[Vision2020] global mean temperature

[Paul Rumelhart]

After some earlier discussion on global warming, I scanned the web 
looking for graphs of global temperature.  I found several, but nothing 
really good.  So, since I was in need of yet another life-draining 
outside-of-work project, I decided I'd go find the data and run some 
analyses on it myself.

I found a dataset from the National Climatic Data Center, part of the 
National Oceanic and Atmospheric Administration.  The data files I used 
are located here: ftp://ftp.ncdc.noaa.gov/pub/data/ghcn/v2/

I used the "adjusted" datasets, which have stripped known bad data.

So, a little c programming (and not a few Mountain Dews) later, I had a 
graph of global mean temperature.  The  datasets I started with are 
averaged for each month, some data going back to the early 1700's.  I 
took the monthly values, multiplied them by the number of days per 
month, averaged them out for the year, and averaged the data from all 
the global temperature stations (many hundreds or even thousands of them 
- I haven't counted them) and produced a nice graph that shows the data 
for years 1970 through 2009 (attached).

I expected something more straightforward, what I got was an enigma.  I 
see a rise in temperature in 1989 - 1993, a slightly smaller peak around 
1998, and a severe temperature drop in 2006, which returns almost to 
it's previous temperature by 2009.  Obviously, 2009 is not finished, so 
that data point will probably change.

I otherwise see a gentle sloping upwards over the course of the last few 
decades.  This presumably reflects the presumed global warming and is 
part of what is used to predict future temperature.  I'm looking forward 
to reading up on methodologies used for determining global mean 
temperature, and what sophisticated statistics they use to get there.  
It's been 20 years since my last statistics course.

Once my program becomes a little more sophisticated, I'm going to plot 
Moscow's weather vs. the global mean, since the U of I has data in the 
file going back (iirc) to 1893.  I'm also going to plot min and max 
annual temperatures, and plot the monthly values close up around 2006 to 
see what happened.

So, what happen in or around 1991, 1998, and 2006?  I'm wondering if 
2006 was affected by La Nina or a volcano eruption or something.  I'm 
going to plot each month's values per year around that time period too, 
to see if I can find out if there's some bad data lurking in the file 
somewhere or if it's consistent.

Anyway, a work in progress.

Paul
-------------- next part --------------
A non-text attachment was scrubbed...
Name: global_mean_temp_adj_1970_to_2009.png
Type: image/png
Size: 17298 bytes
Desc: not available
Url : http://mailman.fsr.com/pipermail/vision2020/attachments/20091019/ebdbc395/attachment.png