In order to cover a substantial amount of stratospheric intrusions during the project Influence of Stratosphere-Troposphere Exchange in a Changing Climate on Atmospheric Transport and Oxidation Capacity (STACCATO), coordinated measurements were carried out at several locations in central and southern Europe, based on quasi-operational stratosphere-troposphere exchange forecasts. In this context, lidar measurements of tropospheric ozone were performed at Thessaloniki (23°E, 40.5°N), Greece, for the investigation of stratosphere-to-troposphere transport (STT) events over the southeastern Mediterranean region, during 2000–2002. The study of STT in this area is of particular interest not only because of its geographic location, which is more southern than the typical position of the polar front jet, but also because of the sparseness of detailed studies in this area. A summary of the main characteristics of the STT events that were detected by lidar during the investigation period reveals a tropospheric ozone increase of the order of 10% between 4.5 and 6.5 km above sea level, a coincident decrease in relative humidity, and elevated values of potential vorticity, thus providing a direct indication of the occurrence of stratospheric air in the middle troposphere. No direct effect on surface ozone was recorded. Furthermore, the majority of the STT cases detected reveal a common pathway of the stratospheric air masses that reached Thessaloniki originating from the North Sea. In addition, the event of 9 January 2001, during which the clearest and strongest descent of stratospheric air occurred, is further analyzed. An ozone-rich layer of the order of 60–90 ppbv between 5 and 6.5 km above sea level is clearly depicted, which was the result of the intermixture of originally stratospheric and boundary layer air originating from North America.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
 The increase of tropospheric ozone concentration has been the topic of numerous investigations during the past years [e.g., Bojkov, 1986; London and Liu, 1992; Logan, 1994; De Muer et al., 1995; Oltmans et al., 1998] as it is considered to be extensively influenced by human activities. Analysis of near-surface ozone data, mainly in the Northern Hemisphere, has shown in many cases episodes of high ozone concentrations, which can either, result from local or long-range transport mechanisms, or be of stratospheric origin. In the latter case, stratospheric air can get irreversibly mixed into the troposphere and thus cause an increase even of surface concentrations after isentropic transport across the tropopause. Such transport is usually followed by non-conservative processes such as diabatic cooling or heating and small-scale turbulent mixing along the surface of the intruded stratospheric filaments [Holton et al., 1995].
 However, only a few studies investigated so far the role of STT to the tropospheric ozone budget in the eastern Mediterranean, which is a region of great interest as it presents among the highest levels of background tropospheric ozone around the globe [Zerefos et al., 2002]. In southeastern Europe and mainly in the eastern Mediterranean, the ozone phytotoxicity EU limit (32 ppbv) is exceeded throughout the year and the human health protection limit (53 ppbv) is exceeded most of the time during summer, as evidenced from both measurements at rural stations and model calculations [Kourtidis et al., 2002; Kouvarakis et al., 2002; Zerefos et al., 2002]. Long-range transport of ozone is manifested in air masses originating from northern directions and passing over sunlit SE Europe and the eastern Mediterranean. These air masses being rich in ozone and precursors form a regional ozone background, which, together with the emissions from local sources, maintain regional ozone levels between the highest in Europe. Recent results from the PAUR II project confirm that a large portion of this ozone is beyond local emission controls and even very drastic reductions of local emissions will not result in attainment of the phytotoxicity limit and the health protection limit [Zerefos et al., 2002; Kourtidis et al., 2002; Kouvarakis et al., 2002]. Hence, this region of SE Europe presently appears to be the most problematic area in Europe for controlling regional background tropospheric ozone. These results underline the significance of the understanding and quantification of the natural ozone controlling mechanisms, such as STT, to the tropospheric ozone budget in the climatically sensitive area of eastern Mediterranean.
