Htis is a slight cross-post from another thread: nbut, here also seems to be a good place for it:
Does this look like a warming trend that has stopped?
http://www.internationalskeptics.com/forums/imagehosting/1449447af6df244038.png
running a cusum (discussed in originating post) you can see that the temperature started unambiguously rising about the 1920's
http://www.internationalskeptics.com/forums/imagehosting/1449447af6bdf2c579.png
Now tried with a higher "running temperature" for the cusum target:
Target 0.225°C above LTA (Long Term Average {mean}):
this was appropriate for the 1980's but it is now warmer than that, and although there isn't yet enough evidence to be certain, the gradient seems to be getting steeper, which would indicate an increase in the rate of temperature change.
http://www.internationalskeptics.com/forums/imagehosting/1449447af6bdf7ce49.png
I like using cusums, as they are pretty robust, and they account for historic data.
An upwards slope means the parameter is running above "target", a downwards one lower, and a level one at target.