Atmospheric Physics 
System Nordsee 
27 
One prominent departure from the climatological trajectory consists in the SSW-ly 
course in winter 2006 when meridional winds were dominant on the whole, even 
though directional persistence was as low as 0.5:9.2 or 5% in February. Also of note 
are a strong and persistent (68%) W-ly flow in and most untypical of April, distinctly 
weak SW-ly winds in July, and the intense zonal circulation from November 2006 
through January 2007 with mean scalar wind speeds beyond 15 m/s (Jan.) and per 
sistences up to 90% (Dec.). In July 2007, W-ly winds reached scalar and vectorial 
speeds 2 and 3 times as high as those of July 2006. The NW-ly course of the trajectory 
in September and November was in contrast to climatology and prevalent SW-ly winds 
of the previous year. These sizeable differences in the seasonal evolution of wind 
conditions, not only between both years, but also in comparison to climatology, do not 
vanish but become blurred on inspection of annual mean scalar (8.4, 8.9, 8.6 m/s) 
and vectorial wind speeds (4.0, 4.2, 3.5 m/s) for 2006, 2007, and 1971 - 2000. On an 
nual time scales, the perhaps most conspicuous distinctions are due to wind direction 
(242°, 270°, 253°). 
Gales (p. 92 sqs.) 
Gale identification employs a peak-over-threshold technique on the gale index G* de 
fined in terms of wind-speed and vorticity indices as carried over from the weather 
typing scheme. Original thresholds were specified for a UK Met Office daily mean 
sea level pressure dataset based on one-per-day operational analyses. By contrast, 
NCEP/NCAR Reanalysis I MSLP data used here constitute daily means of 4 fields 
per day. To account for the smoothing effect arising from averaging, thresholds had to 
be re-adjusted as to not underestimate gale frequencies by some 30%. To this end, 
a Generalized Pareto Distribution was fitted to the tail of G* within 1971 - 2000. The 
gale index thus turned out Pareto-Type-2 distributed with upper bound at 69.2 hPa 
and associated limits of wind speed (V) and vorticity (Q of 42.9 m/s and 1.05 rpd, re 
spectively. The new thresholds for the 3 classes >Gale< (G), »Severe Gale< (SG), and 
»Very Severe Gale< (VSG) were estimated for excess probabilities of 0.1,0.02 & 1/365 
at 28.3, 36.6 & 44.6 hPa. Corresponding limits of V and C, are 17.5,22.7 & 27.7 m/s and 
0.43, 0.55 & 0.67 rpd. 
The 2 most severe gales occurred on Dec. 2, 1966 (G* = 55.3 hPa) and Feb. 16, 1962 
(G* = 53.7 hPa, >Vincinette<, Hamburgian storm surge disaster). The December storm 
also holds the record vorticitywise (0.78 rpd), while as yet >Vincinette< produced the 
highest wind speed (31.4 m/s). Re-calibrated gale identification was further used in 
compiling catalogues of monthly and seasonal frequencies of classified and total gale 
days from 1947/48 through 2009/10. These catalogues for gale years (shifted by 6 
month vis-à-vis the calendar year) served as the basis of assessment of gale frequen 
cies during gale years 2005/06 and 2006/07. 
Apart from the cyclone >Uriah< that gave rise to a record summer storm surge in the 
Ems region on June 26, 2007, all gales of both, the 2005/06 and the 2006/07 season, 
fell into the core period October to March (02M). Monthly gale frequencies were con 
stantly below average throughout 02M in 2005/06 such that, taken together at 26 gale 
days of which only 2 were classified SG, this season ranks unusually low. In compari 
son, counts of G (40) and SG (12) were about twice and 6 times as high, respectively, 
during 02M 2006/07. At 54 gale days altogether 2006/07 ranks 3 rd place behind gale 
years 1994/95 (56) and 1989/90 (57). New records were established in fall (OND) and 
December (27 & 14). Gale frequencies in December and January (14 + 12) were sur 
passed just once, viz. 1982/83 (7+21), and went along with record high values above