Roger wrote:

Those who see a definite trend already are cherry picking their data.
Given a few more years the downturn will become evident if it is
anything other than a transient event. The figures below are for the
smoothed curve 1850 - 2060 using the inverted around 2005 which shows
the downturn clearly and centred on 2005. (I hope no one picks these
figures up and takes the forward projection as gospel.)

What smoothing technique did you use?

This is how your data looks in graphical form he

http://farm4.static.flickr.com/3256/...bbd292db_o.jpg