Surface weather charts of the Swiss Meteorological Institute (SMI) at 00 UTC on a 30th April, and b 1st May 1982. A cold front impinges upon the Alps from the north, whereupon a lee cyclone develops over the Gulf of Genoa 

Contexts in source publication

Context 1

... modi®cation of a cold front by the Alps is one of the main topics of Alpine mountain meteorology. For instance, many of the objectives of the Mesoscale Alpine Programme (MAP) are directly or indirectly related to the passage of a cold front in the Alpine region (Binder and Sch ar, 1995). The in ̄uence of the Alps on the propagation and structure of the front, is usually associated with a multitude of mesoscale ̄ow features (Smith, 1985; Tafferner and Egger, 1992; Sch ar et al., 1998). Ahead of the cold front, there is often southerly prefrontal foehn in the Alpine valleys (Seibert, 1985). As the cold front reaches the Alps, the winds turn into northerly or north- westerly directions, and the cold front experiences retardation and modi®cation (cf. Davies, 1984; Heimann and Volkert, 1988; Hoinka et al., 1990; Egger and Hoinka, 1992). The low-level ̄ow becomes essentially blocked and experiences ̄ow splitting on the windward side (Pierrehumbert and Wyman, 1985; Chen and Smith, 1987; Binder et al., 1989). Westward de ̄ected cold-air masses are responsible for a cold-air outbreak into the Mediterranean sea between the Alps and the Pyrenees (Pettre, 1982). The resulting mistral wind is then further intensi®ed by the channeling along the Rhone valley. At the south-western Alpine tip, the mistral separates from the Alpine massif leading to the formation of an initially shallow cyclonic orographic vortex (Aebischer and Sch ar, 1998). If the approaching upper-level trough is able to overtake the retarded surface cold front, a deep lee cyclogenesis event may result over the gulf of Genoa (e.g., Buzzi and Tibaldi, 1978; Bleck and Mattocks, 1984; Tafferner, 1990). A part of the cold-air ̄ow is driven over the Alps through major mountain passes and gaps in the Alpine topography, setting-up a north foehn ̄ow in the Po valley. Finally, the eastward de ̄ection of cold air around the Alps leads to the formation of the north-easterly bora wind, which ̄ows over the Dinaric Alps and sometimes reaches as far west as the Po valley (Smith and Sun, 1987). In a few documented cases, the separation of the bora from the Alpine massif may also lead to the formation of a shallow anticyclonic orographic vortex over the Adriatic sea (Steinacker, 1984a, 1984b). In this study detailed consideration is given to such a frontal interception event that occurred towards the end of the Alpine Experiment (ALPEX) between April 29 and May 1, 1982. This spectacular event involves a very strong cold front and exhibits almost all of the aforementioned mesoscale ̄ow features (Kuettner, 1980). It has been investigated in several previous case studies (Steinacker, 1984a, 1984b). An overview on the ̄ow situation is provided by the surface weather charts presented in Fig. 1. On April 30, 00 UTC, the parent low pressure system is located over Sweden, and the associated cold front is about to reach the Alps. The front is progressing at high speed and its interception by the Alps leads to mistral, lee cyclogenesis, north foehn and bora. On May 1, only 24 hours after the front has reached the Alps, the lee cyclone has already reached the mature stage and is located between Sicily and Italy. The ®rst objective of the current study is to assess the quality of the ALPEX-IIIb reanalysis data set provided by the German Weather Service (DWD). It covers the whole of the ALPEX Special Observing Period SOP, 1 March±30 April 1982 and the following 8 days which were included due to a strong south foehn case occurring brie ̄y after the end of the SOP (Gerhard, 1994). Due to its late production, the reanalysis has not been exploited in the immediate post- ALPEX research, and the current study appears to be the ®rst to make detailed use of