Detecting Extratropical Cyclones of the Northern Hemisphere with Single Shot Detector

In this paper, we propose a deep learning-based model to detect extratropical cyclones (ETCs) of the northern hemisphere, while developing a novel workflow of processing images and generating labels for ETCs. We first labeled the cyclone center by adapting an approach from Bonfanti et al. in 2017 and set up criteria of labeling ETCs of three categories: developing, mature, and declining stages. We then gave a framework of labeling and preprocessing the images in our dataset. Once the images and labels were ready to serve as inputs, an object detection model was built with Single Shot Detector (SSD) and adjusted to fit the format of the dataset. We trained and evaluated our model with our labeled dataset on two settings (binary and multiclass classifications), while keeping a record of the results. We found that the model achieves relatively high performance with detecting ETCs of mature stage (mean Average Precision is 86.64%), and an acceptable result for detecting ETCs of all three categories (mean Average Precision 79.34%). The single-shot detector model can succeed in detecting ETCs of different stages, and it has demonstrated great potential in the future applications of ETC detection in other relevant settings.

Intraseasonal variability of ocean surface wind waves in the western South Atlantic: the role of cyclones and the Pacific South American pattern

Extratropical cyclones are known to generate extreme significant wave height (swh) values at the ocean surface in the western South Atlantic (wSA), which are highly influenced by intraseasonal scales. This work aims to investigate the importance of intraseasonal timescales (30–180 d) in the regional climatology of waves and its atmospheric forcing. The variability is explained by analyzing the storm track modulation due to westerly winds. These winds present timescales and spatial patterns compatible with the intraseasonal component of the Pacific South American (PSA) patterns. The analyses are made using ECMWF’s ERA5 from 1979 to 2019 and a database of extratropical cyclones based on the same reanalysis. Empirical orthogonal function (EOF) analyses of the 10 m zonal wind and swh are used to assess the regime of westerlies and waves in the wSA. The EOF1 of the 10 m zonal wind (u10) presented a core centered at 45∘ W and 40∘ S, while the EOF2 is represented by two cores organized into a seesaw pattern with a center between 30–40∘ S and another to the south of 40∘ S. Composites of cyclone genesis and track densities as well as swh fields were calculated based on the phases of both EOFs. In short, EOF phases presenting cores with a positive (negative) u10 anomaly provide a favorable (unfavorable) environment for cyclone genesis and track densities and, therefore, positive (negative) swh anomalies. The modulation of the cyclone tracks is significant for extreme values of the swh. The spatial patterns of the EOFs of u10 are physically and statistically consistent with 200 and 850 hPa geopotential height signals from the Pacific, indicating the importance of the remote influence of the PSA patterns over the wSA.

Characteristics of extratropical cyclones and precursors to windstorms in northern Europe

Extratropical cyclones play a major role in the atmospheric circulation and weather variability and can cause widespread damage and destruction. Extratropical cyclones in northern Europe, which is located at the end of the North Atlantic storm track, have been less studied than extratropical cyclones elsewhere. Our study investigates extratropical cyclones and windstorms in northern Europe (which in this study covers Norway; Sweden; Finland; Estonia; and parts of the Baltic, Norwegian, and Barents seas) by analysing their characteristics, spatial and temporal evolution, and precursors. We examine cold and warm seasons separately to determine seasonal differences. We track all extratropical cyclones in northern Europe, create cyclone composites, and use an ensemble sensitivity method to analyse the precursors. The ensemble sensitivity analysis is a novel method in cyclone studies where linear regression is used to statistically identify what variables possibly influence the subsequent evolution of extratropical cyclones. We investigate windstorm precursors for both the minimum mean sea level pressure (MSLP) and for the maximum 10 m wind gusts. The annual number of extratropical cyclones and windstorms has a large inter-annual variability and no significant linear trends during 1980–2019. Windstorms originate and occur over the Barents and Norwegian seas, whereas weaker extratropical cyclones originate and occur over land areas in northern Europe. During the windstorm evolution, the maximum wind gusts move from the warm sector to behind the cold front following the strongest pressure gradient. Windstorms in both seasons are located on the poleward side of the jet stream. The maximum wind gusts occur nearly at the same time as the minimum MSLP occurs. The cold-season windstorms have higher sensitivities and thus are potentially better predictable than warm-season windstorms, and the minimum MSLP has higher sensitivities than the maximum wind gusts. Of the four examined precursors, both the minimum MSLP and the maximum wind gusts are the most sensitive to the 850 hPa potential temperature anomaly, i.e. the temperature gradient. Hence, this parameter is likely important when predicting windstorms in northern Europe.

