Extra-Polar Climate Impacts on Antarctic Sea
Ice: Phenomenon and Mechanisms
Xiaojun Yuan
Lamont-Doherty Earth Observatory of Columbia University
61 Rt. 9W
Palisades, NY 10964
Submitted to Antarctic Science
December 2003
1
Abstract
Many remote and local climate variabilities influence on Antarctic sea ice at different
time scales. The strongest sea ice teleconnection at the interannual time scale was found
between ENSO events and a high latitude climate mode named Antarctic Dipole. The
Antarctic Dipole is characterized by an out-of-phase relationship between sea ice and
surface temperature anomalies in the South Pacific and South Atlantic. The Antarctic
Dipole manifests itself and persists seasons after being triggered by the ENSO forcing.
This study investigates what mechanisms transport tropical signals to Polar Regions and
why ADP anomalies persist at high latitudes. The study summarizes current knowledge
regarding ENSO/polar region teleconnection and puts a few isolated teleconnection
mechanisms into a coherent scheme. The hypothesised scheme suggests that the heat flux
due to the mean meridional circulation of regional Ferrel Cell and regional anomalous
circulation generated by stationary eddies are the two main mechanisms responsible for
the formation/maintenance of the Antarctic Dipole. The changes in the Hadley Cell and
jet stream in the subtropics associated with the ENSO link the tropical forcing to these
high latitude processes. Moreover, these two mechanisms operate in phase and are
comparable in magnitude. The positive feedback between the jet stream and stationary
eddies in the atmosphere, positive feedback within the air-sea-ice system, and seasonality
all reinforce the anomalies, resulting in long-lasting Antarctic Dipole anomalies.
Key word: teleconnection, Rossby wave, meridional circulation, jet stream, heat flux, the
Southern Hemisphere.
2
1. Introduction
The cryosphere in the Southern Ocean is an active component in the global climate. It is
also influenced by the many local, regional and remote climate variabilities at time
periods ranging from synoptic to geological scales. Since satellite observations became
available in the 1970s, many studies have been devoted to investigating the interannual
variability of sea ice and its covariation with global climate. A number of such studies
suggested that the Antarctic sea ice fields linearly covary with the El Niño/Southern
Oscillation (ENSO) phenomenon at the tropical Pacific (Chiu, 1983, Carleton, 1989,
Gloersen 1995, Simmonds and Jacka, 1995, White and Peterson, 1996, Yuan et al., 1996,
Smith et al., 1996, Ledley & Huang, 1997, Carleton et al., 1998, Yuan and Martinson,
2000, 2001, Harangozo, 2000, Kwok and Comiso, 2002, Martinson and Iannuzzi, 2003).
These ENSO related sea ice variabilities mainly occurred in key regions, such as the
Ross, Bellingshausen and Weddell seas, as well as the Southern Indian Ocean. Due to the
limitation of the sea ice time series, some of these earlier studies were not able to
evaluate the significance of the relationship or found that the relationship was
insignificant. With the accumulation of sea ice time series, sea ice/ENSO teleconnection
is found statistically significant in more recent studies, such as in Simmonds and Jacka,
(1995), Yuan and Martinson, (2000) and Martinson and Iannuzzi (2003).
The Antarctic sea ice was also found linked to some preferred climate patterns. For
example, Yuan and Martinson (2000) showed a few most common teleconnection
3
patterns between Antarctic sea ice and global surface air temperature (SAT) in addition to
the ENSO teleconnection. For example, sea ice is linked to an out-of-phase temperature
pattern between the Drake Passage and the central tropical Pacific, and the ice is related
to a regional teleconnection pattern in temperature between the southeast Pacific and
Weddell Sea. The sea ice also has some statistically significant relationships with other
extra-polar climate indices, such as the tropical Indian Ocean sea surface temperature
(SST), tropical land precipitation and Pacific North America (PNA) index. Though, the
ENSO teleconnection stands out as the most significant link of Antarctic sea ice with
extra-polar climate variability.
Although ENSO influences have been found in sea ice of many regions around the
Antarctic, only recently have studies identified the ENSO signal in Antarctic sea ice with
remarkable temporal and spatial coherency. Two types of climate modes in southern
subpolar regions were suggested linking to the ENSO variability at the tropics. White and
Peterson (1996) reported that the sea ice anomaly, as well as anomalies in sea surface
temperature, pressure and wind fields, propagates eastwards in a coherent pattern at
ENSO time scales, referred to as the “Antarctic Circumpolar Wave”. Yuan and Martinson
(2000, 2001), on the other hand, found that a quasi-stationary wave in sea ice, SST, and
SAT in the western Hemisphere of the Antarctic is strongly linked to the ENSO
variability and also dominates the interannual variability of the sea ice field. This
stationary wave was named the Antarctic Dipole (ADP).
Obviously, Antarctic sea ice is not the only component in the Southern Ocean influenced
4
by the ENSO variability. Many aspects of atmospheric circulation are altered by the El
Niño/La Niña events at the tropics. As reviewed by Carleton (2003), ENSO related
changes in the atmosphere mainly occur in the South Pacific ranging from mid to high
latitudes. These include changes in (1) sea level pressure (SLP), troposphere height,
stream function and surface winds (van Loon and Shea 1985, Mo and White, 1985,
Krishnamurti et al., 1986, Trenberth and Shea, 1987, Karoly 1989, Sinclair et al., 1997,
Kiladis and Mo, 1998), (2) the strength of the subtropical jet (STJ) and polar front jet
(PFJ) (Chen et al, 1996, Bals-Elsholz, et al, 2001), (3) precipitation and temperature
observed by weather stations in the South Pacific and surrounding continents (Smith and
Sterns, 1993, Cullather et al., 1996, Mashall et al., 1998, Noone et al. 1999, Bromwich et
al., 2000), (4) meridional and zonal heat fluxes (Carleton and Whalley, 1988, Chen et al.
1996, Cullather et al., 1996, Kreutz et al., 2000, Liu et al., 2002), (5) transient eddy
activity (Carleton and Carpenter, 1990, Carleton and Fitch, 1993, Sinclair et al., 1997,
Carleton and Song, 2000).
These extra-tropical atmospheric changes are likely
responsible for transporting the anomalous climate signal from the tropics to Antarctic
sea ice fields.
