Climatology of Cyclogenesis Mechanisms in the Mediterranean

Isabel F. Trigo Climatic Research Unit, University of East Anglia, Norwich, United Kingdom

Search for other papers by Isabel F. Trigo in
Current site
Google Scholar
PubMed
Close
,
Grant R. Bigg School of Environmental Sciences, University of East Anglia, Norwich, United Kingdom

Search for other papers by Grant R. Bigg in
Current site
Google Scholar
PubMed
Close
, and
Trevor D. Davies Climatic Research Unit, University of East Anglia, Norwich, United Kingdom

Search for other papers by Trevor D. Davies in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A general climatology of the main mechanisms involved in Mediterranean cyclogenesis is presented. A diagnostic study of both composite means and case studies is performed to analyze processes occurring in different seasons, and in different cyclogenetic regions within the same season. It is shown that cyclones that developed over the three most active areas in winter—the Gulf of Genoa, the Aegean Sea, and the Black Sea—are essentially subsynoptic lows, triggered by the major North Atlantic synoptic systems being affected by local orography and/or low-level baroclinicity over the northern Mediterranean coast. It is also suggested that cyclones in two, or all three, of these regions often occur consecutively, linked to the same synoptic system. In spring and summer, thermally induced lows become progressively more important, despite the existence of other factors, such as the Atlas Mountains contributing to lee cyclogenesis in northern Africa, or the extension of the Asian monsoon into the eastern part of the Mediterranean. As a consequence, the behavior of Mediterranean cyclones becomes modulated by the diurnal forcing; the triggering and mature stages are mostly reached by late afternoon or early nighttime, while cyclolysis tends to occur in early morning.

 Current affiliation: Departamento de Física, Faculdade de Ciências de Lisboa, Lisboa, Portugal.

Corresponding author address: Isabel Franco Trigo, Departamento de Física, Faculdade de Ciências de Lisboa, Campo Grande, Ed. C8, Piso 6, 1749-016 Lisboa, Portugal. Email: iftrigo@fc.ul.pt

Abstract

A general climatology of the main mechanisms involved in Mediterranean cyclogenesis is presented. A diagnostic study of both composite means and case studies is performed to analyze processes occurring in different seasons, and in different cyclogenetic regions within the same season. It is shown that cyclones that developed over the three most active areas in winter—the Gulf of Genoa, the Aegean Sea, and the Black Sea—are essentially subsynoptic lows, triggered by the major North Atlantic synoptic systems being affected by local orography and/or low-level baroclinicity over the northern Mediterranean coast. It is also suggested that cyclones in two, or all three, of these regions often occur consecutively, linked to the same synoptic system. In spring and summer, thermally induced lows become progressively more important, despite the existence of other factors, such as the Atlas Mountains contributing to lee cyclogenesis in northern Africa, or the extension of the Asian monsoon into the eastern part of the Mediterranean. As a consequence, the behavior of Mediterranean cyclones becomes modulated by the diurnal forcing; the triggering and mature stages are mostly reached by late afternoon or early nighttime, while cyclolysis tends to occur in early morning.

 Current affiliation: Departamento de Física, Faculdade de Ciências de Lisboa, Lisboa, Portugal.

Corresponding author address: Isabel Franco Trigo, Departamento de Física, Faculdade de Ciências de Lisboa, Campo Grande, Ed. C8, Piso 6, 1749-016 Lisboa, Portugal. Email: iftrigo@fc.ul.pt

1. Introduction

The spatial distributions and general cyclogenesis mechanisms in the Northern Hemisphere have been analyzed by Petterssen (1956), in his climatology of surface cyclones. Over the Mediterranean region, he identified two main centers of activity in winter, over the western and eastern basins, respectively; and one main center over Iberia in the summer. However, the Mediterranean Sea and its immediate environs experience a high spatial variability of weather conditions, leading to large arid areas (e.g., Thornes 1998, 2–4) and, yet, still accommodating the greatest annual precipitations totals in Europe, in the Dinaric Alps [see Fig. 1 for location; Radinovic (1987)]. Such high variability suggests that a rather more detailed study of Mediterranean cyclone distributions and genesis mechanisms than that of Petterssen (1956) is appropriate now that adequate data are available. Certainly there is growing interest in these particular weather systems, since the number of studies of Mediterranean cyclone formation has increased in recent years, and—in recognition of the high spatial variability over the region—most have focused on case studies occurring in specific areas, such as in the Sahara (e.g., Thorncroft and Flocas 1997), in the Aegean Sea (e.g., Flocas and Karacostas 1996), and in the lee of the Alps (e.g., Buzzi and Tibaldi 1978; Gomis et al. 1990). However, there has been no such study for the Mediterranean Basin (Fig. 1). This paper addresses that deficiency through the study of the main cyclogenesis processes using a detailed surface cyclone database (Trigo et al. 1999, hereafter TDB), complemented by upper-level information, thus allowing an examination of the interactions between lows developing in different parts of the basin and large-scale features.

As suggested by TDB, the traditional four meteorological seasons (December–February, DJF; March–May, MAM; July–August, JJA; September–November, SON) do not fit well the patterns of cyclone occurrence in the Mediterranean, mainly due to its high intermonthly variability. The annual distributions of cyclone events per 105 km2 (i.e., the cyclones' spatial densities) in the most active cyclogenetic regions (defined in TDB, and shown in Table 1) between 1987 and 1996 are represented in Figs. 2a (western Mediterranean) and 2b (eastern Mediterranean); only cyclone events lasting at least 12 h have been considered. It should be noted that the values shown in Table 1 represent the total number of events detected in each area and, hence, differ from the values per unit area represented in Fig. 2 for the same months. As expected in an area where cyclogenesis is highly determined by geographical features, the lower curve in both diagrams, which corresponds to the whole basin, reflects a much lower number of cyclone events than any of the other curves. The relatively small annual cycle over the whole basin reflects the completely different seasonalities in the separate cyclogenetic regions. For example, there is a peak of activity in spring over the Sahara, in summer over Iberia, and a smoother annual cycle over the Black Sea. The western Mediterranean has a much more marked seasonality than the eastern basin. Frequency of occurrence should not be simply equated with strength. An example is over the Gulf of Genoa where, although cyclones are a constant feature over the whole year, they are generally deeper, and have more severe weather in winter than during the summer, when they are, in fact, more frequent. Hence, in order to overcome the problems presented by these differing seasonalities in the various regions, the analysis in the present work is performed for representative winter, spring and summer months, namely January, April, and August (see TDB, section 3, for argument). Since October is characterized by a rather sudden transition from summer to winter conditions, which are established by November (TDB; HMSO 1962), the traditional autumn period is considered to be part of the winter season.

The data used in this work, including the database of Mediterranean cyclones, and the diagnostic techniques are briefly described in section 2. The diurnal cycle of cyclogenesis/cyclolysis is analyzed in section 3, where we will show that the development of Mediterranean lows becomes progressively more dependent on the time of day throughout spring and summer. This is a feature that clearly separates winter from spring/summer cyclogenesis processes.

To achieve our goal of a global description of cyclogenesis mechanisms in the Mediterranean, we will then perform a composite analysis for the most active regions in January (section 4), and in April and August (section 5), respectively. For each of these two cases, the composite results will be complemented by case studies, in order to confirm the physical reality underlying the compositing procedure and its interpretation.

2. Data and methodology

The present work is based on the analysis of a subset of an 18-yr (1979–96) climatology of Mediterranean cyclones derived from European Centre for Medium-Range Weather Forecasts (ECMWF) Re-Analyses [1.125° × 1.125° horizontal and 6-hourly temporal resolution; Gibson et al. (1999)]. Cyclones were detected by identifying 1000-hPa height minima (thresholds for corresponding sea level pressure and pressure gradient were used). When cyclogenesis occurs over elevated terrain, as in the region to the south of the Atlas Mountains (Fig. 1), the 1000-hPa level may be under the ground surface. In such cases the ECMWF postprocessing algorithm extrapolates geopotential height below the surface based on the assumption of hydrostatic balance, where the underground temperature is estimated using the standard atmosphere lapse rate (6.5 × 10−3 K m−1); discussion of the data assimilation and quality monitoring of ECMWF Re-Analyses may be found in Uppala (1997) and Kållberg (1997). Geopotential height at 1000 hPa might then be underestimated over regions with high surface elevation, in late spring and summer, and overestimated during winter and overnight with clear sky and calm conditions. However, diagnosis of 10-m wind and 10-m vorticity [obtained from National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalyses] tend to confirm most of our results.

The cyclone tracking was based on a nearest neighbor search in the previous field, within an area defined by imposing thresholds to the maximum cyclone velocity. If no cyclone was found within that area, cyclogenesis was assumed to have occurred (see the appendix for further details; a full description of the detecting and tracking techniques may be found in TDB). The resolution of the ECMWF dataset proved to be adequate for a climatology of Mediterranean systems, including the detection of the initial stages of development, often associated with subsynoptic scales (e.g., Buzzi and Tibaldi 1978). The spatial resolution of the data strongly limits the detection of systems with radii of less than 150–200 km (Fig. 3 in TDB).