Comparing Frontal Structures of Extratropical Cyclones in the Northwestern Pacific and Northwestern Atlantic Storm Tracks

The frontal structures of extratropical cyclones developing in the Northwestern Pacific storm track are relatively poorly understood compared with those in Europe and the Atlantic Ocean, for which representative conceptual models have been developed. In this paper, the structures of cyclones and their associated fronts in the Northwestern Pacific (NP), as well as in the Okhotsk Sea and Sea of Japan (OJ), are examined at their developing and mature stages using Japanese 55-year reanalysis dataset. Furthermore, the frontal structures in the NP are compared with those in the Northwestern Atlantic (NA). At the time of maximum deepening rate, cyclones in the NP are accompanied by strong warm and cold fronts, whereas cyclones in the OJ are more frequently accompanied by cold fronts than by warm fronts and tend to have stronger cold fronts than warm fronts. The weaker warm fronts than cold fronts to the east and northeast of cyclones in the OJ is likely due to the cyclones developing to the north and away from the region where the horizontal gradient of environmental potential temperature is strong. A comparison between mature cyclones in the NP and NA shows that the warm fronts in the NA tend to extend northeastward, whereas those in the NP extend more southeastward. These differences in warm fronts between NP and NA are suggested to be due to the difference in the horizontal structures of the warm currents between NP and NA.

Cyclones of Subtropical Origin in the Southwest Pacific: a Climatology and Aspects of Movement and Development

<p>A comprehensive study on cyclones of subtropical origin (STCs) in the Southwest Pacific is carried out. A brief history of the damage caused by STCs in New Zealand between 1990 and 2005 is given. It shows that approximately 2 to 3 times a year STCs come into the vicinity of New Zealand, mostly affecting the North Island and causing predominantly flood damage. A climatology is compiled with a cyclone track database covering 21 years, providing an overview of the behaviour and characteristics of STCs in this region. Distinct annual and seasonal patterns in frequency, tracks and intensity are revealed. Some of these patterns resemble those of tropical cyclones, in particular those undergoing extratropical transition, while others resemble those of extratropical cyclones in this region. In addition, it is shown that there is a significant increase in the number of summer STCs, which coincides with an increase in sea surface temperatures in the area. The structure and processes involved in the development of STCs are investigated in more detail using data from the United Kingdom Meteorological Office (UKMO) global model spanning 5 years (1999 to 2003). An analysis of the upper-level flow shows that STCs are steered into midlatitudes by upper-level baroclinic waves, m general through interaction with an upper-level trough. Differences in the structure and development of STCs can be attributed to the fact that upper-level baroclinic waves are able to propagate far into the sub tropics in this region. This is also the reason for the existence of three types of STCs, when differentiating by characteristics of their development process. Type 1 STCs are very similar to extratropical cyclones in structure and development. The structure and the development process of Type 3 STCs resemble more those of tropical cyclones. The initial development of Type 2 STCs is similar to that of Type 3, but they then undergo a transition, found to be very similar to that of tropical cyclones undergoing extratropical transition. Interseasonal variations in the upper-level flow over the Southwest Pacific are reflected in the behaviour and characteristics of STCs and subsequently the occurrence of the three types of STCs. During the colder seasons baroclinic waves frequently propagate relatively far into the subtropics in this region. This means STCs not only have a high chance of being picked up by an upper-level trough and undergoing extratropical transition, they are also able to actually form in the vicinity of a trough. Thus, during that time most STCs tend to be either Type 1 or 2. On the other hand, during summer, when baroclinic waves only occasionally propagate into the subtropics, there is a higher frequency of Type 3 STCs. In terms of weather-related threats to New Zealand, the interaction with an upperlevel trough is the cause for STCs coming into the vicinity of New Zealand, while the high rain rates that accompany them, and that are the cause for the extensive, mostly flood-related, damage, are attributed to their place of origin.</p>

Cyclones of Subtropical Origin in the Southwest Pacific: a Climatology and Aspects of Movement and Development

