Mesoscale weather features over the Mediterranean: Part 1


Surface synoptic observations, even from a high-resolution network of stations, may not reveal the arrangement or detail of mesoscale weather features. High-resolution satellite imagery, typically with a horizontal scale of about 1km, allows many smaller-scale features to be recognised. In many cases, features visible in imagery from geostationary satellites, such as the second-generation Meteosat orbiter and its Spinning Enhanced Visible and Infrared Imager (SEVIRI), may be tracked almost continuously.

This article illustrates the fine detail apparent in satellite imagery of the Mediterranean area.

Mesoscale cloud lines

Amongst the most notable of small features revealed by satellite imagery are mesoscale cloud lines, which have a width of between about 1 and 10km, but lengths of tens to hundreds of kilometres. The cloud lines may be formed by convection or turbulent mixing and examples of both forms are discussed here. Cloud lines with a width of less than about 3km are rarely resolved by infra-red radiometry from meteorological satellites, so that no more than a ‘smear’ of lighter shade may be evident in imagery. However, if the resulting convection is deep, cirriform anvils may be seen. By day, high-resolution visible imagery usually allows these features to be seen clearly.

Convective cloud lines tend to form where there is little or no other low cloud in comparatively dry air masses in which instability may only be released by additional forcing mechanisms. Inspection of radiosonde profiles for these events typically shows that their formation is in otherwise stable dry air, often with a lapse rate (γ ) marginally less than 9.8 degCkm–1 (the dry adiabatic lapse rate – DALR) in the boundary layer. Cloud lines of this type are frequently observed over the Mediterranean.

The variations in the length of shadows cast by convective cloud lines confirm that many are formed as ‘ropes’ of adjacent cumulus clouds.1 Three notable cases are shown in Figures 1–3.

Figure 1.

False-colour image from Meteosat-9/SEVIRI showing a baroclinic convective cloud line south of Rhodes (cirriform tops highlighted blue) at 1200 utc on 19 February 2009. Another mesoscale convective cloud line can be seen west of Cyprus. In the west of the image, mountain waves with a wavelength of about 5km can be seen in the cyclonic circulation in the lee of Sicily and the ‘toe’ of Italy. (© EUMETSAT 2009.)

Figure 2.

High-resolution2broad-band visible image of the eastern Mediterranean from Meteosat-9/SEVIRI at 0615 utc on Saturday, 14 February 2009. Over the east of Cyprus, fog and stratus can be seen across the Mesaoria in this early-morning image. Cumulus clouds are developing over the mountain ranges of Lebanon and southern Syria. (© EUMETSAT/Met Office.)

Figure 3.

False-colour image (emphasising the 0.8μm visible wavelength) from NOAA-16 at 1229 utc on 4 December 2003, showing narrow cloud lines between northern Corsica and Toulon (France), as well as to the east and north of Corsica. Cirriform clouds forming along the crest of the Alps and extensive low clouds (stratus with a top near 500m) beyond the Alpine lee slopes and over northern France, north of the frontal zone over northern Italy, are also visible in Figure1. The AVHRR instrument on this satellite gives a resolution of 1.1km at the nadir ( (© 2003 RSGB/NOAA.)

Rope clouds around Cyprus and Corsica

Occasionally, convective cloud lines can be seen to form along weakly baroclinic zones; two examples are shown in Figures 1 and 2. These cloud lines formed as narrow bands extending south from cold fronts over Turkey.

On 19 February, a long line of convective cloud could be seen (Figure 1) in association with a zone of falling wet-bulb potential temperature (θw). In this case, the cumulonimbus clouds were approximately 20km wide and initially extended about 450km south from Rhodes. However, the total length and development of the line diminished as it moved east and thundery rain was observed later in the day only between the north of Cyprus and Turkey. Extensive cirriform clouds can be seen streaming away from the cloud line, indicating deep convection along much of the line. Contour heights decreased gradually through, and behind, this cloud line in the middle troposphere. Cloud tops were at approximately 7km and were carried east by winds of around 35ms–2.