it. The value of this reanalysis is assessed by comparing a pair of simulations driven by the ALPEX and the ECMWF reanalysis (ERA) data sets, respectively. The second objective of the study is to provide a detailed high-resolution account of the associated mesoscale ̄ow systems in the Alpine region. Hence, a high-resolution simulation is driven by the ALPEX-IIIb reanalysis in a doubly- nested mode. In our analysis of the simulations which cover the period from 29 April 00 UTC to 1 May 1982 06 UTC, we focus on the transient development of synoptic and mesoscale features and wind systems. The analysis includes time series of vari- ables at selected locations, as well as trajectory calculations employing the methodology of Wernli (1995). Special emphasis is further given to the potential vorticity (PV) perspective (Hoskins et al., 1985), considering upper-level PV advection (Bleck and Mattocks, 1984; Tafferner, 1990), low-level orographic generation of PV anomalies and PV banners (Aebischer and Sch ar, 1998), as well as diabatic generation and ampli®cation of mid-tropospheric PV streamers by condensational heating (e.g., Morgenstern and Davies, 1999). Finally, we assess the sensitivity of the simulations to the blocking effect of the Alps. Several parameters are expected to have a direct effect on the ability of the Alps to block and split the incident ̄ow and thereby to contribute to and interact with frontogenesis and cyclogenesis: According to Pierrehumbert and Wyman (1985), increasing the mountain height intensi®es the upstream blocking of the low-level air. Georgelin et al. (1994) found a signi®cant increase in upstream blocking due to the increase of surface friction. Besides, latent heating can favor a ̄ow over a mountain instead of around it due to the reduced buoyancy in ̄icted by condensation processes (Schneidereit and Sch ar, 2000). In the following section we display the numerical model and present a brief discussion of the driving reanalysis products. In Sect. 3 we compare EM 56 km resolution simulations driven by the ALPEX-IIIb and the ERA reanalysis, respectively. Section 4 provides a detailed analysis of associated mesoscale ̄ow systems using a 14 km resolution HM simulation. Sensitivity experiments regarding the effects of surface friction, orographic height and diabatic processes follow in Sect. 5. The conclusion can be found in Sect. 6. The numerical simulations of this study were performed using the numerical modeling system originally developed by the German Weather Service (DWD). The low-resolution version of the model is referred to as the Europa-Modell (EM). It has a horizontal grid spacing of 56 km and was until recently used operationally by the DWD with a domain covering all of Europe as well as the northern Atlantic. The high-resolution model (HM) is one-way nested into the EM, has a horizontal resolution of 14 km and has been derived from the EM in a collaboration between the DWD and the Swiss Meteorological Institute (SMI). These institutions have run the HM in order to provide high-resolution short-range weather forecasts. A detailed description of the model can be found in Majewski (1991), and Schrodin (1995). The model is based on the hydrostatic approxima- tion on a regular latitude-longitude grid with rotated pole. It contains advanced physical parameterizations for radiation (Ritter and Geleyn, 1992), surface layer processes (Louis, 1979), boundary layer turbulence (Mellor and Yamada, 1974; M uller, 1981), and grid-scale cloud micro- physics of the Kessler type including a parameterization of the ice phase and moist convection (Tiedtke, 1989). For the HM simulations, a gravity wave absorber is employed at p 20 hPa level (see Klemp and Durran, 1983; Bougeault, 1983). The implementation of this absorber in pressure coordinates is by Herzog (1995). At the lateral boundaries, the models are driven using relaxation boundary conditions (Davies, 1976). The relaxation zones have a width of 8 grid points, and the ...