The composite analysis and case studies presented here are performed for the 1987–96 period, for which upper-level ECMWF reanalyses were also available. These include geopotential height, temperature, relative vorticity, and vertical motion, at 1000, 850, 700, 500, 400, 300, and 200 hPa. The data cover the area from 24.75° to 50.625°N, and 15.75°W to 45°E (Fig. 1).

The objective of the composite analysis is to characterize the main conditions that favor (a) the triggering and (b) deepest phases of lows formed over the relatively most active areas at particular times of the year. The selection of cases for compositing is based on one of the following criteria, representative of incipient and mature stages, respectively, of cyclone development:

  1. 1stP time—situations (within the Mediterranean cyclone database described above) corresponding to the first detection of a surface cyclone; only lows generated within one of the cyclogenesis areas defined in Table 1, and with a minimum life cycle of 12 h, are considered; and

  2. PminP time—situations corresponding to the minimum central pressures reached by any of the lows described above.

The total numbers of composite cases per region are represented in Table 1.

The variables selected for the composite analysis include geopotential height and potential temperature, which will provide useful information about characteristic depths of the lows, as well as of the main (larger scale) features that are most frequently associated with them. Meridional vertical cross sections (approximately perpendicular to the northern and southern Mediterranean coastal lines, and to the main mountain barriers in the region; Fig. 1) of relative vorticity and static stability will give insight into the pertinent dynamical processes involved.

3. Seasonality of the cyclogenesis processes

The peak of the winter season in the Mediterranean occurs around January, when the midlatitude cyclone belt has usually reached its southernmost position (e.g., HMSO 1962). By then, synoptic lows and troughs, steered along the meanders of the polar front jet, are more likely to directly influence Mediterranean cyclogenesis and weather, with the highest impact over the northern coast. High values of low-level baroclinicity may also be observed during the winter, mainly along the northern Mediterranean coastline, due to the strong sea–land temperature contrast (e.g., HMSO 1962). Although all these conditions favor the development of cyclones, the highest number of cases detected in TDB did not occur in the winter season. It is in winter, however, that the most intense, deepest, and more persistent cyclones are observed, frequently associated with wet and/or severe weather conditions (Radinovic 1987).

The increased importance of local conditions, in particular local instabilities and convection, in determining the development and damping of Mediterranean cyclones, constitutes the most prominent feature that marks the transition between winter on one hand, and spring and summer months on the other. As a consequence, cyclogenesis and cyclolysis are characterized by progressively more pronounced diurnal cycles, as indicated in Fig. 3, which shows the percentage of first and last detections of cyclones at each observation time. The distribution of cyclones' first/last detections is fairly uniform through the day in December and January (Figs. 3l and 3a). However, from February onward, there is an increased tendency for events to be first (last) detected at 1800 UTC (0600 UTC). By August (Fig. 3h), the diurnal asymmetry reaches its peak, with relative frequencies of nearly 50% concentrated at the observation times mentioned.

Using data with coarser temporal (12-hourly) and spatial (2.5° × 2.5° grid) resolution, as well as different methods to detect and track Mediterranean lows, Alpert et al. (1990) obtained results similar to those described above. Increasing the 12-h threshold for cyclone duration also has little impact on the diurnal cycles described above. Such consistency reinforces the hypothesis that the diurnal cycles are not a spurious result of the technique or data used. This particular characteristic of Mediterranean cyclogenesis, which could be used as a criterion to determine the beginning of spring and the end of summer in the Mediterranean, will be addressed again in section 5.

To a great extent the location of the most cyclogenetic areas in the Mediterranean is determined by the local geography. As shown in the next section, this is particularly evident during winter, when the Alps are one of the main contributors to the triggering of Genoa cyclones. Also, warm surface potential temperature anomalies over the Aegean and Black Seas, which are partially or totally surrounded by colder land, form cyclonic potential vorticity anomalies concentrated at the surface (Bretherton 1966; Hoskins et al. 1985). These seas become two of the most active areas in the Mediterranean winter. The spring/summer processes will be further analyzed in section 5.

4. January

a. Composite analysis

The Gulf of Genoa is by far the most active area in the whole Mediterranean Basin between November and February (TDB). The intense weather conditions that are often observed at this time have attracted attention to Alpine cyclogenesis (e.g., Dell'Osso and Tibaldi 1982; Dell'Osso and Radinovic 1984; Tafferner 1990). Although the Aegean and Black Seas are also regarded as two of the most cyclogenetic areas in the eastern Mediterranean winter (e.g., HMSO 1962; Radinovic 1987; Alpert et al. 1990; TDB), there are fewer studies of the main cyclogenesis mechanisms in these areas. There are, however, some studies of severe cases affecting the Black Sea and the Middle East region, which tend to occur between November and March (e.g., Lee et al. 1988; Ziv and Alpert 1995; Tayanç et al. 1998). It should be noted that where reference is made below to factors observed at 1000 hPa over land these are also found in surface fields. Below, we examine the Gulf of Genoa, the Aegean Sea, and the Black Sea regions in turn.

1) Gulf of Genoa

The anomalies of geopotential height, relative to the January climatology, at 1000 and 500 hPa, are shown in Figs. 4a,b for Gulf of Genoa 1stP and PminP times, respectively. The statistical significance of each composite was evaluated using two-tailed Student's t-tests (e.g., Wilks 1995), so that regions where the composite mean is statistically different from the January mean at the 99% confidence level area are shaded.

Geopotential anomaly fields of the 1stP time composite (Fig. 4a) clearly indicate that the first stages of Genoa cyclogenesis are associated with a large negative height anomaly over central/northern Europe. This well-defined depression has the characteristic horizontal and vertical scales of a synoptic midlatitude perturbation (e.g., Holton 1992); its horizontal scale is of the order of 1000 km and, although not shown, the upper trough extends at least up to the 200-hPa level. Southward of this depression center, it is possible to detect the development of a clear low-level secondary trough at 1000 hPa centered over northern Italy, where the air is forced to flow around and/or over the mountains. This is the picture accepted in most studies of Genoa cyclogenesis, where cyclones are usually described as orographically induced by the Alps, when the eastward movement of an upper-level trough is blocked by the mountains (e.g., Egger 1988; Pichler et al. 1990; Aebischer and Schar 1998). Several numerical experiments have been conducted that show that much of the cyclogenesis over the Gulf of Genoa would be very weak, or even absent without the forcing effect of the Alps (Zupanski and McGinley 1989; Tafferner 1990; Alpert et al. 1996, among others).

As the main synoptic system continues to shift eastward, Genoa cyclones mature. The PminP time geopotential anomalies (Fig. 4b) show that the 1000-hPa minimum associated with surface cyclogenesis (Fig. 4a) is now elongated along the Italian peninsula. This is in accordance with the two main routes for Genoa cyclones found by TDB, which cross either the northern part of Italy, or move southeastward, along the Adriatic coast. The vertical profiles of relative vorticity at 10° and 15°E for the PminP time (Fig. 4c) clearly show the formation of a cyclonic cell in the lee side of the Alps, associated with the vertical separation of the isentropes in the lee subsidence. During the mature stage, the upper and lower cells are already coupled (as described by Buzzi and Tibaldi 1978).

In order to look at the larger-scale picture of 1000-hPa height field over Europe, composites for Genoa 1stP time were also computed using NCEP–NCAR reanalyses data (Kalnay et al. 1996) available 6-hourly, on a 2.5° × 2.5° grid, and over a larger domain than the ECMWF data. The same 1987–96 period was considered. The full composite field and the anomalies with respect to the NCEP–NCAR 1987–96 climatology are shown in Figs. 5a and 5b, respectively. Although there is considerably less definition over the Genoa area due to the coarser spatial resolution of this dataset, it is now possible to identify Genoa cyclones' parent low, associated with the area with negative height anomalies, statistically significant over central and northern Europe.

2) Aegean Sea

The Alps are not the only topographic barrier perturbing the main synoptic flow. It is clear in Fig. 4 that there are other secondary troughs, particularly over the Balkans, that may contribute to further cyclogenesis events over that region. In fact, the 1stP time composite anomalies of geopotential height obtained for lows forming over the Aegean Sea (Fig. 6a) show several features of note:

  1. The 1000 hPa anomaly field has a two-core structure, with a height minimum over the Aegean and a secondary one over the Ionian Sea (∼38°N, 17°E).

  2. The influence of Mediterranean orography, in particular of the Dinaric Alps along the northeastern Adriatic coast, is quite clear, inducing secondary troughs at 1000 hPa.

  3. At 500 hPa, the geopotential minimum is centered over eastern/northern Europe, with a trough extending to the west of the Aegean Basin.

The importance of upper-level troughs in the development of Mediterranean and, in particular of Aegean Sea, cyclogenesis has already been mentioned in previous studies (Jacobeit 1987; Flocas and Karacostas 1996). The Alps seem to be the primary factor perturbing mobile troughs crossing central Europe, and influencing the triggering of cyclones over the most active region during the winter season. In fact, the Ionian minimum mentioned above may be considered a residual of Alpine cyclogenesis, either as a mature Genoa cyclone that has followed a southeastward path along the Adriatic coast, or as a low from southern Italy (also an area with relatively high winter cyclogenesis, usually influenced by Italian orography). This is emphasized further by the lagged composites shown in Fig. 7. The development of the 1000-hPa minimum over the Aegean Sea (Fig. 7, lower panels) is preceded by the movement of the 1000-hPa system, over central-southern Italy in the 12-h-lag composite, and of the 500-hPa trough toward the Balkans.