<p>A comprehensive study on cyclones of subtropical origin (STCs) in the Southwest Pacific is carried out. A brief history of the damage caused by STCs in New Zealand between 1990 and 2005 is given. It shows that approximately 2 to 3 times a year STCs come into the vicinity of New Zealand, mostly affecting the North Island and causing predominantly flood damage. A climatology is compiled with a cyclone track database covering 21 years, providing an overview of the behaviour and characteristics of STCs in this region. Distinct annual and seasonal patterns in frequency, tracks and intensity are revealed. Some of these patterns resemble those of tropical cyclones, in particular those undergoing extratropical transition, while others resemble those of extratropical cyclones in this region. In addition, it is shown that there is a significant increase in the number of summer STCs, which coincides with an increase in sea surface temperatures in the area. The structure and processes involved in the development of STCs are investigated in more detail using data from the United Kingdom Meteorological Office (UKMO) global model spanning 5 years (1999 to 2003). An analysis of the upper-level flow shows that STCs are steered into midlatitudes by upper-level baroclinic waves, m general through interaction with an upper-level trough. Differences in the structure and development of STCs can be attributed to the fact that upper-level baroclinic waves are able to propagate far into the sub tropics in this region. This is also the reason for the existence of three types of STCs, when differentiating by characteristics of their development process. Type 1 STCs are very similar to extratropical cyclones in structure and development. The structure and the development process of Type 3 STCs resemble more those of tropical cyclones. The initial development of Type 2 STCs is similar to that of Type 3, but they then undergo a transition, found to be very similar to that of tropical cyclones undergoing extratropical transition. Interseasonal variations in the upper-level flow over the Southwest Pacific are reflected in the behaviour and characteristics of STCs and subsequently the occurrence of the three types of STCs. During the colder seasons baroclinic waves frequently propagate relatively far into the subtropics in this region. This means STCs not only have a high chance of being picked up by an upper-level trough and undergoing extratropical transition, they are also able to actually form in the vicinity of a trough. Thus, during that time most STCs tend to be either Type 1 or 2. On the other hand, during summer, when baroclinic waves only occasionally propagate into the subtropics, there is a higher frequency of Type 3 STCs. In terms of weather-related threats to New Zealand, the interaction with an upperlevel trough is the cause for STCs coming into the vicinity of New Zealand, while the high rain rates that accompany them, and that are the cause for the extensive, mostly flood-related, damage, are attributed to their place of origin.</p>

A global analysis of the dry-dynamic forcing during cyclone growth and propagation

Mechanisms driving the intensification and propagation direction of extratropical cyclones are an active field of research. Dry-dynamic forcing factors have been established as fundamental drivers of the deepening and propagation of extratropical cyclones, but their climatological interplay, geographical distribution, and relatedness to the observed cyclone deepening and propagation direction remain unknown. This study considers two key dry-dynamic forcing factors, the Eady growth rate (EGR) and the upper-level induced quasi-geostrophic lifting (QGω), and relates them to the surface deepening rates and the propagation direction during the cyclones' growth phase. To this aim, a feature-based cyclone tracking is used, and the forcing environment is climatologically analysed based on ERA-Interim data. The interplay is visualized by means of a forcing histogram, which allows one to identify different combinations of EGR and QGω and their combined influence on the cyclone deepening (12 h sea-level pressure change) and propagation direction. The key results of the study are as follows. (i) The geographical locations of four different forcing categories, corresponding to cyclone growth in environments characterized by low QGω and low EGR (Q↓E↓), low QGω but high EGR (Q↓E↑), high QGω and low EGR (Q↑E↓), and high QGω and EGR (Q↑E↑), display distinct hot spots with only mild overlaps. For instance, cyclone growth in a Q↑E↑ forcing environment is found in the entrance regions of the North Pacific and Atlantic storm tracks. Category Q↓E↑ is typically found over continental North America, along the southern tip of Greenland, over parts of East Asia, and over the western North Pacific. In contrast, category Q↑E↓ dominates the subtropics. (ii) The four categories are associated with different stages of the cyclones' growth phase: large EGR forcing typically occurs earlier, during the growth phase at genesis, while large QGω forcing attains its maximum amplitude later towards maturity. (iii) Poleward cyclone propagation is strongest over the North Pacific and North Atlantic, and the poleward propagation tendency becomes more pronounced as the deepening rate gets larger. Zonal, or even equatorward, propagation on the other hand is characteristic for cyclones developing in the lee of mountain ranges, e.g. to the lee of the Rocky Mountains. The exact location of maximum QGω forcing relative to the surface cyclone centre is found to be a good indicator for the direction of propagation, while no information on the propagation direction can be inferred from the EGR. Ultimately, the strength of the poleward propagation and of the deepening is inherently connected to the two dry-dynamic forcing factors, which allow cyclone development in distinct environments to effectively be identified.