Data Comparisons

GISS August 2009: 0.51C down from July.

28 Comments

Like UAH, GISSTemp dropped in August. I was expecting an El Nino rise. Here’s the graph with trends since 1990 and 2001:

Figure 1: GISSTemp, dropped relative to July, 2009.
Figure 1: GISSTemp, dropped relative to July, 2009.

Other details:

  1. We’re still waiting for El Nino to finally push the trend since 2001 into positive territory; it remains just barely negative at 0.1C/century. The trend since 1990 is still well below that predicted by models in the AR4.
  2. The August 2009 value is higher than last August’s reading of 0.36K
  3. The August 2009 value is cooler than August 1998’s reading of 0.63K.
  4. This is the 6th warmest August ever recorded by GISSTemp and pretty average for the century so far.

28 thoughts on “GISS August 2009: 0.51C down from July.”

  2. The 30 day running average of the SOI is at 0 and rising sharply which does not bode well for the El Nino’s future duration or intensity. See

    http://www.bom.gov.au/climate/enso/

    What is the value of the anomaly needed that will cause the trend since 2001 to become positive.

  3. Chuck–The temperature anomaly required to make the trend positive depends on the duration of the “high” anomalies. If you look at the blue trend line, you’ll see today’s report pulled the trend down. (The way to tell is to notice the July temperature falls below the trend line.)

  4. Interesting that GISS map shows plenty of warmth in Arctic, and mixed warm/cool anomalies. NCEP last 30 and 90 days shows less warmth in Arctic, and Antarctic dominated by warm anomalies up to 20 degrees.

  5. On holidays now, I won’t be able to analise what changed in GISS old temperature records until friday. Regards.

  6. Maybe it is time for the AGW community to stop hyping the monthly stats? The pattern of their behavior- to screech out on every media system about tiny monthly increases, while ignoring the tiny monthly decreases that inevitably follow, is only making their case less and less credible.

  7. Hunter–
    I think over the past year the AGW community has been not hyping monthly stats. 🙂

    Earlier this summer, Joe Romm did hype the predicted warming associated with El Nino.

  9. Lucia,
    August anomalies down across the board and a trend that remains negative in spite of a generally healthy El Nino? SH temperatures are down and Arctic ice extent is approaching 2005-2006 values. WUWT has comments by both Spencer and Whitehouse on other factors that are probably at play.

    Meanwhile, on an anecdotal note, here on the West Coast the wood stove was fired up last night. Last time it was on was June 12. A full 3 months of no heating…!

  10. Last December, Hansen predicted new record temperatures accompanying the next El Nino in 2009 or 2010, or is “seems likely” an out?

    http://data.giss.nasa.gov/gistemp/2008/
    “Given our expectation of the next El Niño beginning in 2009 or 2010, it still seems likely that a new global temperature record will be set within the next 1-2 years, despite the moderate negative effect of the reduced solar irradiance.”

    Acknowledgment of the sun? Hmm…..

  12. Tetris–
    El Nino is expected to build. (I’m hoping for my tomatoes to not freeze. It’s been a cold growing season here.)

    AndrewFL–
    I turned of Akismet, which was causing trouble with my comments. But… there are still problems. There was a security warming, but I’d upgraded. I’m trying to figure out what’s going on.

  13. Lucia [19626]
    I hope so too! The growing season here has been abysmal: very little rain and temperatures well below the norm. Green red peppers and green tomatoes mid-September…
    Time will tell whether this El Nino has legs. 🙂

  14. People — don’t wish an El Nino on us it may be good for your tomatoes but it is terrrible for Australia’s wheat crop (too dry) and so gives me a very bad year 🙁

  15. Andrew Kennett–
    I guess I’ll hope for El Nino to end for your sake then. I don’t really need the tomatoes; the super market has decent ones. (Home grown are better though.)

  18. OT: I finally succeeded in replicating Santer’s T2LT anomalies for GISS AOM Run 1. Now I can extend the time series out beyond Dec. 1999 and compare with all the observational data available. Very excited.

  21. Oh yeah, I forgot that bit. Although even in an inactive season a big one can still hit us. Or so the NHC has been beating into our heads! 🙂

  22. Aslak–
    I think the southeast including Florida has a love/hate relationship with cyclones. On the one hand, no one wants to be hit by a hurricane. On the other hand, many of the states need the fresh water from the rain associated with tropical cyclones. (We near Lake Michigan can draw from the lake, but not everyone is that lucky.)

  23. Lucia-that’s my understand as well, although in South Florida, we are mostly metropolitan and so don’t reap the agricultural benefits, instead we mostly sow property damage by having expensive beach houses.

    I live a little more inland though.

  24. Solar wind adds energy to Earth’s magnetosphere

    SOme of you may be interested:
    UCLA atmospheric scientists have discovered a previously unknown basic mode of energy transfer from the solar wind to the Earth’s magnetosphere. The research, federally funded by the National Science Foundation, could improve the safety and reliability of spacecraft that operate in the upper atmosphere.
    “It’s like something else is heating the atmosphere besides the sun. This discovery is like finding it got hotter when the sun went down,” said Larry Lyons, UCLA professor of atmospheric and oceanic sciences and a co-author of the research, which is in press in two companion papers in the Journal of Geophysical Research.
    http://www.spacedaily.com/reports/Scientists_Discover_Surprise_In_Earth_Upper_Atmosphere_999.html

  25. Lucia,

    How are error margins for trends in autocorrelated data calculated? I have found the formula to work out standard deviations in a trend calculated through least squares regression. But autocorrelation is a stumbling block.

    It seems to me that what autocorrelation does (in effect; not literally) is reduce the number of data points, which would tend to increase the sd in the trend. Is this correct?

  26. David,
    Take the regression residuals and calculate the AR(1) coefficient with the following formula:

    r = sum(e(t-1)*et(t))/sum(e(t)^2)

    e(t-1) is e(t) lagged behind one observation. Let’s say e(t) has 100 values. e(t – 1) would be e(values 2:100) and e(t) would be e(values 1:99). Multiply the standard error for the slope estimate by sqrt((1+r)/(1-r)).

    Here’s an unpublished paper by Douglass Nychka (NCAR) on the correction for autocorrelation. The way I’ve seen Nychka’s method applied is the way I described above. But the method in his paper uses a slightly different equation for adjusting the SE and he also uses the adjusted degrees of freedom in the calculation of the critical t-value. I briefly looked into other ways and I found general least squares (gls), but I haven’t read enough enough about it to really say anything. I suggest you look into it.

    Hope this helps.

  27. Chad,

    I am still not clear what ‘e’ is in the above. ‘expected value’ popped into my head, but that does not make sense. Sorry: I’m not up with this stuff.

    David

  28. After reading the paper, I am still no clearer on what e is. It is a random variable with mean 0, and an unknown sd (I think). Is it the error for each individual data point?

Comments are closed.