Deeply unstable clouds are seen west of Cyprus on 14 February 2009 in Figure 2. In this case, the cloud line formed in association with a short-wave trough at 500mbar ahead of a broad upper trough over central parts of the Mediterranean. A marked line of convective clouds could be seen stretching south from the Gulf of Antalya. In this case, as in the case described in Galvin (2009), the line was on the leading edge of a zone of decreasing θw, associated with a slight veer of the wind in the boundary layer. The release of instability yielded a line of cumulonimbus clusters about 10km wide, extending 250km south, and continued as a narrower line of towering cumulus clouds, yielding a total cloud-line length of about 450km. The cumulonimbus clouds were separated by about 60km, but each brought occasional thunderstorms. Further south, there were isolated showers as far south as the Troodos Mountains of Cyprus, although by midday most of the rain and thunder were seen over the mountains of southern Turkey, where insolation generated more widely-spread cloud development. Cumulus congestus was observed at RAF Akrotiri on the south coast of Cyprus around 1400 utc as the dew point rose to 15°C, but there was no rain. It can be seen that the cumulonimbus clouds are spreading out towards the northeast in strong winds at around 5km altitude. The cirrocumulus spreading away from the cumulo nimbus was itself unstable, as reflected in the radiosonde profile for Ísparta (Turkey) at 0000 utc on 14 February (not shown). A few cumulonimbus tops may have reached the tropopause, near 9km. Only with additional forcing was deep instability possible given sea temperatures of 15–19°C over the eastern Mediterranean; a stable layer at 1500m with dry air above also restricted the development of convective cloud.

Two other narrow cloud lines, east of the baroclinic cloud line, can also be seen in Figure 2: one close to the Turkish coast and the other off the southwest coast of Cyprus. Each was formed by turbulent mixing where katabatic drainage (mountain) winds met moist air over the sea. Their formation is similar to that described in O'sullivan and Galvin (2008) and they are a common feature off the mountainous coasts of the eastern Mediterranean, in particular overnight, when mountain winds converge with the prevailing flow over the open sea.

In Figure 3, cloud lines can be seen around the island of Corsica, west of a slow-moving frontal system over northern Italy on 4 December 2003. Figure 4 shows the profile from Ajaccio, Corsica for 1200 utc on 4 December, which suggests a lifting condensation level (LCL) near 900m with γ = 10.8 degCkm–1 from the surface for the cloud-line indicated by the arrow in Figure 3. This is reflected in the widespread development of cumulus clouds over the mountains of Corsica. However, the sea temperatures (∼16 °C) in the Gulf of Genoa are likely to have been somewhat lower than overland temperatures in this example, so with γ ≈ 8.8 degCkm–1 no convective cloud would be expected to develop. Dew points around 9 °C, as well as low relative humidity above the surface, further restricted any cloud development. Once again, additional forcing mechanisms were necessary for the cloud lines to form; convergence appears to have been the most important mechanism in this case. At 1200 utc, there was a centre of low pressure near Corsica, supplying positive vorticity to aid ascent. Northwest of Corsica, as a result of the partial blocking of low-level winds by the Alps, there appears to have been convergence of northwesterly winds with comparatively moist northeasterlies, resulting in the cloud line indicated by the arrow. The other cloud lines in this image are more likely to have had a baroclinic origin, associated with small changes in wet-bulb potential temperature (θw) in the boundary layer, as well as convergence.

Figure 4.

Radiosonde profile for Ajaccio, Corsica at 1200 utc on 4 December 2003. (Courtesy University of Wyoming, Department of Atmospheric Science.)

Further to the southwest, cloud-line development was more pronounced as there was a greater depth of instability; cirriform tops reveal the presence of a few cumulonimbus clouds, which the Nîmes (France) radiosonde profile (Figure 5) reveals must have deepened to reach at least 4500m. For this cloud line, γ ≈ 10.5 degCkm–1 in the boundary layer to the LCL over sea temperatures of 16 °C. Although most cumulus cloud tops would have been capped at 2000m (in very dry air), sea temperatures were sufficient to produce isolated tops to 6000m. Note that the width of this cloud line associated with convergence along a trough to the southwest was proportionately greater than was the case further to the northeast, the convective clouds having an apparent width of 5–20km, rather than the 1–8-km width of the other lines.