 In addition the study of STT in this region is of particular interest because of its location to the south of the typical position of the polar front jet and along the axis and at the end of a relatively frequent path way of stratospheric intrusions [Stohl et al., 2000; Gerasopoulos et al., 2001], but also due to the sparseness of relevant detailed studies in this area. In previous studies, Varotsos et al.  linked the existence of ozone layers in vertical ozone profiles above Athens with stratospheric intrusions over Greece, while Kentarchos et al. [1998, 1999] first identified and simulated such an event for this geographical area using a coupled chemistry-general circulation model. Earlier, Ancellet et al.  speculated on the possible impact of STT over Greece arising from the development of stratospheric filaments moving from central to southeastern Europe based on forward trajectory analysis.
 The current study presents an evaluation of the STT events that reached Greece during the past two years (2000–2002) and were measured by the AUTH (Aristotle University of Thessaloniki) differential absorption lidar (DIAL) system. Additionally, a detailed description of one of the events is presented. This study provides for the first time a more detailed picture of STT over the Eastern Mediterranean region.
2. Measurement Data
 In the framework of the EU project STACCATO (Influence of stratosphere-troposphere exchange in a changing climate on atmospheric transport and oxidation capacity), coordinated measurements at different stations in central and southeastern Europe, were carried out, in order to cover a substantial amount of STT cases. Hence, quasi-operational stratosphere-to-troposphere transport forecasts has been calculated daily since November 2000, at ETH Zurich, using forecast data from the ECMWF. This provided helpful information to the measurement groups for the set-up and planning of measurement periods. Trajectories, which have been calculated for the Atlantic/European region, reside initially in the stratosphere i.e. with potential vorticity (PV) values larger than 2 potential vorticity units (pvu, 1 pvu = 10−6 m2 K kg−1 s−1), and descend within a period of four days by at least 300 hPa into the troposphere. The graphical results were then disseminated to the experimental STACCATO partners. The principal goal of the measurements was to provide comprehensive data on stratosphere-to-troposphere intrusions, along their rather frequent pathway from the North Sea to the Mediterranean Sea, passing over Central Europe (P. Zanis et al., Forecast, observation and modeling of a deep stratospheric intrusion event over Europe, submitted to Atmospheric Chemistry and Physics, 2003).
 The differential-absorption lidar measurements of ozone performed at Thessaloniki (23°E, 40.5°N, 60m asl) were carried out by the use of the AUTH-DIAL system [Papayannis et al., 1999]. The lidar system is based on a frequency quadrupled pulsed Nd:YAG laser and the Raman shifting technique in deuterium gas. It emits simultaneously three wavelengths (266 nm, 289 nm and 316 nm) at the repetition rate of 10 Hz, using a single low pressure Raman cell. The optical receiving system is based on a 50 cm Newtonian telescope, which is directly coupled to a concave grating spectrometer. The system is able to perform ozone measurements in the free troposphere (2–12 km) with high temporal (5–30 min) and spatial resolution (300–1000 m). In order to improve the range of the measurements and to decrease the statistical error, 1–2 hours mean O3 profiles were calculated with accuracy of the order of 5–10% in the troposphere.
 To assist the ozone lidar vertical profile measurements, profile measurements of meteorological parameters, such as temperature and relative humidity, were deduced by the operational radio-sounding measurements performed at Thessaloniki at 12 h UTC by the National Hellenic Meteorological Service. Relative humidity (RH) was used in order to obtain additional indication concerning the origin of the air parcels, even though it should be kept in mind that it cannot be regarded as an ideal stratospheric air tracer since dry air masses may also originate from the upper troposphere.
 The profile measurements were further supported by continuous surface measurements of ozone with a Dasibi 1008 RS ozone analyser and daily measurements of the cosmogenic radionuclide 7Be at Livadi station. Livadi (23°15°E, 40°32°N) is a semi-elevated, rural village located on the eastern peak of Mt. Hortiatis, at an elevation of 850 m asl, situated 50 km from the city of Thessaloniki.
