1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
NORTH PACIFIC RESEARCH BOARD FINAL REPORT
Biophysical Moorings
NPRB Project B52 Final Report
Phyllis J. Stabeno1, Jeffrey M. Napp2, Terry Whitledge3
1
NOAA Research, PMEL, 7600 Sand Point Way NE, Seattle, WA, 98115, 206-526-6453,
23
Phyllis.stabeno@noaa.gov
24
2
25
3
26
27
28
29
30
31
NOAA Fisheries, AFSC, 7600 Sand Point Way NE, Seattle, WA, 98115
UAF/SFOS, Institute of Marine Science, PO Box 757220
October 2013
i
32
Abstract
33
This project resulted in collection of three years of nearly continuous data at four mooring sites,
34
and hydrographic and plankton sampling around the sites. Major findings from these data
35
include:
36

37
38
Identified warm and cold years in southeastern Bering Sea - warm (cold) years have little
(extensive) ice in March – April.

Identified transition from strong interannual variability in warm/cold years before 2000
39
and multi-year variability (stanzas) after 2000 - stanzas of warm years resulted in
40
decreased concentrations of large zooplankton -populations rebounded during cold
41
period.
42

Identified sharp division between northern and southern shelves at ~60 °N.
43

During cold years, the monthly-mean currents over the southeastern shelf are largely
44
45
westward, while in warm years the flow is northward during winter.

46
47
Flow on southern middle shelf is complex and not completely dominated by winds winter bottom currents at M2 are complex, summer bottom currents at M4 are southward.

48
Temperature dominates stratification in the south, and temperature and salinity are
equally important in the north.
49

The strength of stratification is not related to warm or cold years.
50

Approximately 1 m of ice melts over the southern (northern) shelf during ice advance
51
52
(retreat).

