Meteosat-11 Airmass RGB, 14 April 2020

Retour d’est or extended Bora flow along Po Valley

13 April 2020 12:00 UTC-15 April 12:00 UTC

Meteosat-11 Airmass RGB, 14 April 2020
Meteosat-11 Airmass RGB, 14 April 2020

Examples of retour d’est or extended Bora flow along Po Valley, could be seen in satellite imagery in April 2020.

Last Updated

03 February 2021

Published on

02 February 2021

By HansPeter Roesli (Switzerland)

Bora winds don't just blow over the Adriatic Sea, they are strong enough to even penetrate the Po Valley and blow up to its end in Lombardy, Piedmont and Italian Switzerland. This phenomenon is also known as ‘retour d’est’, i.e. back flow from the east. The retour d’est used to be a forecasting/nowcasting challenge. In particular, the temporary formation of the cloud deck would impair the quality of many temperature and sunshine forecasts. Observational hints were scarce and, beyond lacking 3D-resolution of the regional orography, early numerical models dried up in the lower layers early in forecast time.

Nowadays, as proved by this case, numerical models are very capable in predicting this phenomenon, even with lead times of a couple of days.

On 13 April the tail of a long cold front, linked to a depression over Scandinavia, passed over Central Europe. It crossed the Alps in the early morning of 14 April, generating a weak north-foehn situation over the alpine southside and the Po Valley. On the sequence of Airmass RGBs from Meteosat-11, between 13 April 12:00 UTC and 14 April 12:00 UTC (Figure 1), the cold front was marked with long cirrus streaks, which hid from sight what was going on at lower levels.

Figure 1: Meteosat-11 Airmass RGB, 13 April 12:00 UTC-14 April 12:00 UTC.

However, when the cirrus shield left Austria and Slovenia, cloud development occurred in its back, over northeastern Italy (convective cloud) and on the Adriatic Sea (cloud arc). Of particular interest is the cloud arc (two green arrows on the Airmass RGB at 10:45 UTC (Figure 2).

Meteosat-11 Airmass RGB, 14 April 2020
Figure 2: Meteosat-11 Airmass RGB, 14 April 10:45 UTC

The evolution of this arc can be forward/backward tracked on the looped sequence of close-up Airmass RGBs between 08:00 UTC and 12:30 UTC (Figure 3), as it moved out from under the frontal cirrus off the Croatian coast and arrived over the Italian coast, south of the Po estuary, some four hours later (average speed ~25 km/h).

Figure 3: Meteosat-11 Airmass RGB, 14 April 08:00-12:30 UTC. Looped animation.

Sentinel-3B passed over the scene at 09:33 UTC when the arc still was somewhat hidden by the cirrus. Using the previous Airmass RGB imagery as guidance, the arc could be detected through the semi-transparent cirrus veil of the Sentinel-3B OLCI True Colour RGB (green arrows on Figure 4).

Sentinel-3B OLCI True Colour
Figure 4: Sentinel-3B OLCI True Colour RGB, 14 April 09:33 UTC.

With the same OLCI-B data granule the foehn flow can be rendered visible by displaying the ratio between channels Oa19 (NIR 900 nm) and Oa17 (NIR 865 nm). Oa19 is in the near-infrared solar water vapour absorption band (weak absorption), Oa17 is in an adjacent atmospheric window band. The ratio measures the so-called Solar Moisture Transmittance (SMT), with values in the range from about 0.6 (humid - dark grey) to 1 (dry - bright). The image in Figure 5 shows two bright-dry foehn swaths (red arrows) that still persisted over the central Po Valley at 09:33 UTC.

Sentinel-3B OLCI NIR
Figure 5: Sentinel-3B OLCI ratio of two NIR channels.

Turning back to the previous discussion, it appears that the cloud arc in Figure 4 was driven by a first Bora outbreak. According to Figure 3, after 12:00 UTC, another cloud arc, marking a second Bora outbreak, developed over the Venice area. It soon overran the first Bora outbreak, as illustrated by two Natural Color RGBs from the SVI-bands of VIIRS on NOAA-20 and Suomi-NPP, at 12:07 UTC and 12:52 UTC, respectively (Figure 6).

Figure 6: NOAA-20 and Suomi-NPP Natural Color RGBs, 14 April, 12:07 UTC and 12:52 UTC, respectively.

In the following hours the Bora tongue moved up to the end of the Po Valley, replacing the warmer drier foehn air. There was enough temperature contrast in the IR10.8 channel of Meteosat-11 to enhance and track the Bora's push up to the formation of a cloud deck at night in the uppermost part of the valley, as shown in the image sequence between 12:00 UTC and 01:00 UTC on the following day (Figure 7).

Figure 7: Meteosat-11 infrared, 14 April 12:00 UTC-15 April 01:00 UTC.

Further details may be gleaned from Meteosat-8 HRV imagery, chosen over Meteosat-11 as the increased scattering from its viewing angle gives more contrast. The single frames of the animation were enhanced according to the varying illumination between 09:00 UTC and 17:30 UTC (Figure 8).

Figure 8: Meteosat-8 HRV, 14 April 09:00-17:30 UTC.

In both the IR10.8 and HRV animations a bright spot appears around 17:00 UTC, just east of the Lake Como (red boxes on Figure 9). These strong signals were from tiny lenticularis clouds that had temporarily formed in the otherwise clear sky of the mid-level north-foehn flow. Confirmation of the clouds comes from my personal observation at Locarno, some 50 km west of the lake.

Meteosat-11 HRV and Infrared
Figure 9: Meteosat-11 HRV (top) and infrared (bottom), 14 April, 17:00 UTC.

The arrival of the foehn air and the penetration of the Bora tongue were both also well marked in three consecutive radiosounding profiles of Milano-Linate (Figure 10):

  • 14 April 00:00 UTC – ‘normal’ profile with westerly wind in the boundary layer
  • 14 April 12:00 UTC – extremely dry air under northerly winds up to 500hPa – north foehn
  • 15 April 00:00 UTC – humid air with inversion at 800hPa under sustained easterly winds – Bora T
Milan radiosonde
Figure 10: Radiosounding profiles of Milano-Linate

The GFS 0.5° analysis and forecast of wind and temperature at 950 hPa on 14 April at 12:00 UTC and 15:00 UTC (Figure 11) respectively give an appraisal of the wind and thermal fields close to the ground at 950 hPa. They compare well to the Bora tongue shown by the underlaid HRV images of Meteosat-8.

Meteosat-8 HRV with GFS analysis overlaid
Figure 11: Meteosat-8 HRV with GFS T-wind analysis overlaid, 14 April 12:00 UTC (top) and 15:00 UTC (bottom).

The meteorology behind retour d’est

Dynamically the back flow is linked to a cold front that has crossed the easternmost outcrops of the Alps and the Dinaric Alps, which leads to north-foehn conditions south of the central and western parts of Alpine arc. In these cases, the rather weak foehn currents are short-lived and soon get overpowered by the Bora winds that blow through the Ljubljana gap (the transition area between the Alps and Dinaric Alps in the area of Trieste and Ljubljana) over the Adriatic Sea. Thus, the Bora winds occur in the back of the cold front.

When the Bora winds penetrate the Po Valley the dry foehn air is pushed westward and undercut by the cooler, less dry Bora air. In addition, in the westward direction the Po Valley acts as a funnel to the Bora air. The valley tapers between the Apennines range and the southward curving Alps, and has only a minor outlet at its end across the West Ligurian hills that leads down to the Gulf of Genoa.

Under these dynamical-mechanical forcings clouds form along the arc-like south-alpine slopes. Depending on the strength of the Bora flow and the high pressure building up aloft, the cloud deck consists of either stratocumuli or altocumuli. In extreme cases there might be some drizzle (winter) or shower/thunderstorm (warmer seasons).

Over Italian Switzerland the event frequently follows a well defined diurnal cycle with additional local forcing from differential heating, lifting and mixing:

  • Break-down of the north foehn in the late morning.
  • Onset of valley, initially gusty, winds supported by the Bora.
  • Cloud formation in the evening, starting along the mountain crests.
  • Dissolution of the (broken) clouds by the afternoon of the following day latest.