 In addition to the measurements, the 500 hPa geopotential height maps were examined in order to understand the prevailing weather conditions during the cases investigated. Supplementary information was obtained from total column ozone maps from the World Meteorological Organization (WMO) Total Ozone Mapping Center hosted by the Laboratory of Atmospheric Physics at Thessaloniki.
 PV is a semi-conservative stratospheric tracer and thus high PV values associated with lower stratosphere-upper troposphere region are not expected to survive into the lower troposphere due to mixing. ECMWF global analysis data were used to calculate on the PV 310 K and 315 K isentropic surfaces were used, as well as vertical cross section of PV. To decide on the stratospheric origin of the air masses, the dynamic definition of the tropopause based on a threshold value was used. Threshold values reported in the literature range from 1.0 pvu [Danielsen, 1968] to 3.5 pvu [Hoerling et al., 1991] depending partly on the synoptic situation and the geographical location. We used the rather common value of 1.6 pvu [Stohl et al., 2000]. In addition to PV, vertical cross sections of RH and O3, also from ECMWF analysis data, were also inspected, since both are considered as useful stratospheric tracers [e.g., Elbern et al., 1997].
 Finally, three-dimensional 3 or 5-day back trajectories calculated with ECMWF global analysis data were also used to trace the origin of air masses. Air masses that arrived at Thessaloniki from 200 to 800 hPa were examined, with vertical resolution of 100 to 25 hPa, depending upon the case investigated.
3. Characteristics of STT Events Detected at Thessaloniki by Lidar During 2000–2002
 By the use of the aforementioned measurements and based on the fact that stratospheric air has certain characteristics, which can be used for its detection in the troposphere, a description of the performed lidar measurements and a summary of the main characteristics of the STT events that were detected by the AUTH lidar system, during the 2000–2002 period, is attempted.
 During the two years of the STACCATO project 45 warning alerts were issued for Thessaloniki, based on the quasi-operational stratosphere-to-troposphere transport forecasts. Since clear sky conditions are necessary for the performance of lidar measurements, these were not feasible for every warning. In 24 of them O3 DIAL measurements were performed and 8 ozonesondes were launched at Thessaloniki during the period under investigation. Additionally, 46 daily measurements of 7Be were carried out at Livadi together with continuous surface ozone measurements. Radiosonde data from Thessaloniki were available for almost all warnings.
 Detailed examination of the lidar ozone profiles revealed 8 cases during which the tropospheric ozone profiles showed an increase, in agreement with the STT forecast. As seen in Figure 1, all of these lidar measurements were performed from November until March. The tropospheric ozone profiles, from 2 to 10 km height, reached a maximum value of 80 ppbv (at 5.6 km) in one case (Figure 1c) whereas in all the others did not exceed the value of 70 ppbv. A tropospheric ozone increase and a coincident decrease in relative humidity is evident during almost all cases, thus indicating the intrusion of stratospheric air into the troposphere, as further investigated below.
 During these 8 warning alerts 7Be concentrations generally lay within 1σ of the monthly climatological mean values derived from previous measurements at Thessaloniki over the last 14 years and only for one case (Figure 1a) the measured concentrations exceeded 2σ with a value higher than 8 mBq m−3, which is a typical threshold value for intrusion events [Sladkovic and Munzert, 1990; Scheel et al., 1999; Zanis et al., 1999; Stohl et al., 2000]. Less prominent ozone-peak structures over the night were also found in a few cases, but since such peaks are not very uncommon during the two years period, it is not easy to discriminate whether they are due to intrusion events or horizontal transport processes. In general, during these 8 warnings no clear effect on surface measurements at Livadi station was revealed and this is expected if one takes into account the altitude that the layers of ozone rich dry air were observed.
 The 500 mb geopotential height maps were also examined for these 8 cases, in order to conclude on the prevailing synoptic situation. Greece was affected mostly (in 4 cases) by an upper trough over north Europe moving towards south-southeastern Europe and during 2 cases by being at the southern tip of an upper-level cold trough that moved from the north Atlantic cyclonically over central Europe. This result is also depicted in Figure 2, where a typical example (22 Feb 2001) of the pathway of the stratospheric air masses that reached Thessaloniki is presented. Finally, only in 2 cases the ozone increase could be attributed to a cut-off low system that was present over the Balkans.
 However, in order to conclude with certainty whether during the cases presented in Figure 1 an intrusion of stratospheric air into the troposphere did occur, the origin of the air masses that reached Thessaloniki at the time of the lidar measurements had to be carefully examined. Back-trajectory calculations were performed for every case and the main results of its correlation with the vertical profiles in Figure 1 are presented in Table 1. In two cases (Figures 1d and 1e) the trajectories calculated with the analysis fields did not confirm the STT forecasts that gave some sign of a possible intrusion. The increase in ozone was attributed to the ascent of low tropospheric air, probably rich in anthropogenically produced ozone. In addition, in two other cases (Figures 1a and 1g) the increased ozone concentration was due to the descent of upper tropospheric air, underlining again the difficulty in capturing predicted STT events due to their short lifetime. However, it is evident that during the remaining four cases (Figures 1b, 1c, 1f, and 1h) intrusion of stratospheric air took place with clearly increased ozone, decreased relative humidity and high potential vorticity values.
Table 1. Correlation Between the O3 - RH Profiles Presented in Figure 1 and Information Based on Coincident Air Mass Back Trajectory Analysis
- quazi-horizontal motion in middle and upper trop. (5–10 km)
- descent from upper trop. (7.7 km) to the lower (2.5–4 km), depicted as a decrease in RH (lower than 15%)
29 November 2000
- strong ascent of lower tropospheric air (below 2 km) to the upper trop. (7–8 km)
- descent from upper trop. to 4.5–5 km and from the strat. to 5–6 km. Both were depicted as ozone-rich layers with RH values lower than 25% and high PV values
9 January 2001
- strong ascent of boundary layer air (below 1.5 km) to the upper trop. (6–7.2 km)
- very clear descent from the strat. to the middle trop. (4.6–5.6 km), depicted as an intense increase in ozone and PV with RH values lower than 10%
- descent from the upper to the low trop. (3.5–4.5 km)
4 February 2001
- ascent of low tropospheric air (below 2 km) to the upper and middle trop. (4–9 km), depicted as an increase in ozone. Surface air reached Thessaloniki at 7.7 km and was depicted as an intense ozone increase
- quazi-horizontal motion in the low trop. (below 3 km)
13 February 2001
- ascent of low tropospheric air (below 2 km) to the upper and middle trop. (4–9 km), depicted as an increase in ozone. Surface air reached Thessaloniki at 5.6 and 8.7 km and had the same effect in ozone with relatively high humidity values (up to 55%)
22 February 2001
- descent from the strat. to 5.5–7 km, depicted as an increase in ozone and PV with RH values lower than 10%
- strong ascent of boundary layer air (below 1.5 km) to the middle trop. (4–5 km) depicted as an increase in RH
- strong descent from the upper to the low trop. (3–4 km) depicted as an intense decrease in RH (below 15%)
15 March 2001
- ascent from the low trop. (2 km) to the upper trop. (6–9 km), depicted as an increase in ozone
- ascent from boundary layer air to the low trop. (3–4 km)
- strong descent from the upper to the low trop. (2 km)
22 November 2001
- descent from the lowermost strat. (8.7 km) to the middle trop. (4.2–6 km) depicted as an ozone increase with low RH (lower than 25%) and relatively high PV values
- descent from the middle (5.6 km) to the lower trop. (3 km) with relatively high RH values
 In order to comment on the common characteristics of these 4 detected STT events their ozone and relative humidity profiles were averaged, after being normalized, so as to remove the seasonal variation of tropospheric ozone background. The results are presented in Figure 3a, where a single ozone-rich layer in the middle tropospheric region, from 4.5 to 6.5 km, is clearly depicted. Although in previous studies [Eisele et al., 1999] descending ozone-rich structures in lidar were compared and it was found that most of them reached a level of 3 km, the observed ozone-rich layer did not reach altitudes lower than 4.5 km and was approximately up to 2 km in width. The ozone concentrations at that level exhibited more than 10% increase and no secondary maxima was found.
 In Figure 3b the average vertical profiles of relative humidity on the same, previous and next day of the lidar measurements, are presented for comparison to be easily visualized. It is evident that during the STT days (days when lidar measurements were performed) the relative humidity profiles depicted a sharp decrease above 4.5 km height, which is in good agreement with the lower height of the observed ozone layers. The decrease in the relative humidity profile is very pronounced during STT (with values that stay below 25%), while for the previous and next days the average RH values were around 50%. The low humidity values at the level where ozone-rich layers were observed provide us with strong evidence of the stratospheric origin of the air masses at that level. Figure 3b confirms also the short duration of the STT events detected at Thessaloniki, since the relative humidity profiles 24 hours before and after the examined days do not show any disturbance. This was not unexpected since stratospheric air is not expected to stay more than a day, or even a few hours. In addition Thessaloniki is typically at the end of the intruding tongues of stratospheric air and thus the ozone-rich layers were more vulnerable to dilution or mixing with the ambient air.
4. Case Study: 9 January 2001
 Taking into account that the troposphere over southeastern Mediterranean was affected by STT events during the investigation period (2000–2002) and in order to understand the parameters that may contribute to the evolution of these episodes, one of the aforementioned 4 events is described in more detail. It occurred on 9 January 2001 and, according to Table 1 it was the “clearest” of these 4 STT events, since it was characterized by a particularly pronounced descent of the stratospheric air.
 On 9 January 2001, 20:00 UTC an ozone-rich layer was observed below Thessaloniki's tropopause, giving direct evidence for the evolution of a stratospheric intrusion. As seen in Figure 4 a distinct layer of increased ozone density was clearly depicted from 5 to 6.5 km, with ozone values that reached 90 ppbv, when the average tropospheric ozone concentration that day was approximately 60 ppbv. Radiosonde measurements revealed that although one day before the event the relative humidity was more than 40%, on 9 January the layer from 5–6.5 km - characterized by higher ozone concentration in the lidar profile - had a distinct low humidity level, with values below 10% (Figure 4, dashed line).
 Weather maps revealed the evolution of an upper trough at 500 hPa, which moved from northern UK towards central Europe and finally lost its strength, leaving Europe with approximately uniform pressure conditions. During this movement it had a direct effect on the total ozone values, which was evident in World Meteorological Organization total ozone maps. The relative difference of total ozone from the previous day is presented in Figures 5a and 5b, for 8 and 9 January, respectively. In Figure 5a, a 20% increase ozone band was over central Europe and northern Italy, affecting Greece. The absolute difference from the previous day (not shown) depicts that during that day the total ozone over Thessaloniki increased more than 50 Dobson Units. This increase was also evident during the following day (Figure 5b), when the band had moved northeasterly but its remains stayed over the Aegean Sea and caused an additional 10% increase in total ozone value.
 The strong pressure gradients present on 9 January especially on the southeastern flank of the cut-off low, were associated with a large undulation of the jet stream on the 310 K isentropic surface. The trough was quite strong and continued its movement over Europe without breaking into vortices, although its southern tip seems to have been cyclonically influenced. After examining in detail the corresponding PV maps it was apparent that the trough with PV values higher than 1.5 pvu reached only the north of Greece on 9 January 2001. Its southernmost location was observed in the evening of the day and afterwards it started to turn cyclonically over Eastern Europe. Within 18 hours the trough had ended its southward movement at that level and had already started moving northeasterly.
 The south-north (latitudes from 30° to 50° N) vertical cross-sections of PV at 12:00 and 18:00 UTC on 9 January revealed the development of a deep tropopause fold down to 750 hPa at 37°N crossing the latitude of Thessaloniki at a level between 550 and 650 hPa (Figure 6), which corresponds to the altitude where high ozone values were measured with lidar (Figure 4). Figure 6b indicates that PV exceeded the 2 pvu threshold down to 450 hPa height at about 42°N and the 1.5 pvu threshold down to 550 hPa height at about 41°N, thus implying intrusion of stratospheric air down to 550 hPa. Above 40°N the PV ranges from 1 to 1.5 pvu at the layer between 550 and 650 hPa indicating that air of stratospheric origin has been mixed with tropospheric air. In both Figures 6a and 6b, the low humidity line (dash and dotted line) is in agreement with the radiosonde results in Figure 4, confirming that an extremely dry layer of air was present over Thessaloniki at the level of 600–400 hPa on 9 of January.
 In order to investigate the origin of the air masses 3-D air mass trajectories were examined (Figure 7). As seen in Figure 7a the ozone-rich air had originated from North America, reached the Balkans in four days, after their pass over the Atlantic and North Europe. The situation was complex since there was strong evidence that the air masses of originally stratospheric and boundary layer air were mixed. This fact led to the detailed study of the trajectories with minimum spatial step 25 hPa and for 96 hours before their arrival at Thessaloniki. Figure 7b shows that the air masses that reached Thessaloniki at 500–600 hPa had indeed stratospheric origin (bold lines). Lower than 600 hPa and up to 700 hPa the air masses arrived after their descent in upper troposphere (moderate lines). At that layer the air had ascended from the low troposphere (800 hPa). From 400 to 475 hPa the air had had a rapid ascent from the boundary layer to the stratosphere (from 980 hPa to 280 hPa) and finally descended, having thus characteristics of both boundary and stratospheric air. The strong mixing between the troposphere and the stratosphere had as a result the occurrence of the ozone-rich layer, with both stratospheric (natural) and boundary layer (presumably anthropogenic) origin.
 Enhanced levels of tropospheric ozone have been detected in the Southeastern Mediterranean region, although its location is southern than the typical position of the polar front jet. Due to the sparseness of detailed studies in this area the current study of STT events is of particular interest.
 A detailed study of lidar measurements based on quasi-operational stratosphere-to-troposphere transport forecasts revealed 4 STT events. All were characterised by a tropospheric ozone increase of the order of 10% between 4.5–6.5 km, coincident decrease in relative humidity (below 25%) and elevated values of PV, giving thus direct indication of the occurence of stratoshperic air. No direct effect on surface ozone was recorded. The short duration of all STT events detected at Thessaloniki was evident, since the relative humidity profiles 24 hours before and after the examined cases did not show any disturbance.
 The STT event on 9 January 2001 was chosen for further investigation, since it was the one with the clearest and strongest descent of stratospheric air. Thessaloniki was along the eastern flank of a low-pressure system present over north-central Europe. Lidar measurements depicted a well-defined enhanced ozone layer at 5–6.5 km with peak values above 90 ppbv. The vertical cross section of PV revealed relatively high PV values which did not reach areas south of 40°N. 3-D trajectories demonstrated intermingling of originally stratospheric and boundary layer air originating form North America. Contrary to the fact that air from 500 to 600 hPa had clearly stratospheric origin, air at 400–475 hPa had both stratospheric and boundary layer characteristics. The strong mixing between tropospheric and stratospheric air had as a result the occurrence of the observed ozone rich layer at Thessaloniki, with both natural and presumably anthropogenic origin, underlining the need for systematic ozone DIAL measurements over the Eastern Mediterranean region, in order to enable the quantification of the role of STT in the tropospheric ozone budget of the area.
 This study was carried out within STACCATO (contract EVK2-CT1999-00050), a project funded by the European Commission under the Fifth Framework Programme.