If ice is present after mid-March, spring phytoplankton bloom is related to ice retreat; if
53
ice is absent after mid-March, a spring bloom occurs in May-June, a fall bloom often
54
occurs in late September.
55
The extensive mooring data provided information for model validation.
56
57
Key words: Bering Sea, shelf domains, currents, fluorescence, zooplankton, nutrients,
58
stratification, warm-cold years, north-south structure, sea ice
59
60
Citation: Stabeno, P. J., N. B. Kachel, C. W. Mordy, J. M. Napp, and T. E. Whitledge. 2013.
61
Four Biophysical Mooring sites on the eastern Bering Sea shelf. NPRB Project B52 Final Report,
62
204 pages.
63
64
ii
65
Table of Contents
66
Abstract
ii
Keywords
ii
Citation
ii
Table of Contents
iii
Study Chronology
1
Introduction
3
Objectives
4
Manuscripts
12
Published Manuscripts
13
Chapter 1
14
Chapter 2
15
Chapter 3
16
Chapter 4
18
Chapter 5
19
Chapter 6
20
Chapter 7
21
Chapter 8
22
Chapter 9
23
Chapter 10
25
Chapter 11
26
Submitted Manuscripts
Chapter 12
27
28
Introduction
29
Data and Methods
31
Results
33
Conclusions and Discussion
43
Literature Cited
49
Chapter 13
52
Introduction
53
Methods
56
Results
64
Discussion
73
iii
Conclusions
80
Literature Cited
82
Chapter 14
98
Introduction
99
Data and Methods
101
Results
106
Discussion
110
Conclusions
118
Literature Cited
119
Chapter 15
142
Introduction
143
Methods
146
Results and Discussion
150
Conclusions
164
Literature Cited
167
Papers in Preparation
179
Paper 1
179
Paper 2
182
Conclusions
186
Management or Policy Implications
188
List of Publications
189
In Preparation
189
Submitted/in press
189
Published
190
Outreach
193
Websites
193
Presentations in schools
193
Conference Presentations
194
Acknowledgements
199
67
68
69
70
iv
71
72
List of Tables
Table 1.
Table 2.
Table 3.
Table 4.
Table 5.
Table 6.
73
74
75
Study Chronology
Design of Moorings
Deployments at M2
Deployments at M4
Deployments at M5
Deployments at M8
1
5
6
7
8
9
List of Figures
Figure 1. Location of the four biophysical moorings at the 70 m isobath.
Figure 2. The annual Bering Sea Report Card from the Ecosystem Considerations
Chapter, including 10 key indicators of the status of the Bering Sea.
Figure 3. Depth averaged temperature measured at M2 during the last 18 years and
the temperature anomaly, with the annual cycle (1995-2009) removed.
Figure 4. Eastern Bering Sea and M2 study site where the acoustic instruments
collected zooplankton volume backscatter data.
Figure 5. ADCP-derived zooplankton volume backscatter. Top – 2007, Middle –
2009, Bottom – 2010.
Figure 6. Mean currents during summer (May – September) derived from satellitetracked drifters drogued at 45 m. There are at least 10 independent
estimates of velocity in each grid cell.
Figure 7. Bottom currents at M2 divided into octents determined by wind
directions (e.g. for 0 – 45 °T the vector is the average velocity when
winds are toward that direction). Note currents are not 45 – 90 ° to the
right of the winds (lower left). Number of points in each bin shown in
lower right.
Figure 8. Bottom currents at M5 divided into octents determined by wind
directions (e.g. for 0 – 45 °T the vector is the average velocity when
winds are toward that direction). Note currents are not 45 – 90 ° to the
right of the winds (middle panel). Number of points in each bin shown
in bottom.
Figure 9. Bottom currents at M4 divided into octents determined by wind
directions (e.g. for 0 – 45 °T the vector is the average velocity when
winds are toward that direction). Note currents are not 45 – 90 ° to the
right of the winds (middle panel). Number of points in each bin shown
in bottom
Figure 10. Schematic of the pathway for management action for walleye pollock
3
11
11
180
181
183
183
184
185
188
76
77
v
78
Study Chronology
79
This report documents the activities, analytical methods, results and findings of Project B52 of the
80
Bering Sea Integrated Ecosystem Research Program. The project started in 2008 and ended in April
81
2013. Relationships identified in this analysis were used to inform or parameterize FEAST (B70,
82
Aydin et al.). In addition, there were strong collaborations with several NSF funded components of
83
the Bering Sea Project (NSF Project 1107250, Synthesis, Mordy et al.; NSF Project
84
0732640 Hydrographic Structure and Nutrients over the Eastern Bering Sea Shelf During
85
Summer, Mordy et al.; NSF Project 0732430 Impacts of Sea-ice on Hydrographic Structure and
86
Nutrients over the Eastern Bering Sea Shelf, Mordy et al.; and NSF project 0732534, Downscaling
87
global climate projections, Bond et al.). There were no no-cost extensions to the program. Progress
88
reports were submitted biannually and can be found on the North Pacific Research Board website.
89
90
91
Table 1. Study Chronology
2008-2010 Tasks, Assignments, Timeline
What
Who/Ship
Prep equipment, design and Stabeno, Napp,
build moorings, develop
Whitledge, Mordy
cruise plans.
(SNWM)
Recover/deploy moorings
NOAA ship Miller
M2 and M4 (ice permitting) Freeman
SNWM
Process and disseminate
SNWM
data
Data Delivery to data
manager
Send plankton samples to
Napp
Poland to be counted
Deploy moorings at M5 and
USCG Healy
M8
SNM
Recover/Deploy moorings
Miller Freeman
M2, M4, M5 and M8
and/or Melville
SNWM
Process and disseminate
data
Data Delivery to data
manager
Sent plankton samples to
Poland to be counted
Prepared equipment and
built moorings
Failed to deploy moorings at
M2
SNWM
Napp
Oscar Dyson,
Stabeno
Dates
2008
Other information
May 1 – May 10,
2008
Conducted
CTD/bongos along
southern hydro. line
May 15-Nov 15
Nov. 2008
phys/chem.
Sep 2009 zoop
Jun 1, 2008 – Jun
1, 2009
July, 2008
Sep, 2008
Oct – Dec, 2008
March 2009
phys/chem
Jan 2010 zoop
Oct 1, 2008 – Oct
1, 2009
Jan – Feb. 2009
March 2009
The actual time for
processing varies.
CTD/bongos along
the 5 primary hydro.