This study investigates and summarizes spatial and temporal characteristics of ENSO
influences in Antarctic sea ice with an emphasis on the ADP region. I also investigate
mechanisms that transport the tropical anomalous signal to high latitudes. The results
reveal that one tropical forcing can trigger two different atmospheric processes (altering
mean meridional circulation in the South Pacific and South Atlantic, and initiating the
stationary Rossby Wave) that both contribute to generating Antarctic Dipole anomalies in
5
high latitudes. The contributions from these two processes are in phase and comparable,
resulting in long-lasting ENSO anomalies in the air-sea-ice coupled system in the western
Hemisphere of the Antarctic. A coherent ENSO-ADP teleconnection scheme is
introduced in this study.
2. Data and Process
This study used satellite-observed sea ice concentrations (SIC) from 1978 to 2002.
Space-born passive microwave measurements of polar sea ice over more than twenty
years have provided observations of sea ice concentrations with a typical spatial
resolution of 25 km and near daily temporal resolution. These enable us to study sea ice
variabilities from synoptic to decadal time scales. In this study, I used approximately 24
years of monthly sea ice concentration data derived from brightness temperatures by the
bootstrap algorithm (Comiso, 1999).
To investigate teleconnections between Antarctic sea ice and global climate, I used
several variables from the National Centers for Environmental Prediction (NCEP)National Center for Atmospheric Research (NCAR) reanalysis data (Kalnay et al. 1996,
Kistler, et al. 2001), which include monthly SAT at the 1000mb, sea level pressure
(SLP), vector winds at 10m height, temperature and meridional winds at all pressure
levels, and zonal wind at 300mb. All of these variables except vector winds at 10m are in
class A of the reanalysis output, indicating that the variables are most strongly influenced
by observed data with minimal model impact. Vector winds at 10m are classified into
class B, indicating that the variable is strongly influenced by the model even though
observational data exist. Besides inadequate observational data input into the reanalysis
6
in the Southern Ocean, a few known problems and errors exist in the NCEP/NCAR
reanalysis, as discussed in Kistler et al. (2001). Among them, the PAOBs error has the
most impact on the Southern Ocean assimilation. PAOBs are estimates of the SLP
produced by Australia. NCEP/NCAR reanalysis unfortunately shifted 180˚ of longitude
in the use of the data for 1979-1992. The problem is significant on a synoptic scale but
becomes small in monthly mean fields (Kistler et al., 2001). To evaluate the reliability of
the NCEP/NCAR reanalysis in the Southern Ocean in the context of climate study,
especially teleconnection study, Yuan and Martinson (2000) replaced NCEP/NCAR
reanalysis SAT by a couple of observational and simulation datasets, such as the Jones
monthly near-surface temperature anomaly (Jones, 1994, Parker et al., 1994, 1995) and
the NASA Goddard Institute for Space Studies global surface air temperature anomaly
data (Hansen and Lebedeff, 1987, Reynolds and Smith, 1994), and repeated their
calculations on the EOF, correlation and significance test. They found no significant
differences in the results. Therefore, the NCEP/NCAR reanalysis monthly data in the
Southern Hemisphere were considered reliable observations for the purpose of studying
interannual variability in the paper.
In addition, I also used monthly Reynolds and Smith (1994) SST anomalies from
1981 to 2003, which were blended from ship, buoy and bias-corrected satellite data. To
focus on the interannual and decadal variability, all data were filtered by a low-pass
Gaussian filter to minimize sub-annual variability. Based on the Niño3.4 index from
NCEP, I selected five El Niño events (1982-1983, 1986-1987, 1987-1989, 1991-1992,
1997-1998) and five La Niña events (1984-1985, 1988-1989, 1995-1996, 1998-1999,
1999-2000) for composite analysis in the study.
7
3. The ENSO footprint in the Southern Ocean
Recent studies (Yuan and Martinson, 2000, 2001, Renwick, 2002) suggested that
Antarctic sea ice responds to ENSO forcing in a coherent spatial pattern – the ADP.
Warm ENSO events usually generate warm temperature and less sea ice anomalies in the
Pacific center of ADP, but cold temperature and more sea ice anomalies in the Atlantic
center of the ADP. In contrast, cold ENSO events are associated with cold temperature
and more sea ice in the Pacific center of the dipole but warm temperature and less sea ice
anomalies in the Atlantic center of the dipole. To examine the extent of global ENSO
impact in temperature fields, Liu et al. (2002) constructed the difference between the El
Niño annual SAT composite and the La Niña annual SAT composite. They revealed that
temperature anomalies in the ADP region are the largest ENSO temperature anomalies
outside of the tropical Pacific. It can be concluded that the ADP is the ENSO footprint at
southern high latitudes. However, high latitude processes significantly modulate this
ENSO anomaly. The ADP has its own high latitude characteristics in space and time.
3.1
A Coupled Climate Mode in the Air-Sea-Ice System
In general, the air-sea-ice system is highly coupled in the Southern Ocean. The ADP
represents such a coupled climate mode, particularly at the ENSO time scale. To illustrate
relationships between the atmosphere and cryosphere, I constructed monthly composites
of SAT, SIC, ice edge, SLP and surface wind anomalies for warm and cold ENSO
8
events, respectively. As an example, figure 1 shows the differences between El Niño
composites and La Niña composites in temperature and sea ice fields in May after the
ENSO matured at the tropics. The largest ENSO anomalies appear in the ADP region.
Clearly SAT and ice fields synchronize responses to the ENSO with warm temperature
and less ice in the Pacific and cold temperature and more ice in the Atlantic for a warm
event. The main ENSO impact on sea ice concentration field occurs near the ice edge.
The corresponding SLP anomaly field is characterized by a high (low) anomaly pressure
center in the Bellingshausen Sea for the warm (cold) events. The prevailing high pressure
center during ENSO warm events brings warm air from lower latitudes to the polar
region east of the Ross Sea and Amundsen Sea, and cold air from the Antarctic continent
to the open ocean in the Weddell Sea, creating dipole anomalies in these two regions
simultaneously (figure 1c). An exactly opposite situation occurs for ENSO cold events
(figure 1d). Such observed anomalies in SAT, SIC, SLP, and surface wind suggest a
thermodynamic causality relationship between the atmosphere and cryosphere.
3.2
Temporal Evolution
Temporal evolution of ENSO events and associated development of ADP anomalies are
examined by composites of SLP, SAT, SST and SIC anomalies. For example, figure 2
displays the seasonal evolution of La Niña events (illustrated by La Niña composites in
respective fields) starting from the northern fall when a cold SST anomaly has built up in
the tropical Pacific. The Southern Ocean SAT, SST and ice field have not responded to
the forcing in a coherent pattern yet, though a low SLP center develops in the Amundsen
9
Sea. During the next season (December of year0 to February of year1), the cold events
mature with large SAT and SST cold anomalies in the tropical Pacific. In the Southern
Ocean, the low SLP center moves eastward, while negative SAT/SST anomalies in the
South Pacific and positive SAT/SST anomalies in the Weddell Sea start to establish. Sea
ice cover is at its minimum in this austral summer season, though a weak positive
anomaly east of the Ross Sea and a weak negative anomaly in the Weddell Sea already
exists in the SIC field. In the following three seasons, the low SLP center sets in the
eastern South Pacific and persists. In the mean time, the La Niña phase of ADP
anomalies in the SAT, SST and SIC fields are fully developed, amplified and persist
while the ENSO anomaly at the tropics is demised. The strongest anomalies in
temperature and ice fields occur in the austral winter. The evolution of the ADP during
El Niño events is quite similar to that of La Niña events, but with opposite phase
anomalies. The high-pressure center in the Bellingshausen Sea established during warm
events only persists two seasons after an El Niño matures at the tropics contrasting three
seasons in the La Niña phase. The high-pressure center is weakened from austral fall to
austral winter after warm ENSO events, consistent with the finding of Garreaud and
Battisti (1999). Despite a larger variability from event to event in the El Niño phase
(compared to the La Niña phase), ADP anomalies in temperature and ice composite fields
are still amplified and persist three seasons after warm events mature in the composite
analysis (figure 3).
Apparently, the ENSO anomaly at the tropics changes the atmospheric circulation and
starts to affect pressure and temperature fields at southern high latitudes in austral
10
summer. As the Southern Hemisphere approaches austral winter, the ENSO anomaly in
the air-sea-ice coupled system amplifies and persists. It is likely caused by some positive
feedback among the atmosphere, ocean and cryosphere. For example, low ice
concentrations in centers of the ADP allow more heat flux to be released from the ocean
to atmosphere, which will further enhance warm SAT anomalies and melt more sea ice.
The ADP evolves and develops its high latitude characteristics in the air-sea-ice coupled
system after the tropical forcing is reduced. Therefore, the ADP is not just a reflection of
the ENSO variation at high latitudes.
The ADP indices were then constructed to examine their temporal variability
quantitatively. The SIC index (DPIice), SAT index (DPISAT) and SST index (DPISST) for
the ADP were calculated by subtracting the mean monthly SIC, SAT and SST anomalies
averaged in the Atlantic center (40˚-60˚W, 60˚-70˚S) from those in the Pacific center
(130˚-150˚W, 60˚-70˚S), respectively. The power spectra of the Niño3 index and ADP
indices reveal different characteristics of ENSO signals at the tropics and high latitudes.
In general, all three ADP indices have a strong variability at the 3 years period. The
DPISAT also peaks at the four and five years periods, while the DPIice has large interannual
variance in the four years period too. Moreover, strong decadal variability exists in all
three ADP indices and is a unique characteristic of the ADP. It is worth mentioning that
these three ADP indices were calculated from independent data sources. The consistency
of energy peaks in the interannual frequency band of the 3-5 years periods and the
decadal period from all three sources suggests that those peaks likely represent the
reality, even though twenty some years of time series are not long enough to examine the
11
decadal variability. On the other hand, the power spectrum of Niño3 has strong peaks in
four and five years periods and a relatively weaker peak at the 2.5 years period.
Apparently, ENSO and ADP closely share interannual variance in periods from 3-5 years,
although they do not mirror each other exactly. (figure 4).
3.3
Associated Atmosphere Circulation
Among many known ENSO responses in the atmosphere of the Southern Hemisphere, a
few described below likely influence the sea ice variability in the ADP region. A distinct
climate feature in the South Pacific is the presence of a split jet stream during austral
winter. The jet stream in the upper-level flow near Australia and New Zealand splits into
the subtropical jet (STJ) centered near 25˚S and polar front jet (PFJ) centered around
60˚S. Although the locations of the STJ and PFJ do not change much, the strength of
these jets change in response to ENSO cycles (Karoly, 1989, Kitoh, 1994, Chen et al.,
1996, Bals-Elisholz, 2001). For example, Bals-Elsholz et al. (2001) found that the slip jet
is significantly correlated with the Southern Oscillation Index. In a case study on a
complete warm-to-cold ENSO cycle from 1986-1989, Chen et al (1996) revealed that the
strong STJ and weak PFJ in the South Pacific are associated with the warm phase of the
ENSO, while the weak STJ and strong PFJ are associated with the cold phase of the
ENSO. Here I expand their analysis to include the entire Southern Ocean, five ENSO
warm events and five cold events for a composite analysis (Figure 5). During the warm
ENSO phase, the strong STJ centered near 25˚S between 160˚E and 120˚W extends
across the South Pacific to South America. This STJ turns southward in the South
12
Atlantic and becomes a single jet stream there. The PFJ is very strong in the South Indian
Ocean but becomes quite weak in the South Pacific. The split jet ends in South America.
During the cold ENSO phase, the PFJ is greatly enhanced in the South Pacific and
extends all the way across the tip of South America, where it joins the single jet stream in
the South Atlantic. The STJ in the South Pacific is much weaker, and turns southward
and joins the PFJ in the southeastern Pacific. The result is consistent with the earlier
studies in the South Pacific. It is worth mentioning that even the double-jet is not
profound in the southeastern Pacific. The location and structure of the jet there varies
significantly from the El Niño scenario to the La Niña scenario.
Blocking highs present another circulation characteristic in the Southern Ocean
influenced by ENSO cycles. The blocking events, usually identified by persistent
anomalous high-pressure centers that interrupt the mean zonal flow, frequently occur in
both the Southwest Pacific and the Southeast Pacific (Sinclair, 1996). The blocking highs
in the Southeast Pacific have been reported to associate with ENSO warm events in the
tropics (Rutllant and Fuenzalida, 1991, Renwick, 1998, Renwick and Revell, 1999).
Renwick and Revell (1999) suggested that the anomaly in the tropical outgoing longwave
radiation forces an extratropical wave response that results in enhanced persistent high
centers in the Southeast Pacific. On the other hand, low pressure centers occur more
frequently in the Southeast Pacific in response to the ENSO cold events, as shown in
Kiladis and Mo (1998). These low-pressure centers persist even longer after cold events
than high-pressure centers after warm events (figure2, figure3).
13
In a more recent study, Liu et al. (2002) calculated correlations between the Niño3 index
and mean meridional heat flux and found that polarward heat flux has opposite anomalies
in the South Pacific and South Atlantic basins. They claimed that the stronger Ferrel Cell
in the Pacific and the weaker Ferrel Cell in the Atlantic in response to the ENSO warm
events cause such heat flux distribution.
Figure 6 summarises the current knowledge about the ADP mode at southern high
latitudes and associated atmospheric circulation patterns. For ENSO warm events, warm
temperature and less sea ice occur in the Pacific center of the dipole while cold
temperature and more sea ice occur in the Atlantic center of the dipole simultaneously. A
persistent high-pressure center exists in the Bellingshausen Sea accompanying dipole
anomalies in temperature and ice fields for warm events. The subtropical jet is
strengthened and polar front jet is weakened in the South Pacific. At the same time the
strong jet stream swings polarward in the South Atlantic. In addition, polarward heat flux
due to the regional mean meridional circulation is strengthened in the South Pacific but
weakened in the South Atlantic. The ADP anomalies and associated atmospheric
circulation patterns for the La Niña scenario are just mirror images of that for the El Niño
scenario.
4. Mechanisms for the formation of the ADP
4.1
Mechanisms for the Polar-tropical teleconnection
A number of studies (van Loon and Shea, 1987, Karoly, 1989; Mo and Higgins, 1998;
14
Kiladis and Mo, 1998, Garreaud and Battisti, 1999) have attributed ENSO teleconnection
to the stationary Rossby Wave that propagates the tropical signal to southern high
latitudes. The Rossby Wave train usually responds to the change of the tropical
convection due to anomalous heating (Hoskins and Karoly, 1981, Mo and Higgins,
1998). This wave train is one predominant circulation pattern in the Southern Hemisphere
(Kidson, 1999) with timescales ranging from synoptic (Mo and Ghil, 1987), to
intraseasonal (Mo and Higgins, 1998), to interannual (Kidson, 1988). The wave train
also forms the Pacific and South America (PSA) pattern (Mo and Higgins, 1998), the
counterpart of the PNA pattern in the Northern Hemisphere. Karoly (1989) illustrated the
Rossby Wave and PSA pattern in the early phase of warm events based on the composite
of three El Niño events. In response to the warm ENSO events, the PSA pattern consists
of a low pressure center east of New Zealand, a high pressure center in the subpolar
region of the Southeast Pacific, and another low pressure center in South America and the
South Atlantic. The high-pressure center is consistent with frequent blocking highs in this
region (Ruttland and Fuenzalida, 1991, Sinclaire, 1996, Renwick, 1998) and consistent
with the persistent anomalous high pressure center discussed in section 3 (figure 3) in
response to the tropical warming. Evidently, the Rossby Wave can effectively transport
tropical signals to high latitudes. However, it is not the only mechanism for the tropicalpolar teleconnection.
The mean atmospheric circulation also plays a role in linking climate variabilities at the
tropics and in polar regions. As discussed above, the jet stream in the South Pacific varies
in response to the ENSO variation. Recent studies, such as Chen et al. (1996) diagnostic
15
case study and Rind et al. (2001) modeling study attributed the changes in the jet stream
to the tropical-polar teleconnection. Rind et al. pointed out that the jet stream shifts
equatorward in the South Pacific and polarward in the South Atlantic in response to the
same warm anomaly in the tropical Pacific. That alternates the storm distribution in these
two oceans, leading to less cyclones in the South Pacific subpolar region and more
cyclones in the South Atlantic subpolar basin. Indeed, ENSO signals have been observed
in the transient eddy activity in the Southern Ocean (Carleton and Carpenter, 1990,
Carleton and Fitch, 1993, Sinclair et al., 1997, and Carleton and Song, 2000). However,
Simmonds (1996) and Liu et al. (2002) suggested that cyclone distribution is not directly
related to sea ice distribution. The changes in the storm track could influence the ice field
indirectly. For example, strengthened PFJ in the South Pacific during cold events
increases cyclone activity in the polar region and reinforces an anomalous low-pressure
center in the Southeast Pacific (Chen et al, 1996). On the other hand, the variation of the
Ferrel Cell in response to the ENSO forcing (as mentioned above) is another example of
the mean circulation contributing to the tropical/polar teleconnection.
Convergence/divergence of meridional eddy heat fluxes likely influence the regional
meridional circulation (Liu et al, 2002).
4.2
Mechanisms for the formation of the ADP – a coherent scheme
All above-mentioned atmospheric processes influence sea ice distribution and/or
contribute to the formation of the ADP. The questions remaining are if they are
independent or interrelated, and if one process dominates others? I believe that all these
processes are interrelated and work in phase to generate and maintain ADP anomalies in
16
temperature and ice fields of southern high latitudes. Figure 7 shows a schematic
illustrating how the tropical anomalous signal is transported to the polar region through
an atmospheric bridge and creates ADP anomalies during an ENSO warm event. In the
scenario of an El Niño event, the warm SST in the tropical Pacific enhances the tropical
convection and meridional thermal gradient from the equator to the pole in the Pacific.
This strengthens and contracts the Hadley Cell (Rind et al. 2001). As a consequence, the
subtropical jet is strengthened and the storm track is shifted equatorward in the South
Pacific. The same tropical warming shifts the zonal circulation eastward so that its
descending branch takes place in the tropical Atlantic, which relaxes and expands the
Hadley Cell there. Consequently, the storm track shifts polarward in the South Atlantic
(Rind et al., 2001). The changes in the jet stream and the regional Hadley Cells
apparently result in an enhanced Ferrel Cell in the South Pacific and weakened Ferrel
Cell in the South Atlantic. Therefore, more (less) heat is transported into the polar region
in the South Pacific (South Atlantic) in the lower level atmosphere (Liu et al., 2002).
Therefore, the variation of polarward heat transport due to the regional mean meridional
circulation directly contributes to the formation of out-of-phase anomalies in the South
Pacific and South Atlantic. Meanwhile, the same tropical warming can trigger the
stationary Rossby wave train that results in the PSA pattern in mid-high latitudes. In
response to the warm event, an anomalous high-pressure center of the PSA sets in the
Bellingsgausen Sea and creates a regional circulation that brings warm air from lower
latitudes to the polar region in the South Pacific and cold Antarctic air out to the open
ocean in the Weddell Sea. This regional anomalous circulation is clearly capable of
generating ADP anomalies simultaneously in these two basins. Moreover, the weakened
17
PFJ means less cyclone activity in the Southeast Pacific region, which encourages
persistent high-pressure anomalies in this polar region. In addition, frequent cyclones in
the Atlantic basin, which are usually accompanied by ridges in the upstream, favor the
high pressure center in the Bellingshausen Sea too. This not only implies that the
variation of the jet stream indirectly contributes to the ADP formation, but also suggests
that the positive feedback among the jet stream, storm distribution and stationary eddy
activity in this particular region prolongs and maintains the high pressure center of the
PSA.
The atmospheric circulation in the La Niña scenario (Figure 8) is just the opposite of the
El Niño scenario. The tropical cooling relaxes the Hadley Cell and weakens the Ferrel
Cell in the Pacific. The zonal circulation shifts westward in the tropics due to the
shrunken warm pool, so the descending branch is located in the eastern tropical Pacific.
Therefore, the Hadley Cell is relatively strong in the South Atlantic due to the warm SST
at the tropics and lack of competition from the descending branch of the zonal
circulation. The strong Ferrel Cell is observed in the South Atlantic too. As a
consequence, the PFJ is enhanced in the South Pacific and the jet stream shifts
equatorward in the South Atlantic, creating an alternating storm distribution in these two
basins. The changes in the regional Ferrel Cell generate more polarward heat flux in the
Atlantic and less polarward heat flux in the Pacific, creating ADP anomalies for the cold
ENSO phase. The Rossby wave train creates a cold phase PSA pattern with a lowpressure center in the Bellingshausen Sea. This low-pressure center is reinforced by the
rich cyclone activity associated with the strong PFJ there. The regional anomalous
18
circulation generated by this persistent low pressure center also directly contributes to the
formation of the ADP cold phase anomalies.
Therefore, the changes of polarward heat flux due to the changes in the regional Ferrel
Cell and the regional anomalous circulation created by the persisting high/low centers
associated with the Rossby wave train in the southeast Pacific are the two critical
mechanisms for the formation of the ADP. These two processes are both initiated by the
ENSO events at the tropics. The changes in the Hadley Cell and Jet Stream link these two
high-latitude processes to the tropical forcing. In addition, the storm distribution in
response to the changes of the Jet Stream reinforces anomalous pressure centers in the
Southeast Pacific, representing the positive feedback between the mean flow and
stationary eddies, and contributing to the persistence of anomalous pressure centers.
4.3
The relative importance of the two ADP formation mechanisms
To evaluate the relative importance of two mechanisms that form the ADP anomalies, the
meridional heat transports due to these two processes are investigated here. The total
meridional heat transport from the atmosphere can be represented as the following:
Total meridional heat transport =< V T > + < V * T* > + < V 'T'> ,
(1)
where the right-hand side terms represent the heat transport by the mean circulation,
†
stationary eddies and transient eddies, respectively, and <> indicates the zonal mean
(Peixoto and Oort, 1992). Since the sea ice distribution seems unrelated to transient
eddies (Simmonds, 1996, Liu et al, 2002), the emphasis here is given to the heat
transports induced by the mean meridional circulation and stationary eddies. The heat
19
flux due to the mean meridional circulation ( V T ) is calculated by monthly meridional
wind and monthly temperature subtracted by a global mean temperature. The stationary
†
eddy heat flux ( V * T*) is calculated by the monthly mean meridional wind and
temperature minus their zonal mean respectively at each grid point. In terms of zonal
†
mean heat transport climatology, the mean circulation transports an order degree more
heat than stationary eddies do (Peixoto and Oort, 1992). However, the interannual
variability of the mean circulation heat flux is comparable to that of the stationary eddy
heat flux. Figure 9 displays ENSO composites of meridional heat fluxes due to these two
processes in the western Hemisphere. In the El Niño scenario, the mean circulation heat
flux is negative (polarward) in the subpolar regions north of the Ross Sea and Amundsen
Sea and positive (equatorward) north of the Bellingshausen Sea and Weddell Sea. The
stationary eddy heat flux has the same phase of anomaly pattern and similar magnitudes
of the anomalies as the mean circulation heat flux in these regions, except its anomaly
centers occur further south than that of the mean heat flux. In the La Niña scenario, the
anomalies flip sign and anomaly centers shift westward compared to the ENSO warm
phase in both mean heat flux and stationary eddy heat flux. Although the magnitude of
the mean circulation heat flux is slightly higher than that of the stationary eddy heat flux,
it cannot dominate the contribution from stationary eddies, in contrast to the zonal mean
climatological heat fluxes. These two processes actually supplement each other: the mean
circulation heat flux dominates the subpolar regions while the stationary eddy heat flux
dominates the polar seas. They both influence the sea ice variability.
4 Summary and discussion
20
The Antarctic Dipole, a predominant high latitude mode in the western Hemisphere of the
Southern Ocean, is closely related to the ENSO variation at the tropics. It is characterized
by out-of-phase anomaly centers in the South Pacific and South Atlantic synchronized in
the atmosphere, cryosphere and ocean. In response to warm ENSO events, warmer
temperature and less sea ice occur in the South Pacific, and colder temperature and more
sea ice occur in the South Atlantic, vice versa for cold ENSO events. These ENSO
related anomalies in temperature and sea ice usually persist three to four seasons after
ENSO events mature at the tropics. Two dynamic processes in the atmosphere are critical
in the formation of the ADP. First, the regional mean meridional circulation, namely the
regional Ferrel Cell, creates enhanced (weakened) polarward heat flux at the lower
atmosphere in the South Pacific and weakened (enhanced) polarward heat flux in the
South Atlantic in response to warm (cold) ENSO events. Second, the stationary Rossby
wave creates a persistent anomalous high (low) pressure center in the vicinity of the
Bellingshausen Sea in response to warm (cold) ENSO events, generating an anomalous
regional circulation that directly contributes to creating opposite phase anomalies in the
South Pacific and South Atlantic simultaneously. Both processes are initiated by the
tropical forcing. Variation of the Hadley Cell and jet stream in response to ENSO events
connects the tropical forcing and these two high latitude processes. Moreover, these two
processes work in-phase to produce the dipole anomalies. Their contributions to the
dipole formation in terms of heat flux are comparable in magnitude. The positive
feedback between the mean circulation and stationary waves in the atmosphere, such as
weak (strong) polar front jet in the southeast Pacific encourages the anomalous high
21
(low) pressure center during ENSO warm (cold) events, adds an enhancement to the
dipole formation mechanism. On the other hand, the positive feedback within the air-seaice system also reinforces the existing anomalies. For example, warmer air temperatures
melt sea ice and allow more heat to be transported from the ocean to the atmosphere,
which warms air temperature even more. These positive feedbacks in the atmosphere and
within the air-sea-ice coupled system are likely the reasons for the long-lasting anomalies
in the ADP regions after the tropical forcing is diminished.