Figure 5.

Radiosonde profile for Nîmes, France, at 1200 utc on 4 December 2003. (Courtesy University of Wyoming, Department of Atmospheric Science.)

For the cloud line indicated by the arrow in Figure 3, the very low humidity of the air above 1850m (RH < 50%, falling to around 9% at the potential maximum cloud-top level), would have prevented most cloud tops reaching this level. Close examination of Figure 3 reveals a shadow about 4km wide. At 1230 utc the Sun's elevation was 21° at 43°N, 8°E (, indicating a cloud top at about 1600m (the level at which humidity is falling rapidly and there is inhibition to saturated ascent).3

Most narrow lines of convective cloud appear to form over the sea (where there is no orographic or strong diurnal influence and there is a ready source of moisture). In many cases, the cumulus clouds reach only modest altitude, although there may be the potential for showers to develop if moist instability reaches freezing level. Given the scarcity of observations over the sea, it is not known whether the cumulus cloud lines discussed here produced precipitation, although this may be assumed to have occurred from the cumulonimbus clouds. Even where the convection is limited to the middle troposphere, there is often the potential for precipitation.

Dendritic cumulus

Cumulus clouds, formed at the top of a well-mixed boundary layer reaching about 1000m over the seas south of Cyprus, could be seen to have developed in a tree-like form at 1115 utc on 30 July 2009 (Figure 6). Not only were there branching cloud lines, but the spreading of the convection around the dendritic cumulus lines resulted in stratocumulus cells with an appearance of leaves. This imagery has a resolution of 1km and shows that the branches were about 3km across with most ‘leaves’ somewhat smaller, although locally the stratocumulus covered 5/8 to 7/8 of the sky in zones as much as 30km across.

Figure 6.

Unusual dendritic convection over the eastern Mediterranean, observed at 500m resolution from Aqua/MODIS at 1115 utc on 30 July 2009. (Courtesy MODIS Rapid Response Team at NASA/GSFC.)

The dendritic lines are likely to have been the result of convergence, although clearly this was a complex case, the underlying cause of which is not known.

Dust ‘clouds’

Significant dust4 is commonly carried across the eastern Mediterranean either from the northern Sahara desert, or from the desert river valleys of the Middle East.

The dust is most commonly lifted to around 1500m and greatest concentrations are usually above about 600m, where the wind is often strong enough for the clay particles to be held in suspension. However, thick dust may sometimes be observed in the very lowest layers of the atmosphere. The differences were illustrated on 24/25 February 2005, when a dust plume was carried northeast across Cyprus from the Sahara. The radiosonde profile at 1100 utc on the 24th from Athalassa (not reproduced) shows that the boundary layer was capped at about 1750m and was unstable to sea-surface temperatures. At this time, the visibility near sea level was good with a thick dust layer affecting the mountain slopes. Pilot-balloon data from Paphos reveal that the wind changed from variable, mainly northeasterly, in the boundary layer early on the 24th to become northwesterly at about 6ms–1 by 1700 utc, associated with the passage of a weakening trough.

Rapid weakening of the trough kept winds light and variable in the boundary layer over the centre and east of the island with the northwesterlies at Paphos soon decreasing. By dawn on the 25th, visibility had decreased below fog limits across much of Cyprus and remained very poor at RAF Akrotiri and Paphos Airport until mid-afternoon. Figure 7 shows that the top of the dust was near the inversion layer, with the mountains visible in bright sunshine. West to northwesterly winds brought improved visibility for a time during the evening but, with light winds throughout the boundary layer, poor visibility returned to Akrotiri overnight. The light winds and stability over land by night allowed the clay particles to settle out during this phase of the event. Visibility improved gradually from the west during the 26th.

Figure 7.

Dust from the Sahara desert associated with a weak cold front (close to the northwestern edge of unstable layer cloud) over the eastern Mediterranean at 0900 utc on 25 February 2005 seen at 2km resolution by Terra/MODIS. The dense area of dust to the north-west of Cyprus, stretching as far as Rhodes, appears to have been trapped by the mountains of Turkey. Its origin may have been similar to the plume north of Mersa Matruh (Egypt), associated with a shallow mesoscale low-pressure area near the coast. Transverse ripples can be seen in thin high cloud to the west of Cyprus. (Courtesy MODIS Rapid Response Team at NASA/GSFC.)