lines (70m isobath, &
4 cross shelf transects)
The actual time for
processing varies.
Failed because of ice
and weather.
1
Submitted data to data
manager
Recover/deploy moorings
M2 and M4
Processed and disseminated
data
Sent plankton samples to
Poland to be counted
Deploy moorings at M5 and
M8
Prepared equipment and
built moorings
Recover/Deploy moorings
M2, M4, M5 and M8
Processed and disseminated
data
Sent plankton samples to
Poland
Recover/deploy moorings
M2 (over half the mooring
was missing)
Deployed moorings at M5,
did not recover moorings at
site.
Oscar Dyson
SNWM
Apr 24 – May 7,
2009
SNWM
May 15 – Sept 30,
2009
July 30, 2009
Napp
Conducted some
hydrographic stations
Actual time for
processing varies
Jun/Jul 2009
R/V Knorr
SW
Stabeno
Jun - Aug. 2009
Napp
Sep 23 – Oct 11,
2009
Oct 2009– Apr,
2010
Nov 2009
Oscar Dyson
SNWM
Apr 23 – May3 ,
2010
SNWM
June 2010
Processed and disseminated
data and data delivered to
data manager
Conducted hydro and
zooplankton survey
SNWM
Apr – Sep 2010
Wecoma
SNWM
Aug 23 – Sep 10,
2010
Recover/Deploy moorings
M2, M4, M5 and M8
Freeman
SNWM
Sep 23 – Oct 3,
2010
Processed and disseminated
data and data delivered to
data manager
Sent plankton samples to
Poland
Synthesis of data and ideas,
complete manuscripts
6 manuscripts were
submitted and accepted for
SNWM
Oct 2010 – Jul
2011
Napp
November 2010
SNWM
2010 - 2012
SNWM
Oct – Dec 2011
Miller Freeman
SNWM
SNWM
CTD/bongos along the
5 primary hydro lines
Actual time for
processing varies
Due to extensive ice
M4, M5 and M8 could
not be reached.
M4 could not be
located and M5 was
not recovered because
of limited space on the
boat.
Conducted last
sampling of 70-m
isobath and 3 cross
lines
M4 could not be
found. It was likely
dragged by ice or a
fishing vessel. Will
continue searching for
it on future cruises.
Actual time for
processing varies
This is an ongoing
task.
2
92
93
publication in Special issue
of DSR II
6 manuscripts published
2 papers published in 2nd
special issue of DSR II
8 manuscripts submitted for
3rd special issue
>3 manuscripts being
prepared for submission to
4th special issue of DSR II
July 2012
In the out years, the prepping, mooring design and construction, and cruise plan development
94
occur during the 3 months before the beginning of the cruise. Similarly the data processing,
95
qcing etc. take approximately 3 months, with the exception of the counting of zooplankton which
96
can take as long as 1 year. Note, the delivery date of data is contingent upon the successful
97
recovery of the moorings.
98
99
Introduction
100
Project B52 is a continuation of a long-term partnership
101
between NOAA and NPRB. List of other NPRB grants
102
which supported the moorings are 203, 315, 410, 517,
103
602, 701, 1202. Moorings have been maintained on
104
the southeastern Bering Sea shelf at four sites: M2
105
(56.9°N, 164.1°W) since 1995, M4 (57.9°N, 168.9°W)
106
since 1996, M5 (59.9°N, 171.7°W) and M8 (62.2°N
107
174.7°W) since 2004 (Fig. 1). This project, together
108
with research by the NOAA program North Pacific
109
Climate
110
(NPCREP), continued these measurements through
111
2013. The moorings, together with observations along
112
114
the 70-m isobath, are core to the long-term observations Fig. 1.
Location of the 4
biophysical
moorings
and the 70-m
on the Bering Sea shelf. B52 continued the time series of
isobath.
temperature, salinity, fluorescence, currents, zooplankton
115
abundance (TAPS-8 at M2), nitrate, and oxygen (at M2) at four mooring sites on the Bering Sea
116
shelf. To avoid damage by ice, most of these moorings were subsurface. During the late spring and
117
summer a surface mooring was deployed at M2, which measured meteorological variables (air
118
temperature, humidity, barometric pressure, wind velocity and solar radiation – PAR).
113
Research
and
Ecosystem
Productivity
3
119
This project supports all five BSIERP hypotheses, by providing critical data to identify long-term
120
physical, chemical, and biological changes in the ecosystem:
121
1. Physical forcing affects food availability: Climate-induced changes in physical forcing
122
will modify the availability and partitioning of food for all trophic levels through bottom-
123
up processes;
124
2. Ocean conditions structure trophic relationships: Climate and ocean conditions
125
influencing water temperature, circulation patterns, and domain boundaries impact fish
126
reproduction, survival and distribution, the intensity of predator-prey relationships, and
127
the location of zoogeographic provinces through bottom-up processes;
128
3.
Ecosystem controls are dynamic: Later spring phytoplankton blooms resulting from
129
early ice retreat will increase zooplankton production, thereby leading to increased
130
abundances of piscivorous fish (walleye pollock, Pacific cod, and arrowtooth flounder)
131
and a community controlled by top-down processes with several trophic consequences;
132
4. Location matters: Climate and ocean conditions influencing circulation patterns and
133
domain boundaries will affect the distribution, frequency, and persistence of fronts and
134
other prey-concentrating features and, thus, the foraging success of marine birds and
135
mammals largely through bottom-up processes;
136
5. Commercial and subsistence fisheries reflect climate: Climate-ocean conditions will
137
change and, thus, affect the abundance and distribution of commercial and subsistence
138
fisheries.
139
140
Objectives
141
1) Provide near continuous, water column temperature, salinity, fluorescence, nitrate, and
142
currents at the four primary mooring sites, plus zooplankton and oxygen at selected sites.
143
Moorings were deployed at all four sites at least once a year. Below are the mooring designs
144
(Table 2) and the list of moorings deployed for each site (Tables 3 – 6), with the date of
145
