Data Comparisons

GISS Temp: Rose to 0.55C in May.

31 Comments

GISS Temp reported their may anomaly. It’s up to 0.09C from April, hitting 0.55C.

Figure 1: GISS Temp May Update.
Figure 1: GISS Temp May Update.

  1. May 2009 is warmer than May 2008 which was 0.42C.
  2. May 2009 is cooler than May 2007, which hit 0.57 C.
  3. The least squares trend since 2000 is 0.103 C/decade; the red corrected uncertainty range is [-0.8C/Century, 2.8 C/century] which is not inconsistent with 2C/century.
  4. The trend since 2001 is -0.2 C/century; the red corrected uncertainty is [-2.0C/Century, 1.6 C/century] which is inconsistent with “about 2 C/century”.

More later when Hadley and NOAA report.

31 thoughts on “GISS Temp: Rose to 0.55C in May.”

  3. It certainly isn’t SST anomaly data that’s that noisy. Global OI.v2 SST anomaly data (the SST dataset GISS has used since 1981) only rose 0.036 deg C from April to May.

  5. Like I commented on another thread, I wonder if the GISTEMP rise from April is due to its Antarctic data. If you look at their polar plot of GISS Analysis land temps only, there is a big round circle of extreme positive anomaly at the south pole. Does that factor in their anomaly calculations? You bet. Is it correct? Nope.

    http://data.giss.nasa.gov/cgi-bin/gistemp/do_nmap.py?year_last=2009&month_last=05&sat=4&sst=0&type=anoms&mean_gen=05&year1=2009&year2=2009&base1=1951&base2=1980&radius=250&pol=pol

  7. VG– Don’t count your chickens before they are hatched. We’ll need at least a few days below 0 to get the monthly averages below zero. Still…. that UAH measurement did cool down smartly over the past few days.

  8. Anthony,

    Oddly enough, there is considerably more noise in the satellite record than the surface record. Or at least the standard deviation of the residuals is much higher than in the surface records. That doesn’t necessarily mean that one is correct and the other is not, just that satellites tend to show higher temperature variability around the trend both short and long-term.

  9. To be more specific, 2 standard deviations of the detrended monthly residuals for all four data sets from 1979 to 2009 are:

    UAH: 0.358546075
    RSS: 0.343499114
    HadCRUt: 0.293970735
    GISS: 0.235874188

  10. Fred Nieuwenhuis: “If you look at their polar plot of GISS Analysis land temps only, there is a big round circle of extreme positive anomaly at the south pole. Does that factor in their anomaly calculations? You bet. Is it correct? Nope.”

    You can access data for Amundsen-Scott station (i.e. South Pole) at http://www.antarctica.ac.uk/met/gjma/temps.html
    It has the mean temp for May 2009 at -51.4C, which is a full 2C above the previous highest value (-53.5C in May 1981)

  11. Anthony, Zeke–
    Yes. The tropospheric measurements are noisier. I don’t know if the troposhere has more “weather noise” or if it’s the measurements. I’m waiting to see the model data Chad has been downloading. Then we’ll at least know what the models say about the relative levels of “weather noise”.

  12. Fred–
    What’s the path to create that image? The map is no longer displayed on the html page, so I clicked a link to the pdf and copied it.

    I wanted to find April’s image so I can see the south pole anomaly for April and also for May 1981. If it’s Amundsen-Scott is 2C above may 1981, the anomoly in May 1981 should have been +6C. Was it?

  13. Perhaps the anomoly level for the Amundsen-Scott station is in the 4-8 C range. But that doesn’t make for an perfect circle anomaly 1000s of kms wide.

  15. Chris–
    Thanks for the link. The +8C matches the Amundsen Scott data. May is -51.4C; further up the list, the averages for May is -57.9. If there is a problem, it’s not GISS. They grab the data that Amundsen Scott lists.

    Maybe they are having a heat wave in the south pole.

  16. Thanks fred–
    Oddly, the image no longer shows if you select the 250 km smoothing radius. With the 1200 km smoothing radius, it currently looks like this:

    Polar Map Created June 12.

    For reference, here’s the 250 km smoothing radius I snagged out of the pdf linked from the page Fred linked:

    Figure 1: Polar Map on June 11.

  18. Interesting (in a bad way) how the 1200 km works on the “left side” of the Antarctic continent. No stations shown on the 250 km smoothing but the 1200 km smoothing, there is a positive 4-8 C anomaly shown between the two negative anomaly stations.

  19. Hmmm…. Oh well. The image doesn’t display for me, but I can get the image of temperatures as a function of latitude.

    Fred– If the +8 is “local weather”, even if it’s overweighted from time to time, it will vary and vanish in three months. We’ll sometimes get negative anomalies over too large of areas. That’s the “measurement” part of the noise. So, even though GISS is sort of conceptually complicated, and sometimes seemingly weird, there is no reason think that doesn’t averages out in the longer run. In principle, all noise — regardless of source– ends up reflected in computed uncertainty intervals.