Most authors state that topography plays a minor role in surface cyclogenesis over the Aegean Basin (e.g., Prezerakos and Flocas 1996). Such cyclogenesis tends to occur when an upper trough moves over a region of weak low-level stability. This is confirmed by the vertical profiles of relative vorticity and static stability (σ = −∂θ/∂p) anomaly fields (Fig. 6b). Since the composite and climatological means of σ are always assumed positive, negative (positive) σ anomalies correspond to weak (high) stability associated with vertical separation (compression) of the isentropes. Figure 6b shows the low-level cyclonic cell over the Aegean Sea (∼37.5°N) developing within a layer of low static stability induced by the relatively warm sea and, at least partially, by separation of the isentropes in the lee of the Balkans orography. However, the influence of the Balkans orography on the flow, which is likely to be associated with the formation of the weak anticyclonic cell within the positive σ anomalies (42°–45°N in Fig. 6b), is much less than that observed in the Alpine region.

3) Black Sea

As Table 1 indicates, cyclogenesis is more frequent in the Black Sea region than it is in the Aegean, and, despite the proximity of these two basins, they are frequently treated as separate cyclogenetic regions [as in, e.g., Alpert et al. (1990) and Flocas and Karacostas (1996)]. However, there is a remarkable similarity between composites of geopotential height obtained for 1stP time for the Black Sea (Fig. 8) and for the Aegean (Fig. 6a). The 1000-hPa height shows an elongated minimum core distributed between the two basins, beneath an upper trough extending over northern and eastern Europe. As in the previous cases, surface cyclogenesis seems to be associated with an upper trough to the west of the region, advecting vorticity toward a relatively warm sea.

4) Basin-scale view

The composites of 850-hPa potential temperature for 1stP time (Fig. 9) confirm that winter Mediterranean cyclones develop within enhanced low-level baroclinicity. In the case of Genoa lows, this seems to be further maintained by cold advection on the western flank of the larger synoptic system toward the relatively warm temperatures over the Mediterranean Sea (Fig. 9a). Most numerical experiments indicate that this is not a triggering factor, since Genoa cyclones are mainly mountain induced. It may, however, be crucial for determining the maximum intensity reached by the lows (e.g., McGinley and Zupanski 1990).

The thermal anomalies for Black Sea lows and, particularly, for Aegean Sea lows exhibit a pronounced baroclinic region, with its main axis passing through these relatively warm seas, between the Balkans and the Anatolian Plateau (∼41°N, 28°E in Fig. 9b). The southwesterly direction of this baroclinic zone, which extends throughout the whole troposphere (not shown), determines the main path of Aegean lows toward the Black Sea, and of Black Sea lows out of the Mediterranean region.

The presence of a mobile baroclinic zone implicit in Fig. 9 for the three regions of study is clearer in Fig. 10, where composites of 850-hPa potential temperature are plotted for 12 h before and after 1stP time. Here anomaly fields are presented instead of the full fields in order to remove the underlying pattern of land–sea-generated baroclinicity. The areas where the gradient of potential temperature anomalies is greater also correspond to high potential temperature gradients. In all three regions of study it is possible to detect the movement and intensification of such gradients; namely;

  1. a relatively weak temperature gradient to the northwest of the Alpine region 12 h before Gulf of Genoa 1stP time, which intensifies and moves in the eastward direction 12 h after;

  2. in the Aegean composites, the area with stronger baroclinicity intensifies and moves from the Adriatic coast (12 h before 1stP time; Fig. 10a) to the Aegean and Black Sea areas 12 h after (Fig. 10b); and

  3. in the Black Sea composites, the negative temperature anomalies over eastern Europe are stronger 12 h after; it is also then that SW–NE baroclinic zone over the Black Sea (also in Fig. 9c) is most intense, conditioning the Black Sea winter cyclone tracks over that area.

The composite analysis shows the link between Mediterranean cyclogenesis in winter and larger synoptic depressions. The Alpine orography plays a crucial role in triggering Genoa cyclones, and in the deepening of the respective upper troughs. Moreover, as Lefevre and Nielsen-Gammon (1995) have shown, the region to the lee of the Alps is a major source of mobile upper troughs. Numerous studies have established northwesterly flow, such as that associated with the upper trough depicted in Figs. 6a and 8, as a precursor of surface cyclogenesis (e.g., Sanders 1986; Lefevre and Nielsen-Gammon 1995; Lackmann et al. 1997, 1999). It follows that cyclogenesis over the Aegean and/or Black Seas occurs when these upper disturbances move over the relatively warm water basins.

As shown below, cyclogenesis over the three study regions of winter cyclone development—all on the northern Mediterranean coast—may occur consecutively, linked to the same large-scale synoptic system. The path of this larger system (and the position of the associated upper trough) seems to be the main factor determining whether or not each cyclogenesis event occurs.

b. Case study

This section focuses on an analysis of the weather systems affecting the Mediterranean during the period 6–9 January 1987, as an example of the interdependence between cyclogenesis events occurring in different areas during the Mediterranean winter. Over this 4-day period a synoptic depression and its associated upper trough, initially located over northern Europe, interacted with local topography and low-level baroclinicity, triggering secondary lows over the three cyclogenetic regions studied above.

The situation near the surface at 0000 UTC 6 January 1987 (Fig. 11a) is dominated by a synoptic depression, with its center north of the domain of study. A secondary trough is forming over the Gulf of Genoa (center A in Fig. 11a). At 500 hPa (Z500; Fig. 11b) there is a pronounced trough extending over the whole eastern part of the Mediterranean down to Egypt. Over the Alpine region, however, the 500-hPa contours are bunched, indicating a strong northwesterly flow over this mountain range. The Genoa low reaches its minimum central pressure by 0000 UTC 7 January (A in Fig. 11c). The movement of this Genoa low seems to be initially determined by the northwesterly upper flow, following the well-known path along the Italian peninsula. It remains stationary when the Ionian Sea is reached (A at 1200 UTC 8 January; Fig. 11e), as if blocked by the Balkans. Instead, another secondary system starts to develop over the Aegean Sea (B in Fig. 11c), following a sequence very similar to that suggested by the previous composite analysis.

The result of the interaction between the larger synoptic system and the baroclinic region over the Aegean Sea is well depicted by the surface charts throughout the rest of the period, until 9 January. The strong 500-hPa northwesterly flow over the Alps, observed from 0000 UTC 6 January, develops into upper trough C (Fig. 11f), in a similar way to that reported in other case studies (e.g., Lackmann et al. 1997, 1999). In this case, the Alpine orography may have acted to further deepen the trough, which, in turn, may be crucial to the maintenance and development of the newly formed Aegean trough B, as it is a major source of upper-level positive advection of vorticity toward the Aegean Sea.

Again in accordance with the composite analysis for the Aegean Sea and Black Sea, a further northeastward extension emerges from the Aegean low (B) in the direction of the Black Sea, by 1200 UTC 8 January (Fig. 11e). An independent Black Sea low (D) is observed 12 h later. The 500-hPa height perturbation (C) mentioned in the paragraph above seems also to play an important role in the development of this newly formed cyclone, transporting positive vorticity into the region.

To obtain a better insight into the dynamics associated with the formation of the Genoa low, the 6 January 1987 (0000 UTC) vertical cross sections of relative vorticity and potential temperature (dotted lines), for 10° and 15°E, are plotted in Fig. 12a. The bold lines show the band of potential vorticity between 1.5 and 3 PVU [1 PVU = 10−6 m2 s−2 K kg−1; Hoskins et al. (1985)] representative of the so-called dynamic tropopause (e.g., Morgan and Nielsen-Gammon 1998). Over the Alps (45°–48°N, both at 10° and 15°E) there is a region of enhanced lower- and midtropospheric baroclinicity, suggested by the pronounced isentropic slopes. There the tropopause dips down to nearly 500 hPa, leading to an intrusion of stratospheric air into the upper troposphere. This type of tropopause fold, often associated with upper-level jet streams and upper-trough development (Shapiro et al. 1987; Davies and Schuepbach 1994; Schuepbach et al. 1999; Lackmann et al. 1999), is in accordance with the tight 500-hPa contours mentioned above.

The effect of topography on the relative vorticity field is well depicted both in the 10° and 15°E cross sections (Fig. 12a); an anticyclonic cell is formed as the isentropes are compressed over the mountains, while a cyclonic circulation develops in the lee side. The mature and incipient stages of the Genoa and Aegean lows are shown by the 17° and 25°E cross sections, respectively (Fig. 12b). The low-level circulation, formed in the lee of the Alps, is now linked with the upper vortex (∼36°N, 17°E), in a similar way to that obtained in the composite analysis (Fig. 4b). On the other side of the Balkans, a cyclonic circulation (around 39°N, 25°E) is being formed over a baroclinic region, indicated by the sloping isentropes, as in the previous case. The fold of the tropopause seems smoother, however, and the effect of the topography is much less pronounced than in the Genoa cyclogenesis.

