –

    Last bets 11.59pm GMT on cut-off day

  31. Anteros–Good idea. For that matter, I should have the script show the cutoff date. Right now I hand enter it which is a problem.

  32. Anteros (Comment #97482)
    June 12th, 2012 at 3:46 am
    Lucia’s system uses what we call ‘normal’ time here in the UK aka Greenwich Mean Time.
    I know – it catches out many people Stateside.
    ———————————-
    Thanks. My fault for trying to observe data up the the very last day. Won’t happen again.

  33. Carrick,

    A couple of questions for you:

    A spike corresponds to broad-band frequency response.

    Do you mean just in this case, or is this a more general statement for frequency distributions?

    The high-frequency roll-off at 20 year^–1 may or may not be “real”… that is it may be associated with UAH

    I assume you mean it may be associated with the Aqua satellite? There is some history of polar orbiting satellite diurnal drift problems as you probably know.

  34. Billc:

    Do you mean just in this case, or is this a more general statement for frequency distributions?

    It’s a general statement. An impulse in the time domain equates to a flat (white-looking) spectrum in the frequency domain. These aren’t perfect spikes (otherwise they’d be Dirac-delta functions). Real-world impulses always have a finite-band-width and a high-frequency roll-off.

    I assume you mean it may be associated with the Aqua satellite? There is some history of polar orbiting satellite diurnal drift problems as you probably know.

    It could be associated with AQUA, issues with the processing software, the result of a loss of short-period coherency, or as I mentioned could represent a real roll-off due to global averaging a high-frequency phenomenon over the globe, or just be a roll-off due to the time-scales of the phenomena itself (I doubt that one, based on the time period of development of events, I’d expect the cut-off to be around 100 day^-1 or so).

    All we can really say is this is a lower-limit on the frequency cut-off of short-period climate associated with ISOs.

  35. Carrick, thanks – my point on the Aqua vs UAH distinction is just to remember we are looking at the UAH Aqua data only as opposed to a daily version of their whole dataset, and the questions directed at Roy are usually oriented toward the entire dataset v/v the processing of multi-satellite data.

  37. Carrick, Dallas and anyone else:

    Why was July 2004 so cold? The UAH anomalies are way cold especially the farther south you go. It’s striking in the Aqua daily plot too. And looks larger-than-life even in Carrick’s plot:

    though since it’s mostly the subtropical and polar SH, ISO’s don’t seem to be to blame.

  38. BillC, Hmmm? After consulting my magical coupled ocean/atmospheric Wattmeter, there appears to be a positive ocean energy imbalance around July of 2004.

    http://i122.photobucket.com/albums/o252/captdallas2/climate%20stuff/MBLmodelversusAQUASSTandCh-6.png

    By zooming in to that year
    http://i122.photobucket.com/albums/o252/captdallas2/climate%20stuff/MBL2004imbalance.png

    The imbalance moved positive at the radiant layer of the atmosphere around 2004.48 but didn’t peak as normal around July.

    This is just a prototype magical coupled ocean atmospheric Wattmeter, but it appears the oceans let off some steam followed by a cloud cover response 🙂

    You can compare 2004 to 2005 here and see the difference.

    http://i122.photobucket.com/albums/o252/captdallas2/climate%20stuff/MBL20042005comparison.png

  41. Not sure if the NH magical burp had much to do with it but it could have. I made a plot (may post it later) of area-weighted anomalies for all the UAH data. This makes a visually more compelling case that it was just a cold month in the SH winter. This happened again in 2010 but was offset by a hot NH summer. So no blip in 2010.

  42. BILLC, I believe the blips themselves are just weather, but the trend of the blips is an indication of climate. It can be hard to sort out what is what.

    I do think I am making some progress with my model though 🙂

    http://redneckphysics.blogspot.com/2012/06/isothermal-boundary-layers.html

    I don’t know if this is an original approach or not, but since I am not set up to do any serious programming, I am simplifying an energy balance model to make it rip on a spread sheet. BTW, I am more interested in Glacial inter-galacial theory and getting rid of the stupid faint sun paradox that climate change.

Comments are closed.