KLIWAS Seite 26 KLIWAS Climatology of North Sea Fronts 5 The impact of the loc al wind field on SST fronts in the German Bight Beside the annual, seasonal, and monthly climatological SST products presented in Part B of this report, also the dependency of frontal position and gradient strength from prevailing wind directions was analysed. Especially for river plume fronts in the German Bight it is known from in-situ and EO data, that the local wind field has an impact on front position and gradient strength. Therefore - as a first approach - the procedure of Jenkinson and Collision (1977) for an objective determination of the ’Lamb Weather Types’ 4 was applied. This procedure allows a classification of daily large-scale weather situations basing on mean sea level pressure (MSLP) data only. This procedure was applied to a North Sea sub-sample of daily MSLP NCEP/NCAR fields 5 . It provides indices for wind and vorticity which are representative for the whole North Sea region. Empirical relations between these indices determine the circulation type (e.g. cyclonic, anti-cyclonic, north-west, etc.) and allow the identification of storm events. The classification of the geostrophic wind direction, corresponds to an 8-point compass (N, NE, E, SE, S, SW, W, NW), a detailed description of the methodology and its application to the North Sea is given by Lowe et al. (2005). The Fig. 8 shows the frequency of easterly (50°-140°) and westerly (230°-320°) geostrophic North Sea winds for the period 1990-2011. In the German Bight these wind sectors are known for periods of persistent wind direction, which is a prior condition, that the local wind can shift the position and properties of fronts significantly. A total of 642 east wind and 2326 west wind days had been detected for this 20-year period. The Fig. 9 shows the monthly distribution of geostrophic east and west wind days for the period 1990-2011. As expected, both wind sectors show a different seasonal cycle. In a first approach, the distribution of the westerly winds has been adapted to that of easterly winds ( Fig. 9). However, thereby the seasonal dependency is not removed, because the effects of changing wind direction on the temporal statistics of fronts can be not distinguished from the effect of the seasonal temperature variations. The observed distribution of all geostrophic wind direction in 1990-2011 is shown in Fig. 10. In order to remove the seasonal cycle of the wind distribution as far as possible, the wind distribution has been adjusted by selecting a uniform distribution regarding the months (Fig. 11). For the month August such an adaption was not 4 http://www.cm.uea.ac.uk/cm/data/lwt/ 5 Reanalysis I, 1948 until present, Kalnay et al. (1996)