deployment and recovery, the number of instruments on the moorings and any comments.
146
Hydrography and net tows were done at the mooring site and at the four corners of the 20 km x
147
20 km box around each mooring site, whenever possible. In spring, with extensive ice some of
148
these stations were skipped because of ice cover.
149
150
4
151
Table 2. The design of the moorings. The extensive sea ice precluded the recovery /deployment
152
of M5 and M8 in May of 2009. To resolve this, a simple mooring with temperature at 15, 10 and
153
5 m was deployed in July at M8 to obtain measurements in the upper column. A summer
154
mooring was also deployed at M5 in July when possible. See Tables 3-6 for details of moorings.
155
Sensors
M4
M5
Summer
& Winter
15, 20, 25,
30, 35, 40,
45, 50, 55,
60, 65, 69
M8
Summer
& Winter
20, 25, 30,
35, 40, 45,
50, 55, 60,
65, 69 m
Summer
Winter
Summer
Winter
Temperature
1, 5, 8, 10,
15, 20, 25,
30, 35, 40,
45, 50, 55,
60, 65, 69
5, 11, 15,
20, 25, 30,
35, 40, 45,
50, 55, 60,
65, 69
11 15 20,
25, 30,
35, 40,
45, 50,
55, 60,
65, 69
Salinity
11, 35, 60
11, 35, 60
11, 35,
60
15, 35, 60
20, 35, 60
Nitrate
Fluorescence
12, 59
11, 22, 44
11, 15,
20, 25,
30, 35,
40, 45,
50, 55,
60, 65,
69
11, 35,
60
12, 59
11, 22,
44
11
11
15
20
CDOM
Zooplankton
biovolume,size
composition
Currents ADCP
Meteorological
variables
156
157
M2
Bottom depth
11 m
35
35
every 4 m
every 4m
every 4m
every 4
m
every 4 m
every 4 m
73
70
70
71
70
surface
73
158
5
159
Table 3. Moorings were deployed twice a year at M2. The first set of moorings in 2008, replaced
160
moorings, which had been deployed the previous September. Moorings indicated by “bsm” were
161
all surface moorings. Those indicated by “bst” contained the TAPS-8 and selection of other
162
instruments. Selected data on bst and bsm were reported back in real time and put on the web.
163
Those indicated by “bsp” are bottom-mounted, upward looking ADCP moorings. Moorings
164
indicated by just “bs” were overwintering moorings with top instrument at 11 m.
Mooring
Deployed
Recovered
M2
Number of
Comments
Instr.
08bsm2a
5/6/2008
9/28/2008
26
07bs2c deployed on Sept. 2007
and recovered May 7, 2008 with
15 instruments
08bsp2a
5/6/2008
Not
recovered
1
07bsp2c deployed on Sept. 2007
and recovered May 7, 2008 with
15 instruments
08bs2c
9/29/2008
4/30/2009
15
08bsp2c
9/29/2008
4/30/2009
11
09bsm2a
4/30/2009
9/25/2009
22
09bst2a
4/30/2009
9/24/2009
7
09bsp2a
4/30/2009
9/25/02009
1
09bs2c
9/26/2009
4/30/2010
4
09bsp2b
9/26/2009
4/27/2010
1
10bsm2a
5/12/2010
10/4/2010
21
10bst2a
4/29/2010
10/3/2010
4
10bsp2a
4/30/2010
10/4/2010
1
10bs2c
10/5/2010
5/19/2011
18
10bsp2b
10/5/2010
5/19/2011
1
The top of this mooring was
trawled and we only recovered
the bottom 4 instruments.
Extensive ice required the 2
week delay of deployment of
surface mooring
165
166
6
167
Table 4. Moorings were generally deployed twice a year at M4. The first set of moorings in
168
2008, replaced moorings ,which had been deployed the previous September. Moorings indicated
169
by “bsm” were all surface moorings. Those indicated by “bsp” are upward looking ADCP
170
moorings. Moorings indicated by just “bs” were overwintering moorings with top instrument at
171
11 m. Mooring 09bsp4b was not recovered and all instruments and data were lost.
172
Mooring
Deployed
Recovered
M4
Number of
Comments
Instr.
5/9/2008
5/2/2009
14
07bs4b deployed on Sept.
2007 and recovered May 7,
2008 with 15 instruments
5/9/2008
9/26/2008
1
08bsp4a
08bs4b
07bsp4b deployed on Sept.
2007 and recovered May 7,
2008 with 1 instruments
9/26/2008
5/2/2009
13
08bsp4b
9/26/2008
5/2/2009
1
09bsm4a
5/2/2009
8/18/2009
19
09bsp4a
5/2/2009
9/27/2009
2
09bs4b
9/27/2009
Not recovered
21
09bsp4b
9/2872009
9/24/2010
1
08bs4a
This surface mooring was
dragged by ice and had to be
recovered early.
10bs4a,
Due to very extensive and persistent ice, these moorings were not deployed. The
10bsp4a
previous fall deployments were expected to supply necessary data.
10bs4b
9/24/2010
5/20/2011
17
10bsp4b
9/24/2010
5/20/2011
2
173
174
7
175
Table 5. Moorings were generally deployed once a year at M5. The first set of moorings in 2008
176
replaced moorings, which had been deployed the previous September. Those indicated by “bsp”
177
are upward looking ADCP moorings. Moorings indicated by just “bs” were overwintering
178
moorings with top instrument at 15 m.
179
Mooring
M5
Deployed
Recovered
9/25/2008
5/29/2009
Number of
Instr.
12
9/25/2008
5/29/2009
1
08bsp5b
09bs5a
7/1/2009
9/29/2009
15
09bsp5a
09bs5b
09bsp5b
10bs5a
10bsp5a
10bs5b
10bsp5b
5/29/2009
9/29/2009
9/29/2009
7/4/2010
7/4/2010
9/27/2010
9/27/2010
9/29/2009
9/26/2010
9/26/2010
9/26/2010
9/26/2010
5/20/2011
5/20/2011
1
13
1
18
1
15
1
08bs5b
180
181
Comments
07bs5b was deployed in Sept.
2007 and recover on 9/25/08
with 11 instruments
07bsp5b was deployed in Sept.
2007 and recover on 9/25/08
with 1 instrument
Weather and time constraints
prevented the deployment of
full water column mooring in
May.
Ice prevented recovery in May.
Ice prevented deployment in
May
8
182
Table 6. Moorings were generally deployed once a year at M8, with a second mooring with
183
thermistors in the upper 20 m deployed in July. The first set of moorings in 2008 replaced
184
moorings, which had been deployed the previous September. Those indicated by “bsp” are
185
upward looking ADCP moorings. Moorings indicated by “bsv” only contained temperature
186
measurements in the upper 20 m. Moorings indicated by just “bs” were overwintering moorings
187
with top instrument at 20 m.
188
Mooring
M8
Deployed
Recovered
7/26/2008
8/31/2008
Number of
Instr.
2
8/31/2008
10/1/2009
11
8/31/2008
10/1/2009
1
7/7/2009
10/1/2009
10/1/2009
9/29/2010
9/29/2010
10/1/2009
9/29/2010
9/29/2010
8/16/2011
8/16/2011
11
2
13
1
08bsv8a
08bs8a
189
08bsp8a
09bsv8a
09bs8a
09bsp8a
10bs8a
10bsp8a
Comments
