Data Comparisons

RSS: November Anomaly Up.

36 Comments

The RSS November anomaly is 0.216 C. This is higher than 0.131 C for Nov. 2007 and 0.181 C for October 2008.

Figure 1: RSS Anomaly for November Up.
Figure 1: RSS Anomaly for November Up.

I may have scooped the others… I guess we’ll see. 🙂

36 thoughts on “RSS: November Anomaly Up.”

  5. JohnM–

    You got the scoop!

    I think SteveM and I are in the same time zone, so it looks like you got me by three hours.

    I usually don’t blog after my husband gets home, so I was all excited when was doing something else, visited RSS and… voila!

    I think this is the first time I posted RSS before either Anthony or SteveM though.

  6. Lucia,

    Actually, I think you should still get the credit. It’s one thing to have a post on it at the top of the blog for all to see, and quite another for a kibitzer to sneak in a reference deep in a thread on someone else’s blog.

    Signing off for now, the kibitzer.

  7. This is quite a positive anomaly, but I think it’ll be the last for this year since talks that La Nina is rolling into town. You should have a “predict the end year anomaly” for 2009 😀 Winner gets an Obama-carbon credit good for one year excursion.

  8. Taking Novl RSS in context:

    2008 will be coolest year of the decade (And noticed by Drudge!)

    “Global average for 2008 should come in close to 14.3C, but cooler temperature is not evidence that global warming is slowing, say climate scientists”

    Global Air Temperature

    “This year is set to be the coolest since 2000, according to a preliminary estimate of global average temperature that is due to be released next week by the Met Office. The global average for 2008 should come in close to 14.3C, which is 0.14C below the average temperature for 2001-07.

    The relatively chilly temperatures compared with recent years are not evidence that global warming is slowing however, say climate scientists at the Met Office. “Absolutely not,” said Dr Peter Stott, the manager of understanding and attributing climate change at the Met Office’s Hadley Centre. “If we are going to understand climate change we need to look at long-term trends.”

    However, the immediate reminder to take this in the long term context:

    Prof Myles Allen at Oxford University who runs the climateprediction.net website, said he feared climate sceptics would overinterpret the figure. “You can bet your life there will be a lot of fuss about what a cold year it is. Actually no, its not been that cold a year, but the human memory is not very long, we are used to warm years,” he said, “Even in the 80s [this year] would have felt like a warm year.”

    See Full Article: 2008 will be coolest year of the decade

  9. No scoop for you Lucia. Getting the data first wins no prize. Misrepresenting that data gets a rap across knuckles. In order to get a linear trend you must have a fixed time start and fixed time finish. My eye tells me you have not run the trend line from Nov 2001 to Nov 2008.

  10. Oh no! Hadley has predicted the year will be the coolest in the Decade. Expect a warm December. No white Christmas for us. 🙂

    Martin– I have picked a convention of starting in Jan 2001 for reasons explained long ago. That’s my “fixed start”. I always update with new data.

    That said, the trend you see is based on anomalies which is supposed to deal with the variations in the annual cycle. So, starting in January and ending in Nov. theoretically should be fine. (At least it doesn’t suffer from the flaw of including the annual cycle in some obvious way.)

    The variations in the annual cycle are large enough that if we used real temperatures, you’d see them rather distinctly. (In models, the swing is about ±1 C or ±2C. If it weren’t taken out, and we did OLS, we’d probably have no significant warming since…. A century? …. Actually, I don’t know. I may need to do that out of curiosity at some point.)

  11. How can the Hadley Center chart attached to the Guardian article be correct. Even based on inflated GISS records 1934 was warmer than the record year of 1998, 1921 was warmer than 2006 and 1931 was warmer than 1999. That chart implies that the temperatures in the 1920’s-1930’s dont even get close to the last 15 years. Something smell fishy.

  12. lucia

    Martin– I have picked a convention of starting in Jan 2001 for reasons explained long ago. That’s my “fixed start”. I always update with new data.

    I have no problem a fixed time time frame but without an X axis time reference your chart makes no sense. What am I to make of the first three data points? Since month on month anomalies are already notionally smoothed it makes no sense to compare Jan temps with November temps because of the angle of earth to the sun.

