Data Comparisons

Hadley SSTs: Up slightly since June.

20 Comments

Hadley seems to record their SST temperature before they report their global temperatures. The HadSST temperature rose from 0.500C in June to 0.512C in July. SST is a major contributor to the global temperature, so expect the global temperature to rise. To view the SST temperature visit HadSST.

20 thoughts on “Hadley SSTs: Up slightly since June.”

  2. Bob–I don’t know if you’ve noticed but temperatures in the lower troposphere have been dropping for about 2 weeks. Those are noisy… but I’m addicted to looking.

  3. If you look at Bobs Graphs it’s immediately obvious that its “not global” by any stretch of the imagination. This actually shows how nonsensical the concept of “global” is. If it was global they MUST ALL go up! This applies to ALL climate measurements! Thanks Bob for keeping the B### honest! BTW SH ice still above anomaly after how many years now? and looks like NH coming bacK?

  4. Lucia,
    The current El Nino is turning out to be “peaky”: high and short, and I’m not sure it will have that much of an effect overall.
    Bob Tisdale’s data is interesting, especially his breakdowns by ocean basin. Take a look at the data for the Arctic basin: I think it helps explain the current developments there in terms of ice extent. As I pointed out over at CA, in mid February of this year ice extent was equal to top years 2002-2003, and between mid-April and mid-May, the 2009 number were the highest on the 9 year JAXA satellite record. If the ongoing slow down in the melt rate holds, 2009 Arctic late summer/fall ice extent numbers could very well wind up closer to 2006 than 2005, and I think that regional SSTs are playing a not inconsequential role in this. That and the powerful effect of wind patterns in 2007 [as discussed by NASA] help explain why 2007 is turning out to be the “statistical outlier” and not the harbinger of an increasingly rapid Arctic melt [as I argued in a post here several months ago].

  5. This el nino neither looks high nor short to me (at least, not that short). It is well below 1997, for example. And yet the global temperatures are much higher. The modelling seems to project a peak of around 1.8, which is well below the peak for the 1997-98 el nino. And it also predicts the el nino to continue until April or May next year. So, maybe one month or two shorter than others.

  6. David,
    By any El Nino standard, this one has peaked. That is what is known as an “evidence based” observation. Should it turn out that this El Nino keeps on going full blast anything like the 1998 one, I’ll stand corrected.

    Any climate related modeling is just that, modeling. Based on overall results to date in any area be it El Ninos or GCMs, most of it isn’t worth the RAM and disc space it uses.

  7. Then can you explain what you mean by ‘peaked’? And ‘high’? After all, if this is the peak, it cannot be high, as the temperatures are quite low compared to other el ninos. Yet you claim that this is both high and peaked. So I am not sure what you mean.

  8. And I do not expect the el nino to go ‘full blast’ like 1998. As I said, the predicted peak is around 1.8, .7 degrees lower than the peak of 1998.

  9. I guess what I am asking is: on what evidence are you basing your two seemingly contradictory claims – that the current el nino has both peaked and is high?

  11. “The HadSST temperature rose from 0.500C in June to 0.512C in July. SST is a major contributor to the global temperature, so expect the global temperature to rise”

    There is, of course, a strong correlation between sea surface temperatures and global surface temperatures. However, to say that we should “expect” the global temperature to rise, following a 0.012°C increase in the sea surface temperature anomaly from one month to the next is vastly over-stating the case.

  12. RW–
    Do you “expect” global temperature to fall or stay the same when 2/3ds of the contribution rose?

    The word “expect” only means “regard something as probable or likely”. Under the circumstances, I think it’s more likely than not that the Hadley global temperature will be higher than June. If the probability of something exceeds half, it’s perfectly ok to say “expect”. I wouldn’t say it’s certain, almost certain etc. But I do expect it will be higher.

  13. I suggest you try this: plot SST against HadCRUT3v. You’ll see a nice strong correlation. Then compute, for each month in each record, the change in anomaly for the previous month. Plot dSST against dHadCRUT3v. Not such a strong correlation, eh? Finally, do that plot for only months in which dSST was 0.012°C or less.

    Based on the final plot, do you see any reason to ‘expect’ the global average temperature anomaly to rise, following a 0.012°C rise in surface temperature anomaly?

  16. RW– As long as the correlation for dHadCrut dSST is positive I “expect” to be postive when I see dSST is positive. If I had was involved in a long term betting game, where every month, I bet $1 dHadCrut was in the same direction as dSST, and someone else took the opposite bet, I would expect to accumulate money at their expense.

    As I said: “Expect”, used casually, does not necessarily imply I think the probability is greater than 50.00001%.

    Of course, now that GISSTemp is in, I can modify my expectation based on more data, can’t I?

  18. If every time the SST rose by 0.012 °C or less, you bet that the global temperature anomaly would rise, you’d end up about even actually. You can see that by plotting the third graph I mentioned above. One should not expect anything of global temperature anomalies, based on a change of 0.012°C in the sea surface temperature anomaly.

  19. Lucia,

    Not sure where to put this comment – picked here because it is vaguely relevant.

    I saw a comment somewhere that the July hadley data, because it was the same as the June number, is likely to be wrong. But the error margins are in some cases different, so it might indeed be the value that hadley have for July.

    And I have a question (I sent it through as the contact page, but it might have got put into spam.)

    How are error margins in trends calculated?

    My ‘David Silly Method’ is:

    2 times the average error for the data points in the series divided by (number of data points – 1).

    The only other method I have discovered in searching the internet is ‘eyeball the graph and draw a couple of lines that seem to be okay’.

    Neither of these methods seems satisfactory. Any help would be appreciated. 🙂

  20. David–
    Hadley posted their other values. It looks like it’s not wrong.
    It’s not so much I thought it was “wrong” per se. But … humans do make errors, I thought it possible they could do a double entry.

    So I thought it wise to wait to see if there was some update. It looks like June and July now have the same Hadley anomaly. (And June is cooler than it used to be!)

Comments are closed.