Sie sind auf Seite 1von 8

Observations of a Quasi-Tropical Cyclone

over the Black Sea

V. V. Efimov, S. V. Stanichnyi, M. V. Shokurov, and D. A. Yarovaya
Marine Hydrophysical Institute, National Academy of Sciences of Ukraine,
ul. Kapitanskaya 2, Sevastopol, Crimea, 99011 Ukraine
Received March 19, 2007
AbstractA rare atmospheric phenomenon in the Black Sea region is described: a mesoscale cyclone
with all main properties of tropical cyclones. The cyclone developed in the southwestern part of the sea
and existed for about five days. General characteristics of the cyclone are presented along with its effect
on thermodynamic structure of the sea upper layer.
DOI: 10.3103/S1068373908040067
Tropical cyclones (TCs) are frequent enough in the tropical areas of the World Ocean and are often
associated with catastrophic weather events. As a rule, tropical cyclones are smaller in size than midlatitude
ones (about 200300 km); strong winds (up to 100 m/s) and a cloud-free central area (an eye) are typical of
them. Tropical cyclones originate and develop only over the warm sea, whose surface temperature, according
to generally accepted estimates, should be not less than 27C. Other typical features of TCs, differing from
those of midlatitude cyclones with their larger scales, also exist [24, 11, 12].
When the satellite cloud images became available, it has been revealed that in the extratropical
atmosphere, too, cyclonic eddies can arise, which can be classified as quasi-tropical for many of their
features. First of all, they also exhibit axisymmetric space structure, they develop only over sea, and the
main mechanism of their generation is the moist air lifting with release of large latent heat amount through
condensation. The quasi-tropical cyclones of different origin arise from time to time over the Mediterranean
Sea [5, 10, 14, 15, 1720, 2325]. In the polar areas, under conditions of cold air intrusions from the land to
the sea, the polar lows develop, which also can be classified quasi-tropical [13, 21, 22].
By the end of September 2005, in the atmosphere over the southwest Black Sea, an intense mesoscale
cyclone developed with all special features of a TC. The cyclone had an eye with a diameter not exceeding
300 km and exhibited almost axisymmetric structure. In the satellite images on September 25 and 26, the
spiral cloud bands are clearly seen; the wind speed in the eddy area, after the QuikScat [28] satellite
measurements, reached 2025 m/s. Though this cyclone did not acquire the disastrous properties of a tropical
hurricane, the hazardous weather with strong winds and surface waves caused stoppage of ships navigating to
Istanbul. The cyclone was located, slightly wandering, over the southwest Black Sea since September 25 till
September 29, then it moved southward and by September 30 left the Black Sea. Over land, the cyclone
quickly decayed.
Consider the cyclone evolution basing on the data available. As a rule, the mesoscale atmospheric processes,
and, in particular, evolution of the cyclone under study, are poorly documented with measurements. Apart
from the satellite cloud images and wind measurements from QuikScat, no other high-resolution data are
available. The operative NCEP/NCAR objective analysis data with 1 1 resolution are used as transmitted
within the framework of international exchange [27]. Also, to study the structure of the cyclone in more detail,
numerical experiments are carried out with the MM5 regional model; their detailed discussion has been
presented in [1].
In Fig. 1a, the surface wind velocity and sea level pressure fields are displayed from the objective
analysis data for 00:00 UTC of September 25. A high pressure area in the north, as associated with a
blocking anticyclone, and a lower pressure zone over the Black Sea region can be seen. Also, despite of
ISSN 1068-3739, Russian Meteorology and Hydrology, 2008, Vol. 33, No. 4, pp. 233239. Allerton Press, Inc., 2008.
Original Russian Text V.V. Efimov, S.V. Stanichnyi, M.V. Shokurov, D.A. Yarovaya, 2008, published in Meteorologiya i Gidrologiya, 2008, No. 4, pp. 5362.
rather coarse space resolution of the objective analysis, two small depressions with pressure about 1010 hPa
can be noted over the Black Sea. The depression in the west part of the sea later gave rise to the
quasi-tropical cyclone. The wind speed in the depression was about 10 m/s and the pressure in its center was
3 hPa lower than at its periphery. The surface wind distribution for the previous analysis time (18:00 UTC
of September 24) does not show any cyclone center, and thus the structure in Fig. 1a can be considered as an
initial stage of the cyclone origination over sea.
Determination of causes and of necessary and sufficient conditions for the tropical cyclone origination
represents a problem which is not yet solved. It is known that only in one of ten cases tropical depressions
develop to the stage of TC. For the Black Sea cyclone under study, a condition favoring its initial
development is represented by a decreased atmospheric stability associated with cold air intrusion into the
Black Sea region.
In Figs. 1c and 1d, the data on vertical stratification of the atmosphere at 00:00 UTC of September 25 in
the western and eastern parts of the Black Sea are displayed. In fact, no radiosonde data are available; the
displayed profiles are taken from the operative objective analysis. The thick dashed line shows height
changes of temperature of the air parcel which is lifting from the sea surface, dry-adiabatically in the lower
layer, up to the condensation level, and then moist-adiabatically. The thick solid lines show the profiles of air
temperature (right) and dew point temperature (left). As it can be seen from Fig. 1c, from the 900 hPa level
and almost up to the tropopause (above the 300 hPa level), the lifting parcel is warmer than the surrounding
air. The atmospheric stratification is thus favorable for convection. As a measure of instability, convective
available potential energy, CAPE, is used, which is represented by the area between the dashed and solid
lines from the condensation level to the equilibrium level in Fig. 1c [9]. The CAPE distribution over the
whole Black Sea area is presented in Fig. 1b. It can be seen that CAPE reaches high values over sea with a
maximum of 1600 J/kg; in the eastern part of the sea, the CAPE values are larger than in the western one.
234 EFIMOV et al.
Fig. 1. The space distribution of (a) wind speed at the 1000 hPa level (m/s) and sea level pressure (hPa); (b) convective
available potential energy (J/kg). The vertical stratification at the points (c) 43 N, 32 E and (d) 43 N, 38 E from the
operational objective analysis data at 00:00 UTC of September 25, 2005. See text for explanations.
In Fig. 1c it can be also seen that the dew point temperature is close to the air temperature in the whole
troposphere, which shows that relative humidity is high. This implies that not only the moist warm air from
the surface layer, but the entire moist tropospheric layer can be involved in the convective development,
with latent heat release due to condensation.
In Fig. 1d, the vertical stratification in the eastern part of the sea is shown. Unlike in its western part,
here, a stable layer occurs from the surface to the 800 hPa level, which hampers convection, as the lifting
parcel acquires negative buoyancy. As a measure of stable layer intensity, convective inhibition energy,
CIN, is used, as represented by the area between the solid and dashed lines from the condensation level up
to the level of free convection [9]. The CIN distribution shows that the stable layer does not occur in the
southwestern part of the sea. This is a possible explanation of the cyclone development in that area.
Convective instability of the atmosphere over the Black Sea developed during several days before the
cyclone formation. Since September 19, in European Russia, a blocking anticyclone occurred, which
existed till the end of September. On September 18, in Spain, a cut-off high-level cyclone with a cold core
developed. The cyclone moved eastward, causing active convection over the Mediterranean Sea due to cold
air advection to the warm sea surface. From September 20 to 24, the cut-off cyclone was blocked in the
Balkan Peninsula by the mentioned anticyclone. All this time, within the cyclone, well-pronounced convec-
tion occurred with anomalous precipitation. By September 25, the cyclone was filled out, and the general
synoptic situation in the region consisted in a very weak circulation at all levels in the free atmosphere,
which can be seen from the height and wind fields at different levels.
It is well known that one of the conditions necessary for TC development consists in absence of strong
wind and especially of vertical wind shear. Evidently, in the case under study this condition is fulfilled. Due
to light winds, the cyclone could stay for a long time at the same place over the warm sea, and as there was
no wind shear, the barotropic structure of the cyclone could develop. One more favorable condition for
origination of the tropical cyclone is represented by existence of a vast reservoir of convective available
potential energy and of background convergence in the lower troposphere, which collects warm moist air
in a certain area. As it can be seen in Fig. 1a, circulation in the high at the north and in the weak cyclone
over the sea favored surface convergence over the Black Sea.
The reservoir of CAPE, as it is mentioned above, is located over the sea (Fig. 1b). The main causes of
its formation are, first, the warm surface of the Black Sea (23C) and, second, relatively cold moist airmass,
which developed in the Balkan cyclone during previous several days.
Numerical computations are carried out to simulate evolution of the cyclone and to study its structure
from the moment of its origination to that of its decay over land. The details of computation and description
of the model have been presented in [1]. The MM5 regional numerical atmospheric model is used [29]. The
grid step of the minimum internal domain is 3 to 10 km; for initial and boundary conditions, the NCEP/
NCAR objective analysis data are used [27]. In Fig. 2a, the results obtained for pressure in the cyclone
center, p
(t), and the maximum surface wind speed V
(t) are displayed in dependence on time. In Fig. 2b,
the maximum radius of the cyclone, R
(t), is shown, which allows estimating its area. R
(t) is a distance
from the cyclone center, at which the azimuth-averaged wind speed at 10 m reaches its maximum [12].
