Conclusion

Simultaneous use of satellite images and NWP output provides interesting insight into the dynamical processes in a developing cyclone. However, one must be very careful during their interpretation: a number of diagnostic parameters using different coordinate systems (p, theta) and vertical cross-sections are needed for this purpose. There is usually little time in the operational shift to follow the development of a cyclone in such detail, although recognition of some characteristics can be useful, mainly during its early development. The main findings of this study were:

  1. The cyclone started to develop far downstream of the target area, along the coast of the United States. Its propagation and evolution were largely determined by a strong upper-level jet, which continuously approached the surface cyclone.
  2. Although the cyclone joined a larger cyclonic circulation in the Northern Atlantic, its pressure was continuously decreasing and there was still significant vertical motion in its environment. A quick look at the mean sea level pressure charts probably does not reveal these facts; one can even get the impression that the cyclone vanishes. Other parameters, like low- and mid-level relative vorticity, low-level (e.g. 925 or 850 hPa) geopotential and storm-relative wind, can be useful to detect whether there is still an independent or even intensifying cyclonic structure.
  3. It is difficult to interpret the flow in the cyclone solely based on satellite imagery. The cloudiness, which is often confined to upper levels (6-7 km height), can be shifted a long way with respect to surface centre of the low and its fronts, mainly due to very strong airflow. It is useful to combine satellite images with pressure and wind analyses on 285, 300 or 315 K theta surfaces, which gives information on both the horizontal and vertical structure of the flow. The presence of dry areas (slots) and onset of sinking motion can be the first signatures of a developing severe event.
  4. The intensification stage of the cyclone is recognizable from the approach of the upper-air jet streak (and of the upper-air PV anomaly) toward the cyclone's centre and by the development of deeper clouds with lower brightness temperatures at their tops on the IR10.8 images. The jet streak can be found ahead of a belt of dry (already stratospheric) air, visible on both WV6.2 and Airmass RGB images.
  5. The structure of the cyclone and its frontal system changed (from Norwegian to Shapiro-Keyser model) during its intensification. Some signs of this were visible on the satellite images as well, e.g., the cloudiness in the occlusion part of the cyclone became more pronounced, as well as the intrusion of drier air close to its centre (related to amplification of the downslope, frontolytic motion at the rear of the cyclonic circulation).
  6. Numerical model results indicate that the onset of high surface wind speed was a result of the superposition of the pressure gradient force effect and vertical transport of momentum from mid- and low-tropospheric levels in an area of strong baroclinicity. This mechanism was indicated by forecasts of downward motion west and southwest of the cyclone centre. The EPV diagnostics and cross-sections in this area also indicate the presence of conditional symmetric instability at mid-levels but its role in generation or intensification of downward motion is uncertain. The location of CSI is also different from the original sting jet concept, which would imply stronger downslope motion due to symmetric instability further eastward or north-eastward and closer to the cyclone centre. However, this original concept is not supported by surface wind gust observations. There were some (though not strong) signatures that the sinking motion accompanied with low relative humidity air approached the surface, e.g., one could see dissipation of low-level clouds at the rear of the cyclone on the HRV and Natural Colour RGB images.
  7. PBL processes have an important role in the transfer of vertical momentum toward the surface. It is useful to know the stratification and wind shear in the low levels of the troposphere, whose footprints can be recognized in the shape and organization of the low-level cloudiness (mainly on the HRV imagery). This would be worthy of extended research in the future. Since soundings are not very frequent (and the processes we study are fast), one could use pseudo-soundings from numerical models and also IASI soundings to check these characteristics. Very different stability conditions can be found in different parts of the cyclones, which can be better displayed if the satellite images are overlapped with lapse-rate or bulk Richardson number contours.
  8. Toward the end of the windstorm, parallel bands of cloudiness appeared in areas where a sting jet would be expected according the original conceptual models. However, it can be misleading to attribute the development of severe winds or sting-jet-like flow to this signature only. There is still a debate among scientists about the importance of symmetric instability in developing very strong surface winds. In the Friederike case, most of the observed gusts (typically between 100 and 150 km/h) were probably not a direct result of this instability. However, it is possible that symmetric instability can sometimes play a role in generating strong downslope motion causing an even more extreme windstorm (e.g. similar to the 16 October 1987 storm over the British Isles). It can therefore be important to follow the location of CSI areas with respect to areas of strong low- and mid-level winds and to estimate the persistence of CSI. Similarly, the presence of deep convection in combination with very strong low-level shear can lead to extreme local events as tornadoes or downbursts (notable example was the Emma windstorm of 1 March 2008). Thus, the convective environment must also be very carefully examined in these kinds of storms.