Storm Ahti causes damage in northern Finland

By Petteri Pyykkö (FMI), Natasa Strelec Mahovic (EUMETSAT) and Ivan Smiljanic (CGI)

The thunderstorms swept over the northern regions of North Ostrobothnia and Kainuu, causing widespread damage to trees and localised flash flooding. One person was reported to have died due to a fallen tree.

A frontal system associated with a low over northern Scandinavia moved across the region in the afternoon hours. The storm was later named 'Ahti' in Finland.

The weather at the time of the storm was exceptionally warm and also very moist. The strongest thunderstorms were associated with the cold front moving in from the west. The front led to a formation of a supercell that hit the city of Oulu.

The cell can be seen best in the infrared channel image (Figure 1 right), depicted as the most red, indicating very cold cloud top temperatures. The cell produced a maximum gust of 112km/h at Oulu Vihreäsaari harbour, and a gust of around 85km/h at the Oulu airport.

The wind felled many trees and ripped off some roofs. At one point 11,000 households were without electricity. Near the airport 30-35 mm hailstones were reported. In the city of Rovaniemi, around 180km to the north east of Oulu, 65.1mm of precipitation was measured within 24 hours. Most of the precipitation fell in association with the MCS visible in Figure 2 (left and right) and Figure 3 (left and right). The overshooting tops are also clearly visible in Figure 2 (right). In total the storm produced almost 17, 000 lighting strikes in Finland.

The airmass (air temperature at 850hPa, at around 1.5km height in the atmosphere) was very warm for the time and location, up to 16°C, south and east of the fronts. Deep layer shear was 72km/h, locally higher, and 0–1km shear was up to 54km/h; supportive of supercells, large hail and even some isolated tornadoes. MLCAPE was 1000–1800j/kg, 250–400m²/s² SREH and mixing ratios were 10–13g/kg according to ECMWF model forecasts.

A comparison of the Natural Colour RGBs from Meteosat-11 SEVIRI and NOAA-20 VIIRS in Figure 4, shows the advantage of high spatial resolution images offered by polar-orbiting satellites compared to lower resolution of the geostationary sensors, especially in higher latitudes, as is the case here over Finland.

The 1km spatial resolution now offered by the VIIRS sensor in polar orbit, only few times a day, will be available from the FCI (Flexible Combined Imager) on board Meteosat Third Generation geostationary satellites every 10 minutes, starting in 2023. However, the problem of countries in high latitudes being on the edge of the satellite view will still remain.

The insight into MTG capabilities in the visible range is provided through a HRV channel of SEVIRI — with its nominal resolution matching that of VIS channels' resolution of 1km. Naturally, the shape and size of one pixel changes with lat-long position — around the Oulu, Finland, that equals to a pixel size of 4.38 x 1.36km (pixel shape and size in Figure 5).

The new generation of satellite images and parallax issues

When the satellite viewing angle is away from the sub-satellite point (nadir), an object’s position is shifted. This is called parallax shift effect (see High latitude simulator for MTG/FCI instrument, section Limitations). This is a very important factor in convective cases at high latitudes. Unfortunately the method used in simulation doesn't take into account that this phenomena is highly visible in geostationary images and also at the swath edges in polar satellite images. At large viewing angles, high clouds move further than low clouds and the distortion of the shape of the original feature is also greater. It should also be remembered that at a very slated view, the satellite actually measures the sides of clouds rather than the tops.

Observing images 6a and 6b, if one takes a look at the thunderstorm cells on the coast to the southwest of the mesoscale convective system (MCS), one can observe parallax shift in the apparent position of cumulonimbus (Cb) clouds. In the MODIS imagery (Figure 6a) the cell appears to be located more inland than in the VIIRS imagery (Figure 6c). Parallax is further illustrated in Figure 6d where the parallax shift of a Cb cloud over the sea to the west of the city of Oulu is also annotated. In the case of polar satellite imagery the parallax effect becomes stronger towards the edges of imaging swaths. Also, the resolution can be seen to significantly worsen towards the swath edges in Figure 6a, because of MODIS' large pixel growth in the scan direction. Whereas, VIIRS has a lot smaller pixel growth from nadir to end of scan, it can be seen that the single cells inland to the southeast of Oulu have a quite clear appearance in Figure 6c.

VIIRS data is a good proxy for what we can except from METimage's upcoming data. The new instrument's imagery will enhance the image resolution, including near the edges of imaging swaths. It can be observed from Figure 6c that single-cell thunderstorms and even overshooting tops are better visible here than in the MODIS imagery (Figure 6a) and SEVIRI's operational Day Microphysical product (Figure 6b). Also, the areas with clear skies are better visible in higher resolution imagery (Figure 6c), which is important for nowcasting purposes.

Figure 7a shows the new Cloud Phase RGB product at the same sampling time as Figure 6b. When comparing these two one can note that cloud tops are easier to distinguish in the Cloud Phase RGB, especially the overshooting tops among the MCS. In the Cloud Phase RGB the colour of the cloud tops also changes from dark blue to light blue, depending on the ice particle size distribution, which makes it easier to evaluate the tops' height/potential severity of thunder clouds. Also, areas of clear skies are very well visible in this product, possibly even more clearly than in the Day Microphysics RGB.

The Cloud Phase Distinction RGB in Figure 7b essentially shows the same information as the Cloud Phase RGB (Figure 7a), but it appears to be very sensitive to the sun's angles of elevation. The highest cloud tops that would appear in different shades of yellow with higher elevation angles, appear too bright to make accurate and meaningful observations about the clouds' structure. The problem appears to be the same in NOAA's recipe for the Day Cloud Phase RGB from the GOES-EAST ABI instrument.

It must be taken remembered that better resolution in new MTG measurements does not remove the effects that appear when the satellite viewing angle is far away from the sub-satellite point (see High latitude simulator for MTG/FCI instrument). Thus, in high latitudes the parallax effect remains the same as it is with current MSG. Figure 8a clearly illustrates the parallax shift when comparing sandwich-ir images from SEVIRI to FCI simulated (Figure 8b). With a high viewing angle the convective cloud seemed to been approximately 45km further northest than it actually was when the MCS was approaching the city of Oulu. Also the shape of the Cb cloud is distorted in the image from the geostationary satellite. In the future, in high latitudes METImage will provide more detailed and accurate measurements compared to FCI, so is the recommended tool for the forecasters. However, the parallax effect can be calculated and easily corrected in physical products like Cloud Top Height (Figures 9a and b) which can useful in monitoring temporal evolution.

Typically, in situ ground observation are located near the coastline or on the continent, and it is seldom (if at all) that observations over the sea are available. Satellite observations provide unique sources of information from pre-convective parameters that are essential in nowcasting. Figure 10 shows an one example of a pre-convective ingredient, total column water vapour from ECMWF (20 June 2021 12 UTC run, coloured fields) and IASI L2 sounding (circles) at 07:57 UTC. The model values and the values from the sounding match relatively well, but there are a few places where they differ. To the west of the city of Oulu, IASI L2 suggests that the water vapour content in the atmosphere is higher than what the model suggests. This might have positively contributed to the formation of an MCS that impacted Oulu and northern Osterborthnia later in the afternoon and evening of 21 June. Along the coast to the southwest of Oulu IASI L2 has lower water vapour values than ECMWF.

Summary

MTG/FCI

• The simulation method used is too simple to give a realistic impression of future FCI imagery at high latitudes (see High latitude simulator for MTG/FCI instrument, section Parallax effects).
• At large viewing angles at high latitudes, high clouds move further than low clouds and the distortion of the shape of the original feature is also greater. At very slated view, satellite actually measures the side of clouds rather than top. Albeit these issues, the new instrument's enhanced temporal and spacial resolution alongside the new RGB products will be an important add-on in forecasting convection in high latitudes.
• Enhanced resolution facilitates the monitoring of developing and moving storms and single-cell tracking
• New channels: Cloud Phase RGB is the best product for monitoring single-cell thunderstorms, cloud top properties and over-shooting tops.

EPS-SG/METimage

• Enhanced resolution towards the edges of imaging swaths (vs AVHRR), comparable to VIIRS
• New channels, Cloud Phase RGB is the best product for monitoring single-cell thunderstorms, cloud top properties and over-shooting tops.
• Instruments will be beneficial in verifying Numerical weather models' performance, for example using Hyperspectral IASI-NG L2 sounder's additional information in pre-convective situations.

Additional content

One killed in Oulu as meteorologists warn of more thunderstorms to come (Helsinki Times)
Watch: Thunderstorm slams Oulu on Monday (yle)
BIG STORM IN OULU | OULUN MYRSKY | FINLAND (Irish in Finland/YouTube)