128 Keith Haines
levels off at about 20 % higher than the wind forced simiulation. Much of this difference seems to involve sea level variability in the quiet regions of the ocean such
as the south Pacific where the OCCAM model has too little variability. In the
active regions, such as the Gulf stream, the Aghulas and the ACC, the OCCAM
(al
30
. .
'0
I
o
-' 0
~
...
"
.,i
·30
110
... .30 .40 ' 50 ' 50 .70 ' 00 ... ... , .. ... '30 2<. 250 "'"
27.
2'"
...
_ude
(b)
30
20
' o
!
~
o
· ' 0
·20
[l;] .,
"
·30
11 0
, .. . 30 .' "
' 50
t 50
'70
' 00
... ... 2. 0 220 230 , ..........
Fig.7.7 Sea level ehange (em) during 20 days from 30th January to 19th February, 1993.
(a) as measured by TOPEX. (b) as predicted by OCCAM model, with assimilation up to
30th January.
model already simulates a realistic level of sea level variability. The reason for the
discrepency in the other regions is not known at the present time.
7.5 Future development priorities
One of the biggest problems with the above approach is that it does not currently
have an adequate error analysis. This is not only important to provide an estimate