Context 2

... initialized on 29th April 1982, 00 UTC. In both cases, the same interpolation procedures are used for the generation of the initial and lateral boundary conditions from the reanalysis ®elds (for further details see Majewski, 1985). Reanalysis and simulations are shown in Figs. 3 and 4 on the 850 hPa and the 500 hPa level, respectively. The use of the ERA and the ALPEX-IIIb reanalysis data set, respectively, results in discre- pancies in the EM simulations. The lee cyclone is captured by both the simulations, but it turns out substantially stronger and more realistic if the ALPEX-IIIb reanalysis is used. This incongruity can be attributed to several factors. Here we focus on differences in synoptic-scale features and in local precipitation rates. One important factor for the development of lee cyclones is the advection of an upper-level PV streamer and its subsequent interaction with the retarded surface front (Bleck and Mattocks, 1984; Tafferner, 1990). Also, misforecasts over Europe are often related to the lack of amplitude and/or structure in the upper-level PV ®eld (Fehlmann and Davies, 1997). Analysis of the two reanalysis data sets demonstrates that indeed there is a pronounced PV streamer crossing the Alps (see later in Fig. 13), which provides the upper-level forcing for the cyclogenetic development in the Gulf of Genoa. However, both the ERA and the ALPEX-IIIb reanalysis (as well as the ALPEX-IIIb simulation) show the progres- sion towards the Alps and the wrap-up over Sardinia of this streamer, whereas the EM simulation using the ERA is unable to produce the wrap-up. In both simulations, however, the PV streamer crosses the Alps at about the same time and with similar amplitude. Hence, it seems unlikely that the shallowness of the ERA lee cyclone can be attributed to lack or erroneous speci®cation of the upper-level PV ®eld in the driving analysis. Another factor that may be important for the development of lee cyclogenesis are moisture processes. Diabatic heating acts by generating a vertically oriented dipole of PV, centered at the level of maximum condensational heating (Hoskins et al., 1985). In the case of Alpine lee cyclogenesis, the lower (positive) part of this dipole can help to strengthen the lower-tropo- spheric vortex, while the upper (negative) part of the dipole may support the cut-off of a pre- existing upper-level PV streamer from its stratospheric reservoir to the north (Morgenstern and Davies, 1999). The diabatic strengthening of an orographically-generated vortex was also observed by Aebischer and Sch ar (1998). This aspect was here analyzed by comparing the precipitation rates of the two simulations, which indeed differ substantially. Especially between simulation times t 32 h and t 42 h, the ALPEX-IIIb simulation produces much stronger precipitation in the Mediterranean region than the ERA run (Fig. 5). Therefore, it is suggested that the differences between the two integrations are driven by different diabatic processes. Finally, the sensitivity of the model simulations with respect to the initialization time of the integration was assessed. To this end, two additional simulations based on ALPEX-IIIb reanalysis data, but starting 12 and 24 h earlier than the one discussed above (i.e., at 00 UTC and 12 UTC on April 28, 1982) were conducted. In both cases the results are very similar as shown in Figs. 3 and 4 for the standard ALPEX simulation. In conclusion, the simulations using the ALPEX reanalysis data set capture the evolution of the synoptic-scale features of the lee cyclone very well. Using ERA data for initial and lateral boundary conditions, however, the lee cyclone is simulated too weak and especially too shallow in its vertical extension, primarily due to differences in the diabatic forcing by precipitation. Therefore, further integrations are conducted based upon the standard EM simulation driven by the ALPEX-IIIb reanalysis starting at 00 UTC on April 29. In this chapter, the weather development in the Alpine region from April 29 to May 1, 1982 is described using observations and corresponding simulations. To account for the small scale of the ̄ow features of interest, the 56 km resolution EM simulation driven by the ALPEX-IIIb reanalysis was re®ned using a 14 km resolution HM simulation. These simulations were initialized at 00 UTC on April 29, 1982 and integrated over 54 hours. The section is arranged according to the approximate temporal occurrence of the different ̄ow phenomena: (1) approaching cold front and ̄ow splitting, (2) cold-air outbreak and mistral, (3) north foehn, (4) cyclogenesis in the lee of the Alps, and (5) formation of an anticyclonic vortex and bora. Unless mentioned other- wise, all simulation diagrams are taken from the HM run. On 28th April 1982, a warm air mass is located over the central parts of Europe. In the course of the following day, the air is replaced by cold air impinging on the Alps from the north. The related