5.3
Storm Climate and its Impacts on Surface Winds in CBHAR
5.3.1 Introduction
Storms can frequently invade the Chukchi-Beaufort seas region to impact both
regional climate and environment (e.g., Zhang et al. 2004). They play fundamental
contributing role in transporting atmospheric heat energy and water to the study area as
well as the entire Arctic Ocean. Surface wind associated with the storms exerts dynamic
forcing to change sea ice and ocean motions, create open leads, and redistributing sea ice
cover. Furthermore, intense storms can induce extreme winds, which may give rise to
high wave surges, cause flooding, coastal erosion, and properties damage, in particular
with retreated sea ice cover (e.g., Lynch et al. 2004). In this section, we will describe the
storm climate in the CBHAR covered study domain, and, especially, analyze its impacts
on surface wind fields.
Synoptic-scale storms are interactively steered by the atmospheric general
circulation pattern. In order to better understand storm climate, we first document the
circulation patterns covering the Beaufort and Chukchi seas. The regional climatological
circulation patterns are generally shaped by influences of the two semi-permanent
systems with the Beaufort high in the north and Aleutian low in the south. Therefore,
climatologically higher pressure appears in the north and lower pressure exhibits in the
south of the study domain throughout the year (Figure 5.3.1). Along with the invasion of
the Beaufort high from January to May, surface pressure increases over the Chukchi and
Beaufort seas. The lower pressure becomes more dominant in the second half of the year
when the Aleutian low gradually strengthens.
To further examine the seasonal cycle of the surface pressure governed by the
Beaufort high and the Aleutian low, we divided the study domain into the northern and
southern parts by the blue line shown in Figure 5.3.1. The SLPs demonstrate the similar
increasing variation from January to March over the both northern and southern
subdomains (Figure 5.3.2). This may suggest an increased influence of intensifying
Beaufort high and a decreased influence of weakening Aleutian low during this time
period. The intensity of the Beaufort high reaches its maximum in March and decreases
afterwards. In conjunction with this, SLPs decreases over the study domain from March
to August. In particular, the SLP decreasing rate accelerates in the northern part of the
domain from May to August. The SLP is about 4 hPa lower in August than it is in May.
Note that, during June-August, the areal-mean of the SLPs over the two sub-regions get
closer. During this period, both the Beaufort high and the Aleutian low are in their weak
state. Starting form September, both Beaufort high and Aleutian low intensifies. The SLP
increases in the north increases but, by the contrast, the SLP remains decreasing changes
in the south until December. The slightly southward shift of the Aleutian low during
December to next January favors the SLP to increase over the study domain (Figure
5.3.3).
Figure5.3.1. Climatological monthly mean sea level pressure (SLP,hPa) over the 31
years from 1979-2009. The blue line shows the divider of the northern and southern
sub-domains for seasonality analysis.
1023
SLP (hPa)
1021
1019
1017
1015
1013
1011
1009
1
2
3
4
5
6
7
8
9
10 11 12
Figure5.3.2. Climatological monthly mean SLP (hPa) over the northern (red) and
southern (blue) sub-domains over the 31 years from 1979-2009.
Figure5.3.3. Climatological monthly mean SLP (hPa) over the 31 years from 19792009 based on the NCEP/NCAR Reanalysis (Kalnay, 1996).
5.3.2 Storm Identification and Tracking Algorithm
A storm identification and tracking algorithm was employed in this study (Zhang
et al., 2004). To well capture the synoptic-scale storms in the high spatial and temporal
resolution CBHAR data, we modified the algorithm for its application in this research.
The modified criteria used in this algorithm are mainly described below.
1) If the SLP at one grid point is lower than its surrounding grid points that are 50 km
away, a cyclone candidate is identified with its center at this grid point. To
minimize impacts of thermodynamically and dynamically induced high-resolution
features on the identification of the synoptic-scale storm, we conducted an areal
smoothing of SLP over 25 points for each grid point.
2) The minimum pressure gradient calculated between the candidate storm center and
the surrounding grid points that are 50 km away is larger than 0.05 hPa (100 km)-1.
3) The SLP gradient between the four surrounding points at the 50 km radius and 100
km radius are compared and at least three of them are negative inward. The
criterion requesting three negative gradients, not all, is mainly to include some
storms with open SLP contours away from the candidate center.
4) If a candidate storm appears within a radius of 600 km at the same time, they are
considered to be one storm. Due to the high latitude and complex surface forcing,
the high resolution CBHAR shows that two separate storms can locate close to each
other. We therefore reduced the radius from Zhang et al.’s original 1200 km to 600
km in this study.
5) The storm has a minimum lifetime longer than 12 hours. If the location of a storm is
within 200 km from the location of a storm one hour earlier, the storm is considered
to move from the previous location. Otherwise, a new storm is considered to
generate.
In addition to the modified criteria according to Zhang et al. (2004), we have
newly added the following two more criteria specifically for our high-resolution data and
complex study area:
6) The candidate storm’s central SLP is less than 1005 hPa.
