Status: Ongoing analysis Material from: Linus
1. Impact
On 2 April, southern Sweden was hit by extreme snowfall on Tuesday, leading to problems in the traffic. The maximum measured snow depth in the south was 42 cm and generally above 30 cm in a region between the big lakes according to observations from SMHI.
At the same time of the snowfall in southern Sweden, cold temperature records for April was broken for 3 stations in the very north of the country. One of the station records go back to 1887 (Kvikkjokk). The coldest temperature was -34.1C in Nikkaluokta, which also was the April record for that station and the coldest April temperature in Sweden since 1991. The temperatures in northern Finland was also very cold, so I chose to do the evaluation for Sodankyla that had a minimum temperature of -29.6C.
SMHI has made a very good summary (in Swedish) for these two extremes. I also noted that they made use of the daily anomaly plot from C3S: https://www.smhi.se/bloggar/vaderleken-2-3336/snoovader-i-soder-koldrekord-i-norr-1.206566
2. Description of the event
The evaluation of the snow will focus on 24-hour snowfall on 2 April in a 0.25x0.25 box over Linkoping (centre point 58.3N, 15.6E).
The evaluation of the cold temperatures will focus on 2-metre temperature at 00UTC on 3 April in a 0.25x0.25 box Sodankyla (centre point 67.4N, 26.6E).
The plots below show analyses of MSLP and 6 hour rainfall from 31 March 00UTC to 3 April 00UTC, every 12th hour.
The plots below show analyses of T850 and z500 from 30 March to 5 April.
3. Predictability
3.1 Data assimilation
3.2 HRES
The plots below show 24-hour snowfall in mm water (2 April 00UTC - 3 April 00UTC) in concatenated short forecasts (first plot) and ENS control forecasts with different lead times.
The plots below show analysis of 2-metre temperature on 2 April 00UTC and forecasts from 30 March from IFS with 9 km resolution,the DestinE run with 4.4 km resolution and AIFS.
3.3 ENS
The plots below shows EFI and SOT for 1-day snowfall (2 April) from different initial dates.
The signal appeared around 3 days before the event, but there was some early signal in a few ensemble members (seen in SOT).
The plot below shows the forecast evolution plot for 24-hour snowfall in mm water on 2 April for points inside a 0.25x0.25 box over Linkoping. Observation - green hourglass, Concatenated 6-hour forecasts - green dot, ENS control –red, , ENS blue box-and-whisker, Model climate – cyan box-and-whisker. Ensemble mean as black diamonds. Triangle marks the maximum in the model climate based on 1800 forecasts.
The forecast seems to have over-predicted the snowfall, but we cannot rule out the risk of under-catching in the observation.
The plot below shows the forecast evolution plot for 2-metre temperature at 00UTC on 2 April inside a 0.25x0.25 box over Sodankyla. Observation - green hourglass, Concatenated 6-hour forecasts - green dot, ENS control –red, DestinE - purple dot, AIFS - cyan dot, ENS blue box-and-whisker, Model climate – cyan box-and-whisker. Ensemble mean as black diamonds. Triangle marks the maximum in the model climate based on 1800 forecasts.
AIFS is warmer than the other deterministic forecasts in the short-range. One can note that DestinE was warmer than all other forecasts for step 0, which is currently under investigation and is believed to be related to the vertical interpolation of temperature. However, all forecasts are much warmer than the observation.
The plot below shows the 2-metre temperature in the forecast from 2 April (Observation - black, ENS control -red, DestinE - purple, ensemble members - cyan). The forecasts missed the amplitude of the night-time cooling.
3.4 Monthly forecasts
3.5 Comparison with other centres
4. Experience from general performance/other cases
Cold temperature:
- No early signal for the extreme cold
- Not a nice forecast evolution
- All forecasts missed the amplitude of the cold.
5. Good and bad aspects of the forecasts for the event