During winter 2008/2009, plumes of desert dust were carried northeast across the Mediterranean Sea relatively frequently. The dust plume of 19 February 2009 shown in Figure 8 also appears at the southeastern corner of Figure 1. In Figure 8 the plume is highlighted using its difference in emission from that of clouds or the surface at the measured wavelengths, the result being shown as pink.5 A deep area of low pressure had formed near the base of an upper trough over Libya, with strong southwesterly surface winds carrying dust from the surface to the north and east. Early in the day the dust plume was associated with an area of rain, remnants of which can be seen over the Troodos Mountains. The rain washed out much of the dust and residents of RAF Akrotiri woke up to find a thin, but relatively even, crust of yellowish dust on their cars and bicycles.

Figure 8.

Dust storm (shown pink) over the eastern Mediterranean at 1200 utc on 19 February 2009 observed by Meteosat/SEVIRI. A particularly dense dust cloud is apparent over Israel and the Palestinian Territories. (© EUMETSAT 2009.)

Whilst major events may be clearly seen in visible imagery, some dust events are on a smaller scale. In these cases, identification is aided by emission-difference processing.

Orographic clouds over Cyprus and Gibraltar

Pennant clouds

During the evening and night of 18 February 2009, a cloud mass near the base of a marked upper trough moved across Cyprus, bringing occasional thunderstorms. The following day, as this mass of cloud diminished, moving into an area of anticyclonic curvature in the middle troposphere, a cap cloud could be seen forming along the ridge of the Troodos Mountains, streaming away to the northeast in southwesterly winds close to the level of the mountain crest. First clearly evident around 1100 utc, this cap cloud persisted for around five hours, its extent gradually diminishing through the afternoon. Figure 9 shows it near its greatest extent at 1245 utc on the 19th, covering much of the centre and east of the island, bringing cloudy, but generally dry, weather, except over some of the highest mountain tops where virga could be seen below the cloud from RAF Akrotiri.

Figure 9.

Cap cloud formed over the Troodos Mountains of Cyprus that persisted for around five hours, seen at 1245 utc on 19 February 2009 using high-resolution broad-band visible imagery from Meteosat-9/SEVIRI. Southwesterly winds across the mountain barrier are leading to the formation of cloud at the crest of the mountains (which appear dark, due to their composition from lava flows and their woodland cover). Towering cumulus clouds can be seen on the south-western slopes of the mountains, indicating a change in stability and a mesoscale cloud line is evident further west. Over southern Israel and Jordan, fine transverse banding can be seen in altocumulus clouds (which are not visible at lower resolution) in the lee of the mountains of Sinai. Medium- and high-level moisture was associated with the transition from the mid-latitude air mass to tropical air, just north of the core of the STJ. (© EUMETSAT/Met Office.)

Although this cloud persisted for only a few hours, its formation has some similarity to the extensive altocumulus and altostratus frequently seen across southern China in winter (Galvin and Walker, 2007). The radiosonde profile for Athalassa, Cyprus (Figure 10) shows that winds were east to northeasterly up to about 1560m, above which they were southwesterly. The convectively mixed moist boundary layer was capped by a modest inversion, which limited cumulus cloud tops to around 1800m where the temperature was close to freezing point. Above this level, the southwesterlies were also moist, so that modest ascent across the Troodos Mountains resulted in extensive altocumulus to above 3500m. The crest of these mountains reaches 1200m or more for 50km of their length, approximately corresponding to the length of the cloud edge shown in Figure 9. It is likely that some of the moisture over the Troodos Mountains was entrained by convection and carried by southerlies at 1500m over the southwest of the island.

Figure 10.

Radiosonde profile for Athalassa, Cyprus, at 1200 utc on 19 February 2009, showing an unstable, moist easterly boundary-layer flow capped by a warmer relatively moist and potentially unstable southwesterly flow within which an altocumulus cap cloud formed over the Troodos Mountains. (Courtesy of University of Wyoming.)