In total, several stages of cyclone development can be indicated. The initial stage covers the period from
00:00 UTC of September 25 to 12:00 UTC of September 26. The maximum surface wind speed was about
15 m/s, the central pressure was about 1010 hPa, and the radius was 100 to 115 km. Note that the initial
stage is characterized by strong asymmetry with pronounced spiral bands of irregular shape, in which, active
convection with high lifting speed and high cyclonic vorticity occurred. The main source of energy at this
stage is CAPE collected from the whole area of the sea. The mechanisms of the eddy enhancing at this stage
are associated with latent heat release due to condensation and with surface friction. The convective release
of latent heat generates positive buoyancy, the air is lifting, and, due to vortex tube stretching, vorticity in
the lower layer increases. The wind cyclonic vorticity leads, due to friction, to Ekman convergence in the
boundary layer, to additional increase in vertical speed, to vorticity increase, and so on. Such a mechanism
of eddy development has been proposed in the theory of CISK (convective instability of the second kind)
for an axisymmetric tropical cyclone [6, 7]. At present, this mechanism is revealed also in irregular
convective cells developing at the stage of tropical cyclone origination [15].
The second stage is the fast growth of the cyclone, from 12:00 UTC of September 26 to 12:00 UTC of
September 27. The wind speed grows quickly to 24 m/s, the central pressure drops to 999 hPa, and the
radius decreases to 65 km. So the cyclone is squeezed almost twice and intensified significantly. Also, it
became more axisymmetric. At this stage, the second mechanism for tropical cyclone intensification, the
so-called wind induced surface heat exchange (WISHE), started acting [7, 8, 11, 12, 26]. In this mechanism,
as well as in the first one, surface friction causes convergence; but the heat source for buoyancy increase is
now not the CAPE reservoir, but the fluxes of sensible and latent heat at the surface, which strongly
increase with increasing wind speed. By the end of the second stage, the maximum fluxes of sensible and
latent heat in the cyclone under study reached 300 and 700 W/m
, respectively. The increased axisymmetry
and squeezing of TCs is associated with angular momentum conservation. The annulus of air converging
to the eddy center conserves its angular momentum; that is why the eddy with higher angular speed should
have a smaller radius, which is observed at the second stage [8, 26].
The third stage, that of quasi-stationary mature cyclone, lasts from 12:00 UTC of September 27 to 12:00
UTC of September 28. During this time, the cyclone slightly deepens, the central pressure drops to 992 hPa.
At this stage, the cyclone becomes practically round, with a constant radius R
of 65 km. Finally, the last
stage of filling and fast decay of the cyclone occurs from 12:00 UTC of September 28 to 00:00 UTC of
September 29 when the cyclone approaches the coast and starts landing.
Consider a horizontal space structure of the cyclone at its quasi-stationary stage of development, that is,
at 12:00 UTC of September 27. In Fig. 3, the satellite cloud images are shown along with the 10-m wind
and sea level pressure contours as obtained from the MM5 simulation with a horizontal resolution of 10 km.
In Fig. 3d, the simulated trajectory of the cyclone center is displayed for the entire period of its occurrence
over sea. As it can be seen in the figure, the model produces almost ideal description of location, shape,
size, and wind speed in the cyclone, as compared with the observations.
The cyclone vertical structure, obtained in the model, is described in detail in [1]. It should be noted here
that the structure of the Black Sea cyclone is similar to that of a TC. First, the similarity feature is repre-
sented by primary azimuthal circulation, which was maximum in the lower layer and decreased with height.
The maximum value of azimuthal speed (27 m/s) was reached in the so-called eyes wall, at a distance of
65 km from the eddy center, where also the maximum lifting with a vertical speed of 0.3 m/s and a convec-
tive precipitation rate of 10 cm/day occurred. In the eye, no convection was observed and the wind was
light. The radial wind component (V
= 6 m/s) was directed to the center in the boundary layer, where air
converged from the periphery to the eyes wall, and from the center with V
= 3 m/s in the upper tropo-
sphere, in the outflow zone. Also, the Black Sea cyclone, like the tropical one, had a warm core (DT = 3C),
236 EFIMOV et al.
Fig. 2. (a) Time changes of (1) pressure in the cyclone center (hPa) and (2) maximum wind speed at the 10-m level over the
surface (m/s); (b) time variations of maximum wind speed radius (km).
high values of humidity (15 g/kg) and of equivalent potential temperature (336 K) in the lower layer in the
center and in the eyes wall.
As it can be seen from the results presented, at the mature stage, the cyclonic vorticity of the surface
wind reaches high values of (12) 10
, which causes high cyclonic vorticity of the tangential wind
stress t at the surface. As a result of Ekman divergence in the upper layer of sea water, the vertical speed w,
as calculated from the standard formula w f =
t rot( / ), where r is the water density and f is the Coriolis
Fig. 3. (a) Cloud satellite image; the distribution of (b) 10-m wind speed (m/s) and (c) sea surface pressure (hPa) from the
MM5 simulation from 12:00 UTC of September 27, 2005, and (d) the cyclone trajectory over the Black Sea. (1) September
25, 00:00 UTC; (2) September 25, 12:00 UTC; (3) September 26, 21:00 UTC; (4) September 28, 6:00 UTC; (5) September 28,
21:00 UTC; (6) September 29, 12:00 UTC.
parameter, reaches 10
m/s (Fig. 4a) at the lower boundary of the mixed layer. Note that typical seasonal
mean values of the vertical speed in the deep part of the Black Sea, after different estimations, are (12)
m/s. As a result of the high vertical velocities, the thermocline exhibits a very fast ascent and even
reaches the sea surface. This patch of cold water under the atmospheric cyclone is clearly seen in Fig. 4c,
in which, the sea surface temperature distribution is represented from the satellite measurements for 19:00
UTC of September 29. In the patchs center, the temperature is more than 10C below its undisturbed value,
which implies that waters from about 30 m depth appear at the surface. The cold water patch generated by
the cyclone remained for a long time (about a month). So, on September 29 and October 13, the temperature
differences are 14C and 34C, respectively, and even on October 23 the difference remained at the level
of 12C. In Fig. 4b, the cross-section of sea surface temperature is shown across the center of the area
covered by the cyclone on October 6; the temperature difference then reached 8C.
Apart from temperature changes, the atmospheric cyclone caused a local lowering of the sea level for
about 25 cm. In Fig. 4b, the sea level difference between October 6 and September 16 is shown for the same
cross-section as for the sea surface temperature on October 6. The estimate is obtained from the satellite
altimeter data. It is in agreement with preliminary results of numerical modeling of the Black Sea circula-
tion for this time period. The Ekman divergence in the sea mixed layer removed warm surface water to the
semicircular southwest coast and thus heightened the sea level there, while lowering it under the cyclone.
The sea level difference between the coast and the west circulation center, typical of this season (20 cm),
increased by other 25 cm due to the atmospheric cyclone, which led to doubling the speed of the Main
Black Sea Current (to the value exceeding 1 m/s) in the southwestern part of the sea.
Thus, the comparatively shortly living, but of compact intensity, cyclonic eddy caused strong response
of thermal and dynamic structure of the upper sea layer, which lasted for long enough time.
238 EFIMOV et al.
Fig. 4. (a) Distribution of Ekman vertical velocity (m/s) at 12:00 UTCof September 27, 2005 ; (b) the sea surface temperature
cross-section through the cyclone center on October 6 (1); the cross-section of sea level differences between October 6 and
September 16 (2) from the satellite data; (c) temperature of the Black Sea surface from the satellite data on September 29,
2005, at 19:13 UTC.
A relatively rare mesoscale phenomenon, anomalous for the Black Sea, is described: a cyclonic eddy
with all features of a tropical cyclone. In [1], it has been shown that the eddy can be classified as a tropical
cyclone because of a number of main features, including the circulation structure and the energy-supply
mechanism, consisting in release of large quantity of latent heat in the convection area due to condensation
of moisture obtained through evaporation from the warm sea surface. This is confirmed by a series of
numerical sensitivity experiments, carried out using the MM5 model. The numerical experiments with
switching off the latent heat release through condensation and with switching off or limiting the surface
heat fluxes gave a negative result the cyclone did not develop at all or turned out to be much weaker that
the observed one.
So, generation of eddies, similar to tropical cyclones in structure and mechanism, but obviously weaker,
is possible over the midlatitude sea if the sea surface temperature is significantly below 27C.
1. V. V. Efimov, M. V. Shokurov, and D. A. Yarovaya, Numerical Modeling of a Quasi-Tropical Cyclone over the
Black Sea, Izv., Fizika Atmos. Okeana, No. 6, 43 (2007) [Izv., Atmos. Oceanic Phys., No. 6, 43 (2007)].
2. A. P. Khain, Mathematical Modeling of Tropical Cyclones (Gidrometeoizdat, Leningrad, 1984) [in Russian].
3. A. P. Khain and G. G. Sutyrin, Tropical Cyclones and Their Interaction with the Ocean (Gidrometeoizdat,
Leningrad, 1983) [in Russian].
4. R. A. Anthes, Tropical Cyclones: Their Evolution, Structure and Effects, Meteorol. Monogr., No. 41 (1982).
5. H. Billing, I. Hauff, and W. Tonn, Evolution of Hurricane-like Cyclone in the Mediterranean Sea, Beitr. Phys.
Atmos., 56 (1983).
6. J. G. Charney and A. Eliassen, On the Growth of the Hurricane Depression, J. Atmos. Sci., 21 (1964).
7. G. C. Craig and S. L. Gray, CISK or WISHE as the Mechanism for Tropical Cyclone Intensification, J. Atmos.
Sci., 53 (1996).
8. K. A. Emanuel, An AirSea Interaction Theory for Tropical Cyclones. Part I. Steady State Maintenance, J. Atmos.
Sci., 43 (1986).
9. K. A. Emanuel, Atmospheric Convection (Oxford University Press, 1994).
10. K. A. Emanuel, Genesis and Maintenance of Mediterranean Hurricanes, Adv. Geophys., 2 (2005).
11. K. A. Emanuel, The Theory of Hurricanes, Annu. Rev. Fluid Mech., 23 (1991).
12. K. A. Emanuel, Tropical Cyclones, Annu. Rev. Earth and Planet. Sci., 31 (2003).
13. K. A. Emanuel and R. Rotunno, Polar Lows as Arctic Hurricanes, Tellus, 41A (1989).
14. J. A. Ernst and M. Matson, A Mediterranean Tropical Storm? Weather, 38 (1983).
15. E. A. Hendriks, M. T. Montgomery, and C. A. Davis, Role of Vortical Hot Towers in the Formation of Tropical
Cyclone Diana (1984), J. Atmos. Sci., 61 (2004).
16. V. Homar, R. Romero, D. J. Stensrud, et al., Numerical Diagnosis of a Small, Quasi-tropical Cyclone over the
Western Mediterranean: Dynamical vs. Boundary Factors, Quart. J. Roy. Meteorol. Soc., 129 (2003).
17. K. Lagouvardos, V. Kotroni, S. Nickovic, et al., Observations and Model Simulations of a Winter Sub-synoptic
Vortex over the Central Mediterranean, Meteorol. Appl., 6 (1999).
18. R. Mayengon, Warm Core Cyclones in the Mediterranean, Mar. Weather Log, 28 (1984).
19. L. Pytharoulis, G. C. Craig, and S. P. Ballard, Study of the Hurricane-like Mediterranean Cyclone of January
1995, Phys. Chem. Earth (B), 24 (1999).
20. L. Pytharoulis, G. C. Craig, and S. P. Ballard, The Hurricane-like Mediterranean Cyclone of January 1995,
Meteorol. Appl., 7 (2000).
21. E. Rasmussen, A Case Study of a Polar Low Development over the Barents Sea, Tellus, 37A (1985).
22. E. Rasmussen and J. Turner, Polar Lows. Mesoscale Weather Systems in the Polar Regions (Cambridge,
Cambridge University Press, 2003).
23. E. Rasmussen and C. Zick, A Subsynoptic Vortex over the Mediterranean with Some Resemblance to Polar
Lows, Tellus, 39A (1987).
24. O. Reale and R. Atlas, Tropical Cyclone-like Vortices in the Extratropics: Observational Evidence and Synoptic
Analysis, Weather Forecasting, 16 (2001).
25. R. J. Reed, Y.-H. Kuo, M. D. Albright, et al., Analysis and Modeling of a Tropical-like Cyclone in the
Mediterranean Sea, Meteorol. and Atmos. Phys., 76 (2001).
26. R. Rotunno and K. A. Emanuel, An Air-sea Interaction for Tropical Cyclones. Part II: Evolutionary Study Using
Nonhydrostatic Axisymmetric Numerical Model, J. Atmos. Sci., 44 (1987).
Reproducedwith permission of thecopyright owner. Further reproductionprohibited without permission.