  13. Edward (#7131) – the 30’s were warm in the US, but not globally. You’re probably thinking of the US temperature record.

  14. MartinGAtkins,

    The X axis labels, at the time I post this, are along the 0 line, not at the bottom of the chart. I assume the X axis labels were present when you first viewed the chart but you may not have noticed them.

    Note also that the data reported by RSS, and for matter also by UAH, GISS, and HadCRU, are in temperature anomaly. While I think the terminology is a bit off, the data represent the variation in the observed temperature from the average value for that month over a specific time period. Thus the chart above tells you that in January of 2001 the temperature was approximately -0.2 C cooler than the average of January temperature.

    Representing the presumed global mean temperature in this manner avoids, or at least reduces, the seasonal effects you mention. It is also a basic tenet of climate science. If you have issues with it you should take them up with RSS, UAH, GISS and HadCRU. Until a better metric is devised for measuring the earth’s response to increased CO2 you’ll have to live with what the climatologists, meterologist, and all them other ‘ologists have chosen as their preferred metric.

  15. Oh!!! I see. Squint and you’ll see the x axis along the 0 line. I’ll try to remember to make them larger.

    There is no particular significance to the first three data points. The temperature anomalies happened to rise after Jan 2001.

  16. lucia (Comment#7137) December 5th, 2008 at 3:23 pm

    Oh!!! I see. Squint and you’ll see the x axis along the 0 line. I’ll try to remember to make them larger.

    There is no need to make them larger. I know where a chart begins and ends.

    There is no particular significance to the first three data points. The temperature anomalies happened to rise after Jan 2001.

    If the first three data points are of no significance then they should not exist on the chart. When you make a linear trend line you start from the beginning at a fixed point and finish at a fixed point.

    What dates do the first three data points represent?

  17. Martin (#7140)

    “What dates do the first three data points represent?”

    I think you’ll find, either by counting the dots or by reading Lucia’s response in 7133, that they represent (on the X axis) the first three months of 2001 (That would be January, February and March.)

    If, “There is no need to make them larger. I know where a chart begins and ends,” – how did you miss the “X axis time reference”?

    If you think about it, a linear trend line monitoring a measurement over time has a fixed, fixed end – in this case Jan 200, and a moving, fixed end – “now”.

    Lucia did not say the first three points had “no significance”, but “no particular significance” – ie they are just points among others, within the time frame under study, contributing to the trend.

  18. Martin– It’s just a graph of temperature anomalies since 2001. Why would I leave three points off and start the graph in April of 2001?

    But, as Michael points out , the points have no particular significance. They have the same level of significance as the other points: They are measurements since Jan 2001. No matter what graph one creates it must start somewhere.

    Out of curiosity, why do those three points bother you?

  19. Either through short sightedness, senility or stupidity or all three, I didn’t notice that you have run the trend line through the eleventh data point of the first year to what will be the eleventh data point of the last year. This of course fore fills the set time start set time finish rule.
    I unreservedly apologise.

  20. Martin,

    I think what you correctly observe about the line passing through the values for Nov 01 and Nov 08 is, in fact, a coincidence.

    I am fairly sure that Lucia has used her usual least squares regression to produce the trend line using all the data from Jan 2001. With that method the trend line does not actually have to pass through any of the measured values at all, but it normally will.

    I suspect you were thinking of another way to show a trend where you simply join up the first and last value to show the change. This is a perfectly respectable thing to do. It is used in the UK for measuring inflation. Clearly, with inflation, it only makes sense if you use eg. November to November.

    Most people do prefer to use a linear regression for climate data as the calculated slope depends on all the data rather than just the starting and ending values. It is critical that seasonality has been removed by using anomalies, as Earle explained, for this method to work sensibly for periods of months that do not comprise complete years.

  21. Jorge is correct. I just put plotted all points since 2001 and used the “linear best fit line” function with EXCEL. The graph is purely descriptive.

  22. That’s why it was so easy to replicate what the lady did and demonstrating for myself once more that the decriers should take some of their own medicine. Nothing fancy or voodoo, but in Excel most is done automatic, and long well able to manipulate trend lines and quickly draw false pictures.