The local dynamical situation observed during the mature phase of the Aegean low is depicted by the 27°E cross section (Fig. 12c) at 0000 UTC 9 January. The cyclonic cell has developed deep into the troposphere and seems to be interacting with stratospheric air, which has dipped down to at least 400 hPa. The surface low pressure center D (Fig. 11g), generated as an extension of the Aegean system toward the Black Sea, corresponds to the cyclonic cell observed at ∼43°N, 37°E (Fig. 12c). By then, the Black Sea circulation has developed up to around 400 hPa, embedded in strong baroclinicity. As the whole Aegean system moves farther toward the Black Sea in the subsequent hours (not shown) the circulation deepens until it eventually moves out of the Mediterranean region by 10 January.

5. April and August

Although spring and summer cyclogenesis events also tend to cluster around a few active regions within the Mediterranean, these are more scattered throughout the whole basin than in winter. The local thermal profiles tend to become dominant toward summer, inducing pronounced diurnal forcings in the life cycles of the lows. The frequencies of first (1stP) and minimum central pressure (PminP) detections throughout the day are plotted in Fig. 13 for April and in Fig. 14 for August, for each of the relevant cyclogenetic areas. The correspondence between the UTC observation times, the local time, and the real period when cyclogenesis (1stP) and PminP may occur is shown in Table 2. Examination of Figs. 13 and 14 reveals that the response to the diurnal forcing varies considerably between cyclogenetic regions, particularly in April. While in spring only Saharan lows exhibit a very pronounced diurnal cycle, with nearly 90% of these lows reaching their minimum pressure at 1800 UTC (Fig. 13a), summer lows forming over the Sahara, Middle East, Iberia, and to a lesser extent the Gulf of Genoa, all tend to follow a similar pattern (Fig. 14). The Mediterranean spring may thus be characterized by a wider range of cyclogenesis mechanisms than summer, when thermal forcing seems to prevail.

Black Sea lows exhibit a variation to the diurnal pattern mentioned above. In spring the diurnal cycle is nearly absent (Fig. 13d), and in summer these lows are predominately triggered at 1800 UTC, but are likely to reach the most mature stage either at 1800 or 0000 UTC (Fig. 14e). In this respect, the most pronounced differences are between winter/spring events and summer events for Black Sea cyclones, in contrast to other cyclogenesis areas in the Mediterranean.

The strong meridional temperature gradient observed along the northern Mediterranean coast during winter shifts down to the southern coast from spring onward. By then, depressions forming over northern Africa often travel eastward along the coastal gradient (Alpert and Ziv 1989; TDB). The strong diurnal cycles of Saharan depressions, and also of Iberian and Middle East lows (Figs. 13 and 14), are however essentially associated with stationary thermal lows, mainly driven by the strong diurnal fluctuations of inland surface temperature.

The relatively smooth diurnal cycles for the Genoa region, remarked on above, reflect the importance of lee cyclogenesis caused by the Alps continuing during spring and summer, since the diurnal fluctuations of surface temperature in the central Mediterranean are also relatively high in April and August (not shown).

a. Composite analysis

The composite analysis presented here will focus on two of the most active areas: one in spring (Sahara) and the other in summer (Middle East). These regions are also characterized by having particularly contrasting cyclogenesis mechanisms between these seasons and those operating in winter.

1) Sahara

The strengthening of the meridional temperature gradient over northern Africa in spring favors the development of Sahara depressions. Their impacts on the Mediterranean region justify the increased interest in these lows, formed mostly on the lee of the Atlas Mountains. They frequently bring strong winds and sandstorms (Alpert and Ziv 1989), with dust transported long distances, affecting the southern parts of Spain, France, and Italy, or Libya and Egypt (Moulin et al. 1998).

The geopotential height anomaly fields for spring Saharan lows are shown in Fig. 15a for PminP. There are two minima near the surface; one over Iberia and the deepest south of the Atlas Mountains. The depression aloft is centered over the Iberian Peninsula and extends southward over the northwestern tip of Africa. Although the criteria to choose the composite cases were based on 1000-hPa height, the composites of relative vorticity (Fig. 15b) suggest the development of relatively strong cyclogenesis south of the Atlas Mountains, extending well into the troposphere. Although not shown, the analysis of 10-m wind (available from NCEP–NCAR 12-hourly reanalyses) also shows signs of diurnal fluctuations in the strength of Saharan lows.

As in the case of Genoa cyclones, Saharan lows seem to be associated with a preexisting depression, developed through the whole troposphere, located northward of a mountainous barrier, in this case the Atlas Mountains (e.g., Egger et al. 1995). The vertical separation of the isentropes on the lee side, and the associated negative σ anomalies (Fig. 15b), are, however, more pronounced in the generation and maintenance of Saharan cyclones. As Fig. 15b clearly shows, the surface cyclone develops within a region with extremely weak static stability up to 650 hPa, and with strong low-level baroclinicity. These seem to be the optimal conditions, found mainly during late afternoon, to trigger the growth of a Saharan low, when an upper trough, such as that observed in the geopotential composites, is perturbed by the mountainous barrier and interacts with the low-level troposphere.

The absorption and multiple scattering of radiation by the airborne dust may significantly amplify the heating of the boundary layer, creating a deep mixed layer and increasing the sea–land temperature gradient. Attempts to quantify the heating rates of the lower atmosphere due to dust during Saharan storms lead to values between 10°C day−1 (Carlson and Benjamin 1980) and 6°C day−1 (Alpert et al. 1998). In the Sahara and other desert regions the role dust plays in the radiative processes of the lower troposphere seems to overcome the lack of moisture and release of latent heat, leading to unexpectedly strong cyclogenesis over these regions (Chen et al. 1995).

A number of different situations may contribute to Saharan cyclogenesis. Although this study is concerned with general considerations, it is worth mentioning the Thorncroft and Flocas (1997) case study where interaction between the polar jet, strongly deflected toward northern Africa with a pattern similar to the upper troughs found in our geopotential composites (Fig. 15a), leads to the highly mobile subtropical jet triggering a cyclone south of the Atlas Mountains. This type of interaction between large-scale features may well be reinforced by local radiative processes to favor the growth of Saharan lows.

2) Middle East

Spring and summer are generally dry in the eastern Mediterranean. Although severe storms may occur in April, these are usually associated with Saharan cyclones that have moved toward the sea (e.g., Barry and Chorley 1998), and are much less frequent than the cyclogenesis events detected over the Middle East region (and also over Cyprus; see TDB). In April the high pressure cells that dominated over Eurasia and the Arabian Peninsula during the winter months have collapsed, being replaced by a low pressure trough, extending to the Sahara from southern Asia (e.g., HMSO 1962). Toward the summer, the surface charts are characterized by an intensification of both the Asian–north African trough and of the Azores anticyclone, which now expands through most of the Mediterranean Basin.

The lowest negative anomalies of the 1000-hPa height obtained for August PminP composites, that is, associated with the deepest phase of the trough's life cycle, are about −20 geopotential meter (gpm), concentrating over relatively small regions in the easternmost part of the domain of study (not shown). The 99% confidence level of the anomaly field, however, shades not only the Middle East region, but also Anatolia, and part of the eastern Mediterranean Sea, including Cyprus, and Egypt. The weak values of the anomaly fields are a direct result of the persistent nature of Middle East lows, as extensions of the Asian trough. The strengthening of this Asian trough, already detected in April mean charts, coincides with the onset of the southern Asian monsoon in mid- or late May. As described by Rodwell and Hoskins (1996), the monsoon also acts as a remote forcing of a strong descending cell observed over the eastern Mediterranean and eastern Sahara during late spring and summer months. Thus, while the Asian trough, associated with the monsoon in the summer, induces low-level and persistent (see TDB) troughs in the Middle East and eastern Mediterranean, it also inhibits their vertical development by favoring strong descent, induced when it interacts with the midlatitude westerlies (see also Hoskins 1996). This explains why very few significant height anomalies are detected at 500 hPa, and the generally dry weather of the eastern Mediterranean in spring and summer.

As in the case of Saharan lows, the histograms of Figs. 13 and 14 for the Middle East reveal the existence of a diurnal cycle in the detection of the first position (1stP) in April, and of both first (1stP) and minimum pressure (PminP) positions, particularly pronounced in August. Although the existence and main characteristics of Middle East lows are determined by planetary-scale phenomena, the expansion of the Asian trough into the region is modulated by local radiative processes in the low-level troposphere.

b. Case study

The situation observed on 21 August 1995 is a good example of the diurnal seesaw observed in summer Mediterranean lows. During this period most of the Mediterranean was dominated by weak surface troughs (Fig. 16), which intensify during the warmest period of the day. This is particularly evident in the total extent and intensity of the eastern trough (D in Fig. 16), and in the development and damping of the surface low over the Iberian Peninsula (A in Fig. 16). By 1800 UTC 20 August (not shown) a low pressure system has developed south of the Atlas Mountains. Both surface and 500-hPa height fields over the subsequent hours have similar patterns to those obtained for Saharan composites (Fig. 15a). As in the 500-hPa height composites the upper trough extends from the Iberian Peninsula toward northwestern Africa, leading it to interact with the low-level baroclinicity along the Atlantic coast and with the Atlas topography, while the upper system also favors the advection of colder air over western Europe. The 1000-hPa system (B in Fig. 16a) south of the Atlas in the 0000 UTC 21 August chart weakens through dawn and morning (Figs. 16c and 16e), and reintensifies only by late afternoon (Fig. 16g).

This Saharan system is a good example of the interaction between upper features and local conditions described above. Despite the presence of the 500-hPa trough over Iberia, the optimal conditions for the triggering and maximum intensity of the low are always reached around 1800 UTC (Figs. 16g and 16h), when static stability is lower. This seesaw continues in the following days, until the damping of the system by 1200 UTC 24 August (not shown).

A relatively weak surface minimum also formed over the Gulf of Genoa (C in Fig. 16), associated with the 500-h Pa trough, which remains relatively weak and stationary throughout the period of study. The Genoa 1000-hPa minimum merged with the eastern Mediterranean trough (D) when it intensified and extended to the west in the afternoon (Figs. 16e and 16g). However, as suggested before, the diurnal fluctuations of Genoa lows are much smoother than over any of the other areas, where thermal processes are dominant; between 0600 and 1800 UTC 21 August 1995 pressure fell by more than 6 hPa to the south of the Atlas Mountains, by more than 4 hPa over central Iberia, by 3–4 hPa over the Middle East, and by less than 1 hPa over the Gulf of Genoa.

The vertical motion (ω) field at 850 and 500 hPa, averaged for the four observation times on 21 August 1995, is presented in Fig. 17. The vertical motion is an indicator of the extent of vertical development of the different systems observed within that time. It is a measure of how deep and strong the convection is and, thus, indirectly of the impacts of each system on local weather. In this sense, it is worth noting the following:

  1. The relatively noisy ω field at 850 hPa (Fig. 17a) is gradually replaced by larger and better defined cells of ascent/descent at 500 hPa (Fig. 17b). Most of the Mediterranean Sea and Middle East is dominated by subsidence at 500 hPa.

  2. Ascent is still significant at 500 hPa over northern Africa around the Atlas Mountains, and over Iberia. This background ascent is likely to favor the development of convection and, thus, support the occurrence of thunderstorms, which are relatively frequent during the summer months (e.g., HMSO 1962).

  3. Also, the ascending cell at 500 hPa over the Alps may be an indication of relatively strong convection associated with the surface system observed in that region.

Most parts of the Middle East and Cyprus are under the influence of the eastern Mediterranean descent. This case study presents one of the most common summer situations in that region, where despite the observed large area of low surface pressure, the conditions remain essentially dry.

6. Concluding remarks

The winter season on one hand, and the spring and summer seasons on the other, exhibit contrasting mechanisms for Mediterranean cyclogenesis. In winter links between synoptic upper troughs and local orography and/or low-level baroclinicity observed over the northern Mediterranean coast are strong. In spring and summer cyclogenesis over land, where thermal troughing plays a major role, becomes more frequent. As a consequence the development and deepening of Mediterranean lows exhibits a higher sensitivity to diurnal forcing.

The geography of the region, namely the high orography skirting the Mediterranean Sea and the existence of embayments and inland seas, determines the relatively small areas where cyclogenesis tends to occur. In the winter this is essentially along the strongly baroclinic northern coast: in the lee of the Alps when an upper trough is blocked by the mountains; over the Aegean and Black Seas when an upper trough moves over the relatively warm water basins. The composite analyses confirm that the strong localization of cyclogenesis events implies that, at least initially, cyclones are of mesoscale size (Radinovic 1987), triggered, however, by synoptic upper disturbances. Moreover, as confirmed by the case study presented above, cyclogenesis over the three regions may occur consecutively as the result of the same synoptic system crossing central Europe.

In spring, the strengthening of the meridional temperature gradient along the northern African coast favors the development of Saharan depressions. These tend to occur on the lee side of the Atlas Mountains, within a region of very weak static stability. Thermal forcing does play an increased role in the genesis and maintenance of Mediterranean lows in spring and, particularly, in summer. As a result the life cycles, particularly of summer cyclones developed over northern Africa, the Iberian Peninsula, and the Black Sea follow the diurnal temperature fluctuations; maximum intensity tends to be reached by late afternoon, and cyclolysis tends to occur mostly by early morning. Also the Middle East trough, which is a semipermanent feature primarily induced by the Asian monsoon acting on a planetary scale (Rodwell and Hoskins 1996), exhibits the same kind of diurnal seesaw associated with the local thermal cycles.

Acknowledgments

The present work was supported by the PRAXIS XXI program (Portuguese Office for Science and Technology), Grant BD/9508/96. The ECMWF dataset was supplied by the British Atmospheric Data Center. The authors are grateful to the anonymous reviewers, whose suggestions helped to improve the manuscript.

REFERENCES

  • Aebischer, U., and C. Schar, 1998: Low-level potential vorticity and cyclogenesis to the lee of the Alps. J. Atmos. Sci, 55 , 186207.

  • Alpert, P., and B. Ziv, 1989: The Sharav cyclone: Observations and some theoretical considerations. J. Geophys. Res, 94 , 1849518514.

  • Alpert, P., B. U. Neeman, and Y. Shay-El, 1990: Climatological analysis of Mediterranean cyclones using ECMWF data. Tellus, 42A , 6577.

    • Search Google Scholar
    • Export Citation
  • Alpert, P., M. Tsidulko, S. Krichak, and U. Stein, 1996: A multi-stage evolution of an ALPEX cyclone. Tellus, 48A , 209220.

  • Alpert, P., Y. J. Kaufman, Y. Shay-El, D. Tanre, A. da Silva, S. Schubert, and J. H. Joseph, 1998: Quantification of dust-forced heating of the lower troposphere. Nature, 395 , 367370.

    • Search Google Scholar
    • Export Citation
  • Barry, R. G., and R. J. Chorley, 1998: Atmosphere, Weather and Climate. 7th ed. Routledge, 409 pp.

  • Blender, R., K. Fraedrich, and F. Lunkeit, 1997: Identification of cyclone track regimes in North Atlantic. Quart. J. Roy. Meteor. Soc, 123 , 727741.

    • Search Google Scholar
    • Export Citation
  • Bretherton, F. P., 1966: Critical layer instability in baroclinic flows. Quart. J. Roy. Meteor. Soc, 92 , 325334.

  • Buzzi, A., and S. Tibaldi, 1978: Cyclogenesis in the lee of Alps: A case study. Quart. J. Roy. Meteor. Soc, 104 , 271287.

  • Carlson, T. N., and S. G. Benjamin, 1980: Radiative heating for Saharan dust. J. Atmos. Sci, 37 , 193213.

  • Chen, S-J., Y-H. Kwo, W. Ming, and H. Ying, 1995: The effect of dust radiative heating on low-level frontogenesis. J. Atmos. Sci, 52 , 14141420.

    • Search Google Scholar
    • Export Citation
  • Davies, T. D., and E. Schuepbach, 1994: Episodes of high ozone concentrations at the earth's surface resulting from transport down from the upper troposphere/lower stratosphere: A review and case studies. Atmos. Environ, 28 , 5368.

    • Search Google Scholar
    • Export Citation
  • Dell'Osso, L., and S. Tibaldi, 1982: Some preliminary modelling results of an ALPEX case of lee cyclogenesis. ALPEX Preliminary Scientific Results, GARP-ALPEX Rep. 7, WMO, 3–19.

    • Search Google Scholar
    • Export Citation
  • Dell'Osso, L., and D. Radinovic, 1984: A case study of cyclone development in the lee of the Alps on 18 March 1982. Beitr. Phys. Atmos, 57 , 369379.

    • Search Google Scholar
    • Export Citation
  • Egger, J., 1988: Alpine lee cyclogenesis—Verification of theories. J. Atmos. Sci, 45 , 21872203.

  • Egger, J., P. Alpert, A. Tafferner, and B. Ziv, 1995: Numerical experiments on the genesis of Sharav cyclones: Idealized simulation. Tellus, 47A , 162174.

    • Search Google Scholar
    • Export Citation
  • Flocas, H. A., and T. S. Karacostas, 1996: Cyclogenesis over the Aegean Sea: Identifications and synoptic categories. Meteor. Appl, 3 , 5361.

    • Search Google Scholar
    • Export Citation
  • Gibson, J. K., P. Kållberg, S. Uppala, A. Hernandez, A. Nomura, and E. Serrano, 1999: ERA-15 description (Version 2). ECMWF Re-Analysis Project Report Series, No. 1, 74 pp.

    • Search Google Scholar
    • Export Citation
  • Gomis, D., A. Buzzi, and S. Alonso, 1990: Diagnosis of mesoscale structures in cases of lee cyclogenesis during ALPEX. Meteor. Atmos. Phys, 43 , 4957.

    • Search Google Scholar
    • Export Citation
  • HMSO, 1962: Weather in the Mediterranean I: General Meteorology. 2d ed. 362 pp.