The need for a relatively moist boundary layer (locally moistened by ascent) and sufficient moisture to allow condensation in the southwesterlies, as well as a ‘cap’ at the top of the boundary layer, is evident. Figure 9 also shows ‘cap’ clouds over the Jebel el Ansariye, Jebel Libam and Jebel esh Sharqi of Syria and Lebanon, showing that sufficient moisture was available in the flow over these mountains to produce similar cloud forms. Although forming a cap over a (modest) mountaintop, these clouds differ somewhat from the Levanter pennant clouds seen on the Rock of Gibraltar.

Levanter clouds are a component of a synoptic-scale low-level east to northeasterly flow through the Strait of Gibraltar. The Levanter is moist, having travelled across the western Mediterranean (Bendall, 1982). As the boundary layer is capped around 1000m, the flow accelerates between converging high ground north and south of the Strait. Above these low-level easterlies, the airstream is usually a dry anticyclonic flow, which may be from the west or southwest.

Most of the cloud that forms in the converging flow is from shallow convection that spreads out beneath the inversion at the top of the shallow moist zone, bringing cloudy skies, in particular along the northeast coast of Morocco, as was the case on 12 March 2009 (Figure 11). On the northern side of the Strait, the flow may be drier, in particular if winds are from the northeast. Cloud characteristically forms over the crest of the Rock of Gibraltar, which reaches almost 400m, forming a pennant as air is forced to rise in the generally stable flow. As the wind is easterly, the pennant cloud extends west from the cap of the rock. In this image, thin ribbons of cumulus clouds can be seen in northeasterly winds over the western Mediterranean, gradually developing and spreading out into stratocumulus, eventually forming an overcast sheet close to and across the northeastern coast of Morocco, into the Strait of Gibraltar, where easterly winds were strong. A drier flow can be seen to the north, but the 250m resolution of the inset image allows a characteristic pennant cloud to be seen (inset, arrowed) over the Rock of Gibraltar.

Figure 11.

Levanter clouds in the western Mediterranean seen at 1km resolution by Terra-MODIS at 1400 utc on 12 March 2009. Thin ribbons of cumulus clouds can be seen in northeasterly winds, gradually developing and spreading out into stratocumulus, eventually forming an overcast sheet close to and across the northeastern coast of Morocco, into the Strait of Gibraltar, where easterly winds were strong. A drier flow can be seen to the north, but the 250m resolution of the inset image allows a characteristic pennant cloud to be seen (inset, arrowed) over the Rock of Gibraltar. (Courtesy of MODIS Rapid Response Team at NASA/GSFC.)

Orographic blocking

Early on 1 January 2010, the effect of the Troodos Mountains on cloud was reversed; the mountains could be seen emerging from an area of cumulus and stratocumulus across the eastern Mediterranean (Figure 12). The boundary layer was very moist across Cyprus with dew points of 13–14 °C, so although air temperatures were near those of the sea surface (∼18 °C), there was extensive convection with cumulus clouds forming with a base near 600m and spreading out into stratocumulus under an inversion near 1300m. The stratocumulus edge can be seen at this level on the slopes of the Troodos Mountains, as well as on the Jebel Libam of Lebanon.

Figure 12.

High-resolution broad-band visible image from Meteosat-9/SEVIRI showing stratocumulus (with cumulus in the west) over the eastern Mediterranean pierced by the Troodos Mountains of Cyprus, the Jebel al Ansariye of Syria and the Jebel Libam of Lebanon at 0530 utc on 1 January 2010. A rope cloud can be seen off the coast of Turkey (beneath turbulent banded cirriform cloud), following the development of a mountain wind from the western Toros Daǧları. (© EUMETSAT/Met Office.)

Further east, where the layer below the inversion was somewhat deeper, wave clouds can be seen in the lee of the Jebel al Ansariye, as well as over the high ground of Israel, Palestine, and Jordan. In this area, early in the day, the near-surface layer was stable, aiding the development of wave clouds. However, wave development (in dry air) was expected over all the moun-tains of the region, even where the inversion was below the level of the mountain tops.

