1 Introduction

In a warming climate, climate hazards are expected to increase in intensity and frequency (AghaKouchak et al. 2020). Knowledge on increased precipitation and intensity is important in climate hazard mitigation. One of the important factors which leads to uncertainty in future trends in precipitation is the dynamics of intense events. The fraction of intense precipitation will increase due to rise in global temperatures (Bennett and Walsh 2015). For every 1 °K rise in temperature intense precipitation increases by 5.9–7.7% (Westra et al. 2013). Since the Arctic is warming at twice the global rate, extreme weather events in the future are expected to be frequent in the Arctic. Global climate model (GCM) simulations for the Arctic suggest positive trends for precipitation, river runoff and evapotranspiration in the future (Rawlins et al. 2010). Under a global warming scenario intense precipitation increases due to enhanced saturation vapour pressure and follows the Clasius-Clapeyron relationship (Poujol et al. 2020). Synoptic circulation systems also impact the intense precipitation events. It is shown that convective precipitation is more sensitive to temperature change than stratiform precipitation (Berg et al. 2013). Many locations over northern Eurasia have recorded increase in convective precipitation (Ye et al. 2017; Chernokulsky et al. 2019). In a changing climate, atmospheric circulation features are expected to change. One of the most conspicuous changes in the Arctic is the reduction in sea-ice area related to regional and global scale atmospheric circulation changes (Trapp and Hoogewind 2018). The reduction in sea-ice area will also result in the interaction of warm ocean waters with the atmosphere resulting in enhanced warming of the Arctic. Even remote climate variability is also attributed to the large-scale atmospheric circulation changes in high latitudes (Piper et al. 2019; Nuncio et al. 2020). An analysis of three decades long information suggests warm air advection intensified by North Atlantic Oscillation (NAO) results in thunderstorm activity in Svalbard (Czernecki et al. 2015). In a warmer and wetter climate, in response to anthropogenic warming, hail occuring days (size) are expected to reduce (increase) over North America (Brimelow et al. 2017). Will the Arctic respond in a similar fashion? Indeed, the Arctic is warming and reduction in sea-ice is expected to increase the evaporation, how this will translate in to hail formation is yet to be ascertained. Thus, large-scale circulation governed by both remote and local drivers together with the thermodynamic response, intense precipitation events in the Arctic will be tough to predict.

One aftermath of intense precipitation events is the formation of Hails. Hail particles form when strong vertical motion in thunderstorms facilitates the accretion of super-cooled liquid water. Large hails can inflict severe damage to property, while small hails can damage agriculture (Púčik et al. 2019). Hailstorm variability and climate change is a nascent field, not many studies are devoted to this. In the northern Europe, climate change is expected to increase hailstorms mainly due to the increase in low level humidity (Rädler et al. 2019). Tropical processes like ENSO is attributed with hail in USA (Allen et al. 2015). There are studies that link severe storms with NAO in Europe (eg. Miglietta et al. 2017). However, in the Arctic such studies are limited. Severe convective storms (SCS) are the precursors for hails. Across the USA, increase in Convective Available Potential Energy (CAPE) is related to increase in severe thunderstorms (Chen et al. 2020). High instability and CAPE were attributed to a recent Canadian thunderstorm (Brown et al. 2020). The study also noted that the increase in CAPE is due to increase in low level humidity. However, occurrence of SCS is quite low in the Arctic. In the present study, we investigate occurrence of hail during 2018–19 December January February (DJF) in Ny Alesund, Svalbard. The purpose of the study is to investigate occurrences of hail with atmospheric circulation patterns and examine atmospheric vertical structure during hail events.

2 Data and methods

As a part of Indian Arctic program, an OTT parsivel is installed in Ny Alesund (Fig. 1). It can detect eight types of precipitation types viz. drizzle, drizzle with rain, dain, rain/drizzle with snow, snow, snow grain, soft hail & hail. It has less maintenance, and its installation and operation are easy. The OTT PARSIVEL2 measures the amount of precipitation by measuring the particle size using a horizontal laser beam. To determine the particle speed, the duration of the signal is measured as soon as a precipitation particle enters the laser beam and ends when it has completely left the beam. The atmospheric vertical structure and clouds are obtained from a microwave radiometer, by Radiometrics corporation. This is a passive remote sensing instrument that provides profiles of atmospheric temperature and humidity by measuring atmospheric brightness temperature in the frequencies between 51 and 59 and 22–30 Ghz, respectively. Totals Total Index is also computed using RAOB software(RAOB users guide, 2018) using the expression T850 + Td850–2 × T500 where T and Td are temperature and dew point at 850 and 500 hPa levels. The total totals index is defined as the sum of two indices, the cross totals and the vertical totals. The vertical totals is given by the formula.