This simple mooring was deployed
to obtain near-surface
temperatures.
07bs8a was deployed in Sept. 2007
and recover on 8/31/08 with 11
instruments
07bsp8a was deployed in Sept.
2007 and recover on 8/31/08 with
1 instruments
190
191
2) Provide derived products: mixed layer depth, timing of the spring phytoplankton bloom and
192
other blooms, heat content, mixed layer depth, and advection at the four sites.
193
Products from the moorings include mixed layer depth, heat content at M2 and M4, temperature,
194
position of the transition between southern pelagic-dominated shelf and northern benthic-
195
dominated shelf, advection, tidal energy, nitrate concentration, the timing of the spring
196
phytoplankton bloom, estimates of zooplankton from TAPS-8 and backscatter from the ADCPs. In
197
addition, a number of annual cycles were derived using the mooring data, including: total heat
198
content (depth averaged temperature) at M2 and M4, vertical temperature structure at M2, bottom
199
and near-surface temperature at each mooring site, mean currents (near surface and bottom) at
200
each site. A number of derived indices of ice cover were develop to better understand the
201
response of the water column and lower trophic level to sea ice.
202
9
203
3) Use mooring and shipboard measurements as indicators of ecosystem status and to apply the
204
observations to metrics/indices developed during Southeast Bering Sea Carrying Capacity
205
(SEBSCC). Possible indices include wind mixing, ocean stratification, mixing events (strength,
206
date, number of), mixed layer temperature, photosynthetically active radiation, sea ice
207
characteristics, extent of the cold pool, timing of the spring phytoplankton bloom, zooplankton
208
biomass and estimated larval advection.
209
EcoFOCI contributes to the Ecosystem Status and Management Indicators section of the
210
Ecosystem Considerations Chapter. Our mooring and shipboard measurements contribute to the
211
following indicators:
212
213
214
215
216
217
1. North Pacific Climate Overview, Bering Sea Regional Highlights contributed by N.
Bond.
2. Eastern Bering Sea Climate – FOCI contributed by J. Overland, P. Stabeno, C. Ladd, S.
Salo, M. Wang and N. Bond.
3. Bering Sea Zooplankton, contributed by J. Napp, P. Ressler, P. Stabeno, and A.
Yamaguchi
218
219
220
EcoFOCI also contributes indicators to the annual Eastern Bering Sea Report Card (Fig. 2).
221
averaged temperature at M2 and the derived anomaly (e.g. Fig. 3) are used to define warm and
222
cold years for the eastern Bering Sea shelf. The chlorophyll-a fluorescence time series at all the
223
moorings provides an index of the timing of the spring and fall phytoplankton blooms, which can
224
be related to both ice-retreat and storminess. The vertical temperature profiles at the moorings
225
provide indicators of mixed layer depth. Sudden increases in mixed-layer depth coupled with
226
winds and fluorescence time series provides an index of how many events during the summer
227
were strong enough to mix nutrients into the euphotic zone.
228
Mooring M8 is located in Region 1 of the Distributed Biological Observatory (DBO). Data from
229
this site provide a range of statistics to characterize that region and how it varies from year to
230
year. Information includes: minimum and maximum bottom temperature, timing of that site
231
becoming well mixed, strength of vertical density gradient, and direction of currents.
Long-term mooring data can be used to create a number of indices. For instance the depth
232
10
Fig. 2. The annual Bering
Sea Report Card from the
Ecosystem Considerations
Chapter, including 10 key
indicators of the status of the
Bering Sea. Ocean
temperature is not one of the
items because of its high
correlations with sea ice
extent in the spring. The
copepod index comes
partially from data collected
as part of this program.
233
234
235
236
237
238
239
240
241
242
4)
243
Use
244
data
Fig. 3. Top panel: Depth averaged temperature measured at M2 during the last 18 years.
Bottom Panel: The temperature anomaly, with the annual cycle (1995 – 2009) removed.
245
from the moorings and shipboard observations to validate and tune the physical and nutrient-
11
246
phytoplankton-zooplankton (NPZ) models that are being developed.
247
The data from the biophysical mooring data were frequently used to calibrate biophysical
248
models of the Bering Sea. These included side-by-side model/data comparisons of the vertical
249
profiles of currents, temperature, salinity, and chlorophyll-a at multiple locations, from hourly
250
to interdecadal time scales. The data were especially useful in demonstrating that initial
251
parameter choices for the model produced unrealistically warm conditions in the Bering Sea. A
252
time series of surface shortwave radiation at M2 demonstrated that shortwave forcing was in
253
fact *not* the source of that bias. Subsequent adjustments to other factors – albedo and bulk
254
flux parameters in particular - yielded a much better fit to the extensive mooring dataset. The
255
data were especially powerful in providing multivariate time series at many locations. These
256
allow the examination of the covariance structure among different locations, and among
257
different variables, which are compared with their model equivalent. Note that this is a far
258
more rigorous method of model-data comparison than is possible with any single time series.
259
260
Manuscripts
261
There are ~30 peer reviewed publications (published, submitted and in preparation) as results of
262
this grant. All published manuscripts are listed with the link to the appropriate web pages from
263
which they can be down loaded and the abstract. The complete texts of manuscripts that are
264
submitted, but not published are presented in Chapters 12 –15. Finally, two manuscripts that are