7) The moving distance during the whole lifetime of a storm is larger than 250 km.
These two criteria remove the complex terrain forced dynamic and thermodynamic lows.
5.3.3 Climatology and Variability of Storm Activity
Following Zhang et al. (2004), we employ four parameters to describe the
storm climate, including the number, duration, and intensity of storm, as well as an
integrative index, the cyclone activity index (CAI), which measures the overall storm
activity in the study region. We will conduct the analsysis at monthly basis. So, the
number of storm is defined as the number of storm trajectories within the study
domain for each month; the duration is the mean time periods over all number of
storms after they enter and before they leave the study domain, or after their
generation and before their distinct for each month; the intensity is the mean
difference between the storm central SLP and the climatological monthly mean SLP
at the corresponding grid points throughout the duration of each storm over all
number of storms for each month. The CAI is the sum of the differences between the
storm central SLP and the climatological monthly mean SLP at corresponding grid
points for all time steps within the study domain for each month. It integrates all
information from the number, duration, and intensity of storm.
We first develop storm climatology for the Chukchi-Beaufort Seas by applying
the modified storm identification and tracking algorithm described above to the CBHAR
data from 1979-2009 (Figure 5.3.4). The storm activity demonstrates an obvious
seasonal cycle. More numerous but weaker storms occur in summer, and less but stronger
storms appear in winter season over our study area. This climatological seasonality is
exactly consistent with previous finding for the whole Arctic Ocean by Zhang et al.
(2004). Considering this contrasting Arctic seasonality to the mid latitudes, Serreze and
Barrett (2008) conducted a follow-up and further found a minimum of the number of
storms in March. The duration of storms displays a much smaller seasonal fluctuations,
characterized by a mean duration about 30 hours throughout the year. The seasonal cycle
of the CAI indicates an overall storm activity intensifies from summer to the late winter
and weakens in spring over the study domain. Comparison of the four panels in Figure
5.3.4 suggests that interplay between the number and intensity of storms makes major
contributions to the CAI seasonal variation.
The seasonal cycle of storm activity is also well corresponding to that of the
regional atmospheric circulation patterns. The weakening storm activity during spring is
in a good agreement with the variation of the Beaufort high and Aleutian low inteisities.
As discussed above, the Beaufort high gets the strongest from March throughout May,
with a maximum SLP reaching 1023 hPa. Simultaneously, the Aleutian low weakens
obviously from March (Figure5.3.3).
Figure 5.3.4 Climatological monthly mean storm count, intensity(hPa), duration
(hour), and CAI (hPa) over the 31 years from 1979-2009 over the study domain.
Based on the climatological seasonal cyclone of storm activity and considering
the seasonal course of the regional atmospheric circulation pattern shown in Figure 5.3.1,
we split the year into four seasons to investigate year-by-year variability of and long-term
changes in storm activity. The four seasons are defined by the early winter (September December, when the Aleutian low gets the strongest); mid winter (January – February,
when the Beaufort high intensifies and the Aleutian low weakens); spring (March-May,
when storm activity is weak); and summer (June-August, when the Beaufort high is in its
weakest state).
The intensity of storms over the study domain exhibits obvious interannual
variability in all four seasons from 1979-2009 (Figure 5.3.5). A slight intensification
during the last 31 years appears in the late winter season, but there is no significant longterm changing trends have occurred. Consistent with the climatological analysis above,
the intensity interannual variability show seasonal differences and are generally stronger
in the late winter than in other seasons. The larger variability variance occurs in the late
winter and the smaller variance in summer.
Figure 5.3.5. The intensity (hPa) of storms over the study domain for the four
seasons from 1979-2009.
5.3.4 Impacts of storms on surface wind
Surface strong winds are always associated with intense storms. To examine how
fluctuation of storm intensity impacts the surface wind, we select strong winds when the
wind speed exceeds the 95th percentile. The computation of the 95th percentile wind
speed was introduced in Section 5.1.1. Both the intensity of storms and the frequency of
strong wind occurrences demonstrate large interannual variability across four seasons
(Figure 5.3.6). The two parameters are well correlated in the seasons of the late winter
and summer, with the correlation coefficients of 0.46 and 0.40 at the significance level of
95%. This suggests that invasion or intensification of storms over the study area may
contribute to an increase in surface pressure gradients, in particular when the Beaufort
high also emerges within the domain, and, in turn, increases the frequency of the strong
wind occurrences in these seasons. The insignificant low correlations show up for the
seasons of the early winter and spring, in particular before the mid 1990s. This could be
attributable to the dominance of stable Beaufort high. Generally, storms may not be able
to play a predominant role in steering surface wind.
Figure5.3.6. Standardized time series of the intensity of storms (red) and the
frequency of the 95th percentile strong wind speed occurrences (black) in each
season.
To further examine impacts of storms, we focus on the intense storms, in
particular, a previous study has indicated that a large proportion of strong wind events are
related to strong storms (Small et al, 2011). We first define the intense storms when its
minimum central SLP is lower than 990 hPa. The climatological number and intensity of
intense storms in each month are shown in Figure 5.3.7. The seasonal cycle of the
number of the intense storms presents differences from that of the all storms. There is
generally smallest number of intense storms from March to June. The number increases
after June and reaches its maximum in January. The intensity has relatively small
amplitude of the seasonal cycle. It also arrives at the strongest in January and gets
weakest in August. Considering these seasonal cycles, we will focus on the early and late
winter seasons in the following analysis.
Figure5.3.7. Climatological number and intensity (hPa) of intense storms for each
month from 1979-2009.
Correlations between the intensity of intense storms and the frequency of strong
wind occurrences are noticeably increased, in particular for the early winter up to 0.44
(Figure 5.3.8a). This changes, comparing with Figure 5.3.6, suggest that the intense
storms is a predominant contributing factor to the occurrence of surface strong winds.
This finding has important implication for assessing climate change impacts. Zhang et al.
(2004) found that there has been an intensifying trend of storm activity over this panArctic. If this also has occurred and will continually occur for our research area, the
recently detected and future projected upward trend of surface strong winds may be
attributed to the changes of storm in a warming climate. In this sense, we particularly
examined the intense storms year-by-year for the last 31 years (Figure 5.3.8). The results
indeed show a increased number of intense storms in the early winter.
Figure 5.3.8. Standardized intensity of intense storms (a, red) and frequency of
strong wind occurrence (b, black). The number of intense storms in the early winter
(c) and late winter (d).
5.3.5
More insight into the impacts of storm on surface winds: A case study
From above analysis, we can find that the variability of intense storms explains
about 20-30 % of the total variance of the surface strong wind variability. Although this
value suggests a predominant role of intense storms in a sense of long-term climate
variability, other factors can also interact with the storms to govern the surface strong
winds. Considering the regional atmospheric circulation patterns, interplay between
storms and the Beaufort high would be the key in shaping the strong wind field, including
three scenarios:
(1) Storms itself bring strong winds when move into or generated within the study
domain;
(2) Intensification of the Beaufort high causes strong winds; and
(3) Storms and the Beaufort high interact to cause strong wind.
To better understand how the above-mentioned three scenarios shape the surface
wind, we conduct a case analysis here. The selected case occurs on 6th – 7th October
1992. An intense storm moved southeastward from Russia to the Bering Strait, and
reached its minimum of SLP on 6th October 1992 (Figure5.3.9a). Strong winds occurred
over the northern Chukchi and Beaufort Seas, with the maximum wind speed exceeding
18 m/s along the Alaska coast of the Bering Strait. Interestingly, a narrow band of strong
wind extends from the Bering Strait into the land along the Brook Range, showing
topographic effects. The storm continually maintained its intensity on 7th October 1992
(Figure5.3.9b). The storm center moved northward to the Chukchi Sea, and the strong
wind also moved north. The Bering Strait strong wind moves northward along the coast.
In this case, the intense storm itself played the predominant role in causing the surface
strong wind. There was not a well-shaped Beaufort high formed.
During the late winter, it is very common that storms strongly interact with the
well-established Beaufort high. This can be exemplified by a case occurring on 21st
January 1991 (Figure 5.3.9c). The storm sits over the Chukchi Sea with the central SLP
of 997 hPa. At the same time, the Beaufort high was located to the north with a central
SLP higher than 1020 hPa. The allocation of the storm and the Beaufort high formed
large SLP gradients that generated 12~16 m/s wind. In this case, the intensity of the
storm is moderate. When the storm is stronger, more extreme wind can be expected. For
example, the case occurring on 19th January 1981 shows a storm over the East Siberia
coastal area with the central SLP of 978 hPa (Figure5.3.9d). The obviously larger SLP
gradients between the storm and the Beaufort high created strong winds with a maximum
wind speed exceeding 18 m/s.
Figure 5.3.9. Weather charts showing interplay between storms and the Beaufort
high in shaping the surface strong winds. Black contours: SLP (hPa); shaded area in
color: wind speed (m s-1); barb: wind vector.
According to the regional circulation analysis above, the Beaufort high reaches
the strongest and the Aleutian low, which represents time integration of daily-based
storms, in spring. Under this circumstance, the Beaufort high predominantly governs
surface winds. The strong wind appeared on 3rd March 1981 showing this case (Figure
5.3.9e). The Beaufort high central SLP was higher than 1032 hPa, and the associated
strong winds were around 16 m/s over the Chukchi Sea.
During summer, the Beaufort high is weak in climatology. There are usually no
winds as strong as in other seasons. The 95th percentile wind speed is less than 9 m/s.
Storms moved into Arctic Ocean or generated over the Arctic Ocean can cause relatively
large winds. For example, a storm occurred over the northern Beaufort Sea on 22nd July
1996 (Figure 5.3.9f). The storm central SLP was 994hPa, and associated large wind
speeds around 9 - 14 m/s.