  • Holton, J. R., 1992: An Introduction to Dynamic Meteorology. 3d ed. Academic Press, 391 pp.

  • Hoskins, B. J., 1996: On the existence and strength of summer subtropical anticyclones. Bull. Amer. Meteor. Soc, 77 , 12871292.

  • Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc, 111 , 877946.

    • Search Google Scholar
    • Export Citation
  • Jacobeit, J., 1987: Variations of trough positions and precipitation patterns in the Mediterranean area. J. Climatol, 7 , 453476.

  • Kållberg, P., 1997: Aspects of the re-analysed climate. ECMWF Re-Analysis Project Report Series, No. 2, 91 pp.

  • Kalnay, E. M., and and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc, 77 , 437471.

  • Lackmann, G. M., D. Keyser, and L. F. Bosart, 1997: A characteristic life cycle of upper-tropospheric cyclogenetic precursors during the Experiment on Rapidly Intensifying Cyclones over the Atlantic (ERICA). Mon. Wea. Rev, 125 , 27292758.

    • Search Google Scholar
    • Export Citation
  • Lackmann, G. M., D. Keyser, and L. F. Bosart, 1999: Energetics of an intensifying jet streak during the Experiment on Rapidly Intensifying Cyclones over the Atlantic (ERICA). Mon. Wea. Rev, 127 , 27772795.

    • Search Google Scholar
    • Export Citation
  • Lee, T. P., S. R. Silberberg, and L. F. Bosart, 1988: A case study of a severe winter storm in the Middle East. Quart. J. Roy. Meteor. Soc, 114 , 6190.

    • Search Google Scholar
    • Export Citation
  • Lefevre, R. L., and J. W. Nielsen-Gammon, 1995: An objective climatology of mobile troughs in the Northern Hemisphere. Tellus, 47A , 638655.

    • Search Google Scholar
    • Export Citation
  • McGinley, J. A., and M. Zupanski, 1990: Numerical analysis of the influence of jets, fronts and mountains on Alpine lee cyclogenesis: More cases from the ALPEX SOP. Meteor. Atmos. Phys, 43 , 720.

    • Search Google Scholar
    • Export Citation
  • Morgan, M. C., and J. W. Nielsen-Gammon, 1998: Using tropopause maps to diagnose midlatitude weather systems. Mon. Wea. Rev, 126 , 25552579.

    • Search Google Scholar
    • Export Citation
  • Moulin, C., and and Coauthors, 1998: Satellite climatology of African dust transport in the Mediterranean atmosphere. J. Geophys. Res, 103 , 1313713144.

    • Search Google Scholar
    • Export Citation
  • Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • Petterssen, S., 1956: Motion and Motion Systems. Vol. 1. Weather Analysis and Forecasting, McGraw-Hill, 428 pp.

  • Pichler, H., R. Steinacker, and A. Lazinger, 1990: Cyclogenesis induced by the Alps. Meteor. Atmos. Phys, 43 , 2129.

  • Prezerakos, N. G., and H. A. Flocas, 1996: The formation of a dynamically unstable ridge at 500 hPa as a precursor of surface cyclogenesis in the central Mediterranean. Meteor. Appl, 3 , 101111.

    • Search Google Scholar
    • Export Citation
  • Radinoivic, D., 1987: Mediterranean Cyclones and Their Influence on the Weather and Climate. PSMP Report Series, No. 24, WMO, 131 pp.

  • Rodwell, M. J., and B. J. Hoskins, 1996: Monsoons and the dynamics of deserts. Quart. J. Roy. Meteor. Soc, 122 , 13851404.

  • Sanders, F., 1986: Explosive cyclogenesis in the west-central North Atlantic Ocean, 1981–84. Part I: Composite and mean behavior. Mon. Wea. Rev, 114 , 17811794.

    • Search Google Scholar
    • Export Citation
  • Schuepbach, E., T. D. Davies, A. C. Massacand, and H. Wernli, 1999: Mesoscale modeling of vertical atmospheric transport in the Alps associated with the advection of a tropopause fold—A winter ozone episode. Atmos. Environ, 33 , 17351744.

    • Search Google Scholar
    • Export Citation
  • Serreze, M. C., F. Carse, R. G. Barry, and J. C. Rogers, 1997: Icelandic low cyclone activity: Climatological features, linkages with the NAO, and relationships with recent changes in the Northern Hemisphere circulation. J. Climate, 10 , 453464.

    • Search Google Scholar
    • Export Citation
  • Shapiro, M. A., T. Hampel, and A. J. Krueger, 1987: The Arctic tropopause fold. Mon. Wea. Rev, 115 , 444454.

  • Tafferner, A., 1990: Lee cyclogenesis resulting from the combined outbreak of cold air and potential vorticity against the Alps. Meteor. Atmos. Phys, 43 , 3147.

    • Search Google Scholar
    • Export Citation
  • Tayanç, M., M. Karaca, and H. N. Dalfes, 1998: March 1987 cyclone (blizzard) over the eastern Mediterranean and Balkan region associated with blocking. Mon. Wea. Rev, 126 , 30363047.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C. D., and H. A. Flocas, 1997: A case study of Saharan cyclogenesis. Mon. Wea. Rev, 125 , 11471165.

  • Thornes, J. B., 1998: Mediterranean desertification. Atlas of Mediterranean Environments in Europe, Wiley, 205 pp.

  • Trigo, I. F., T. D. Davies, and G. R. Bigg, 1999: Objective climatology of cyclones in the Mediterranean region. J. Climate, 12 , 16851696.

    • Search Google Scholar
    • Export Citation
  • Uppala, S., 1997: Observing system performance in ERA. ECMWF Re-Analysis Project Report Series, No. 3, 261 pp.

  • Wilks, D. S., 1995: Statistical Methods in the Atmospheric Sciences. An Introduction. Academic Press, 467 pp.

  • Ziv, B., and P. Alpert, 1995: Rotation of binary cyclones—A data analysis study. J. Atmos. Sci, 52 , 13571369.

  • Zupanski, M., and J. McGinley, 1989: Numerical analysis of the influence of jets, fronts and mountains on Alpine lee cyclogenesis. Mon. Wea. Rev, 117 , 154176.

    • Search Google Scholar
    • Export Citation

APPENDIX

Cyclone Tracking Algorithm

The detection and tracking of Mediterranean cyclones (TDB) are performed using 6-hourly geopotential height at 1000 hPa (Z1000), available from ECMWF reanalyses on a 1.125° × 1.125° grid. The data cover the area from 24.75° to 50.625°N and 15.75°W to 45°E, and the period 1979–96.

A cyclone candidate is identified as a local Z1000 minimum, over a three by three gridpoint area. To be considered a cyclone, this minimum must fulfill two thresholds found empirically: 1) a maximum value of 1020 hPa is required for the central sea level pressure; for that purpose Z1000 was converted to sea level pressure using the hydrostatic balance condition in terms of pcen(hPa) = 0.121 Z1000 (gpm) + 1000 [using the density near the surface ≈1.23 kg m−3 and g = 9.81 m s−2; Peixoto and Oort (1992)]; and 2) the mean pressure gradient, estimated for an area of 9° lat × 11.25° lon around the minimum pressure, must be at least 0.55 hPa (100 km)−1.

The cyclone tracking algorithm is based on a nearest-neighbor search procedure [as in Blender et al. (1997) and Serreze et al. (1997)]: a cyclone's trajectory is determined by computing the distance to cyclones detected in the previous chart and assuming the cyclone has taken the path of minimum distance; if the nearest neighbor in the previous chart is not within an area determined by imposing a maximum cyclone velocity of 33 km h−1 in the westward direction and of 90 km h−1 in any other, then cyclogenesis is assumed to have occurred. Again, these thresholds were determined empirically for this specific region by observing Mediterranean cyclone behavior in Z1000 charts.

Fig. 1.
Fig. 1.

The Mediterranean Sea and topography (m) of the surrounding regions

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 2.
Fig. 2.

Total number of cyclogenesis events detected per 105 km2 within the most active regions in the (a) western and (b) eastern Mediterranean, during the 1987–96 period. The total number of cyclogenesis events detected in the whole basin per unit area is represented by the bottom line

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 3.
Fig. 3.

Relative frequency (%) of cyclogenesis (first detection) and cyclolysis (last detection) events at each observation time, from (a) Jan to (1) Dec

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 4.
Fig. 4.

Anomalies of geopotential height at 1000 (contours every 10 gpm) and 500 hPa (contours every 20 gpm), obtained for Jan composites of Gulf of Genoa cyclones (a) 1stP, and (b) PminP times; (c) vertical cross sections [at 10°; 15°E, as indicated by the solid vertical lines in (b)] of relative vorticity anomalies (contours every 10−5 s−1) obtained for composites of Genoa PminP time. Anomalies within the shaded regions are statistically significant at the 99% confidence level

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 5.
Fig. 5.

Gulf of Genoa composites of (a) 1000-hPa height (contours every 20 gpm) and (b) 1000-hPa height anomalies (contours every 10 gpm) for 1stP time. The shaded regions represent the areas where the difference between the composite field and the Jan climatology is statistically significant at the 99% confidence level

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 6.
Fig. 6.