265
presently in preparation for the fourth special issue are presented with detail currently available.
266
267
12
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
Published Manuscripts
284
(alphabetized by first author)
285
13
286
Chapter 1
287
288
Danielson, S., E. Curchitser, K. Hedstrom, T. Weingartner, and P. Stabeno (2011): On ocean and
289
sea ice modes of variability in the Bering Sea. J. Geophys. Res., 116(C12), C12034, doi:
290
10.1029/2011JC007389.
291
292
http://onlinelibrary.wiley.com/doi/10.1029/2011JC007389/full
293
294
ABSTRACT: Results from a 35 year hindcast of northeast Pacific Ocean conditions are
295
confronted with observational data collected over the Bering Sea shelf within the integration time
296
period. Rotary power spectra of the hindcast currents near NOAA mooring site M2 site fall within
297
the 95% confidence bounds for the observational spectra, except for a high bias in the counter-
298
clockwise rotating component at 10 m depth in the high frequencies (periods <24 h). The model
299
exhibits the most skill in reproducing anomalies of the integrated annual sea ice concentration and
300
monthly subsurface (60 m depth) temperature fields, accounting for 85% and 50% of their
301
observed variability. Analysis of the integrated ice concentration time series reveals evolution in
302
the mean duration of ice-free waters (40 year trend of +6.8 days/decade) and changes in this
303
parameter's variance with time. Correlation and empirical orthogonal function (EOF) analyses
304
reveal the primary temporal-spatial patterns of variability in the temperature and salinity fields
305
over the Bering Sea and northern Gulf of Alaska for near-surface (0–20 m) and subsurface (40–
306
100 m) depth layers. Correlation analysis between the EOF principal components and various
307
climate index and observed time series shows that the Pacific Decadal Oscillation, the North
308
Pacific Gyre Oscillation, and the Bering Sea annually integrated ice area anomalies are important
309
indices of thermohaline variability; the spatial structures of these modes give insight to their
310
potential impacts upon the ecosystem. We identify a number of ecologically and economically
311
important species whose temporal variability is significantly correlated with the identified spatial
312
patterns.
313
314
14
315
Chapter 2
316
317
Hermann, A.J., G.A. Gibson, N.A. Bond, E.N. Curchitser, K. Hedstrom, W. Cheng, M. Wang,
318
P.J. Stabeno, L. Eisner, and K.D. Cieciel (2013): A multivariate analysis of observed and
319
modeled biophysical variability on the Bering Sea shelf: Multidecadal hindcasts (1970-2009) and
320
forecasts (2010-2040). Deep-Sea Res. II, doi: 10.1016/j.dsr2.2013.04.007.
321
322
http://www.sciencedirect.com/science/article/pii/S096706451300146X
323
324
ABSTRACT: Coupled physical/biological models can be used to downscale global climate
325
change to the ecology of subarctic regions, and to explore the bottom-up and top-down effects of
326
that change on the spatial structure of subarctic ecosystems—for example, the relative dominance
327
of large vs. small zooplankton in relation to ice cover. Here we utilize a multivariate statistical
328
approach to extract the emergent properties of a coupled physical/biological hindcast of the
329
Bering Sea for years 1970–2009, which includes multiple episodes of warming and cooling (e.g.
330
the recent cooling of 2005–2009), and a multidecadal regional forecast of the coupled models,
331
driven by an IPCC global model forecast of 2010–2040. Specifically, we employ multivariate
332
empirical orthogonal function (EOF) analysis to derive the spatial covariance among physical and
333
biological timeseries from our simulations. These are compared with EOFs derived from spatially
334
gridded measurements of the region, collected during multiyear field programs. The model
335
replicates observed relationships among temperature and salinity, as well as the observed inverse
336
correlation between temperature and large crustacean zooplankton on the southeastern Bering Sea
337
shelf. Predicted future warming of the shelf is accompanied by a northward shift in both pelagic
338
and benthic biomass.
339
340
15
341
Chapter 3
342
343
Hunt, Jr., G.L., B.M. Allen, R.P. Angliss, T. Baker, N. Bond, G. Buck, G.V. Byrd, K.O. Coyle,
344
A. Devol, D.M. Eggers, L. Eisner, R.A. Feely, S. Fitzgerald, L.W. Fritz, E.V. Gritsay, C. Ladd,
345
W. Lewis, J. Mathis, C.W. Mordy, F. Mueter, J. Napp, E. Sherr, D. Shull, P. Stabeno, M.A.
346
Stepanenko, S. Strom, and T.E. Whitledge (2010): Status and trends of the Bering Sea region,
347
2003-2008. In Marine Ecosystems of the North Pacific Ocean, 2003-2008, S. M. McKinnell and
348
M. J. Dagg (eds.), PICES Special Publication 4, 196-267.
349
350
http://www.pices.int/publications/special_publications/NPESR/2010/NPESR_2010.aspx
351
352
353
HIGHLIGHTS:
o
The Bering Sea experienced four years with low sea ice cover and extraordinarily warm
354
summers (2002-2005), followed by four years with some of the heaviest sea ice cover
355
since the early 1970s and cold summers (2006-2009). During the warm period, integrated
356
water column temperatures were elevated, bottom temperatures were higher, and the cold
357
pool over the southeastern shelf was small and not as cold as in the cold period. During the
358
cold period, integrated water column temperatures were anomalously low, bottom
359
temperatures were below the long term mean, and the cold pool consisted of cold, -1.7 .C
360
water that extended across most of the Middle Shelf Domain with cool waters extending to
361
Bristol Bay and the Alaska Peninsula.
362
o
363
364
Water column stratification varied spatially, and was sometimes stronger in the warm
years and sometimes stronger in the cold years, depending upon location.

