Digital Diatribes

A presentation of data on climate and other stuff

March 2008 Update on Global Temperature Trends

Posted by The Diatribe Guy on March 10, 2008

The NASA global temps are out, located here. I am actually using this source because it seems to suggest higher anomalies than other sources. If the data here actually suggests something other than a warming trend, then it becomes more diffficult for the results to really be argued vehemently.

I have updated my trend analysis on the numbers to include these results, for any interested in scoring at home. The short-term cooling trend is continuing and the negative rate of change in the 120 data point average is also continuing. Here’s the skinny:

The February anomaly was 31 (where 1 = 0.01 degrees Celsius and the anomaly is the deviation for the 1950-1980 time period, which is about 14 degrees Celsius).

This is the same anomaly as January 2008, which at the time was the lowest such anomaly since July 2004. The anomaly’s positive number says it’s still above the base period average, and the base period is relatively close to overall average for the entire period from 1880-current.

The 2-month average anomaly of 31 is the lowest 2-month anomaly average since October-November 2000 (average was 29). The latest 12-month average anomaly of 62.7 is still relatively high by historical standards, but has been moving down over the last couple years. The last time it was at least this cool on a 12-month rolling average basis is the April 2004 – March 2005 period, where the average was 61.4.

This is the lowest February-specific anomaly since 1994. It was the 28th warmest February on record.

With this additional data point, we can map the most recent non-warming/cooling trend from Ocober 2001 – Current. Stated another way, applying a simple linear trend to the data, there has been no warming, and possible cooling, since October 2001. This encompasses 77 data points of anomalies. While this certainly could indicate a peak, or a downward trend, there have been similar periods in the past where we’ve seen this occur. The most recent 77 data-point period of cooling occurred from April 1990 – August 1996. We saw another spurt in the interim before we saw the more recent flattening. I do not point this out as an argument against cooling, but we also must be honest about the data we are looking at. In the event that there is, in fact, an overall warming trend underlying our climate, these cooling/flattening periods occur. Clearly, the longer it continues, the less likely it is that it is a blip, and the more likely it is that something has changed. It’s simply too soon to tell, and all we can do is look at the data. Right now, what we can say is that claims that warming is accelerating are flat-out false, and the most recent few years shows zero indication of warming. We cannot say that it indicates a definite reversal, nor that there may be a longer-term warming trend for which we are simply seeing an aberration.

As previously noted, I have tracked a 120 data point slope of the anomalies (rolling) to see what is occurring as far as a speeding-up/slowing-down of the rate of warming. There has been a consistent decline in the 120 data point slope over the last 70+ slope calculations. The trend in the change of the slope is now -.002774 anomaly change per month. The current 120 point slope is positive, but getting less positive, on average, by that amount over time (the last 6+ years).

March 2008 Warming Change Trend
Graph Explanation: y-axis represents the slope of rolling 120 data point anomalies. x-axis is the month of the slope calculation, where month 1 represents the 10-year period beginning in March 1992, and the final data point is the beginning of the latest 10-year period, beginning in March 1998. The fitted line is the trend line, showing a declining rate of warming (-.002774 anomaly change per month).

Last month, I presented my Required/Predicted Anomalies by month. These would be the anomalies required for that month to hit the trend line in order to continue the trend of decreased rate of warming. My required anomaly for February 2008 was 53.35. The actual anomaly of 31, then, comes in quite a bit lower, and the impact of this was to make the slope of the trend line slightly more negative. As of 2007 year-end, the trend analysis implied a 2008 predicted anomaly of 44.8. The updated analysis after January lowered that to 44.6. Remaining months were last predicted to be: March 47.4; April 46.0; May 37.5; June 31.6; July 24.7; August 38.8; September 65.7; October 55.8; November 59.8; December 43.7.

Because February came in so low, the required anomaly for March to maintain the current slope of the trendline has increased quite a bit. However, because the slope is now more negative, the succeeding anomalies on the trend line are slightly reduced after that. The “trend line anomalies” going forward are:

March 68.1
April 45.9
May 37.4
June 31.6
July 24.6
August 38.8
September 65.6
October 55.7
November 59.8
December 43.6

In laymen’s terms, an anomaly in March of 68.1 continues the decline in rate of global warming at the current level. An anomaly below that will actually indicate that the decline in the warming rate is steepening.

