I happened to download the monthly temperature anomalies records from the University of Alabama in Huntsville the other day and plotted it in Excel. I then did a linear trend that through the data and also did a 12 month moving average. I thought it might be interesting to do the same thing with the data from the Hadley climate research unit and also with the GISS data and plot them on the same graph.
The data is based on satellite data that has been collected since December 1978 through November 2009. I set up the other two data sets to start and end on the same months. You can see the resulting graph above. Please click on it to view it at full size.
Each data set uses a different time period to compute the offset for their anomalies, so each data set is offset from the other two. You will notice that all three data sets tend to resemble one another, but the UAH data shows that sharper transitions even when we are looking at the moving average.
You can generate different slopes depending on what start year you use. If you use 1992, which was a year after the Mount Pinatubo eruption, you end up with a much stronger positive slope. If you select 1998, which was a year of the El Niño, you can actually produce a negative slope. 1979 seems like a pretty mediocre year so it seemed like a good year to start with and it lets me use all of the satellite data.
I thought it might be interesting to take all three temperature data sets and project their trends forward to the year 2100 to see how much difference in temperature each one would predict.
It is interesting to note that all three temperature data sets agreed to less than half a degree centigrade in terms of predicting the temperature for the year 2100.
2010 to 2100 Temperature Change
GISS 1.611
HAD 1.422
UAH 1.143
Range 0.468
It is interesting to note that the worst "prediction" from these three data sets is a warming of 1.6°C for the year 2100.
Here is where I downloaded the data from:
I also uploaded the data in a spreadsheet, so you can download it and draw your own graphs with minimal work. Doing it yourself for the UAH and HADCRU is easy, but I had to massage the GISS data a little to get it into the same format.
Link to Spreadsheet:
2 comments:
Joel, do you know about Wood for Trees? An excellent website that does all the hard work for you in terms of the above, including baselining. Some good examples of what can be done here:
http://www.woodfortrees.org/notes
Thanks Barry, I wish I had known about this site yesterday. I like the look of the graphs in Excel better, but the raw data option would have saved me an hour beating the GISS data into submission. Checking the woodfortrees graphs also made it easy to validate my spreadsheet.
I've added a table showing the temperature in 2100 based on each temperature trend. For my next trick, I'm going to try to guess temperature in 2100 based on CO2 emissions.
Post a Comment