Net primary production and surface chlorophyll a were positively affected by temperature.
365
The size distribution of crustacean zooplankton became smaller in years of warmer
366
temperature.
367

In the warm years of 2002-2005, small neritic species of crustacean zooplankton thrived
368
whereas the medium-large copepod, Calanus marshallae, and the shelf euphausiid,
369
Thysanoessa raschii, were scarce. In the cold years of 1999 and 2006- 2008, both C.
370
marshallae and T. raschii were abundant.
371

Catches of both eastern and western Bering Sea groundfish stocks declined in recent years.
372
In the eastern Bering Sea and Aleutian Islands, Pacific cod, yellowfin sole, flathead sole
373
and Greenland turbot, and particularly walleye pollock, have shown declines while Pacific
374
ocean perch, northern rockfish, rock sole, Alaska plaice, and especially arrowtooth
16
375
flounder (whose biomass has quadrupled since the late 1970s) have increasing biomass
376
trends.
377

Walleye pollock recruitment may be adversely affected by unusually warm conditions;
378
strong year-classes failed to emerge in the warm years of 2002-2005 despite a declining
379
biomass of pollock. Possibly, there were strong year-classes in the cold years of 2006 and
380
2008. In the cold years of 2006-2008, age-0 pollock were in good condition and energy-
381
rich, though not as abundant as they were in the warmer years, when they were less
382
energy-rich.
383

