HadCrut in: Temperature Up in April

HadCrut finally reported: April was warmer than March. Trends computed both starting in 2000 (for those who like positive trends) and since 2001 (for those who like negative ones) are shown below:

Figure X
Figure 1: Temperature Anomalies Since 2000 or 2001, take your pick.

As you can see, all three agencies show negative trends since 2001– though NOAA’s negative trend is so slight we need to show more than 3 significant figures to reveal this. None show negative trends since 2000. For both start years, the trend is less than “about 0.2C/decade” projected by the IPCC. (I’ll say more next week. But for now, I need to go out of town.)

The specific changes for this month:

  1. Hadley: 0.388 C up from 0.362 C in March. So, this is a small jump.
  2. NOAA: 0.6050 C up from 0.5392C. This is a fairly big jump for a monthly temperature.
  3. GISS: 0.44 down from 0.47 in March. This is a slight decline.

What does it mean that one agency reports a decline from March to April and the other two report increases? “Some” like to constantly remind us there is a lot of “weather noise” associated with monthly data. But examining the substantial variability across the agencies tells us there is also a quite a bit of “measurement noise” (a.k.a. “noise- noise”) in monthly data. It is possible to estimate the lower bound in “measurement noise” comparing the various series. This is actually rather important when later comparing the amount of “weather noise” to noise in models to the “noise” in observations of real earth temperatures. But… I’ll defer talking about that and go out of town now. 🙂

18 thoughts on “HadCrut in: Temperature Up in April”

  1. The more likely causes for the differences is the homogenization method and method of merging surface station and SST. IMHO, NOAA’s Smith&Reynolds methodology is crap, and is one of the reasons it’s the outlier.

  2. lucia,

    Which series are you reporting? Here’s what I’m seeing:

    2008/12 0.301 0.321 0.281 0.436 0.166 0.301 0.295 0.438 0.165 0.438 0.165
    2009/01 0.383 0.404 0.362 0.540 0.227 0.383 0.377 0.542 0.225 0.542 0.225
    2009/02 0.356 0.376 0.335 0.504 0.207 0.356 0.349 0.505 0.206 0.505 0.205
    2009/03 0.362 0.381 0.343 0.526 0.198 0.362 0.356 0.527 0.197 0.527 0.197
    2009/04 0.388 0.404 0.371 0.528 0.248 0.388 0.382 0.529 0.247 0.529 0.247

    I.e., up from 0.362 to 0.388, not 0.381 to 0.404. Oh, wait. Now I see it. You are reading from column three? That’s the upper 95% CI. Here’s the data format:

    * Column 1 is the date.
    * Column 2 is the best estimate anomaly.
    * Columns 3 and 4 are the upper and lower 95% uncertainty ranges from the station and grid-box sampling uncertainties.
    * Columns 5 and 6 are the upper and lower 95% uncertainty ranges from the coverage uncertainties.
    * Columns 7 and 8 are the upper and lower 95% uncertainty ranges from the bias uncertainties.
    * Columns 9 and 10 are the upper and lower 95% uncertainty ranges from the combined station and grid-box sampling, and coverage uncertainties.
    * Columns 11 and 12 are the upper and lower 95% uncertainty ranges from the combined effects of all the uncertainties.

  3. Basil–
    Thanks. Fixed.
    The graphs automatically pull from the correct column after I paste the new HadCrut values over the old ones… but my coffee obviously hadn’t hit. I cut and pasted the numbers from the wrong column.

    (My trip to Glencoe is cancelled. So… I guess I can look at some stuff now.)

  4. “for those who like positive trends” “for those who like negative trends”

    Lord, what fools these mortals be! But what of science? What will become of it when we have degenerated to this?

    On the plus side, the sociological papers that could be written on such phenomenon will finally give the liberal arts something meaningful to do.

  5. Andrew_FL: Well, it does sometimes seem there are “some” who are specifically irritated that I would say “+0.2C/decade” is inconsistent with strongly negative observed trends. One gets the impression that they really don’t like the notion that there is some rule we can make any judgment if the current trend happens to negative.

    Though… maybe that’s not what they mean. (Of course, it goes without saying that if the observed trends was +0.2C/decade, that would not be inconsistent with +0.2C/decade.)

  6. “HadCrut finally reported: April was warmer than March.”

    Isn’t April always warmer than March? 😉

    Wouldn’t it be better to say that the April temperature anomaly was higher than March?

    I do find it fascinating how we can’t even agree on what the temperature is yet some claim to know what it will be in 100 years based on model simulations.

  7. Bill-that’s not quite true or fair. Modelers say “this is where we think that temperatures could be given our assumptions about emissions of CO2 in the future-and it is generally admitted to not have a great deal of precision. It is the media which mainly inserts the certainty that it will be precisely so much warmer in a hundred years. Privately scientists will add all sorts of caveats. This phenomenon is sad but common, and not just in AGW-see:
    http://www.phdcomics.com/comics/archive/phd051809s.gif

  8. The world would be so much simpler if these points weren’t all over the place.
    Well, it would be a LITTLE simpler anyway.

  9. I wonder what took Hadley so long? I’ve been checking every day since GISS came out.

    I long ago gave up on NOAA.

  10. RyanO–
    I’ve added NOAA back in because Easterling 2009 used NOAA. So, I was looking at that page recently anyway. Announcing 5 April temperatures in April seems a big much, so I didn’t write a post when NOAA came out. I figured Hadley would be out very soon after. WRONG! heh.

    I do wonder why it took Hadley so long. Depending on funding, there may be a specific person assigned to get stuff out, and they could have gone on vacation. After all, this data isn’t exactly critical to a life safety system.

  11. “I do wonder why it took Hadley so long. Depending on funding, there may be a specific person assigned to get stuff out, and they could have gone on vacation. After all, this data isn’t exactly critical to a life safety system.”

    The way we’ve been following and waiting on this stuff, you’d think it was! 😀

  12. Andrew_FL–
    Admittedly, watching temperatures has become a pastime– sort of like baseball. But we can do it all year around! 🙂

  13. Lucia: It’s just so disappointing to bring up the HadCRUT page and see that the last line starts with “2009/03”. 🙁
    .
    BTW, I looked at that Easterling article – I’m assuming it’s this one:
    .
    http://www.cdc.noaa.gov/csi/images/GRL2009_ClimateWarming.pdf
    .
    Though Table 1 is, IMO, hoaky and without statistical merit, I would interpret it differently than the authors. To me it would indicate that the models have a positive bias compared to observations – NOAA observations, no less – for the 20th Century.
    .
    The final plot is also hoaky.
    .
    Anyway, it’s interesting . . . I hadn’t realized that anyone had recently published a paper using NOAA temps.

  14. “I do find it fascinating how we can’t even agree on what the temperature is yet…”

    Bill Jamison, you just jacked that hanging curveball out of the yard, dude. If The Blackboard were a baseball park, the next lemonade would be on me. I got cha. 😉

    Andrew

  15. Ryan, you say:

    I hadn’t realized that anyone had recently published a paper using NOAA temps.

    Far from NOAA temperatures never being used, NOAA temperatures for the US (and only NOAA temperatures, not GISS, not CRU) were used in the recent EPA Finding. I noted this in a recent CA post. NOAA makes no effort to allow for UHI and NOAA temperatures in the US run noticeably hotter than GISS with the modern-1930s differential being very different in NOAA than in GISS. One of the most useful results of Anthony’s surface studies (a point not acknowledged by the critics of this program) is that it provides an objective reason (based on CRN1-2) to assess the bias in the NOAA data.

  16. There is so much variability in US temperatures, it is difficult to use it for anything.

    For example, this is what US temps look like in the normal monthly anomaly basis that we are used to seeing for global temps.

    http://img14.imageshack.us/img14/3491/usmonthlyanom.png

    Using a 12 month moving average cleans it up enough to begin seeing the trends. US temps are down 1.25C over the last few years. I can’t remember the NOAA reporting that.

    http://img6.imageshack.us/img6/9651/ustempsc.png

Comments are closed.