    To make it round, I stepped through versions with 12-2000 added, whole 2000 added, then 1999, 1998 and so on. Just adding in 2000 was enough to proof the publicity headliner “This Decade” is just another fake to butter the choice public. Is this ethical? The jury is in.

  24. lucia,

    This statement caught my eye:

    “That said, the trend you see is based on anomalies which is supposed to deal with the variations in the annual cycle.”

    What are you saying here? That monthly anomalies are seasonally adjusted? I don’t think so. Anomalies are just deviations from a baseline. So far as I know, no attempt is made to deal with seasonal variation.

    Basil

  26. Jorge (Comment#7151) December 6th, 2008 at 7:21 am

    Martin,

    I think what you correctly observe about the line passing through the values for Nov 01 and Nov 08 is, in fact, a coincidence.

    If it is coincidence then it then it is very fortuitous indeed. If it were not so then I would consider the graph to be in probable error. Average temperatures on a month by month basis since Jan 2001 have been 0.2812C above the anomaly mean. How then can you justify Lucia’s trend line showing an approximate fall from +0.05 above the mean to a -0.05 bellow over the same period? To quote Licia’s header.

    The RSS November anomaly is 0.216 C. This is higher than 0.131 C for Nov. 2007 and 0.181 C for October 20.

    Read RSS November anomaly is 0.216 C. This is a positive number above the zero anomaly but Lucia has the last data point bellow the zero line.

  27. Martin–
    The graph is rebaselined relative to the avearage for the period as indicated below the title. I use this convention because I usually show more than one data set, they all use different metrics and putting them on some common baseline lets readers compare.

  28. Martin,

    I fell into exactly the same trap a while ago by not spotting that Lucia had changed a baseline. It can be very confusing but the fact is that all the anomaly values that are published by different groups are really rather arbitary. The exact values depend on what years are used to calculate a baseline.

    Luckily, the slope of the temperature trend only depends on changes in anomaly values and is not affected by the arbitary baselines. The baselines are important if you want to compare temperatures now with the average temperatures during the baseline period but they have no effect on calculating the slope of the trend over any period.

  31. Jorge (Comment#7180) December 7th, 2008 at 1:41 pm

    Martin,

    I fell into exactly the same trap a while ago by not spotting that Lucia had changed a baseline. It can be very confusing but the fact is that all the anomaly values that are published by different groups are really rather arbitary. The exact values depend on what years are used to calculate a baseline.

    I checked out the actual temp values between few points on the graph and they are consistent with the RSS MSU numbers. Not an easy thing to do when only estimating the exact value by eye from the chart.

    Undaunted, I decided to see what Lucia has done to the zero anomaly line. Using eyeball estimates i have come to the conclusion that it is about 0.278C higher than the RSS MSU. Why is when a woman tidies up you always have to spend hours looking for things?

  32. Why is when a woman tidies up you always have to spend hours looking for things?

    Of course you could just assume that she’s competant!

  33. Martin–
    You could just ask directly. I can provide the graphs using the anomalies the sources use. I’ve just gotten used to doing this as, for the purpose of seeing how the change is relative to other recent changes, the baseline doesn’t matter.

    My husband always just asks me where things are. He agrees 50% with James Thurber who said:

    “I hate women because they always know where things are.”
    Voltaire said the same thing.

    http://www.quotegarden.com/women.html
    http://www.quotationspage.com/quote/583.html

  34. Phil. (Comment#7201) December 8th, 2008 at 1:55 pm

    “Of course you could just assume that she’s competant!”

    I could also “just assume” that Hansen, Mann and many others are competent, but being a skeptic means not just scrutinising those who disagree with my view.

  35. Lucia (Comment#7202) December 8th, 2008 at 2:40 pm

    “I hate women because they always know where things are.”
    Voltaire

    Woman’s faults are many, men have only two.
    Everything they say and everything they do.

  36. Martin,

    “Why is when a woman tidies up you always have to spend hours looking for things?”

    My wife works on the old principle. A place for everything and everything in its place.The only drawback to this scheme is that anything not in its place is as good as lost.

    I work more on the idea that the item most recently used is on the top of the pile. 🙂

Comments are closed.