cold front (Fig. 1) appears in the northern part of Central Europe 14 hours after the start of the simulation. It can easily be located due to wind shear, cold air advection towards south, strong precipitation and an elongated potential vorticity (PV) feature of diabatic origin just north of the Alpine ridge (see later, Fig. 12a). The cold front moves rapidly towards the Alps and reaches the Alpine crest after 24 hours of simulation time. The passage of the cold front over the Alpine crest is marked with a peak in relative humidity and a decrease in temperature (Fig. 6a, c). High values of relative humidity associated with deep convection are found up to 300 hPa (not shown). The same behaviour in relative humidity and temperature is shown in measurements taken from the surface meteorological network of the SMI (ANETZ) in the northern and the southern part of the Alps. With this passage of the front over the Alps, a strong horizontal gradient in equivalent potential temperature develops south of the Alps and subsequently intensi®es considerably (Fig. 7). On the other hand, north of the Alps only a small gradient in equivalent potential temperature is seen. Such a modi®cation of the impinging frontal structure was also reported for another case by an observational study (Hoinka et al., 1990). Figure 7b also shows the retardation of the cold front near the surface. Although the cold front is retarded and deformed by the Alps, it remains visible as a narrow PV belt after crossing the orography (see later in Fig. 12b). Ahead of the cold front, the low-level air to the north of the Alps is blocked. Up to the time t 12 h, i.e., 8 hours before the front reaches the Alpine crest, the wind vectors at 850 hPa show clear signs for ̄ow splitting. The associated splitting of the ̄ow near the eastern part of the Alps is also con®rmed by trajectory calculations (Fig. 8a). A detailed description of the applied trajectory tool is given in Wernli and Davies (1997). The trajectories were selected at time t 12 h in a small volume immediately to the north of the Alps, and at levels between 870 hPa and 850 hPa. The trajectories were subsequently calculated 42 h forward in time and 12 h back- ward in time. Thus, the starting point of each trajectory corresponds to time t 0 h, and is located slightly ahead of the cold front. It can be seen that some air parcels are blocked severely upstream of the Alps and are unable to surmount or detour them. Some other air parcels drift westwards, ̄ow around the Alps and end up in the mistral area and in the cyclonic eddy. Others ®nally are directed along the east of the Alps and follow a cold-frontal track. The air parcels which turn to the right, i.e. towards the Rhone Valley, sink to a level of 960 hPa and rise again to higher levels as soon as they reach the Mediterranean. When the cold front reaches the Alpine crest at time t 20 h, the tendency for ̄ow splitting is transiently reduced. This behavior is in quali- tative agreement with estimates of the dimensionless mountain height N Á H/U , where N is the upstream Brunt-Vais all a frequency, U the upstream velocity (approximately normal to the Alpine chain) and H the height of the Alps. For illustration, Fig. 8b displays the ratio j ~ V j a N at point B (Fig. 2) upstream of the Alps at 850 hPa. Assuming a dimensional mountain height of 2.5 km, the dimensionless mountain height NH/U drops to values below 2. Together with the moist conditions (note the distribution of equivalent potential temperature in Fig. 7) this indicates that during the frontal passage, a substantial fraction of the ̄ow was over the Alps rather than around. However, as soon as stable conditions are restored in the cold air at t 20 h, the cold ̄ow behind the front is forced to ̄ow around the Alps, at least on the Alpine scale. A strong northerly ̄ow between the Pyrenees and the Alps (point A in Fig. 2) starts to develop around time t 10 h (Fig. 6b, solid). It then remains nearly constant at 20 m/s during the next 15 hours. The cold frontal passage is accompanied by a transient increase in relative humidity (Fig. 6a, solid). The stronger blocking of the cold air behind the cold front and the resulting channeling of the air along the Rhone valley between the Alps and the Pyrenees leads to a further increase of the mistral up to 27 m/s. The maximum of the mistral is located at about 900 hPa and is separated from the calm air in the lee of the Alps by a region of strong horizontal shear. The related cold-air outbreak also implies a decrease in temperature by about 6 C (Fig. 6c, solid). At the time of its maximum extent ( t 36 h), the impressive jet reaches as far south as Sicily (Fig. 9). Behind the front, cold air is advected towards and ®nally blocked by the Alps. The air, therefore, piles up to the north, spills over the Alps through the mountain passes, and induces north foehn. The associated strong winds replace the earlier almost windless situation immediately to the ...