Another important element in the development of the ADP is the seasonality of the
Southern Hemisphere. ENSO events mature at the tropic in austral summer when sea ice
cover is at its minimum. The weak temperature gradient across the air-sea interface limits
heat and momentum fluxes between the atmosphere and ocean. Although an anomalous
ENSO signal could already be transferred to southern high latitudes through the
atmospheric bridge, the ADP anomalies are weak. When the Southern Hemisphere
approaches austral fall and winter, air temperature drops and sea ice grows, air-sea-ice
interaction intensifies. Consequently, the ADP anomalies start to amplify because of
positive feedback within the air-sea-ice coupled system even though the tropical forcing
is demised. Therefore, seasonality plays a critical role in the long-lasting ADP
phenomenon.
Even though ENSO’s impact on Antarctic sea ice is the main focus here, the sea ice field
is also regionally influenced by other climate modes at mid-high latitudes of the Southern
Ocean. The Southern Annular Mode (SAM), marked by a zonally symmetric but out-of-
22
phase pressure anomalies between mid and high latitudes, is a dominant climate
variability in the Southern Hemisphere (Thompson and Wallace, 2000), explaining 50%
of the monthly SLP variance over the Antarctica (Gong and Wang, 1999). This climate
mode has its share of impact on the sea ice field, particularly in the South Indian Ocean
and South Atlantic Ocean (Visbeck, personal communication). The Semi-annual
oscillation (SAO) describes another zonally symmetric mode in the southern extratropics, which is characterized by the twice-yearly enhancement in meridional gradients
of temperature and pressure fields (van Loon, 1984, Simmonds and Jones, 1998, Walland
and Simmonds, 1999). The atmospheric convergence line with a strong half-year cycle
exerts significant influences on the seasonal asymmetric behavior of ice extent: slowly
advancing equatorward in the winter and fast retreating in summer (Enomoto and
Ohmura, 1990). The SAO also influences the open water areas within the sea ice pack
(Watkin and Simmonds, 1999). In addition, the quasi-stationary wave-3 pattern in the
southern mid-high latitudes, a predominant winter mode in pressure/wind fields (van
Loon, 1972), actively interacts with the sea ice field beneath. Yuan et al., (1999) showed
that three southerly branches of the wave-3 pattern coincide with three northward
maximum extent of sea ice edge, indicating the role of the wave-3 pattern in advancing
the ice edge. Moreover, the wave-3 pattern is positively coupled with the sea ice
distribution, promoting eastward propagation of the ice edge maxima and providing
preferred locations for cyclonegenesis in the open ocean north of the ice cover. All above
mentioned high-latitude climate modes, together with remote influences from the tropics,
form atmospheric forcing exerted on sea ice from above. The polar ocean with huge heat
contest, on the other hand, influences the ice from below (Martinson, 1990). More
23
importantly, the interaction among atmosphere, cryosphere and ocean ultimately
determines the complex variability of the Antarctic sea ice field.
Acknowledgement
The author thanks D. Rind, M. Ting and I. Simmonds for their insightful discussions on
climate dynamics. C. Li’s computational assistance is greatly appreciated. Special thanks
go to the National Snow and Ice Center for their prompt updates of the microwave sea ice
concentration data. This study is supported by NASA grants NAG5-12586 and JPLCIT
1216483. Lamont contribution number is xxxx.
24
References
Bals-Elsholz, T. M., E. H. Atallah, L. F. Bosart, T. A. Wasula, M. J. Cempa, and A. R.
Luro, The winter Southern Hemisphere split jet: Structure, variability, and
evolution. J. Climate, 14, 4191-4215, 2001.
Bromwich, D. H., A. N. Rogers, P. Kallberg, R. I Cullather, J. W. C. White, and K. J.
Kreutz, ECMWF analyses and reanalyses depiction of ENSO signal in Antarctic
precipitation. J. Climate, 13(8), 1406-1420, 2000.
Carleton, A.M, Antarctic sea-ice relationships with indices of the atmospheric circulation
of the Southern Hemisphere. Climate Dynamics, 3, 207-220, 1989.
Carleton, A. M., Atmospheric teleconnections involving the Southern Ocean. J. Geophys.
Res., Vol. 108, No. C4, 1029/2000JC000379, 2003.
Carleton, A. M. and D. Whalley, Eddy transport of sensible heat and the life history of
synoptic systems: A statistical analysis for the Southern Hemisphere winter,
Meteorol. Atmos. Phys., 38, 140-152, 1988.
Carleton, A. M and D. A. Carpenter, Satellite climatology of “polar lows” and broadscale
climatic associations for the Southern Hemisphere, Int. J. Climatol., 10, 219-246,
1990.
Carleton, A. M and M. Fitch, Synoptic aspects of Antarctic mesocyclones, J. Geophys.
Res., 98, 12,997-13,018, 1993.
Carleton, A. M., G. John, and R. Welsch, Interannual variations and regionality of
Antarctic sea-ice-temperature associations, Ann. Glaciol., 27, 403-408, 1998.
25
Carleton, A. M. and Y. Song, Satellite passive sensing of the marine atmosphere
associated with cold air mesoscale cyclones, Prof. Geogr. 52, 289-306, 2000.
Chen, B., S. R. Smith, and D. H. Bromwich, Evolution of the tropospheric split jet over
the Southern Pacific Ocean during the 1986-89 ENSO cycle. Mon. Weather Rev.,
124, 1711-1731, 1996.
Chiu, L. S., Antarctic sea ice variations 1973-1980. Variations in the Global Water
Budget, A. Street-Perrott, M. Beran, and R. Ratcliffe, Eds. Reidel, 301-311, 1983.
