date: Wed Nov 2 13:35:27 2005
from: Keith Briffa <k.briffaatXYZxyz.ac.uk>
subject: Re: optimal detection
to: Gerard van der Schrier <schrieratXYZxyzi.nl>
just a note to say I am reading these messages and happy you are continuing along the lines
you are - and especially appreciate the continuing work on the long PDSI data sets.
At 13:19 02/11/2005, you wrote:
Keith & Tim,
I've applied the optimal detection to the observed trend in sea-level over the 1990 as
observed from satellites. The routine comes up with a best-guess estimate, which is
negative, suggesting that the meridional heat transport at 30N decreased. This is
consistent with some of the results we heard at the RAPID science meeting. The heat
transport is still normalized, so I can't say yet how large (in percentage) the decrease
The problem is: the best-guess amplitude is too small (or the error bars too large) to
say with some confidence that the result is significantly different from zero.
I actually expected the sea-level variability in HadCM3 to be less strong compared to
the real world, feeding my hope that we might be able to detect a response significantly
different from zero, but this turns out not be true.
There are still one-or-two things I could try. E.g., I've taken now the *difference* in
heat transport between 30N and 70N. Heat transport at 70N is ca. 10% of that at 30N.
Moreover, the HadCM3 model seems to overestimate meridional heat transport at high
latitudes. It might be a good idea to relate sea-level changes to heat transport at 30N
Professor Keith Briffa,
Climatic Research Unit
University of East Anglia
Norwich, NR4 7TJ, U.K.