(a) As in Fig. 4a but for Aegean Sea lows. (b) Vertical cross sections [(at 25°E, as indicated by the solid vertical line in (a)] of relative vorticity anomalies (contours every 10−5 s−1) and static stability [contours every (0.005 K) 100 hPa−1] obtained for composites of Aegean 1stP time

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 7.
Fig. 7.

Geopotential composites at (a) 1000 (contours every 15 gpm) and (b) 500 hPa (contours every 50 gpm) for 1stP time (lower panels) and 12h before 1stP time (upper panels). Anomalies within the shaded regions are statistically significant at the 99% confidence level

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 8.
Fig. 8.

As in Fig. 6a but for Black Sea lows

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 9.
Fig. 9.

Composites of potential temperature (contours every 2°C) at 850 hPa, obtained for Jan 1stP time for (a) Gulf of Genoa, and (b) Aegean and (c) Black Sea cyclones; light (dark) shaded regions are significantly colder (warmer) than the climatology, at the 99% confidence level

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 10.
Fig. 10.

Composite anomalies of potential temperature (contours every 1 K) at 850 hPa, obtained for (a) 12 h before and (b) 12 h after of 1stP time for Jan Gulf of Genoa, Aegean Sea, and Black Sea cyclones. Anomalies within the shaded regions are statistically significant at the 99% confidence level

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 11.
Fig. 11.

The contours represent the reanalysis of sea level pressure (left column; contours every 2.5 hPa) and 500-hPa height (right column; contours every 50 gpm) between 0000 UTC 6 Jan 1987 and 0000 UTC 9 Jan 1987. The shading represents 500-hPa geostrophic wind (m s−1; as indicated by the legend on the right of each panel). The respective date and time are shown on the top of each diagram

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 12.
Fig. 12.

Vertical cross sections, through the systems depicted in Fig. 11—(a) at 10° and 15°E, 0000 UTC 6 Jan; (b) at 17° and 25°E, 0000 UTC 7 Jan; and (c) at 27° and 37°E, 0000 UTC 9 Jan—of relative vorticity (contours every 2.5 × 10−5 s−1), and potential temperature (contours every 5 K; dotted lines). The bold lines enclose the 1.5–3 PVU potential vorticity band. The respective longitude and time are shown on the top of each diagram

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 13.
Fig. 13.

Relative frequency of 1stP and PminP events per observation time, for each of the most cyclogenetic areas in Apr

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 14.
Fig. 15.
Fig. 15.

(a) Anomalies of geopotential height at 1000 (contours every 10 gpm) and 500 hPa (contours every 20 gpm), obtained for Apr composites of Saharan cyclones PminP type; (b) vertical cross sections [at 5°W; solid vertical line in (a)] of relative vorticity anomalies (contours every 10−5 s−1) and static stability [contours every 0.005 K (100 hPa)−1] obtained for composites of Saharan PminP type. Anomalies within the shaded regions are statistically significant at the 99% confidence level

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 16.
Fig. 16.

Reanalysis of sea level pressure (left column; contours every 2.5 hPa and shaded according to the legend on the right of each panel) and 500-hPa height (right column; contours every 50 gpm) between 0000 and 1800 UTC on 21 Aug 1995. The respective date and time are shown on the top of each diagram

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Fig. 17.
Fig. 17.

Average vertical motion (contours every 7.5 × 10−2 Pa s−1), at (a) 850 and (b) 500 hPa on 21 Aug 1995. Solid (dashed) contours represent descending (ascending) air

Citation: Monthly Weather Review 130, 3; 10.1175/1520-0493(2002)130<0549:COCMIT>2.0.CO;2

Table 1.

Number of cyclogenesis events detected in each of the most cyclogenetic regions for Jan, Apr, and Aug, between 1987 and 1996. Absence of a figure means the region is not a major area for cyclogenesis during the season in question

Table 1.
Table 2.

Correspondence between the observation time, the local time, and the real range of times over which the 1stP and PminP events occur

Table 2.
Save
  • Aebischer, U., and C. Schar, 1998: Low-level potential vorticity and cyclogenesis to the lee of the Alps. J. Atmos. Sci, 55 , 186207.

  • Alpert, P., and B. Ziv, 1989: The Sharav cyclone: Observations and some theoretical considerations. J. Geophys. Res, 94 , 1849518514.

  • Alpert, P., B. U. Neeman, and Y. Shay-El, 1990: Climatological analysis of Mediterranean cyclones using ECMWF data. Tellus, 42A , 6577.

    • Search Google Scholar
    • Export Citation
  • Alpert, P., M. Tsidulko, S. Krichak, and U. Stein, 1996: A multi-stage evolution of an ALPEX cyclone. Tellus, 48A , 209220.

  • Alpert, P., Y. J. Kaufman, Y. Shay-El, D. Tanre, A. da Silva, S. Schubert, and J. H. Joseph, 1998: Quantification of dust-forced heating of the lower troposphere. Nature, 395 , 367370.

    • Search Google Scholar
    • Export Citation
  • Barry, R. G., and R. J. Chorley, 1998: Atmosphere, Weather and Climate. 7th ed. Routledge, 409 pp.

  • Blender, R., K. Fraedrich, and F. Lunkeit, 1997: Identification of cyclone track regimes in North Atlantic. Quart. J. Roy. Meteor. Soc, 123 , 727741.

    • Search Google Scholar
    • Export Citation
  • Bretherton, F. P., 1966: Critical layer instability in baroclinic flows. Quart. J. Roy. Meteor. Soc, 92 , 325334.

  • Buzzi, A., and S. Tibaldi, 1978: Cyclogenesis in the lee of Alps: A case study. Quart. J. Roy. Meteor. Soc, 104 , 271287.

  • Carlson, T. N., and S. G. Benjamin, 1980: Radiative heating for Saharan dust. J. Atmos. Sci, 37 , 193213.

  • Chen, S-J., Y-H. Kwo, W. Ming, and H. Ying, 1995: The effect of dust radiative heating on low-level frontogenesis. J. Atmos. Sci, 52 , 14141420.

    • Search Google Scholar
    • Export Citation
  • Davies, T. D., and E. Schuepbach, 1994: Episodes of high ozone concentrations at the earth's surface resulting from transport down from the upper troposphere/lower stratosphere: A review and case studies. Atmos. Environ, 28 , 5368.

    • Search Google Scholar
    • Export Citation
  • Dell'Osso, L., and S. Tibaldi, 1982: Some preliminary modelling results of an ALPEX case of lee cyclogenesis. ALPEX Preliminary Scientific Results, GARP-ALPEX Rep. 7, WMO, 3–19.

    • Search Google Scholar
    • Export Citation
  • Dell'Osso, L., and D. Radinovic, 1984: A case study of cyclone development in the lee of the Alps on 18 March 1982. Beitr. Phys. Atmos, 57 , 369379.

    • Search Google Scholar
    • Export Citation
  • Egger, J., 1988: Alpine lee cyclogenesis—Verification of theories. J. Atmos. Sci, 45 , 21872203.

  • Egger, J., P. Alpert, A. Tafferner, and B. Ziv, 1995: Numerical experiments on the genesis of Sharav cyclones: Idealized simulation. Tellus, 47A , 162174.

    • Search Google Scholar
    • Export Citation
  • Flocas, H. A., and T. S. Karacostas, 1996: Cyclogenesis over the Aegean Sea: Identifications and synoptic categories. Meteor. Appl, 3 , 5361.

    • Search Google Scholar
    • Export Citation
  • Gibson, J. K., P. Kållberg, S. Uppala, A. Hernandez, A. Nomura, and E. Serrano, 1999: ERA-15 description (Version 2). ECMWF Re-Analysis Project Report Series, No. 1, 74 pp.

    • Search Google Scholar
    • Export Citation
  • Gomis, D., A. Buzzi, and S. Alonso, 1990: Diagnosis of mesoscale structures in cases of lee cyclogenesis during ALPEX. Meteor. Atmos. Phys, 43 , 4957.

    • Search Google Scholar
    • Export Citation
  • HMSO, 1962: Weather in the Mediterranean I: General Meteorology. 2d ed. 362 pp.

  • Holton, J. R., 1992: An Introduction to Dynamic Meteorology. 3d ed. Academic Press, 391 pp.

  • Hoskins, B. J., 1996: On the existence and strength of summer subtropical anticyclones. Bull. Amer. Meteor. Soc, 77 , 12871292.

  • Hoskins, B. J., M. E. McIntyre, and A. W. Robertson, 1985: On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Meteor. Soc, 111 , 877946.

    • Search Google Scholar
    • Export Citation
  • Jacobeit, J., 1987: Variations of trough positions and precipitation patterns in the Mediterranean area. J. Climatol, 7 , 453476.

  • Kållberg, P., 1997: Aspects of the re-analysed climate. ECMWF Re-Analysis Project Report Series, No. 2, 91 pp.

  • Kalnay, E. M., and and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc, 77 , 437471.

  • Lackmann, G. M., D. Keyser, and L. F. Bosart, 1997: A characteristic life cycle of upper-tropospheric cyclogenetic precursors during the Experiment on Rapidly Intensifying Cyclones over the Atlantic (ERICA). Mon. Wea. Rev, 125 , 27292758.