Comiso, J. Bootstrap sea ice concentrations for NIMBUS-7 SMMR and DMSP SSM/I.
Boulder, CO, USA: National Snow and Ice Data Center. Digital media, 1999,
updated 2002.
Cullather, R. I., D. H. Bromwich, and M. L. Van Woert, Interannual variations in
Antarctic precipitation related to El Niño – Southern Oscillation, J. Geophys.
Res., 101, 19,109-19,118, 1996.
Enomoto, H. and A. Ohmura, The influence of atmospheric half-yearly cycle on the sea
ice extent in the Antarctic. J. Geophys. Res., 95, C6, 9497-9511, 1990.
Garreaud, R. D. and D. S. Battistti, Interannual (ENSO) and interdecadal (ENSO-like)
variability in the Southern Hemisphere tropospheric circulation, J. Climate, 12,
2113-2123, 1999.
Gloersen, P, Modulation of hemispheric sea-ice cover by ENSO events. Nature, 373,
503-506, 1995.
Hansen, J. E. And S. Lebedeff. Global trends of measured surface air temperature. J.
Geophys. Res., 92. 13345-13372, 1987.
26
Harangozo, S. A., A search for ENSO teleconnection teleconnections in the west
Antarctic Peninsular climate in austral winter, Int. J. Climatol., 20, 663-678,
2000.
Hoskin, B. J. and D. J. Karoly, The steady linear response of a spherical atmosphere to
thermal and orographic forcing. J. Atmos. Sci., 38, 1179-1196, 1981.
Jones, P.D., Hemispheric surface air temperature variations: a reanalysis and an update to
1993. J. Climate, 7, 1794-1802, 1994.
Kalnay, E., M. Kanamitsu, R. Kistler, W. Collins, D. Deaven, L. Gandin, M. Iredell, S.
Saha, G. White, J. Woollen, Y. Zhu, M. Chelliah, W. Ebisuzaki, W. Higgins, J.
Janowiak, K. C. Mo, C. Ropelewski, J. Wang, A. Lettma, R. Reynolds, R. Jenne,
and D. Joseph, The NCEP/NCAR 40-year reanalysis project, Bulletin of the
American Meteorological Society, 77: 437-471, 1996.
Karoly, D. J, 1989: Southern Hemisphere Circulation Features Associated with El NiñoSouthern Oscillation Events. J. Climate, 2, 1239-1252.
Kidson, J. W., Interannual variations in the Southern Hemisphere circulation, J. Climate,
1, 1177-1198, 1988.
Kidson, J. W., Principal modes of Southern Hemisphere low frequency variability
obtained from NCEP-NCAR reanalyses. J. Climate, 12, 2808-2830, 1999.
Kiladis, G. N. and K. C. Mo, Interannual and intraseasonal variability in the Southern
Hemisphere, in Meteorology of the Southern Hemisphere, Am. Meteorol. Soc.
Meteorol. Monogr., vol., 27, edited by D. J. Karoly and D. G. Vincent, pp307337, Am. Meteorol. Soc., Boston, Mass., 1998.
27
Kistler, R., E. Kalnay, W. Collins, S. Saha, G. White, J. Woollen, M. Chelliah, W.
Ebisuzaki, M. Kanamitsu, V. Kousky, H. van den Dool, R. Jenne, and M. Fiorino,
The NCEP-NCAR 50-year reanalysis: monthly means CD-ROM and
documentation. Bull. Am. Meteor. Soc., 82, 247-268, 2001.
Kitoh, A., Tropical influence on the Southern Pacific double jet variability. Proc.
NIPRSymp. Polar Meteor. Glaciol., 8, 34-45, 1994.
Kreutz, K. J., P. A. Mayewski, I. I. Pittalwala, L. D. Meeker, M. S. Twickler , and S. I.
Whitlow, Sea level pressure variability in the Amundsen Searegion inferred from
West Antarctic glaciochemical record, J. Geophys. Res., 105, 4047-4059, 2000.
Krishnamurti, T.N.,S-H Chu, and W. Iglesias, On the sea level pressure of the Southern
Oscillation. Arch. Met. Geoph. Biocl. Ser. A 34, 385-425, 1986.
Kwok, R. and J. C. Comiso, Southern Ocean climate and sea ice anomalies associated
with the Southern Oscillation. J. of Climate, 15, 487-501, 2002.
Ledley, T. S. And Z. Huang, A possible ENSO signal in the Ross Sea. Geophysical
Research Letters, Vol. 24, 3253-3256, 1997.
Liu, J., X. Yuan, D. Rind and D. G. Martinson, Mechanism Study of the ENSO and
Southern High Latitude Climate Teleconections. Geophysical Research Letters,
Vol. 29, No. 14, 1029/2002GL015143. 2002.
Martinson, D.G., Evolution of the Southern Ocean Winter Mixed Layer and Sea Ice:
Open Ocean Deepwater Formation and Ventilation, J. Geophys. Res. 95, C7,
11641 - 11654, 1990.
28
Martinson, D. G. and R. A. Iannuzzi, Spatial/temporal patterns in Weddell gyre
characteristics and their relationship to global climate. J. Geophys. Res., Vol. 108,
No. C4, 1029/2000JC000538, 2003.
Marshall, G. J., J. Turner, and W. D. Miners, Interpreting recent accumilation records
through an understanding of the regional synoptic climatology: An example from
the southern Antarctic Peninsular, Ann. Glaciol. 27, 610-616, 1998.
Mo, K. C. and M. Ghill, Statistics and dynamics of persistent anomalies, J. Atmos. Sci.,
44, 877-901, 1987.
Mo, K. C. and R. W. Higgins, The Pacific-South America modes and tropical convection
during the Southern Hemisphere winter. Mon. Weather Rev., 126, 1581-1596,
1998.
Mo, K. C. and G. H. White, 1985: Teleconnections in the Southern Hemisphere. Mon.
Wea. Rev., 113, 22-37.
Noone, D., J. Turner, and R. Mulvaney, Atmospheric signals and characteristics of
accumulation in Dronning Maud Land, Antarctica. J. Geophys. Res., 104, 19,19119,211, 1999.
Parker, D.E., Jones, P.D., Bevan, A. and Folland, C.K., Interdecadal changes of surface
temperature since the late 19th century. J. Geophys. Res., 99, 14373-14399, 1994.
Parker, D.E., Folland, C.K. and Jackson, M., Marine surface temperature: observed
variations and data requirements. Climatic Change, 31, 559-600, 1995.
Peixoto, J. P. and A. H. Oort, Physics of Climate. American Institute of Physics, New
York, pp520, 1992.
29
Renwick, J. A., ENSO-related Variability in the Frequency of South Pacific Blocking,
Mon. Weather. Rev., 126, 3117-3123, 1998.
Renwick, J. A., Southern Hemisphere circulation and relationships with sea ice and sea
surface temperature. J. Climate, 15 (21), 3058-3068, 2002.
Renwick, J. A. and M. J. Revell, Blocking over the South Pacific and Rossby Wave
Propagation, Mon. Weather Rev., 127, 2233-2247, 1999.
Reynolds, R. W. and T. M. Smith, Improved global sea surface temperature analyses. J.
Climate, 7, 929-948, 1994.
Rutllant, J. and H. Fuenzalida, Synopic aspects of the central Chile rainfall variability
associated with the Southern Oscillation, Int. J. Climatol., 11, 63-76, 1991.
Simmonds, I. and T. H. Jacka, Relationships between the Interannual Variability of
Antarctic Sea Ice and The Southern Oscillation. J. Climate, 8, 637-647, 1995.
Simmonds, I., Climatic role of Southern Hemisphere extratropical cyclones and their
relationship with sea ice, Papers and proceedings of the royal society of
Tasmania, 130, 95-100, 1996
Simmonds, I. and D. A. Jones, The mean structure and temporal variability of the
semiannual oscillation in the southern extratropics. Int. J. Climatol., 18, 473-504,
1998.
Sinclair, M. R., J. A. Renwick, and J. W. Kidson, Low-frequency variability of Southern
Hemisphere sea level pressure and weather system activity, Mon. Weather Rev.,
125, 2531-2543, 1997.
30
Smith, S.R and C. R. Stearns, Antarctic Pressure and Temperature Anomalies
Surrounding the Minimum in the Southern Oscillation Index, J. Geophys. Res., 98
13071-13083, 1993.
Smith, R. C., S. E. Stammerjohn, and K. S. Baker, Surface air temperature variations in
the western Antarctic Peninsula region. Foundation for Ecological Research West
of the Antarctic Peninsula, Antarctic Research Series, Vol. 70, 105-121, 1996.
Thompson, D. W. J. and J. M. Wallace, Annular modes in the extratropical circulation,
part I, Month-to-month variability. J. Climate, 13, 1000-1016, 2000.
Trenberth, K. E. and D. J. Shea, On the evolution of the Southern Oscillation, Mon.
Weather Rev., 115, 3078-3096, 1987.
van Loon, H. Wind in the Southern Hemisphere. Meteorology of the Southern
Hemisphere, Meteor. Monogr. No. 35, Amer. Meteor. Soc., 59-86, 1972.
van Loon, H., The Southern Oscillation, part 3, Associations with the trades and with the
trough in the westerlies of the South Pacific Ocean. Mon. Weather Rev., 112, 947954, 1984.
van Loon, H., and J. C. Shea, The Southern Oscillation, part 4, The precursors south of
15˚S to the extremes of the oscillation. Mon. Weather Rev., 113, 2063-2074, 1985.
Walland, D. and I. Simmonds, Baroclinicity, meridional temperature gradients, and the
southern semiannual oscillation. J. Clim., 12, 3376-3382, 1999.
Watkin A. B. and I. Simmonds, A late sprint surge in the open water of the Antarctic.
Geophys. Res. Lett., Vol. 26, No. 10, 1481-1484, 1999.
31
White, B. W. and R. G. Peterson, 1996: An Antarctic circumpolar wave in surface
pressure, wind, temperature and sea ice extent. Nature, 380, 699-702.
Yuan, X., M. A. Cane and D. G. Martinson, Climate Variation -- Cycling Around the South Pole,
Nature, 380, 673-674, 1996.
Yuan, X., D. G. Martinson and W. T. Liu, Effect of air-sea-ice interaction on Southern Ocean
subpolar storm distribution. Journal of Geophysical Research, 104, D2, 1991-2007, 1999.
Yuan, X. and D. G. Martinson, Antarctic sea ice extent variability and its Global
Connectivity, Journal of Climate, 3, 1697-1717, 2000.
Yuan, X. and D. G. Martinson, The Antarctic Dipole and its Predictability. Geophys. Res.
Lett, Vol. 28, No. 18, 3609-3612, 2001.
32
Figure Captions
Figure 1. Differences between El Niño composite and La Niña composite in the SAT
anomaly (a), sea ice concentration and ice edge anomalies (b) in May after ENSO events
matured at the tropics. The white (black) line represents ice edge composite for the warm
(cold) ENSO events. The El Niño composites (c) and La Niña composites (d) of the SLP
and surface wind anomalies are plotted for the same month.
Figure 2. Seasonal composites of SLP (mb), SAT (°C), SST (°C) and SIC (%) anomalies
for five La Niña events (between 1984 and 2000) sequencing from the austral spring
before the events matured to the austral spring after the events matured.
Figure 3. Same as figure 2, except for El Niño scenario.
Figure 4. Power spectra of DPIice (a), DPISAT (b), DPISST (c), and Niño3 index (d).
Figure 5. El Niño composite (a) and La Niña composite (b) of zonal winds (m/s) at
300mb for the season of September to November.
Figure 6. SST anomaly composites (°C) for the El Niño scenario (a) and La Niña
scenario (b). The composites were the results of averaging SST from May before ENSO
events matured to the following April, and over five El Niño events and five La Niña
33
events, respectively. Schematic jet stream (STJ and PFJ), persistent anomalous high and
low pressure centers, and heat fluxes due to mean meridional circulations are marked in
corresponding SST composites.
Figure 7. Schematic atmospheric circulation pattern in response to a warm ENSO event
superimposed on the El Niño SST composite.
Figure 8. Schematic atmospheric circulation pattern in response to a cold ENSO event
superimposed on the La Niña SST composite.
Figure 9. El Niño and La Niña composites of heat flux anomalies at low-level atmosphere
by mean meridional circulation (a for El Niño, c for La Niña), and stationary eddies (b for
El Niño, d for La Niña) in Km/s. The heat fluxes were averaged from 1000mb to 850mb.
Negative values indicate polarward heat fluxes in the Southern Hemisphere.
34
35
36
37
38
39
40
41
42
43