Atmosphere 2022, 13, 1634 
13 0f 21 
3.5. Future Scenarios 
3.5.1. Future Development of Meteorological Conditions 
Results from the previous sections show that both, wind direction (from Southeast) 
and wind speed are important to trigger ELWs. The question arises of whether these 
conditions might change in the future. We tackle this question by analysing the frequency 
of SE weather types, gale classes, as well as effective wind in four climate scenarios based 
on SMHI-LENS. 
Figure 12a shows the distribution of the occurrence of the derived LWT “SE” per year 
for ERA5 and the 50-member SMHI-LENS ensemble for the historical period (1971-2000) 
and the far future (2071-2100) in SSP1-2.6, SSP2-4.5, SSP3-7.0, and SSP5-8.5, respectively. In 
these plots, two medians are significantly different (p < 0.05) if their confidence intervals 
(triangular markers) do not overlap. Therefore, it can be concluded that SMHI-LENS 
historical is consistent with ERA5 regarding the median frequency for the LWT “SE”. 
With climate change according to SMHI-LENS, the frequency of LWT SE systematically 
shifts towards lower values, with SSP5-8.5 being significantly (p < 0.05) lower than the 
nistorical equivalent. 
- 
00 
50 
(a) Occurrence of LWT "SE" 
a0} 
5 601 
S 
"40| 
lan 
zZ 20 
{b) Occurrence of Gales 
f 
de’ 
or” PO 
0 {c} Occurrence of v_eff >= 95th percentile 
OU m—m—m—————— m— —A FU ou os A 
540 
5 
220 
) E— 
Legend 
ERAS5 (1971-2000) 
— EC-Farth3 historical (1971-2000) 
— EC-Earth3 SSP1-2.6 (2071-2100) 
EC-Earth3 5SP2-4.5 (2071—2100) 
EC-Earth3 SSP3-7,0 (2071-2100) 
EC-Farth3 SSP5-8.5 (2071—2100) 
V eff (142°) 
Figure 12. Boxplot of the occurrence of (a) the Lamb Weather Type (LWT) “SE”, (b) Gales, and 
{c) effective wind speeds > 95th percentile for ERA5 (light grey) and the 50-member SMHI-LENS 
ensemble (black) for the historical period (1971-2000) and respective runs for each the SSP-Scenarios 
12.6 (blue), 24.5 (orange), 37.0 (light red), and 5—8.5 (dark red) for the far future (2071-2100). 
Boxplots display the median (black dot), the interquartile range (25-27th Percentile, box), the extremes, 
Le., approximately +2.7 sigma and 99.3 coverage, of the distribution (whiskers), and outliers (circles) 
Notches, depicted as triangles around the median, correspond to q2 — 1.57 (q3 — ql)/sqrt (n) and 
q2 + 1.57 (g3 — g1)/sqgrt (n), where q2 is the median (50th percentile), g1 and qg3 are the 25th and 75th 
percentiles, respectively, and n is the number of observations. 
Figure 12b shows the distribution of the occurrence of gales (G, SG, and VSG) per 
year for ERA5 and SMHI-LENS again for the historical period (1971-2000) and the far 
future (2071-2100). A significant decrease for all scenarios is apparent, while the inter- 
scenario differences are considerably smaller. While the “drop” in gales might seem strong, 
investigations of near future time-slices (2015-2044 /2041-2070) reveal smaller decreases. 
In fact, the decrease already emerges towards the end of the historical period (1985-2014; 
not shown). 
Figure 12c finally shows the distribution of the occurrence of effective winds being 
above the 95th percentile of the climatological probability distribution function of the 
historical period. In ERA5, this 95th percentile approximately matches the value of 10 m/s 
and we have seen in Figure 11 that the majority of ELWs were subject to effective winds