Digital Diatribes

A presentation of data on climate and other stuff

August 2008 Update on Global Temperature – RSS

Posted by The Diatribe Guy on August 7, 2008

Well, I still don’t have a predictive spreadsheet put together yet for the RSS anomalies, although now that I have the HadCrut spreadsheet completed I will be moving on to the satellite data next. Until I look at it in more detail, here’s a quick look at RSS data.

The latest RSS anomaly is 0.147. This is the highest reading since November 2007, and yet it is lower than July 2007 by more than 2 tenths of a degree. Overall, It is the 11th warmest July of the 30 data points on record. Of the more recent data, it is cooler than the last three years in July, and is the 4th coolest of the last ten years. It now extends to 11 months the year-over-year anomalies that are lower than the previous year’s. The last such consecutive stretch dates back to the period beginning April 1999 and ending February 2000.

The cooling trend now extends back to March 1997. So, while this anomaly is higher than the previous few, it actually had the impact of adding two months to the cooling trend on the front-end and adding the new month to the back end. This puts the current cooling trend for RSS at 11 years, 5 months.

I will keep it at that for the moment. I anticipate having a spreadsheet analysis completed by the end of the month and will look at it more closely at that time.

Advertisements

3 Responses to “August 2008 Update on Global Temperature – RSS”

  1. wow, good job

  2. I’ve only just come across this blog, and I really like your presentations, but would greatly welcome a link or links to the data that you use (the original numbers!).

    The reason for this is that I also analyse climate data, generally monthly averages and also annual averages, but my “technology” starts out very differently from what one sees published in various blogs! I can do all the line fitting “instantly”, together with plots of the fitted line and confidence intervals for the line itself and for a future individual value, something one does not find readily in most software, but which is trivial in mine.

    However, I have another general comment regarding climate data (who hasn’t?) This is the apparent fascination of analysts for a straight line model, when all the indications are that climate data are certainly not “linear” wrt time. Of course, in time series one is interested in de-trending for good reasons, but with a linear fit the residuals (in regression parlance) or de-trended values (in time series parlance) have exactly the same characteristics in respect of their lack of fit to the line model as do the original data All that differs is that a hypothetical trend has been subtracted from the undisturbed data values.

    In my analyses I address initially the possibility that the data may come from a population that consists, at least partially, of stable regimes, punctuated by occasional (and presumably randomly spaced wrt time) step changes. The steps can be very abrupt indeed – manifesting themselves in many instances over the course of one or two months, or may be gradual (that is, a rising trend) occurring over several months or sometimes years.

    There is a very simple process that facilitates the detection of patterns of these sorts in noisy (climate) data, and once the occurrence of one or more step changes is admitted as being a possibility ones view of climate change tends to be radically affected.

    I could enlarge greatly on these ideas, but would prefer to do this via email since it is vital to be able to use numerous diagrams to illustrate the scheme with real data from around the world.

    Any comments you have would be welcomed!

    Robin

  3. Diatribical Idiot said

    Robin, thanks for your comments.

    I always (or almost always) include a link to the data in my posts. In the above post, there is a link embedded in the first line (the bolded, blue text).

    As regards top the line-fitting, I completely agree. I use linear trends as a starting point in analyzing how slopes are changing over time. I present the different trend values as a basis of comparison, but in any projections of future anomalies I do, I am actually using weighting values against the second difference of the changing slope values to determine projected anomalies.

    As far as the above post, it’s a fairly simple analysis, just comparing recent trends, and I understand the limitations to that. But it is still an interesting point of comparison with historical trend values, faults and all.

Leave a Reply

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

WordPress.com Logo

You are commenting using your WordPress.com 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: