When you look at global temperature
graph like HadCRUT3, the evidence is now gone.
Not exact matches
I
like this
graph where you can see the development of the 15 - year linear trend along the 50 - year linear trend (not centered, the last 15 and 50 years,
HADCRUT3).
(That's not quite what your temperature
graph looks
like, but it is what my downloaded
Hadcrut3 temperature data looks
like.
The SH in my
graph shows a local high anomaly in 1945 and resumed general warming in 1950,
like the
hadcrut3 analysis.