Low cloud

As increasingly warm moist air spreads across the eastern Mediterranean in spring and summer, stratus is seen quite frequently. Figure 13 shows extensive stratus across much of the eastern Mediterranean at 0330 utc on 13 June 2009 – about the time when the area covered by stratus reaches a peak in early summer as air with dew points of 21–22 °C spreads west over sea temperatures of 22–24 °C. The thickest stratus can be seen between Cape Kiti and Cape Gata along the south coast of Cyprus, where the water-surface was anomalously cool.

Figure 13.

High-resolution broad-band visible image from Meteosat-9/SEVIRI showing extensive banks of stratus across much of the eastern Mediterranean at 0330 utc on 13 June 2009. Fog, locally lifting into low cloud, can be seen over the Mesaoria of Cyprus, the Nile delta and Jordan valley. Notable cloud lines can be seen where land breezes have developed from the coasts of Israel, Lebanon and southwest Cyprus. (© EUMETSAT/Met Office.)

Fine structure can be seen in the cloud area in this high-resolution image. Stratus often appears to exhibit a cellular or banded structure in the eastern Mediterranean similar to that often seen in stratocumulus, suggesting overturning within the saturated layer which, at this time, reached only about 350m, but was moving over comparatively warm water. Later in the summer, the boundary layer is often deeper, so this cloud form is more readily identified as stratocumulus, formed as air moves from cooler waters across increasingly warm seas. However, typical wind speeds rarely generate a mixed layer more than about 700m deep: in these cases the cloud base is observed between 350 and 550m over sea-surface temperatures of around 27 °C.

Lines of stratus are also evident at high resolution close to the coasts of Israel, Lebanon, Syria, and southwest Cyprus. All along these coasts, land breezes were observed – originating in the many coastal mountain ranges of the region. Typical surface wind speeds were between 1ms–1 and 3ms–1 and the drainage flows can be seen to extend to between 10 and 25km offshore at this time when near their maximum extent. The development of offshore winds led to the clearance of low cloud from most coastal areas, although it remained rather hazy. The air was so humid that fog had formed earlier in the night over the Mesaoria of the east of Cyprus as temperatures fell to around 19°C.6 Although it had lifted to form stratus by the time of this image, the cloud top shows fewer structural features than the stratus offshore. This area is prone to fog in moist airstreams, whenever easterly winds develop late in the day and overnight. Fog can also be seen over the Nile delta and along the valley of the River Jordan.

In Part 2 of this paper, turbulence along the sub-tropical jet stream, wave clouds and an image of snow cover will be presented.


Rob Dempster and Paul Arbuckle (Met Office) compiled a wealth of useful information about the dust event of 24–26 February 2005. The comments of two anonymous reviewers were very helpful.

  • 1

    The proximity of these clouds and the typical width of the lines may allow these clouds to occupy more than 4/8 of the sky when observed from the ground. (This is a characteristic of convection associated with non-random formation mechanisms, including those of mountains and coasts.) Assuming a typical cumulus cloud base at 800m and a typical width of 3.5km, a cloud line occupies about 130° of the 180° arc of a surface observer's view (i.e. 6/8 cloud cover).

  • 2

    The resolution of the HRV channel on Meteosat-9 (used in Figures 2, 9, 12 and 13) is about 1.5km over Cyprus.

  • 3

    Note that there is a probable high error in this figure, due to a lack of clarity in the shadow and the resolution of the image.

  • 4

    The general term ‘dust’ refers to crystals of clay, defined by the maximum length of their longest axis: 2μm. Larger particles may be carried for short distances by strong winds (often bouncing along near the surface), but crystals of clay are flat and may be carried long distances by winds of about 8ms–1 or more.

  • 5

    Wavelengths of 12.0μm, 10.8μm, and 8.7μm are used to highlight dust (as well as contrails and thin clouds), which is given the colour pink in false-colour imagery. Dust lying on the surface is also highlighted.

  • 6

    The fog point of 19°C was some 6 degC higher than that suggested by the 1100 utc radiosonde profile from Athalassa – near the head of the Mesaoria.