Digital Diatribes

A presentation of data on climate and other stuff

July 2008 Update on Global Temperature – GISS

Posted by The Diatribe Guy on July 9, 2008

Ah, the feeling of clicking the refresh button and finding a brand new anomaly is like nothing else…  That is, if you’re a complete nerd with nothing better to do with your time.   Thankfully, we losers exist in the blogosphere in order to keep our paltry audiences abreast of current trends in the data so that you non-nerds and non-losers don’t have to.  You’re welcome.

I’ve briefly touched on the satellite data, but the current data I track most closely is the GISS and NCDC data.  I plan on moving on to add HadCrut to the list, before finally doing the same with RSS and UAH satellite data.   So much to do, and on top of that, the company college football maniac just reminded me to let him know when I have my college predictive models updated.  Gosh, I haven’t even started those updates yet.   But I digress.

Anyway, as always, GISS global temperature data can be located here, in case you don’t trust me and want to check my numbers. The June 2008 anomaly was 26. This drop is relatively in line with what we saw with the Satellite data, which will make the skeptics momentarily happy… well, as happy as skeptics get when talking about Hansen or anything related to him.

Here’s the top-level skinny on what that number means: It is the lowest June anomaly since 1996. It is, however, the 18th highest anomaly of the 129 June anomalies in the record. So pick your spin there. Overall, it is the fourth lowest anomaly regardless of month since January 2001.

Interestingly, while last month there were a number of consecutive months of year-over-year declines, that isn’t true this month. The old algorithmic adjustment method employed by our friends at NASA have struck and the March 2008 anomaly slightly increased and the March 2007 anomaly slightly decreased, and voila – the numbers now tell us that March 2008 was warmer than March 2007. That’s what the numbers showed in April, and then they dropped in May, and then dropped again in June, and now they increased this month. Don’t worry, though, it all makes perfect sense if you do a few shots of an adult beverage where the word “proof” can be found on the label. All that said, though, even with that March blip, 10 of the last 11 monthly anomalies are lower than the previous year’s anomaly.

Looking at the 12-month average anomaly, which has dipped down to 42.1, we are now at the lowest 12-month average level since the period ending October 2001 (41.9). This means a couple things – the short-term trend is definitely in a cooling phase from previous record-high levels. Many believe, due to the missing sunspots of the current Cycle 23/24, that this cooling trend will continue. Whatever the future holds, all I know is that recent data backs up a cooling trend. That said, the data also shows that the current annual average anomaly isn’t unheard of in the last few decades, where everyone agrees that there has been a warming phase. So while current indications show one thing, only time will add to the conclusivity of the actual underlying trend.

That leads me to my observations on the actual anomaly versus my predicted anomaly, which is based on optimal weighting of the changes in slope patterns, and then predicts future changes to slopes, which then derives the anomaly needed to achieve it. The predicted anomaly for June GISS was about 48, depending on the method and average used. This differential is quite large, indicating one of three things:
(1) It’s an aberration, and these things just happen occasionally, (2) my model sucks, or (3) there is an underlying change in the data trends that is not properly reflective of slope changes and optimal weigting based on a rolling previous 11-year period.

If #1 is true, we can expect to see more accurate values going forward, with occasional aberrations like this. The aberration could be either up or down. If #2 is true, then we can expect to see monthly deviations in either direction that swing wide enough where any predictive value of the results is highly questionable. If #3 is true, then we will see a continued bias on one side of the prediction, indicating that something else is going on that represents a change from the norm. Since this is a very recent exercise, it’s too early to tell, but so far I have over-projected the anomaly in all three months. This last month, for the second month in a row, every single methodology came in too high. This may indicate that a fundamental change has occurred such that the trend lines are shifting down at a faster pace than the optimal historical weights suggest they should. But time will tell.

As always, I present the overall historical trend from inception of the data, just to put things in perspective. Long-term trends have a slope value of 0.0467 change in anomaly per month, which represents warming of 0.560 degrees per Century. Yes, this even includes what we’ve heard is the unusual warming of the last 30-40 years.

Also, I continue to take a look back to when we can date the most recent period of cooling temperatures. It seems like almost every month lately this period not only adds the most recent month but also backs up a month. This happened again with the June anomaly. Last month we had the trend line starting in February 2001, extending through May 2008 (88 months). The trend line now extends back to January 2001, extending through June 2008 (90 months, or 7 and a half years).

So, now let’s look at what happened to our various period trend lines.

Rolling 60-month Trend
Current slope = -0.1694. This is the steepest negative slope since the 60-month period ending May 1995 (-0.1721).

The most recent peak trend value was the period ending May 2007 (0.2051). However, to put things more in perspective as to how the short-term trends have changed, the period ending February 2004 had a slope value of 0.5747. So we’ve seen quite a change. Here is the current trend of slope values:

This isn’t a completely unexpected circumstance, as can be seen by this long-term chart of the changes in 60-month slope values over time.

Based on my optimal weighting predictive methodology, the 60-month trend is expected to continue a steady decline over the next year, at least. The predicted slope 12 months from now is -0.423.

So as to not overload your browser with graphs, I will verbalize the results of the other trend lines here, and then you can hit the “read on” link if you’d like to peruse the associated charts.

120-month trend lines
Current slope: 0.1311

Current slope trends/observations: Fell this month after a blip upward during the last three months. It is still higher than the value of the period ending 5 months ago. Prior to this stagnation, the previous time it reached this level was the period ending April 1998.