  20. How does the ‘noise’ relate to the noise in the source data? We know that the original data is ‘massaged’ in the ground record for various reasons and the sat data is based on sensor data that is converted to a temperature. Can you infer the measurement uncertainty from what amounts to secondary data at best?

  21. BarryW–
    I don’t know any slam dunk way to separate measurement noise from weather noise. The lower bound has to be the difference between measurements from two groups measuring the same thing. So, the difference between Hadley and GISS can help us estimate the lower bound. But measurement noise could be higher than that estimate because, after all, GISS and Hadley draw from overlapping measurements.

  22. Oh– and I’m using “measurement” loosely. The error isn’t necessarily an error from the thermometer itself, but due incomplete sampling, poor thermometer locations, processing to obtain the full product from the individual measurements etc.

  23. I used Fred’s map link changing the inputs to polar projection and 250 km smoothing.

    It seems GISS only smooths the data if there is a station within 250 km of a 2×2 grid cell and then they draw a 1200 km block of grid cells. I don’t see the justification for going 1200 km.

    The closest active station to the Amundsen S. Pole station is Vostok at -78.45S, 106.87E, 1284 km from the pole. That would be the red block overlapping the S. Pole burgundy at about 1 o’clock.

    GISS provides links to the data at the bottom of the map once you have drawn it.
    Here is a portion of their data that may give an idea how they do their thing.

    All anomalies from S. Pole up to here are equal to 5.9833
    i j lon lat array(i,j)
    138 5 95.00 -81.00 5.9833
    139 5 97.00 -81.00 5.9833
    Here Vostok at -78.45S 106.87E starts to be combined with the grid cell evidently weighted by distance from pole.
    Vostok is 323 km from center of cell, so they must be calculating 250 km smoothing from cell edge which is at 237 km from -80S 99E. The actual anomaly for Vostok by itself starts at -79S and is 2.8318.
    140 5 99.00 -81.00 3.1268
    141 5 101.00 -81.00 2.9865
    142 5 103.00 -81.00 2.9865
    143 5 105.00 -81.00 2.9865
    144 5 107.00 -81.00 2.9865
    145 5 109.00 -81.00 2.9865
    146 5 111.00 -81.00 2.9865
    147 5 113.00 -81.00 2.9865
    148 5 115.00 -81.00 2.9865
    149 5 117.00 -81.00 3.1268
    Back to just using S. Pole anomaly until reaching -79S.
    150 5 119.00 -81.00 5.9833
    151 5 121.00 -81.00 5.9833

    By the way, the map is rotated 90 degrees counter-clockwise. East is at the top in the S. Pole view.

  25. At -51.4 degrees this is the warmest May on record! This Antarctic reading looks very suspect-does anyone have the airline schedule for flights to the south pole? Perhaps a midnight BBQ party has taken place next to the Stevenson Screen at Amundsen-Scott!

  26. If the gloss of a worldwide AGW crisis is being given credibility because -51.4 degrees is the ‘warmest’, then the crisis promoters are running out of rational things to talk about. By focusing only on tiny changes that come from data points that represent circles of 600 – > 1000 kms. in radius, do they really think we are getting much useful information?
    Just how much of a crisis is there if the ‘record high’ is -51.4?

  27. hunter–Evidence for global warming isn’t based on measurements at the Amundsen-Scott station alone. We’re just discussing it because it’s anomalously warm this month, and it seems to contribute disproportionately to the GISS computed global average compared to other thermometers. But, it’s important to recognize that sometimes it’s anomaloously cool; at those times, it will drag GISSTemp down.

  28. hunter,

    Its only people here (many of whom are rather skeptical of temperature measurements, and probably not folks you would consider “crisis promoters”) who are highlighting the Amundsen-Scott station temperatures.

    In general, the downside to GISS’s weighting method for arctic regions is you can have a disproportionate impact on global temperatures from a single anomalous station. Thats not to say that the temperatures measured are necessarily wrong, just that they might bear double checking.

  29. Lucia and Zeke,
    If my note gave an impression that I was calling posters here ‘crisi promoters’, that was a mis-impression for which I will take full responsibility and apologize for. As a lay person, this site, with its reasonableness, excellent information and open discussion is one of the best sites I am aware of.
    Lucia has made a firm ‘lukewarmer’ out of me.
    In no way was my intention to lump people here with the fear mongering antics of so much of the AGW community.

  30. Pingback: Fielding gets no answer « An Honest Climate Debate
  31. Pingback: The Blackboard » VG predicts UAH June anomaly will go negative.

Comments are closed.