Using actual January/February anomalies and the trend line anomalies, the updated 2008 average anomaly would be 44.4. The last time we had a 12-month average anomaly at least that low was April 2000-March 2001 (43.5). The last Calendar year with an average that low was 2000 (41.8). This would be the 12th warmest year on record, which is sure to be the way the results are communicated. In reality, such a result would be perfectly consistent with the idea that we are in a cooling cycle, or at the very least, not in a warming cycle.


18 Responses to “March 2008 Update on Global Temperature Trends”

  1. Sheaks3 said


    Interesting post. If you do not mind me asking, what is your background? I find it odd that more people are not doing analysis like you did.

  2. Diatribical Idiot said

    I’ve always been a math guy, and work as an actuary. I just have an interest in taking a look at different trends in the numbers. I am not a climatologist, which is why I don’t delve too much into that aspect of things, but I can take a good look at the numbers to see what’s actually occurring on a trending basis.

  3. Sheaks3 said

    I was discussing with a poster at Watts up with that and he made a statement that a 5 year trend on temperature anomallies are meaningless. You demonstrate a sound analysis of trends. I think the manitude of the trend is as important as the direction. I not sure if that is lost on most people. He was trying to refute Anthony’s point about 3 out of 4 matching. Anthony’s analysis tells a story and I do not think he was listening.

  4. Diatribical Idiot said

    I wouold tend to agree that a straight 5-year anomaly of straight temperature trends doesn’t necessarily point to a firm conclusion. My analysis is a little different than that, in that it is a trend of 120 data point slopes. Also, my trends don’t necessarily say we are in a cooling phase, just that the rate of warming is slowing, not accelerating as some others claim. It doesn’t make a statement about future expectation.

    All that said, the last 7 years where the trend is flat to cooling may not say with definitiveness that we are going to start heading into another direction, but at the same time it cannot be ignored that we have not continued the warming trend, either. You can’t have it both ways.

  5. Sheaks3 said

    Thanks for your time.

  6. John A. Brown said

    I’m sure I’m not speaking out of turn here however, I’m certain you’ll agree that by placing figures in front of a mere layman like myself you’ll sell any point of view you wish. This certainly seems to be the case with both sides of the argument.

    The Global Warming alarmist like the IPCC along with the world’s politicians as well as global media argues that global warming is on a rapid increase. The thing is any sceptics who dare to question their wisdom are shown graphs as well as mind boggling figures with the statement, “The figures back our findings.” Whilst the Global Warming deniers as well as sceptics also produce graphs as well as mind boggling figures, again with the statement, “The figures back our findings.”

    Now both of the aforementioned camps either in my view cherry pick data vindicating their own political agenda or knowing that the data is comprehensive and complex use the same data set thus baffling the mere layman convinced that it doesn’t matter as they wont spot it.

    As you have no allegiances to any of the previously mentioned camps, although still somewhat baffled I would be more inclined to believe your interpretation of the figures. Although I interpreted your explanation of the figures as a levelling off of warming trends I can’t ignore the fact that you explain everything whilst discount nothing. This is refreshing as for years both global warming camps have more or less explained nothing whilst discounting everything to the contrary thus producing an air of apathy where no one believes either’s argument.

    I would appreciate your view on this highly contested topic.


  7. Diatribical Idiot said

    I appreciate your perception that I am willing to take a balanced view of the data. However, to be up-front, I admit that I have major concerns with the proponents of anthropogenic Global Warming (AGW) from a political standpoint, and the realism versus the cost on how to address the issue, even if there is merit to it.

    I do agree that both sides will tend to ignore certain evidence. Also, there are four traditional data sets that are used to look at the figures, and while they basically track each other quite well, there can be aberrations that are jumped on. In addition, a single month or two, whether high or low, says little about overall trends.

    I will generalize, but here is what I see as the difference between the two camps: The AGW proponents are generally model-driven. Results of numerous statistics are fed into computer models, and the assumption is made that the data that best fits the past will best predict the future. Thus, recent history is extrapolated into the future, and they can validly claim that the model fits past results, and thus is indicative of the future. The skeptics will tend to look at actual results, in combination with more longer-term cycles, and the conclusion – again with good reason – is that past models simply have not performed well, different cycles affect the weather, making recent trends not necessarily indicative of the future, and the current trends indicate that things just aren’t really getting a whole lot worse.

