System Nordsee 
293 
List of Figures 
Atmospheric Physics 
Fig. 2-1: Sea level pressure (hPa) for the very severe gale >Orkun< of March, 18 2007 together 
with grid positions of weather-typing scheme 41 
Fig. 2-2: Percent frequency distribution of daily circulation states for the period 1971 -2000 
together with classification criteria. Boundary lines (green) separating directional from 
hybrid (/*//* = ±1), and hybrid from rotational sectors (Ç*/V* = ± 2). Thresholds G% r for 
elliptical gale shells (blue): 28.3 (gale), 36.6 (severe gale), 44.6 hPa (very severe gale); half 
axes: a* = 2G% r , b* = a*/2. Top and right axes: V = 0.62 V*, £ = 2.72 £*. 42 
Fig. 2-3: Air circulation around high (A) and low pressure systems (C) 46 
Fig. 2-4: Percent frequencies of reduced weather types for 2006 (blue) and 2007 (red) along 
with empirical distributions for 1971 -2000 (stacked columns). Percentile range Min to 
Max, internal interval limits at 10,25,50 (median, continuous stepline), 75 and 90%. . .52 
Fig. 2-5: As Fig. 2-4, except for monthly absolute frequencies (days) 53 
Fig. 2-6: Scaled histograms (stacked) of run length (rl) of weather types (w) in 2006&2007. Base 
%-scale: 365 d. Fleight of column blocks: rl x number of runs r(w,rl). Black figures: r(w,rl), 
blue: r(rl). + Symbols: rl x r(rl) x 365/10958 d for 1971 - 2000 55 
Fig. 2-7: Cumulative empirical vs. geometric survival function (PP-Plot) for weather-type runs 
(5438 altogether) during 1971 -2000 (10958 d). p is the inverse of the mean life time (cf. 
Table 2-10), 1 - p the probability of surviving. Powers (1-p) k yield the crosshairs lateral po 
sition due to the survival function S (= 1-GeoCDF). The topmost position (k = l) gives the 
percentage of runs or episodes that reach a minimum age of 2 days 58 
Fig. 2-8: Weather-type transitions, annual climatology. The arc length of a ring section reflects 
the frequency fin days of weather type X (marginal sum of countmatrix C in Table 2-11). 
The segments within section X depict stacked frequencies of transitions Y^X^X (column 
entries); they are colored after the preceding weather type and given in ascending order; 
the leftover portion of section X is associated with self transitions Y=X^X. Transitions 
X^Y^X(row entries), colored after the succeeding weather type, are identified by the so 
ckets of Bezier ribbons. The ribbons have color X, iff(X ~^Y)> f(Y -» X), otherwise color Y. 
62 
Fig. 2-9: Same as Fig. 2-8, except for 2006 and 2007. 63 
Fig. 2-10: Classified seasonal sea level pressure distributions (hPa). Left: UKMO climatology 
(1971 - 2000), contour interval 1 hPa. Middle: same as left, except for NCEP/NCAR. Right: 
NCEP/NCAR minus UKMO, contour interval 0.25 hPa. JFM, AMJ,... = winter, spring, ...All 
distributions displayed at the horiz. resolution inherent in the UKMO dataset (5x 10 deg.). 
70 
Fig. 2-11: Climatological departures (1971 - 2000) from »observed« sea level pressure at WMO 
station 01492 Oslo-Blindern (59°56'34"N, 10°43' 15"E, elevation: 94 m) for UKMO, NCEP/ 
NCAR, and ERA-40 MSLP at 60° N, 10° E. 71 
Fig. 2-12: NCEP/NCAR climatology (1971 - 2000) of seasonal height distributions (gpdm) of the 
925 hPa level. This pressure level has about the same height as the Scandes in the model's 
orographic representation of this mountain range (600 - 700 m). Isohypses play the same 
role as isobars in the MSLP distributions of Fig. 2-10 72