    • Search Google Scholar
    • Export Citation
  • Lackmann, G. M., D. Keyser, and L. F. Bosart, 1999: Energetics of an intensifying jet streak during the Experiment on Rapidly Intensifying Cyclones over the Atlantic (ERICA). Mon. Wea. Rev, 127 , 27772795.

    • Search Google Scholar
    • Export Citation
  • Lee, T. P., S. R. Silberberg, and L. F. Bosart, 1988: A case study of a severe winter storm in the Middle East. Quart. J. Roy. Meteor. Soc, 114 , 6190.

    • Search Google Scholar
    • Export Citation
  • Lefevre, R. L., and J. W. Nielsen-Gammon, 1995: An objective climatology of mobile troughs in the Northern Hemisphere. Tellus, 47A , 638655.

    • Search Google Scholar
    • Export Citation
  • McGinley, J. A., and M. Zupanski, 1990: Numerical analysis of the influence of jets, fronts and mountains on Alpine lee cyclogenesis: More cases from the ALPEX SOP. Meteor. Atmos. Phys, 43 , 720.

    • Search Google Scholar
    • Export Citation
  • Morgan, M. C., and J. W. Nielsen-Gammon, 1998: Using tropopause maps to diagnose midlatitude weather systems. Mon. Wea. Rev, 126 , 25552579.

    • Search Google Scholar
    • Export Citation
  • Moulin, C., and and Coauthors, 1998: Satellite climatology of African dust transport in the Mediterranean atmosphere. J. Geophys. Res, 103 , 1313713144.

    • Search Google Scholar
    • Export Citation
  • Peixoto, J. P., and A. H. Oort, 1992: Physics of Climate. American Institute of Physics, 520 pp.

  • Petterssen, S., 1956: Motion and Motion Systems. Vol. 1. Weather Analysis and Forecasting, McGraw-Hill, 428 pp.

  • Pichler, H., R. Steinacker, and A. Lazinger, 1990: Cyclogenesis induced by the Alps. Meteor. Atmos. Phys, 43 , 2129.

  • Prezerakos, N. G., and H. A. Flocas, 1996: The formation of a dynamically unstable ridge at 500 hPa as a precursor of surface cyclogenesis in the central Mediterranean. Meteor. Appl, 3 , 101111.

    • Search Google Scholar
    • Export Citation
  • Radinoivic, D., 1987: Mediterranean Cyclones and Their Influence on the Weather and Climate. PSMP Report Series, No. 24, WMO, 131 pp.

  • Rodwell, M. J., and B. J. Hoskins, 1996: Monsoons and the dynamics of deserts. Quart. J. Roy. Meteor. Soc, 122 , 13851404.

  • Sanders, F., 1986: Explosive cyclogenesis in the west-central North Atlantic Ocean, 1981–84. Part I: Composite and mean behavior. Mon. Wea. Rev, 114 , 17811794.

    • Search Google Scholar
    • Export Citation
  • Schuepbach, E., T. D. Davies, A. C. Massacand, and H. Wernli, 1999: Mesoscale modeling of vertical atmospheric transport in the Alps associated with the advection of a tropopause fold—A winter ozone episode. Atmos. Environ, 33 , 17351744.

    • Search Google Scholar
    • Export Citation
  • Serreze, M. C., F. Carse, R. G. Barry, and J. C. Rogers, 1997: Icelandic low cyclone activity: Climatological features, linkages with the NAO, and relationships with recent changes in the Northern Hemisphere circulation. J. Climate, 10 , 453464.

    • Search Google Scholar
    • Export Citation
  • Shapiro, M. A., T. Hampel, and A. J. Krueger, 1987: The Arctic tropopause fold. Mon. Wea. Rev, 115 , 444454.

  • Tafferner, A., 1990: Lee cyclogenesis resulting from the combined outbreak of cold air and potential vorticity against the Alps. Meteor. Atmos. Phys, 43 , 3147.

    • Search Google Scholar
    • Export Citation
  • Tayanç, M., M. Karaca, and H. N. Dalfes, 1998: March 1987 cyclone (blizzard) over the eastern Mediterranean and Balkan region associated with blocking. Mon. Wea. Rev, 126 , 30363047.

    • Search Google Scholar
    • Export Citation
  • Thorncroft, C. D., and H. A. Flocas, 1997: A case study of Saharan cyclogenesis. Mon. Wea. Rev, 125 , 11471165.

  • Thornes, J. B., 1998: Mediterranean desertification. Atlas of Mediterranean Environments in Europe, Wiley, 205 pp.

  • Trigo, I. F., T. D. Davies, and G. R. Bigg, 1999: Objective climatology of cyclones in the Mediterranean region. J. Climate, 12 , 16851696.

    • Search Google Scholar
    • Export Citation
  • Uppala, S., 1997: Observing system performance in ERA. ECMWF Re-Analysis Project Report Series, No. 3, 261 pp.

  • Wilks, D. S., 1995: Statistical Methods in the Atmospheric Sciences. An Introduction. Academic Press, 467 pp.

  • Ziv, B., and P. Alpert, 1995: Rotation of binary cyclones—A data analysis study. J. Atmos. Sci, 52 , 13571369.

  • Zupanski, M., and J. McGinley, 1989: Numerical analysis of the influence of jets, fronts and mountains on Alpine lee cyclogenesis. Mon. Wea. Rev, 117 , 154176.

    • Search Google Scholar
    • Export Citation
  • Fig. 1.

    The Mediterranean Sea and topography (m) of the surrounding regions

  • Fig. 2.

    Total number of cyclogenesis events detected per 105 km2 within the most active regions in the (a) western and (b) eastern Mediterranean, during the 1987–96 period. The total number of cyclogenesis events detected in the whole basin per unit area is represented by the bottom line

  • Fig. 3.

    Relative frequency (%) of cyclogenesis (first detection) and cyclolysis (last detection) events at each observation time, from (a) Jan to (1) Dec

  • Fig. 4.

    Anomalies of geopotential height at 1000 (contours every 10 gpm) and 500 hPa (contours every 20 gpm), obtained for Jan composites of Gulf of Genoa cyclones (a) 1stP, and (b) PminP times; (c) vertical cross sections [at 10°; 15°E, as indicated by the solid vertical lines in (b)] of relative vorticity anomalies (contours every 10−5 s−1) obtained for composites of Genoa PminP time. Anomalies within the shaded regions are statistically significant at the 99% confidence level

  • Fig. 5.

    Gulf of Genoa composites of (a) 1000-hPa height (contours every 20 gpm) and (b) 1000-hPa height anomalies (contours every 10 gpm) for 1stP time. The shaded regions represent the areas where the difference between the composite field and the Jan climatology is statistically significant at the 99% confidence level

  • Fig. 6.

    (a) As in Fig. 4a but for Aegean Sea lows. (b) Vertical cross sections [(at 25°E, as indicated by the solid vertical line in (a)] of relative vorticity anomalies (contours every 10−5 s−1) and static stability [contours every (0.005 K) 100 hPa−1] obtained for composites of Aegean 1stP time

  • Fig. 7.

    Geopotential composites at (a) 1000 (contours every 15 gpm) and (b) 500 hPa (contours every 50 gpm) for 1stP time (lower panels) and 12h before 1stP time (upper panels). Anomalies within the shaded regions are statistically significant at the 99% confidence level

  • Fig. 8.

    As in Fig. 6a but for Black Sea lows

  • Fig. 9.

    Composites of potential temperature (contours every 2°C) at 850 hPa, obtained for Jan 1stP time for (a) Gulf of Genoa, and (b) Aegean and (c) Black Sea cyclones; light (dark) shaded regions are significantly colder (warmer) than the climatology, at the 99% confidence level

  • Fig. 10.

    Composite anomalies of potential temperature (contours every 1 K) at 850 hPa, obtained for (a) 12 h before and (b) 12 h after of 1stP time for Jan Gulf of Genoa, Aegean Sea, and Black Sea cyclones. Anomalies within the shaded regions are statistically significant at the 99% confidence level

  • Fig. 11.

    The contours represent the reanalysis of sea level pressure (left column; contours every 2.5 hPa) and 500-hPa height (right column; contours every 50 gpm) between 0000 UTC 6 Jan 1987 and 0000 UTC 9 Jan 1987. The shading represents 500-hPa geostrophic wind (m s−1; as indicated by the legend on the right of each panel). The respective date and time are shown on the top of each diagram

  • Fig. 12.

    Vertical cross sections, through the systems depicted in Fig. 11—(a) at 10° and 15°E, 0000 UTC 6 Jan; (b) at 17° and 25°E, 0000 UTC 7 Jan; and (c) at 27° and 37°E, 0000 UTC 9 Jan—of relative vorticity (contours every 2.5 × 10−5 s−1), and potential temperature (contours every 5 K; dotted lines). The bold lines enclose the 1.5–3 PVU potential vorticity band. The respective longitude and time are shown on the top of each diagram

  • Fig. 13.

    Relative frequency of 1stP and PminP events per observation time, for each of the most cyclogenetic areas in Apr

  • Fig. 14.

    As in Fig. 13 but for Aug