    As an actuary, I can fully appreciate a modeling approach. One of the largest drawbacks of that approach, however, is that an accurate model not only needs to consider emissions and temperatures, but it must also consider solar cycles. These solar cycles are generally not properly accounted for, in my opinion. There is an interesting piece of work I recently found that I think shows the complexity in the solar cycles:

    I guess what it boils down to for me is, there is so much uncertainty, and enough compelling evidence that AGW is not nearly as serious as others tend to make it, even if there is some measurable contributory effect. The long-term trend over 128 years is 0.6 C warming per century – I don’t argue the point. Opinions are all over the map as to how much of that is natural and how much is not. Given the uncerrtainty, and the overall relatively slow trend rate, it seems unreasonable to commit an overage of resources towards a hypothetical problem when there are very pressing, immediate issues to address in the world.

    In full disclosure, I have frequent conversations with another actuary who is very capable who is convinced through his own analysis that AGW is real phenomenon and a real concern. Just goes to show that two people with similar backgrounds can look at the same information and come up with two different conclusions.

  8. John A. Brown said

    Thanks for your response.

    Again even within your response you have laid your cards on the table whilst at the same time keeping the reader mindful that there is always two sides to every story.

    Just a pity that all sides in the Global Warming argument assume that everything is black and white when everything has grey areas. Some more than others.

    Thanks again,


  9. Dan said

    I attempted the same thing you did recently. I used temperature anomalies from satellite data to put together the charts. The data I used came from the RSS set.

    Going back to January of 1998 it showed pretty much a flat trend. And in the last 7 years a significant cooling trend. My analysis was very simple. I just loaded the anomalies into Excel and plotted the points and used the Excel trend feature to plot the trend out. I’m a bit of a novice when it comes to that program so I’m sure I could have done a better job with practice.

    Ok perhaps my presentation wasn’t quite as objective as yours was either but I think they both show pretty much the same type of results. How could I have improved the trending? Or is that a class in itself?

  10. Diatribical Idiot said

    I took a look, and there is nothing wrong with your trend line, nor using the Excel functionality. The main issue is the error term on the data. If you calculate the r-square (also an excel function for you) on the data group you can see the “fit.” An r-squared of 1.00 is a perfect fit and the closer you get to zero reduces the credibility of the trend line. Also, the more data points you have to show the trend, the more reasonable it is to believe such a trend, in fact, exists.

    Going back to 1880, there are a number of 10-year periods that will show flattening or cooling. So, we can’t get too carried away with rash conclusions. What I am doing here is a little different. I calculated a slope on rolling intervals of 10-years of data. The slope indicates the rate of warming (or cooling) over that ten year period. Over the last 73 sets of rolling 120-month data points, there has been a continued reduction in the slope calculation. The slope itself would still indicate a warming trend, but the graph above shows that the calculated warming trend has slowed considerably. Should the trend continue, we may eventually see an actual cooling trend in the 10 year GISS data. We’re not there yet. However, the trend line since October 2001 in the GISS data is flat to cooling, but the r-squared measurement of that data has a much lower r-squared value.

    By taking this extra step, I have increased the r-squared value to 81%, which is fairly robust. So, I feel confident that it is, in fact, a current trend rather than random fluctuations. It does not mean the trend will continue into perpetuity, and as I state above, at this point it only says that warming rate is declining, not that we are actually cooling.

  11. Diatribical Idiot said

    One other thing. For those interested in this analysis, I do intend to continue to present this on a monthly basis. I’m hoping to dig into the data a bit more and refine things a bit as well. Again, I’m not a climatologist, so I won’t be incorporating climatology into long-range models. I will be looking at history, current trends, and presenting what it means in the near-term should trends continue.

  12. Dan said

    Thanks for the feedback. Out of curiosity how would I get the r-squared value closer to 1.0?

  13. Dan said

    Disregard that last comment. I think I understand the r-square value better now that I’ve messed around with a few trends. I plotted July 07 through February 08 as a test. The points show a very clear downward trend. Without even putting up a trend line the trend is obvious. When I plotted the trend line it had an r-square value of 0.9312. The harder it was to just look at the data and see the trend the lower the r-square value seemed to be. Over the 10 year period with 120 points and dramatic temperature changes the r-square value was extremely low. I assume this is where improvements could be made by going with 2 or 3 month averages versus month to month?

