Nordseezustand 2003 
197 
C.3 Concluding Remarks 
by the birds of a feather flock together rule, however, nature couples maritime and 
continental wind characteristics according to the complementary principle opposites 
attract which results in a hybrid wind climate. 
While within-regime differences among the seasonal distributions are significant be 
yond doubt, the cross-regime differences for the winter and summer distributions 
(Fig. 3-19, p. 75, bottom) were felt to deserve statistical confirmation. To this end, a 2-di 
mensional Kolmogorov-Smirnov test (Press et al. 1992) was performed on the un 
binned seasonal sample distributions. The K-S test statistic >D< is the maximum differ 
ence in the fraction of data within 1 of 4 natural quadrants around the maximizing data 
point in the velocity/direction plane which is found by ranging over both all data points 
and associated quadrants. For the 2 winter distributions D attains a maximum of 19 % 
in the quadrant northeast of approx. (8 hPa, 200°). Similarly, for the summer samples, 
D = 17 % again in the quadrant NE of about (5 hPa, 190°). Both, the wintry and the 
summery distributions turn out to be significantly different at p-levels next to certainty, 
viz. < 5 x 10" 8 % and < 7 x 10" 6 %, respectively. The D-values are in good agreement 
with cumulative differences within the ± 0.5 % contours of Fig. 3-19,p. 75 (bottom) which 
amount to ± 18 % for winter (JFM) and ± 15 % for summer (JAS). As an aside, the 
complementary difference patterns of Fig. 3-19, p. 75 imply that annual wind statistics 
would not have produced any cross-regime differences to speak of due to mutual can 
cellation over the annual cycle. 
A quantitative summary of Fig. 3-19, p. 75 is supported by Table 3-9, p. 76. Notably differ 
ent decadal SST regimes (- A(R 2 ) + A(R 3 ) « 1 K) are sustained by hybrid couples of 
wind distributions. The cold regime is consistent with an intra-annual shift from conti 
nental winds in winter (cold) to maritime winds in summer (cold). A regime shift in SST 
is accompanied by a consistent reversal of the seasonal wind climate. 
The most recent cold to warm shift in the late 1980s went along in winter with a 61 % 
increase in strong maritime winds on the expense of a 40 % decrease in moderate 
continental winds (»ConWin« => »MarWin«). Conversely, strong maritime winds 
dropped by 26 % in summer due to an overall increase of continental winds by 59 % 
(»MarSum« => »ConSum«). While changes in direction and velocity were about bal 
anced in winter, directional changes clearly surpassed velocity changes in summer. 
C.3 Concluding Remarks 
The bi-stability of SST regimes casts doubt on the usefulness of standard CliNo clima 
tologies in climatic mishmash regions such as the North Sea where the implicit as 
sumption of climatic stationarity is violated. As a consequence, the »significant-depar- 
ture-from-normal alert« is on next to constantly (cf. Fig. 3-18, p. 72) because the alarm 
device is not »aware« of a significant change in normality. A solution might consist in 
judging the significance of anomalies from mixed probability density functions that ac 
count for the bi-stability phenomenon. 
The findings as to hybrid and reversed wind distributions during different SST regimes 
were obtained for the cold period 1978 - 1987 and the recent warm episode. The ro 
bustness of these results should be examined e. g. by way of extending the analysis 
to the long SST time series of Helgoland Roads. 
Finally, it would appear useful to explore how the changeable wind characteristics 
across the North Sea associate with large scale (anomalous) features of the general