129
T1000 12r1 DJF 1986-1988 (15) conl=1.
ZSOO 12r1 DJF 1986-1988 (3) conl=2.
r.,.,...,,..,,,..,..,...,.,,---------,:...,.-----, I3S"E
c)
NH clusters (MTP 1990)
1986187 VS. 1988189 12r1 (3)
60~---------------------,
55
50
45
l40
>0 35
u
i 30
::::I 25
I 20
15
10
5
o+--~
cl.1
cl.2
cl.3
cl.4
cl.S
Figure 21: a), b) Difference between DJF 1000hPa temperature and 500hPa height (respectively) from two (3-member) 120-day ensembles of the T63L19 ECWMF atmosphere
model. The first ensemble was run with observed SSTs for the winter 1986/87, the second
ensemble was run with observed SSTs for the winter 1988/89. c) Cluster frequency of
pentad fields with the DJF period for 1986/87 ensemble (left hand bars) and 1988/89
ensemble (right hand bars) . Clusters are based on Molteni et al (1990) analysis.
decreased others. Fig 20c shows the impact of the SST changes on the
climatic frequency of the Molteni et al (1990) regimes. These frequency
statistics have been based on 5-day mean 500hPa height fields. It can be
seen, comparing 1988/89 with 1986/87, that the frequency of both clusters
1 and 2 have decreased, and the frequency of cluster 5 has increased. This