  14. Diatribical Idiot said

    Yep. Smoothing the data by combining points is one way to go. Using a rolling average is another (meaning, you use, for example, the average of 12 data points, and then your next average is calculated by dropping off the first point and including the next point.)

  15. Robert said


    Thanks for a dispassionate analysis of the trend.

    For my years, and for the science I understand, I have had the impression for some time that the weather is not what it should be. But that things change I understand.

    There seems to be a gap in the analysis regarding the trend of warming (or not) relative to the measured carbon dioxide, as some deniers want to claim, the carbon dioxide trend follows the temperature trend, suggesting that warmer temperatures precede increased carbon dioxide, an not the other way around. Which would indicate that the cause for global temperature change is something other than carbon dioxide (sunspot cycles, as some would have it).

    There are may voices quick to be vitriolic on this subject, and has been noted by others above, difficult to trust their analysis. For one, if there is cause to expect significant climate change, I would like to know. If the data suggest otherwise, I would like to otherwise get on with my life.

    Do you have data to generate an analysis of that relationship? It would also be informative to know if recent carbon monoxide measurements show a change, as well.

  16. Diatribical Idiot said

    Thanks for the post, Robert. I have not specifically looked at that relationship. The reason why I am not is because, in my opinion, until all factors in climate change are knowable and determinable, and until there is an absolute certainty in the correlation versus causation relationships, it is ripe with error. And I am not limiting the potential error to one side – it goes both ways.

    Carbon Dioxide is most definitely a greenhouse gas, for example. Without that effect, the dark side of the earth would plunge to desperate temperatures each night. One of my personal issues, though, in the Carbon Dioxide debate is the difficulty in assessing the impact of an incremental addition of Carbon Dioxide into the atmosphere. The fist of the issues is that it also must be measured as an increment to water vapor, which itself is volatile. Taken by itself, I am not sure that it is meaningful. The second issue can be demonstrated as follows: Suppose I have a house built with one-inch walls and no insulation. During the winter, there will be virtually no heating efficiency. Now, I coat my entire house with a 1-inch layer of insulation. Next winter, the efficiency is worlds better, but still not as efficient as needed. Each year, we add an inch of insulation. Each year, the efficiency improves to a point, but at an increment much smaller than the year before. At some point, there will be no discernible effect.

    At what point does the same happen with Carbon Dioxide? And at what point is the effect completely counteracted by the reduction in the sun’s ability to penetrate? I have seen studies all over the map on this, and I have no comfort level with the answer. Some sources (generally the more pro-AGW sources) suggest an incremental reduction of only 9%, while other sources (on the other side) suggest it is logarithmic, which means it flattens a lot quicker. I don’t know the answer.

    The problem with trying to study that is the near impossibility of isolating it from effects of the Sun, volcanism, and any other of the myriad impacts on climate.

    Because of all this, and while I have some interest in the modeling of climate, the models have almost always proven to be poor predictors. Thus, I’m more interested in just reviewing the data to see what the long-term and short-term trends are. I am also interested in seeing – in general – what the different cycles (such as those discussed by the cited Landscheidt study above) tell us about the impact of the cycles. In the case of the Landscheidt review, it is clear that solar effect, while addressed by the IPCC, is too simply addressed compared to reality.

    Correlation vs. causation may be too often thrown out without thought, but it is an issue. I have seen studies, for example, where the methane output of bacteria increases exponentially with temperature. The “deniers” generally base their opinion not on a current scientific analysis of the issue, but there is evidence that, in past cycles of CO2 spikes, temperature increased first. This is also probably too simple, because this seems to happen ever 100,000 years or so, but I don’t think there is supporting evidence that says every time temps increase, so does CO2. This tells me, that this adds in a combination of other factors, perhaps the ability of plants on some natural cycle to absorb Carbon Dioxide.

    I have no looked at all into Carbon Monoxide impacts.

    In the meantime, I have a few other analyses I want to look at in the data trends. But all I am demonstrating is what has occurred and what is likely to occur in the short term. Anything past a year is probably not predictable specifically from the analysis I am doing, unless I can isolate multiple cyclic trends in the data. (I have read that a study of Chinese scientists using harmonics claims to have done this, so I may have to track this down and see if I can make heads or tails out of it.)

  17. AztecBill said

    The following graph shows the 30 year anomolies of 1910-1940 and 1977-2007.

  18. […] #1: 3/08 Global Temps […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: