Extratropical cyclones (ECs) have a crucial importance for shaping the weather and climate of midlatitudes. ECs dominate the daily variability of weather and cause the majority of extreme weather events like extreme surface winds and precipitation. Extreme surface winds caused by ECs cause substantial losses in the insured property when they hit the land and have, therefore, been closely studied. They have been studied both climatologically and in terms of case studies of individual impactful ECs. 

However, ECs that come to land are only a small portion of all ECs. ECs most frequently occur over the midlatitude oceans in the regions called the storm tracks. Even though the highest frequency of occurrence of ECs is over the storm tracks, comparatively less attention was paid to the ECs which cause extreme winds over these areas.

This thesis studies the ECs that cause extreme surface winds over the oceanic storm tracks, in the areas where the occurrence of ECs is the highest. More specifically, the papers in the thesis focus on the ECs and the extreme surface winds over the North Atlantic, North Pacific and Southern Ocean. We analyze various datasets, from observations and reanalysis products, to those produced by our climate modeling experiments. 

One of the key results of the thesis is that the extreme surface winds in the storm track areas are larger in the Northern Hemisphere than in the Southern Hemisphere, even though the Southern Hemisphere is known to have the stronger average surface winds. More specifically, the North Atlantic has the largest extreme surface winds, followed by the North Pacific and the Southern Ocean. The hemispheric difference in extreme surface winds exists because the extreme winds are stronger around winter ECs in the Northern Hemisphere, and the strongest in the North Atlantic.

Another important result from the thesis is that the ECs that cause the most extreme surface winds have a similar large-scale development in all major storm track basins. This development is characterized by the presence of a pre-existing downstream EC, which helps to create an environment conducive to the development of extreme wind-causing ECs. In addition, we connect the differences in extreme surface winds between the basins to differences in the mid-tropospheric Eady growth rates. The extremes of Eady growth rates, as well as the percentage of explosively deepening ECs, are higher in the Northern Hemisphere than in the Southern Hemisphere. Finally, the modeling studies we do show that the main drivers of the differences in the extreme surface winds (and the Eady growth rates) are spatial distributions of sea-surface temperatures and orography.

While the Southern Ocean is known to have the strongest annual-mean surface winds globally, it remains unclear whether surface wind extremes are stronger there than over the Northern Hemisphere basins. We address this question by analyzing reanalysis and satellite data sets and employing feature tracking to associate cyclones with surface winds. Consistent with previous work, we find the highest annual-mean and median winds over the Southern Hemisphere. However, we find a statistically distinguishable hemispheric asymmetry in extreme surface windspeeds, with the Northern Hemisphere having stronger extremes. The stronger extremes in the Northern Hemisphere are driven primarily by extreme windspeeds occurring during winter and in proximity to cyclones (within a 1,000 km radius around objectively tracked cyclone centers). Large-scale differences between basins likely play a role in shaping hemispheric asymmetries, as the Northern Hemisphere has higher extreme windspeeds above the boundary layer (700 hPa) and higher extremes of midtropospheric Eady growth rates.

This study investigates the role of large-scale atmospheric processes in the development of cyclones causing extreme surface winds over the central North Atlantic basin (30 to 60° N, 10 to 50° W), focusing on the extended winter period (October–March) from 1950 until 2020 in the ERA5 reanalysis product. Extreme surface wind events are identified as footprints of spatio-temporally contiguous 10 m wind exceedances over the local 98th percentile. Cyclones that cause the top 1 % most intense wind footprints are identified. After excluding 16 (14 %) of cyclones that originated as tropical cyclones, further analysis is done on the remaining 99 extratropical cyclones (“top extremes”). These are compared to a set of cyclones yielding wind footprints with exceedances marginally above the 98th percentile (“moderate extremes”). Cyclones leading to top extremes are, from their time of cyclogenesis, characterised by the presence of pre-existing downstream cyclones, a strong polar jet, and positive upper-level potential vorticity anomalies to the north. All these features are absent or much weaker in the case of moderate extremes, implying that they play a key role in the explosive development of top extremes and in the generation of spatially extended wind footprints. There is also an indication of cyclonic Rossby wave breaking preceding the top extremes. Furthermore, analysis of the pressure tendency equation over the cyclones' evolution reveals that, although the leading contributions to surface pressure decrease vary from cyclone to cyclone, top extremes have on average a larger diabatic contribution than moderate extremes.

This study examines the characteristics of several model parameter perturbation methodologies for ensemble simulations of cloud microphysical processes in convection. A simplified 1D model is used to focus the results on cloud microphysics without the complication of feedbacks to the dynamics and environment. Several parameter perturbation methods are tested, including non-stochastic and stochastic with various distributions and parameter covariance. We find that an ensemble comprised of different time-invariant parameters (non-stochastic) exhibits little bias, but small spread. In addition, its behavior does not respect the time evolution of convection through its various phases. Stochastic parameter (SP) methods in which no inter-parameter covariance is applied produce greater spread, but significant bias. The bias is particularly large for lognormal parameter perturbation distributions. The ensemble spread is retained and the bias reduced when time-varying parameter covariance is applied. In this case, the SP scheme is able to adapt to the time and state-dependent covariance structures and produce ensemble characteristics that are consistent with the specific microphysical processes operating at any given time. The results suggest that SP schemes would benefit from inclusion of parameter covariances, and specifically those that vary with the state of the system. It also suggests that a Normal or LogNormal SP scheme with no covariance may significantly impact the ensemble bias. Finally, the results indicate that high temporal and spatial resolution observations may be needed to characterize the variability in parameter values and covariance.