Most recent peak 120-month slope value: Period ending May 2002 (0.3363)

Predicted changes in slope over the next 12 months: No real change over the next quarter. A trend down over the subsequent five months, with a jump upward in month 8, followed by further declines after that. Projected slope value 12 months from now is 0.0849.

180-month trend lines
Current slope: 0.1808

Current slope trends/observations: Continued a downward trend. The previous period at this low of a value is the period ending February 2005 (0.1785).

Most recent peak 180-month slope value: Period ending March 2007 (0.2600)

Predicted changes in slope over the next 12 months: Continual decline to a 180-month slope level of 0.1271 in 12 months.

240-month trend lines
Current slope: 0.1598

Current slope trends/observations: Fell this month after a blip upward during the last three months. It is back to the level of the period ending 4 months ago. Prior to this stagnation, the previous time it reached this level was the period ending January 2007. So, overall, there really has not been a whole lot of change in the 240-month slope value in the last year and a half.

Most recent peak 240-month slope value: Period ending March 2004 (0.1755)

Predicted changes in slope over the next 12 months: A slow and somewhat steady decline, with a couple small reversals along the way, to a value in 12 months of 0.1525.

300-month trend lines
Current slope: 0.1559

Current slope trends/observations: Fell this month after a blip upward during the last three months. Fell this month after a blip upward during the last three months. It is back to the level of the period ending 4 months ago. Prior to this stagnation, the previous time it reached this level was the period ending June 2006.

Most recent peak 300-month slope value: Period ending February 2007 (0.1632)

Predicted changes in slope over the next 12 months: Not much change at all over the next 12 months, and in fact a slight increase in slope value is anticipated. This is because the period beginning 25 years ago has some very low anomalies that depress the front end of the trend line. There are enough consecutive low anomalies that the predicted anomalies – while trending lower in the short term – still exceed the front-end anomalies. The slope in 12-months is projected to be 0.1564.

360-month trend lines
Current slope: 0.1375. The 30-year slope is often used by AGW proponents as the most indicative slope of future trends. At least, this was the case prior to all these cooler anomalies. But assuming it actually is relevant, it would suggest a trend of 1.65 degrees per century at the current slope value.

Current slope trends/observations: This is the lowest 30-year trend value since the period ending December 2002.

Most recent peak 360-month slope value: Period ending June 2005 (0.1502)

Predicted changes in slope over the next 12 months: A slow but steady decline in slope value is expected, with a projected slope value in 12 months of 0.1275.

For the chart warriors, here are all the relevant charts sans additional commentary:


9 Responses to “July 2008 Update on Global Temperature – GISS”

  1. CM'Blog said

    Hi. I’m looking to build up incoming links for my blog. Would you like to exchange blogroll links with me? If yes, please visit: and leave your URL there. The purpose of my blog is to generate debate on anything and everything that matters.

  2. Smokey said

    I just ran across your blog courtesy of WattsUpWithThat. I like it!

    Great charts — I am a big believer in visual aids. The general public’s eyes glaze over when someone starts explaining things like R squared correlations or thermocouple hysteresis. But put a chart in front of them like this, and they tend to sit up straight and pay attention. A picture or a chart makes it easy for anyone to see what’s happening.

    Same here, when the media starts arm-waving about the [temporarily] hazy air in Northern California from the usual summer brush fires. Volcanoes put a lot more crap into the air in a short time than the entire North American continent’s annual forest fires — which throughout geological history happened every summer from lightning strikes. Some seeds can’t even germinate without a fire.

  3. Diatribical Idiot said

    Thanks, Smokey. And your point on the wildfires is well taken, particular with regard to the article posted by Anthony Watts on history and wildfires in California. I bookmarked that article for future reference:

  4. Patrick Hadley said

    Thanks for doing all this work. I find these graphs fascinating.

    The rolling slope graphs are particularly informative. You mention that the 5 year slope trend is at its lowest since 1995. Well of course the Pinatubo eruption in 1991 was held responsible (by climate scientists) for that cooling. I believe that the El Chichon volcano in 1982 was blamed for the dip in five year trend that is the previous low on your 5 year trend chart. There was also a big volcano in 1963, at Agung, which came before another of the low points on that graph. Among the recent low points in the five year slopes only the dip that came around 1976 seems to be missing a convenient volcano.

  5. Jack Simmons said

    Just for fun, do you think there might be any value to doing an 11 year trend chart?

    11 years is the average length of the solar cycle; it would be interesting to compare such a trend chart with the solar cycles.

    Great charts by the way.

  6. Diatribical Idiot said

    I’m guessing that the 11-year trend chart won’t look substantially different from the 10-year, but it should be no big deal to put it together and check it out. I recently pulled the sunspot data into a spreadsheet as well, so maybe I’ll actually try to get a little more sophisticated and see if I can match up varying trend lengths that correspond to the lengths of the specific cycles. I’ll add that to my wish list. I also finally downloaded the add-ins to do more statistical analysis, such as Fourier analysis on some of the data. Then again, I want to get the spreadsheets done for the other temperature measures.

    If only I had no job.. 😉 All these kids actually want a place to sleep and food to eat. They’re so demanding.

  7. Jack Simmons said

    It is hard to concentrate when the background is contaminated with the cries of hungry children.

  8. spurljub said

    I agreed with you

  9. Thanks for sharing the info. I found the details very helpful.

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 )

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s

%d bloggers like this: