1 Ianos - A hurricane in the Mediterranean

3 Lagouvardos Ka, A. Karagiannidisa. S. Dafisa,b, A. Kalimerisc, and V. Kotronia

4 a
Institute for Environmental Research, National Observatory of Athens, Greece

5 b
LMD/IPSL, CNRS UMR 8539, École Polytechnique, Université Paris Saclay, ENS, PSL Research University,

6 Sorbonne Universités, UPMC Univ Paris 06, Palaiseau, France

7 c
Laboratory of Environmental Physics, Energy, and Environmental Biology, Department of Environment,

8 Ionian University, Zakynthos, Greece

9
10
11
12

13 Corresponding author: Dr. Kostas Lagouvardos, [email protected]

14

15

Early Online Release: This preliminary version has been accepted for
publication in Bulletin of the American Meteorological Society, may be fully cited,
and has been assigned DOI 10.1175/BAMS-D-20-0274.1. The final typeset
copyedited article will replace the EOR at the above DOI when it is published.

© 2021 American Meteorological Society

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC
16 ABSTRACT

17 During 15-21 September 2020, an intense medicane, named Ianos, formed over the warm

18 Mediterranean Sea. Following a path of approximately 1900 km, Medicane Ianos affected Greece

19 resulting in four casualties and devastating damages in the western and central parts of Greece.

20 Persistent gale force 1-minute winds up to 44 ms-1 and wind gusts up to 54 ms-1 were recorded in

21 Cephalonia island (Ionian Sea), while record-breaking amounts of accumulated rainfall have been

22 recorded in several Ionian Islands, as well as in parts of Central Greece. Analysis of the available

23 observations showed that Ianos was the most intense medicane ever recorded in the Mediterranean.

24 This paper aims at investigating the genesis and evolution of the medicane, based on in situ

25 observations, satellite measurements and model analyses. Towards that objective, Meteosat Second

26 Generation (MSG) SEVIRI imagery, combined with lightning data permitted to follow the

27 evolution of convective activity during the various phases of Ianos. This investigation is

28 complemented with upper-air model analyses in order to evaluate the synoptic environment within

29 which Ianos has formed and was sustained during 7 days. Finally, the Global Precipitation Mission

30 Core Observatory satellite (GPM-CO) overpasses over Medicane Ianos provided invaluable

31 information about its 3-D structure, especially during its most intense phase.

32

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
33

34 1. Introduction

35 The Mediterranean Sea occasionally hosts intense cyclones with characteristics similar to tropical

36 cyclones, such as a frontless and axisymmetric structure, deep convection close to their centers with

37 a windless “eye” and sustained wind speeds close to hurricane strength. Such cyclones are often

38 named “medicanes”. Although most of their lifetime is spent over the sea, when they reach the

39 coasts of Mediterranean countries they can be very destructive. Recent case studies have

40 investigated the formation and intensification mechanisms of these cyclones which are mainly

41 attributed to baroclinic processes (e.g., Mazza et al., 2017; Fita and Flaounas, 2018) and to intense

42 surface fluxes (Miglietta and Rotunno, 2019). Noyelle et al. (2019) suggested that the higher the

43 sea-surface temperature (SST) during a medicane, the higher is the probability for tropical transition

44 and thermodynamic intensification of the medicane, similarly to tropical cyclones. Thus, in a

45 warming climate there is a special concern about the future changes of medicanes’ intensity and

46 frequency of occurrence, with several independent studies showing an increasing trend of stronger

47 medicanes in the near future (e.g., González‐Alemán et al., 2019; Koseki et al., 2020).

48 Between 2016 and 2020, four strong medicanes have developed in the Ionian Sea (Dafis et al.,

49 2020) and have caused damage in Greece, with the most recent and most destructive case being the

50 Medicane Ianos between 15 and 20 September 2020. In October 2016, Medicane Trixie battered the

51 south coasts of Greece with excessive rainfall up to 180 mm in less than 12 hours in Crete and wind

52 gusts stronger than 30 ms-1. One year later, in November 2017, Medicane Numa hit the west coasts

53 of Greece resulting in severe flooding, landslides and excessive precipitation. Medicane Zorbas in

54 September 2018 was one of the most costly weather-related disaster in Greece in the past decade,

55 bringing hurricane-force wind speeds in South Greece and causing severe flooding in Central

56 Greece (up to 506 mm of rainfall in 48 hours).

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
57 Ianos initially emerged as a surface cyclone close to a thunderstorm cluster in the Gulf of Sidra on

58 15 September 2020, off the coasts of Libya (Fig. 1). Global and regional model guidance showed a

59 highly possible transition of the surface cyclone to a powerful medicane, a fact that led the METEO

60 unit at the National Observatory of Athens (NOA) to name the medicane “Ianos” in the morning of

61 16 September, following NOA’s practice adopted since 2017 to name weather events that are

62 expected to produce high socio-economic impacts (https://www.meteo.gr/namedstorms.cfm). Based

63 on satellite imagery (namely on SEVIRI imagery, various scatterometers’ data and GPM-CO data),

64 Ianos was evolved into a powerful medicane on 17 September 2020 when it started gradually to

65 affect Greece with storm surges accompanied by high coastal waves, high rainfall amounts and

66 devastating winds, that resulted in four fatalities and extended infrastructure damages and landslides

67 in the Ionian Islands and in Central Greece. Zekkos et al. (2020) reported that the torrential rainfall

68 resulted in a large number of landslides and debris flows and more specifically in Central Greece

69 the authors mapped more than 1400 landslides in 2 days.

70 The objective of this study is to highlight important aspects of the development of this high-impact

71 medicane based on the analysis of the available in-situ and satellite observations along with model

72 derived field aiming to better understand its formation and evolution mechanisms. The physical

73 processes behind the formation and behavior of medicanes have been intensively studied during the

74 last decade (e.g., Tours and Romero, 2013; Picornell et al., 2014; Fita and Flaounas et al., 2018). As

75 the observational and modeling tools are continuously updated and improved, the scientific

76 community is able to utilize new information towards improved understanding of medicanes. Based

77 on previous studies regarding the intensity and impact of medicanes (Winstanley, 1970; Miglietta et

78 al., 2013 – their Table 1; Cioni et al., 2016; Dafis et al., 2020 – their Supporting Information S1),

79 our analysis offers insights into the formation and evolution on the strongest medicane ever

80 recorded. The main novel element of the present work is the use of the state-of-the-art observational

81 tools (like the GPM-CO Microwave Imager and the Dual-frequency Precipitation Radar) which can

82 be used to analyze the convective nature of Medicane Ianos in high detail. A secondary, yet

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
 83 important, novelty is the fact that it is actually the first time that GPM-CO captures a medicane in

84 its mature state, allowing us to observe its shape, cloud bands and precipitation regime at that stage.

85 The added value of the analysis, is mostly the new information gained from these observations,

86 which can definitely improve our knowledge and help us prepare against and mitigate the effects of

87 such high-impact weather phenomena.

88 2. Data and tools

89 The METEO unit at NOA operates the NOAAN network of 430 automatic surface weather stations

90 in Greece that is monitoring all basic meteorological variables at 10-min intervals (Lagouvardos et

91 al., 2017). Due to the absence of a homogenized operating weather-radar network, the weather

92 station network allowed the real-time monitoring and analysis of the spatio-temporal distribution of

93 rainfall produced by Ianos. Moreover, an additional local network that covers the Ionian Islands

94 (Kallimeris et al., 2016) provided higher temporal resolution (at 1-min intervals) weather

95 observations of the Ianos’ core.

96 The synoptic conditions during the event were investigated using the Final Operational Global

97 Analysis of the National Center for Environmental Prediction (NCEP FNL). FNL analyses are

98 available at 0.25° x 0.25° spatial and 6 h temporal resolution. Meteosat Second Generation Spinning

99 Enhanced Visible and Infrared Imager (MSG SEVIRI) brightness temperatures on water vapor at

100 6.2μm (WV6.2) and infrared at 10.8μm (IR10.8) channels together with lightning data provided by

101 ZEUS Very Low Frequency (VLF) lightning detection network (Κotroni and Lagouvardos, 2016)

102 permitted to construct the full path of Ianos as well as to scrutinize its convective nature.

103 GPM, initiated by NASA and the Japan Aerospace Exploration Agency (JAXA) as a global

104 successor to TRMM, comprises a consortium of international space agencies (e.g., CNES, ISRO,

105 NOAA, EUMETSAT). It consists of a constellation of research and operational satellites, where the

106 NASA/JAXA GPM Core Observatory (GPM-CO) satellite, carrying the radar / radiometer, serves

107 as a reference standard to unify precipitation measurements. The GPM-CO encompasses a multi-

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
108 channel conical-scanning microwave imager – the GPM Microwave Imager (GMI) (Newell et al.,

109 2015; Draper et al., 2015) and a Dual-frequency Precipitation Radar (DPR) (Kojima et al., 2012;

110 Furukawa et al., 2016). GMI has 13 microwave channels that measure brightness temperature (Tb)

111 sensitive to precipitation type ranging from light to heavy rain, snow and ice. The reader can find

112 information about the GMI channels in Hou et al. (2014). DPR consists of a Ku- and a Ka-band

113 frequency radar and can offer 3-D data of precipitation particles inside the cloud systems.

114 Overpasses of the GPM-CO satellite provided invaluable data of the 3-D structure of Ianos during

115 its mature stage.

116 3. Path of Medicane Ianos – in-situ rainfall and wind observations

117 The area of genesis of Medicane Ianos was over northern Libya and the surface cyclone could be

118 first tracked by satellite imagery when it passed over the Gulf of Sidra, at the beginning of its 7 days

119 long life. Figure 1 shows Ianos’ path between 15 and 21 September 2020, when, after a route of

120 approximately 1900 km, it made landfall in the Mediterranean coasts of Egypt. This path was

121 inferred by the combined analysis of sea-level pressure (SLP) provided by FNL and tracking of the

122 cyclonic curvature or the closed circular patterns of clouds based on MSG SEVIRI imagery. Ianos

123 formed over the warm waters of the Gulf of Sidra, where SST exceeded 28°C (Fig. 1) based on

124 satellite observations from the Ultra-high Resolution Sea Surface Temperature Analysis product

125 (Buogiorno Nardelli et al., 2013). According to the daily SST climatology built from 21 years

126 (1985-2005) of AVHRR Pathfinder data (Marullo et al., 2007), the SST anomaly in the Ionian Sea

127 was higher than +2 K. Ianos acquired its tropical characteristics, like the symmetric frontless

128 structure, over the anomalously warm Ionian Sea at around 0300 17 September 2020.

129 Luckily, Ianos’ core was sampled by a weather station in Palliki, Northwest Cephalonia (Fig. 1)

130 while 4 weather stations recorded the evolution of surface pressure, 10-m wind and rainfall within

131 50 km from Ianos’ “eye”. It is worth mentioning that such observations are very rare given the

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
132 small size of medicanes. The previous most recent case when a medicane “eye” was sampled by

133 surface stations was in 2014, in Malta during Μedicane Qendresa (Cioni et al., 2018).

134 Hurricane-force winds, resulted in extended damage in the Ionian Islands (see Fig.1 for the location

135 of the most affected areas), fortunately with no losses of human lives. The minimum SLP of 984.3

136 hPa, as well the maximum wind speed and gust was recorded in Palliki, Cephalonia when at the

137 same time the weather station in Skinari, North Zakynthos, reported 989.1 hPa minimum SLP and

138 42.0 ms-1 and 30.2 ms-1 maximum 1-minute wind gust and mean wind speed, respectively (Fig. 2).

139 These extreme values took place a few hours after a sharp wind direction change during torrential

140 rainfall inside Ianos’ “eyewall” cloud bands between 0000 and 0400 UTC 18 September 2020 (Fig.

141 2). The highest 1-min mean wind speed, derived by wind measurements at a sampling rate of 1 Hz,

142 was 44.1 ms-1 while the corresponding wind gust (maximum recorded wind speed over the same

143 interval) was 54.2 ms-1. When looking at the 10-min average records, the corresponding highest

144 values were 39.3 ms-1 for the mean wind speed and 47.6 ms-1 for the wind gust. In addition, the

145 weather station in the northern-most part of Cephalonia in Antipata (Fig. 1) reported a record-

146 breaking daily accumulated rainfall of 644.7 mm. These values are the highest ever recorded during

147 a medicane and some of the highest ever recorded in Greece.

148 The rainfall distribution in Greece during the 48 hours ending at 1200 UTC 19 September 2020,

149 shows two distinct areas of maximum rainfall (Fig. 3a), one in the Ionian Islands and one in Central

150 Greece (the latter detailed with a zoom in Fig. 3b). The 48-h (17 and 18 September 2020)

151 accumulated rainfall in the Ionian Islands reached 769 mm and 250 mm in Cephalonia and

152 Zakynthos islands, respectively. During the same period, a very strong easterly flow was established

153 over Central Greece, associated with the low-level medicane flow in the Ionian Sea which resulted

154 in local strong low-level convergence zones and long-lasting rainfall with extreme rain rates. On 18

155 September 2020 the daily accumulated rainfall exceeded 250 mm in Central Greece. The high

156 amounts of precipitation in Central Greece resulted in widespread flooding, extensive damage in

157 rural and urban areas, and unfortunately 4 fatalities. Figure 3 shows the 48-h accumulated rainfall at

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
158 the stations of Pertouli (P), Karditsa (K) and Mouzaki (M) in Central Greece (see Fig. 3b for the

159 location of the stations). The highest value was recorded in Pertouli (P), at the station located at the

160 highest elevation among the three stations (1170 m), with a total amount of 317 mm. Mouzaki (M)

161 and Karditsa stations (M and K, respectively) recorded 268 and 212 mm of rain within the same

162 time period.

163

164 4. Satellite imagery

165 Figure 4 provides four snapshots of MSG SEVIRI satellite infrared imagery (IR 10.8 μm) in the

166 time period 16-18 September 2020. Infrared imagery is combined with the brightness temperature

167 difference between the 6.2μm water vapor and the 10.8 μm infrared channels, which can be used as

168 an indication of convective activity (Olander and Velden, 2009; Dafis et al., 2020).

169 At 1200 UTC 16 September 2020 (position #4 in Fig. 1) Ianos had not yet lost its extratropical

170 characteristics, but it was characterized by high convective activity on the western flank of the

171 cyclone center, as inferred by the close to zero and positive values of the WV-IR brightness

172 temperature differences (Fig. 4a). The presence of convection is also supported by the cloud-to-

173 ground (CG) lightning activity observations, provided by ZEUS network (yellow marks in Fig. 4a).

174 The characteristic whirl of clouds is clearly observed 24 hours later, at 1200 UTC 17 September

175 2020 (Fig. 4b). Convection is evident on the western flanks of the medicane (Fig. 4b), but it is

176 slightly weaker than 24 hours earlier. Lightning activity presents a wide spread around the vortex

177 but now it is located further away from the center.

178 At 2230 UTC 17 September (this instance was selected as it almost coincides with the GPM-CO

179 overpass discussed later in Section 6) the cloud walls around the center of the cyclone retain heights

180 reaching the tropopause as suggested by the negative but close to zero or positive values of the WV-

181 IR brightness temperature differences (Fig. 4c). Feidas and Giannakos (2010) state that positive or

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
182 close to zero values of that difference indicate tall opaque clouds approaching or breaking the

183 tropopause barrier. However, lightning activity in the specific instance is absent. .

184 During the period of intense rainfall over Central Greece, at 1800 UTC 18 September (Fig. 4d)

185 medicane Ianos has lost its symmetry and convection is limited mainly over parts of Central Greece,

186 with localized lightning activity. During the next 48 hours, while Ianos was following a path

187 towards Crete and the coasts of Egypt (see Fig. 1) it maintained a quasi-symmetric structure around

188 its centre, with scattered lightning activity, mostly far from its center. The full path of Ianos, as

189 inferred by satellite imagery and concurrent lightning activity is provided in the Supplementary

190 material.

191 5. Synoptic-scale setting of Ianos

192 The synoptic environment, into which Medicane Ianos formed over the Gulf of Sidra, was

193 associated with a cut-off low dominating the Central Mediterranean. The low is clearly identified at

194 the 500 hPa level (Fig. 5a). At this stage (1200 UTC 16 September), as already stated, Ianos was not

195 yet a medicane, as indicated by a pronounced vortex tilt, namely the eastward shift of the surface

196 low-pressure center with respect to the center of the cut-off at 500-hPa. Twenty-four hours later, at

197 1200 UTC 17 September 2020, both low and upper-level lows were vertically aligned, a typical

198 synoptic pattern associated with a mature medicane (Fig. 4b). The same structure is evident at 2300

199 UTC 17 September, when Ianos’ center was offshore the coasts of western Greece (Fig. 4c),

200 moving very slowly towards south (Fig. 5d). It should be noted however, that due to its low spatial

201 resolution FNL analysis is not able to realistically depict the minimum sea-level pressure values

202 within Ianos’ center, which were recorded by the surface weather stations networks, discussed

203 earlier in Section 3.

204 Analysis of the potential vorticity (PV) field at the 335 K isentropic level revealed that an upper

205 level intrusion of dry air of stratospheric origin approached the vortex at 1200 UTC 16 September.

206 This PV streamer approached the area over the low center and wrapped around it (Fig. 6a). These

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
207 upper-level disturbances set the precursor environment for the subsequent evolution of a cyclone

208 and favor its intensification during the first phase of its formation (Cioni et al., 2016; Miglietta et al.

209 2017; Flaounas et al., 2020).

210 At 1200 UTC 17 September, Ianos was already a medicane and the PV structure presents the

211 characteristic alignment throughout the troposphere, with the creation of a PV “tower” in the center

212 of the medicane (Fig. 6b). It is worth mentioning that high upper-level PV values vanished near the

213 area around the medicane center, due to the fact that a major part of it has been eroded at earlier

214 stages by deep convection within the cloud bands around the medicane.

215 Inspection of the effective precipitable water (EPWAT) and integrated vapor transport fields

216 indicate advection of moist air masses related to Ianos. The EPWAT is calculated by the subtraction

217 between the FNL-derived precipitable water and the vertically integrated saturation deficit.

218 Therefore, maps of EPWAT permit to identify the areas with the highest tropospheric moisture

219 content that can lead to high rainfall amounts. During the lifetime of Ianos, extreme values of water

220 vapor in the troposphere were transported from the South Ionian Sea towards continental Greece

221 and high rainfall amounts were reported in many places. Figure 7 shows the integrated tropospheric

222 water vapor transport as well as the EPWAT at the same four instances as in Fig. 5. According to

223 these analyses, the EPWAT exceeded 50 mm offshore Western Greece between 1200 UTC 17

224 September and 0000 UTC 18 September (Figs. 7b, c). A value of 50 mm is considered as extreme

225 for the Mediterranean area while it is more common in the tropics (Sudradjat et al., 2005). In the

226 evening hours of 18 September 2020 (Fig. 7d) high values of EPWAT are also evident over Central

227 Greece. During this timeframe, the highest accumulated rainfall was recorder over this area, as

228 discussed earlier in Fig. 3.

229 6. GPM measurements

10

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
230 a. GMI data

231 During the lifetime of Ianos, there have been five overpasses of GMI covering the area of interest:

232 at 1340 and 2320 UTC 16 September, 2230 UTC 17 September, 1240 UTC 19 September 2020 and

233 2130 UTC 20 September 2020. Thus the first two overpasses took place before Ianos entered its

234 phase as a medicane (around 0300 UTC 17 September 2020) while the third overpass, when Ianos

235 was an intense medicane, located very close to the western coasts of Greece.

236 The vertically polarized at 89 GHz (V89 GHz) Tb imagery of GMI on 1340 UTC 16 September

237 2020 (Fig. 8a) shows a Tb decrease that reaches 129 K near the top of the cloud bands. Since that

238 channel is strongly affected by ice and snow scattering, low Tbs indicate high amounts of frozen

239 hydrometeors at the mid/high levels of the troposphere, which are usually related to deep

240 convective clouds and lightning. Indeed, lightning activity over the western flank of Ianos, as

241 shown in Fig. 4a, was very high. High rainfall rates at the surface are indicated by the Tb increase in

242 V10 GHz (Fig. 8b), a channel which is sensitive to radiative emission by liquid water. It should be

243 noted however that lower frequency channels suffer from lower spatial resolution (approximately

244 19×32 km for V10 GHz against 7x4 km for V89GHz) and therefore although they permit to follow

245 the rainfall intensity changes during the medicane evolution, they do not provide details on the

246 rainband structure within the medicane.

247 At 2320 UTC 16 September 2020, when Ianos was still in its development phase, a lower Tb

248 depression (min Tb value = 163 K) on the cloud bands around the western flanks of the cyclone

249 center is evident in the V89 GHz channel (Fig. 9a), suggesting weaker convective activity than 12

250 hours earlier. Such a weakening of the convective activity before the TLC phase of a medicane has

251 been identified and discussed in previous medicane studies, like the December 2005 case (Fita and

252 Flaounas, 2018; Dafis et al., 2020). This finding is also in agreement with ZEUS lightning

253 observations that presented a significant decrease during the elapsed 12 h. The V10 GHz Tbs (Fig.

254 9b) is also lower around the cyclone center, a fact that supports a decrease of surface rainfall rates

255 relatively to the previous GMI overpass.

11

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
256 Ianos became a powerful medicane around 0300 UTC 17 September. During the third GMI

257 overpass over the area of interest, at 2230 UTC 17 September 2020, the cyclone center is found

258 very close to the Ionian Islands. The V89 GHz channel (Fig. 10a), reveals a rotating pattern of ice

259 hydrometeors around the center of the medicane, but V89 GHz Tb values are still higher and V10

260 GHz Tb values are still lower than the previous day (Fig. 8), indicating a gradual shift to a less

261 convective status of the rainfall regime. As evidenced in the V10 GHz channel, the Tb increases

262 close to the cyclone center (Fig. 10b), with respect to the previous overpass (Fig. 9b), providing

263 thus a clear indication of persistent rainfall over the sea and the Ionian islands. This finding is

264 supported by the high EPWAT values that are provided by FNL (Fig. 7c) and the available rain-

265 gauge measurements.

266 It is worth noting that GPM-CO overpasses over medicanes are very rare. Marra et al. (2019)

267 presented an analysis of Medicane Numa that also developed over the Ionian Sea between 15-19

268 November 2017 and affected Southern Italy and Greece. The authors identified strong convective

269 activity during the development phase of Numa and it was detected to be weaker after the maximum

270 intensity of the cyclone. For what it concerns the Tb values in the V89 GHz channel, Marra et al.

271 (2019) reported values as low as 158 K during the development stage of Medicane Numa, a value

272 substantially higher than the value of 129 K found over the western flank of Ianos during its

273 developing stage (Fig. 8a). During the medicane phase, Numa was showing only small portions of

274 Tb depressions over its western flank (Marra et al., 2019 – their Figs. 7-8), while Ianos showed

275 large areas of Tb depressions over its western flanks, with no lightning activity, in contrast to its

276 developing phase. Regarding the spatial distribution of convection, Ianos shares similarities with

277 Medicanes Rolf (Dafis et al., 2018), Trixie and Zorbas (Dafis et al., 2020) as convective activity

278 was significantly reduced more than 150 km away from the medicane center, but it was increased

279 close to the core of the system within the first 50-100 km when the cyclone exhibited more

280 symmetry and higher intensity. These results provide a different evolution for the distribution and

281 intensity of convection with the three medicanes between 2003 and 2006 that were examined by

12

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
282 Claud et al. (2010) and Miglietta et al. (2013), during which deep convection and precipitating

283 clouds were not detected close to the center during the maximum cyclone intensity, despite the

284 impressive cloud-free “eye” formation.

285 b. Precipitation radar

286 The near-surface DPR corrected reflectivity (normal scan) at 2230 UTC 17 September 2020 is

287 presented in Fig. 11, overlaid on the MSG SEVIRI infrared imagery (already shown in Fig. 4c).

288 High reflectivity values dominate the ring around the medicane “eye”, with the exception of the

289 southeastern part of it. The spiral band around Ianos’ center is clearly evident and reminiscent of the

290 cloud structures usually shown in tropical cyclones. The cloud free area, the “eye” of Ianos, has a

291 diameter of ~50 km, as inferred by the DPR measurements.

292 A 3-D view of the reflectivity pattern associated to Ianos is shown in Fig. 12. The cloudbands

293 around the eye are evident, while the surface of maximum height of the 20 dBZ reflectivity

294 indicates the most vigorous cloud development over the southern flank of Ianos, reaching heights

295 up to 11 km. This height is higher than the previous GPM DPR and ground-based radar

296 measurements of Medicane Numa (Marra et al, 2019) when the cloud band tops were as high as 9-

297 10 km. Within this rainband, reflectivity core of more than 45 dBZ was extending up to 5 km and

298 the 20 dBZ isoline was found up to 10 km height, where close to zero and positive values of the

299 WV-IR brightness temperature differences are found (see Fig. 4c). Interestingly, some of the

300 highest values of corrected reflectivity in the northern flank of Ianos’ “eye” in Fig. 11 do not

301 correspond to low Tb values in the V89 GHz channel (Fig. 10a). This can be partly explained by the

302 data provided in Fig. 12, where high reflectivity values are confined in low altitudes in the

303 troposphere, but these low-topped clouds were associated with torrential rainfall in the Ionian

304 Islands as high precipitation rates were recorded during this period (not shown). Moreover V89

305 GHz is strongly affected by ice meteors that are mostly found in the middle and higher parts of

306 troposphere, whilst Fig 11 depicts near-surface reflectivity where the cloud bands are dominated by

13

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
307 liquid water. A more detailed study would be needed to further explore the characteristics of these

308 clouds around Ianos “eye” and their lack of lightning activity.

309 7. Concluding remarks

310 The analysis of all available in-situ and satellite observations, of three GPM-CO overpasses and of

311 gridded analyses permitted to synthesize a coherent picture of Medicane Ianos during its entire

312 lifetime. It is the first time that GPM-CO captures a medicane during its mature stage. The main

313 points of this synthesis are provided in the following:

314  Based on the available data, Ianos was among the strongest medicanes observed in the

315 Mediterranean in the last at least 25 years. Its central pressure was measured to 984.3 hPa

316 and the wind speed reached 54 ms-1. Daily accumulated rainfall during Ianos exceeded 600

317 mm in Western Greece and 300 mm over parts of Central Greece, resulting in four casualties

318 and extensive damage to property and infrastructure caused by floods and landslides.

319  Ianos initially developed close to a thunderstorm cluster in the Gulf of Sidra, over a pool of

320 warm sea-surface waters on 15 September 2020. Its developing phase lasted more than 48

321 hours and at ~0300 UTC 17 September it became a medicane. Ianos followed a path towards

322 Western Greece and finally dissipated over the coasts of Egypt, 7 days after its formation.

323 Its total path over the Mediterranean waters exceeded 1900 km.

324  Satellite infrared imagery, as well as lightning data, permitted to follow the convective

325 activity associated with Ianos, from its developing phase, towards its mature stage when

326 convective activity significantly decreased in the outer parts of the cyclone and intense

327 convection was mostly active close to the medicane core. These results provide a different

328 evolution for the distribution and intensity of convection with the three medicanes between

329 2003 and 2006 that were examined by Claud et al. (2010) and Miglietta et al. (2006), during

330 which deep convection and precipitating clouds were not detected close to the center during

331 the maximum cyclone intensity, despite the impressive cloud-free “eye” formation.

14

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
332  Analysis of global model fields revealed the upper-level precursors to Ianos development

333 and mainly the role of an upper-tropospheric PV streamer on the intensification of the vortex

334 during the initial stage of Ianos formation.

335  High values of EPWAT offshore Western Greece as well as over Central Greece clearly

336 identified the area where the highest accumulations of rainfall occurred.

337  Three GPM-CO overpasses permitted to follow the cloud and precipitation structures

338 associated with Ianos, during its developing and its mature phase. High-frequency Tb

339 revealed details about the precipitation and cloud band evolution of the storm that were not

340 evident in the geostationary imagery. Precipitation radar reflectivity, revealed for the first

341 time the fine structure of cloud bands around a medicane during its mature (or TLC) phase

342 and permitted the identification of certain reflectivity features in the 3-dimensional space

343 during Ianos’ mature stage.

344 Medicanes are rare phenomena but can lead to high impact events in the Mediterranean countries.

345 We believe that such studies contribute to the enrichment of our knowledge and the improvement of

346 our understanding of medicane formation and evolution. Besides the obvious scientific benefits, the

347 increased ability to forecast medicanes and their impacts should be exploited to promote a better

348 societal preparedness and increased coping capacity.

349 Acknowledgments

350 The authors acknowledge EUMETSAT for the provision of MSG SEVIRI data, NASA and JAXA

351 for the GPM data, as well as NCEP, USA for the provision of FNL model analyses. Both

352 observational and reanalysis datasets will be available upon request.

15

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
353 REFERENCES

354 Buongiorno Nardelli B., C.Tronconi, A. Pisano, R.Santoleri, 2013: High and Ultra-High resolution

355 processing of satellite Sea Surface Temperature data over Southern European Seas in the

356 framework of MyOcean project. Rem. Sens. Env., 129, 1-16,

357 https://doi:10.1016/j.rse.2012.10.012.

358 Cioni, G.,Cerrai, D. and Klocke, D., 2018: Investigating the predictability of a Mediterranean

359 tropical‐ like cyclone using a storm‐ resolving model. Quarterly Journal of the Royal

360 Meteorological Society, 144, 1598–1610. https://doi.org/10.1002/qj.3322.

361 Cioni G., P. Malguzzi and A. Buzzi, 2016: Thermal structure and dynamical precursor of a

362 Mediterranean tropical‐ like cyclone. Quarterly Journal of the Royal Meteorological Society.,

363 142, 1757-1766.

364 Claud, C., Alhammoud, B., Funatsu, B.M., Chaboureau, J.P., 2010: Mediterranean hurricanes:

365 largescale environment and convective and precipitating areas from satellite microwave

366 observations. Natural Hazards and Earth System Sciences, 10, 2199–2213.

367 https://doi.org/10.5194/nhess-10-2199-2010.

368 Dafis, S., Claud, C., Kotroni, V., Lagouvardos, K. and Rysman, J.‐ F., 2020: Insight into

369 convective evolution of Mediterranean tropical‐ like cyclones. Quarterly Journal of the Royal

370 Meteorological Society, 146, 4147–4169. https://doi.org/10.1002/qj.3896.

371 Dafis, S., Rysman, J.F., Claud, C., and Flaounas, E., 2018: Remote sensing of deep convection

372 within a tropical‐ like cyclone over the Mediterranean Sea. Atmospheric Science

373 Letters, 19. https://doi.org/10.1002/asl.823.

374 Emanuel, K.A., 2005: Genesis and maintenance of “Mediterranean hurricanes”. Advances in

375 Geoscience, 2, 217–220. https://doi.org/10.5194/adgeo-2-217-2005.

376 Feidas, H. and Giannakos A., 2011: Identifying precipitating clouds in Greece using multispectral

377 infrared Meteosat Second Generation satellite data, Theor. Appl. Climatol., 104:25–42, doi:

378 10.1007/s00704-010-0316-5.

16

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
379 Flaounas, E., Gray, S. L., and Teubler, F, 2020: A process-based anatomy of Mediterranean

380 cyclones: From baroclinic lows to tropical-like systems, Weather Clim. Dynam. Discuss.,

381 https://doi.org/10.5194/wcd-2020-48, (in review).

382 Fita, L., Flaounas, E. Medicanes as subtropical cyclones: the December 2005 case from the

383 perspective of surface pressure tendency diagnostics and atmospheric water budget, 2018:

384 Quarterly Journal of the Royal Meteorological Society; 144, 1028–1044.

385 https://doi.org/10.1002/qj.3273.

386 Hou, A.Y., Kakar, R.K., Neeck, S., Azarbarzin, A.A., Kummerow, C.D., Kojima, M., Oki, R.,

387 Nakamura, K., Iguchi, T., 2014: The Global Precipitation Measurement Mission. Bull. Am.

388 Meteorol. Soc., 95.5, 701-722, doi: https://doi.org/10.1175/BAMS-D-13-00164.1

389 Kalimeris, A., Kolios, S., Halvatzaras, D., Posa, D., DeIaco, S., Palma, M., Skordilis, C., and

390 Myrsilides, M., 2015: The Ionian-Puglia network of meteorological-environmental stations.

391 Geophysical environment and technical description. Odysseus 7, 1–45. (available by

392 http://ionianweather.gr/en/scientific-articles/ionian-meteorology-and-climatology-en.html)

393 Koseki, S., Mooney, P. A., Cabos, W., Gaertner, M. Á., de la Vara, A., and Gonzáles Alemán, J. J,

394 2020.: Modelling a tropical-like cyclone in the Mediterranean Sea under present and warmer

395 climate, Nat. Hazards Earth Syst. Sci. Discuss., https://doi.org/10.5194/nhess-2020-187, (in

396 review).

397 Kotroni V. and K. Lagouvardos, 2016: Lightning in the Mediterranean and its relation with sea-

398 surface temperature. Environmental Research Letters, 11, 034006.

399 Lagouvardos, K.; Kotroni, V.; Bezes, A.; Koletsis, I.; Kopania, T.; Lykoudis, S.; Mazarakis, N.;

400 Papagiannaki, K.; Vougioukas, S., 2017: The automatic weather stations NOANN network of

401 the National Observatory of Athens: operation and database. Geosci. Data J, 4, 4–16.

402 Lagouvardos, K.; Kotroni, V.; Nickovic, S.; Jovic, D.; Kallos, G.; Tremback, C.J., 1999:

403 Observations and model simulations of a winter sub-synoptic vortex over the central

404 Mediterranean. Meteorol. Appl, 6, 371–383.

17

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
405 Marra, A.C., Federico, S., Montopoli, M., Avolio, E., Baldini, L., Casella, D., D’Adderio, L.P.,

406 Dietrich, S., Sanò, P., Torcasio, R.C., and Panegrossi, G., 2019: The Precipitation Structure of

407 the Mediterranean Tropical-Like Cyclone Numa: Analysis of GPM Observations and

408 Numerical Weather Prediction Model Simulations. Remote Sensing, 11, 1690.

409 https://doi.org/10.3390/rs11141690.

410 Marullo, S., Buongiorno Nardelli, B., Guarracino, M., and Santoleri, R., 2007: Observing the

411 Mediterranean Sea from space: 21 years of Pathfinder-AVHRR sea surface temperatures

412 (1985 to 2005): re-analysis and validation, Ocean Sci., 3, 299–310, https://doi.org/10.5194/os-

413 3-299-2007.

414 Miglietta, M.M. and Rotunno, R., 2019: Development mechanisms for Mediterranean

415 tropical‐ like cyclones (Medicanes). Quarterly Journal of the Royal Meteorological

416 Society, 145, 1444–1460. https://doi.org/10.1002/qj.3503.

417 Miglietta, M. M., D. Cerrai, S. Laviola,E. Cattani, and V. Levizzani, 2017: Potential vorticity

418 patterns in Mediterranean “hurricanes”. Geophys. Res. Lett., 44, 2537–2545,

419 doi:10.1002/2017GL072670.

420 Miglietta, M. M., S. Laviola, A. Malvaldi, D. Conte, V. Levizzani and C. Price, 2013: Analysis of

421 tropical-like cyclnes over the Mediterranean Sea through a combined modeling and satellite

422 approach. Geophys. Res. Lett., 40, 2400–2405, doi.org/10.1002/grl.50432.

423

424 Noyelle, R., Ulbrich, U., Becker, N., and Meredith, E. P., 2019: Assessing the impact of sea surface

425 temperatures on a simulated medicane using ensemble simulations, Nat. Hazards Earth Syst.

426 Sci., 19, 941–955, https://doi.org/10.5194/nhess-19-941-2019.

427 Olander, T.L. and Velden, C., 2009: Tropical cyclone convection and intensity analysis using

428 differenced infrared and water vapor imagery. Weather and Forecasting, 24, 1558– 1572,

429 https://doi.org/10.1175/2009WAF2222284.1

430 Picornell, M. A., Campins, J., Jansà, A., 2014: Detection and thermal description of medicanes from

431 numerical simulation. Natural Hazards and Earth System Science, 14(5), 1059-1070.
18

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
432 Senf, F.; Dietzsch, F.; Hünerbein, A.; Deneke, H., 2014: Characterization of Initiation and Growth

433 of Selected Severe Convective Storms over Central Europe with MSG-SEVIRI. J. Appl.

434 Meteorol. Climatol., 54, 207–224.

435 Sudradjat et al., 2005 A. Sudradjat, R.R. Ferraro, M. Fiorino, 2005: A comparison of total

436 precipitable water between reanalyses and NVAP, J. Climate, 18, 1790-2007.

437 Tous, M., and R. Romero, 2013: Meteorological environments associated with medicane

438 development. Int. J. Climatol., 33, 1-14.

439 Weller, E.; Jakob, C.; Reeder, M.J., 2017: Projected Response of Low-Level Convergence and

440 Associated Precipitation to Greenhouse Warming. Geophys. Res. Lett, 44,

441 https://doi.org/10.1002/2017GL075489.

442 Winstanley, D., 1970: The north African flood disaster, September 1969. Weather 25, 390–403.

443 https://doi.org/10.1002/j.1477-8696.1970.tb04128.

444 Zekkos D., Zalachoris G., Alvertos, A. E., Aatya P. M., Blunts P., Clark M., Dafis S., Farmakis I.,

445 Ganas A., Hille M., Kalimogiannis V., Karagiannidis A., Karantanellis E., Khan K.,

446 Kirshbaum D., Kourkoulis R., Kotroni V., Ktenidou O.-J., Lagouvardos K., Loli M.,

447 Makrinikas A., Marinos V., Manousakis J., Nikas K., Panousis D., Papathanassiou G.,

448 Saroglou C., Simopoulos A., Stanley T., Tsavalas A., Valkaniotis S., 2020: The September

449 18-20 2020 Medicane Ianos Impact on Greece - Phase I Reconnaissance Report. Geotechnical

450 Extreme Events Reconnaissance Report, GEER-068, https://doi.org/10.18118/G6MT1T.

451

19

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
452

453

454 Figure 1. Path of Ianos medicane, from 0000 UTC 15 September 2020 up to 0000 UTC 21

455 September 2020. Shading denotes the SST (at 1 K intervals) and the black dots the daily SST

456 anomaly higher than 2 K on 15 September 2020. Symbols for floods, landslides and damaging

457 winds are placed over the most affected areas.

458

20

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
459

460 Figure 2. Time series of mean sea-level pressure, wind and precipitation in (a) Palliki in

461 Cephalonia and (b) Skinari in Zakynthos between 0600 UTC 17 September and 1800 UTC 18

462 September 2020: (i) 10-minute max 10-m wind speed (green line), (ii) max wind gust (green

463 rectangular) (ms-1), (iii) 30-minute mean wind direction at 10 m height (green vectors), (iv) 10-

464 minute minimum sea-level pressure (hPa) and (v) 10-minute accumulated rainfall (blue bars) (mm).

465

21

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
466

22

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
467 Figure 3: (a) Total accumulated rainfall over Greece for the period spanning from 1200 UTC 17

468 September up to 1200 UTC 19 September 2020 (48 hours), and (b) zoom over Central Greece, for

469 the same period.

470

471 Figure 4: MSG SEVIRI infrared brightness temperature (grey scale), and brightness temperature

472 difference between the WV6.2 and IR10.8 channels (color scale), valid (a) at 1200 UTC 16

473 September 2020, (b) at 1200 UTC 17 September 2020, (c) 2230 UTC 17 September 2020 and (d) at

474 1800 UTC 18 September 2020. Yellow dots denote lightning strikes detected by ZEUS network

475 within 60 minutes before and after the satellite nominal time.

476

23

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
477

478 Figure 5: NCEP FNL analysis of 500-hPa geopotential height (shading at 50-gpdam intervals),

479 sea-level pressure (thin black contours at 2-hPa intervals), valid at (a) 1200 UTC 16 September

480 2020, (b) at 1200 UTC 17 September 2020, (c) at 0000 UTC 18 September 2020, and (d) at 1800

481 UTC 18 September 2020.

482

24

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
483

484

485 Figure 6: Potential vorticity (shaded, PVU) and wind barbs at the 335 K isentropic surface, and

486 850-hpa geopotential height (red contours, at 50 gpdam intervals) valid at 1200 UTC (a) 16

487 September and (b) 17 September 2020.

488

25

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
489

490 Figure 7: Integrated tropospheric water vapor transport (kg m-1 s-1, shading and arrows) and

491 effective precipitable water (red contours with 10 mm intervals) at the same four instances as in Fig.

492 5.

493

26

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
494

495 Figure 8: GPM GMI brightness temperature (Kelvin) measured at (a) V89 GHz and (b) V10 GHz

496 around 1340 UTC 16 September 2020

497

27

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
498

499 Figure 9: GPM GMI brightness temperature (Kelvin) measured at (a) V89 GHz and (b) V10 GHz

500 around 2320 UTC 16 September 2020

501

28

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
502

503 Figure 10: GPM GMI brightness temperature (Kelvin) measured at (a) V89 GHz and (b) V10 GHz

504 around 2230 UTC 17 September 2020

505

29

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
506

507 Figure 11: GPM DPR near-surface corrected reflectivity (normal scan) around 2230 UTC 17

508 September 2020

509

30

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.
510

511 Figure. 12: GPM DPR maximum height of 20 dBZ reflectivity isosurface around 2230 UTC 17

512 September 2020 overlaid on MSG SEVIRI IR108 image: (a) view from southwest and (b) view from

513 northeast

514

31

Unauthenticated | Downloaded 10/05/21 06:45 AM UTC

Accepted for publication in Bulletin of the American Meteorological S ociety. DOI 10.1175/BAMS-D-20-0274.1.