$$V_{t} \, = \,T_{850} {-}T_{500}$$

where T is the temperature at 850 and 500 hpa levels. This will give the lapse rate between, 850 and 500 hPa layers, whereas the cross totals is defined as \(C_{T} = T_{{d}(850)}{-} T_{500},\) where Td is the dew point temperature. This takes in to account of the lower layer moisture. As we see later, the synoptic weather pattern showed a high pressure over northern Europe and a cyclonic circulation north of Svalbard. The southwesterly—westerly associated with these pressure transport moisture and heat in to Ny Alesund. Severe weather in the atmosphere could be a result of holding the top air column constant or warming it slightly and adding heat and moisture to the bottom (Miller 1972). An index value of 45–55 suggests moderate severe weather while values greater than 55 indicate strong chances of severe weather (RAOB users guide, 2018).

Fig. 1
figure 1

Location map depicting the measurement location in Svalbard

Full size image

In Svalbard, thunderstorms occurred when the mean value of TT was 39(Czernecki, et al. 2015). Over Europe a high probability of hail occurrence was noticed when TT index exceeded 44 (Siedlecki 2009). CERES cloud data products (Doeling, et al. 2013), ERA5 hourly reanalysis (Hersbach et al. 2020) and NOAA optimum interpolated Sea Surface Temperature(SST) (Reynolds, et al. 2002) were utilised to study the large-scale atmospheric circulation patterns, clouds, aerosol optical depth (AOD) and SST during the precipitation events. Wind patterns at Ny Alesund are investigated using met observations carried out by Alfred Wegner Institute (Maturilli 2019).

We have compared the radiosonde and radiometer profiles during the events studied. Earlier studies indicated a bias between radiosonde and radiometer, particularly during strong winds, which could be due to elevated inversions or radiosonde transported further away from the observation site (Zhang, et al. 2017). The radiosonde derived winds depict strong vertical shear (Figure. S1). The wind speed derived from the radiosonde was in excess of 10 m/s above 1.5 km and the position of radiosonde indicated a strong drift of more than 1.5 km (Figure S2). The translation was faster on 21st February and 13th December than in any other days. The difference in temperature between radiosonde and radiometer started when the wind shear increased at about 1–1.5 km (Figure S3). This is the height at which the horizontal translation of the radiosonde become prominent. This means that radiosonde and radiometer may not be measuring the same airmass. (Bianco, et al. 2017; Xu et al. 2015; Zhang et al. 2017) suggesting the radiosonde may be trapped in a warm air mass as it ascends, but more analysis is required to confirm this. Also, the time of precipitation differs from the launch time of radiosonde, which make a comparison difficult for the present study. A detailed comparison is beyond the scope as well.

3 Results and discussion

3.1 Precipitation intensity during DJF 2018–19

During winter (DJF) 2018–19 maximum intensity in precipitation was noticed during December and February (Fig. 2a). An intense spell was noticed during second and third week of December. The precipitation intensified by 10th December gradually reduced after a fortnight. A second enhancement was noticed by last week of December. January had relatively less precipitation while during February another intense spell started peaking by the third week. It is these two intense spells during December and February that we subject our analysis for hail. Among these precipitation events, hail was noticed on five occasions (Fig. 2b). On 13 December 04–07 UTC, 15 December 17–19 UTC, 18 December 18–22 UTC and 21 February(23-00 UTC). During these days maximum intensity of precipitation was result of hail, except the spell on 18 December 09–14 UTC. The METAR code of precipitation indicates that the hail size could be greater than 5 mm (Figure S4). These type of hails occurred during all the days examined, however, smaller hails (GS) were more frequently observed. It is also interesting to note that peak precipitation was characterised by mixed precipitation types.

Fig. 2
figure 2

a Precipitation during 2018–19 winter b Precipitation during the time period when hail was recorded in OTT Parsivel at Ny Alesund Svalbard. Cyan dots indicate total precipitation intensity, and the red dots indicate hail

Full size image

3.2 Vertical temperature structure

During the intense precipitating events when hails formed, the vertical structure of the atmosphere was characterised by strong warming, mostly in the lower layers (Fig. 3). Totals totals index indicate values ranging between 45 and 55, indicating moderate severe weather in Ny Alesund. In all the cases maximum warming occurred in the lower levels. Winds and temperature exhibit a strong relationship in Ny Alesund. Southerly/southwesterly winds are expected to transport warm and moist air into Svalbard. During 13 December, however, wind direction fluctuated between easterlies to north westerlies (Fig. 4). Temperature at lower levels was maximum when the wind direction was southwesterly between 3 and 5 UTC. During this time totals total index showed high values indicating severe weather in Ny Alesund. The temperature declined further when the wind direction became north westerlies. During 15th December wind was predominantly easterlies, however the air temperature was above zero in the lowest levels until 18 UTC when the wind direction turned north westerly. If we compare both the days, on 13th south westerlies warmed up the atmosphere on two occasions, while no conspicuous warming was noticed on 15 December; the lower atmosphere did not show any co variability with wind direction until 18 UTC. When hail was observed wind was westerly between 15 and 18 UTC. On 18th December however, easterlies were characterised by a continuous increase in temperature, nonetheless during late evening wind turned southerlies, maximum temperature and hail was noticed during this time. A similar relationship was noticed on 21 February also. Therefore, in all the cases hail were noticed when the wind direction was around 250 Degrees accompanied with an increase in atmospheric temperature. High values of totals total, an indication of severe weather, was noted during all the hail events. However, one exception was on the morning of 15 December. Even though, there was an indication of severe weather, hail was noted only when the wind turned south westerly.

Fig. 3
figure 3

Vertical temperature structure and totals totals index when hail was recorded in Ny Alesund. (a) 13-Dec-2018 (b) 15-Dec-2018 (c) 18-Dec-2018 and (d) 21-Feb-2019 Distinct warming can be noticed in the lower layers of atmosphere during the time of hail (please refer Fig. 2). The black curve depicts the total totals index. When hail was noticed totals total index was also high indicating severe weather. Right y axis scales for totals total index

Full size image
Fig. 4
figure 4

Wind in degrees (brown, right axis) and temperature(ºC) (purple, left axis) at 10 m obtained from automated weather station operated by Alfred Wegner Institute in Ny Alesund,Svalbard. Hail was noticed when the wind turned southwesterly—westerly

Full size image

In a study of hails over North America, Johnson and Sudgen (2014) indicated the role of vertical wind shear in the formation of hails. Events analysed in this study were also characterised by vertical wind shear when hails were observed as found in analysis of hourly ERA5 wind data (Fig. 5). Downward vertical velocity near 500 hPa as well as upward vertical velocity at the lower layers were also noted in the reanalysis. Hail formation coincided with intense precipitation when the vertical velocity was upward in lower layers. The radiometer displayed little variation in temperature in the layer 4–7 km height. Thus, warming lower layer while holding upper layers almost constant increased the TT index as observed. High clouds were also observed during the time of intense precipitation indicating severe weather (Figure S5).

Fig. 5
figure 5

Horizontal (vector) and Vertical wind velocities (shading) for the region 10-13E and 78-81N during the occurrence of hails in Ny Alesund. Vertical shear as well as upward(negative) velocities were noticed during the days of strong precipitation and hail

Full size image

Hail production also require formation of hydrometeors like cloud droplets or ice particles (Allen et al. 2019). Smaller water droplets freeze at a lower temperature of about − 40 °C, called homogenous freezing. However, in the presence of nucleating particles like dusts, particles of biological origin or ashes ice formation is triggered at much higher temperature. For example, volcanic ash is an active ice nucleating particle at below  0 °C whereas biological species can be active above − 5 °C (Lamba and Verlinde 2011).

Vertical structure of temperature when hails occurred suggests that − 10 °C isotherm was approximately at 2 km height whereas − 5 °C isotherm was below 1 km. In Ny Alesund, Svalbard, local origin particles mainly constitute CCN during summer, while in autumn and winter, particles of remote origin dominate CCN composition (Jung et al. 2018). Southerly to southwesterly winds travel over the ocean before reaching Ny Alesund, hence it may also pickup particles of marine origin. CERES cloud products analysed for the events depict high aerosol optical depth aligned with the wind vectors indicating transport of aerosols from low latitude (Fig. 6). However, local increase in AOD was also noticed, eg. on 18th December close to Svalbard a relatively high AOD was noticed. Whether the AOD is a result of churning up of the ocean by strong wind or is it all composed of aerosols of distant origin is difficult to ascertain at this stage and needs to be studied separately. Nonetheless, the analysis indicates that wind over ocean can transport/generate aerosols and may potentially impact cloud formation and precipitation in Svalbard.

Fig. 6
figure 6

Hourly Aerosol optical depth (match 55) obtained from CERES cloud data products overlaid with wind vectors at 1000 hPa during the time of occurrence of hails in Ny Alesund. Maximum AOD is aligned with the core of the southerly winds suggesting aerosol transport from the south

Full size image

3.3 Cloud vertical structure and liquid water content in the atmosphere

Microwave radiometer derived cloud extended from 500 to 5000 m when the hail occurred (Fig. 7). The radiometer also showed high liquid water content with in 900 to 700 hPa during the days when hail occurred. Temperature in this layer ranged from − 5 to − 20 °C. Temperature and liquid water profiles suggest the presence of supercooled liquid water between 900 and 700 hPa. When adequate amount of cloud forming particles is injected in to this layers, intense precipitation may happen. Therefore, when hail was observed, southwesterly winds as well as high clouds, liquid water, aerosol transport and conditions for severe weather was noticed. Let us examine the large-scale circulation patterns during the events (Fig. 8). These patterns were characterised by high pressure anomalies over northern Europe and low pressure near Svalbard. The circulation associated with the high pressure pattern over northern Europe result in southerlies over north Atlantic, which transport, heat and moisture. Also, remote source of CCN during winter is mainly associated with Arctic Haze (Jung et al. 2018) and the pressure patterns are conducive for transporting these particles to Ny Alesund as seen in Fig. 6.

Fig. 7
figure 7

Clouds and liquid water obtained from radiometer in Ny Alesund during strong precipitation and hail a 13-Dec-2018 b 15-Dec-2018 c 18-dec-2018 d 21-Feb-2019

Full size image
Fig. 8
figure 8

Hourly mean sealevel pressure (SLP) obtained from ERA5 reanalysis during the time of occurrence of hails in Ny Alesund. All the cases were characterised by a high pressure system over the North western Europe

Full size image

SST variability impact cloud formation. Coupled modelling studies suggest relatively warmer ocean initiate shallow convection and low level cloud formation in the north sea (Fallmann et al. 2017). West of Svalbard West Spitzbergen Current (WSC) make the ocean warmer and develops a west–east SST gradient (Fig. 9). As the wind move over the warm WSC, convection may be initiated followed by cloud formation. Upward vertical velocity observed could be a result of this. The necessary CCN required for this process might be supplied by the winds from the low latitudes. The SST gradient maintain a strong atmospheric temperature gradient in the west to east direction (Fig. 10). Reanalysis suggests this gradient is strong in the lower 2 km of the atmosphere. Therefore, the frontal activity, in the lower atmospheric layer could be more intense when the winds are westerly. All these suggest to the role of WSC in the atmospheric dynamics in the central Arctic. These aspects however require coupled modelling efforts and will be attempted in the future.

Fig. 9
figure 9

NOAA optimum interpolated hourly SST during December 2018 in the North Atlantic The brown colour represents warm SST of the West Spitzbergen current. The current maintain a east–west temperature gradient conducive for cloud formation in the event of winds crossing the gradient

Full size image
Fig. 10
figure 10

Air temperature gradient during December 2018 a The meridional temperature gradient along the longitude of Ny Alesund (11°E) b latitudinal temperature gradient along 79°E. Gradients are strong in the east—west direction (b). There are two distinct gradient in (b), the first one is a positive gradient just west of 0 longitude this is induced by the boundary between coastal Icelandic current and the west Spitzbergen Current (WSC). Another one is between 10 and 15°E longitude. This represents the gradient formed by the cooler land temperature and warm WSC west of it

Full size image

4 Conclusions

Occurrences of hails during the DJF 2018–19 and associated atmospheric and oceanic features were studied in this study. Hails were formed when strong precipitation intensity occurred. It is pertinent to note that intense precipitation was characterised with mixed precipitation types. Vertical shear as well as upward vertical velocity were noticed during the days of hail formation. Total totals index indicates severe weather conditions in Ny Alesund. During this time the atmospheric vertical structure was charcaterised by high liquid water content with in a temperature range of − 5 to − 20 °C. It is found that hails occurred during southwesterly to westerly winds. These winds were part of large-scale pressure systems in the northern Europe and the north Atlantic. The southwesterly- westerly winds gather more moisture as well as nucleating particles along its way over the warm ocean and enhance the precipitation. Earlier studies (Maturilli and Kayser 2017) indicated the positive trend in sealevel pressure over northern Europe and has been linked to various processes, including those from the tropics (Nuncio et al. 2020). The atmospheric circulation is suggestive of the role of marine as well as distant sources for nucleating particles in strong precipitation. The warm WSC could potentially impact the atmospheric dynamics by means of the SST gradient, which could lead to convection or intense frontal activity and strong precpitation and hails.