Age-0 pollock consumed mostly small crustacean zooplankton and smaller pollock in the
384
warmer years, and neither C. marshallae nor T. raschii were important dietary
385
components. In the cold years, their diets were dominated by C. marshallae and
386
euphausiids, and there was less cannibalism.
387

Comparison of fisheries statistics between the eastern and western Bering Sea showed
388
that, overall, the aggregated catches of all major fisheries tended to be positive, suggesting
389
that the two sides of the Bering Sea were responding similarly to shared climate forcing.
390

391
392
Both size-at-age and weight-at-length of groundfish tended to be above the long term
mean in the warmer years, and below the mean in the cold years.

Catches of crabs in the eastern and western Bering Sea declined after the early 1990s
393
(eastern Bering Sea) or the late 1990s (western Bering Sea), but catches began to increase
394
on both sides of the Bering Sea after 2004. The biomass of both red king crab and snow
395
crab have increased in the eastern Bering Sea, although snow crab catches remain low due
396
to conservative catch limits.
397

398
399
Catches of Pacific herring on both sides of the Bering Sea have been stable since 2001,
whereas the Togiak stock in the eastern Bering Sea has declined.

Catches of salmon in both the eastern and western Bering Sea were above average and
400
dominated by pink and chum salmon in the western Bering Sea and sockeye salmon in the
401
eastern Bering Sea. In both the eastern and western Bering Sea, the catches of chinook
402
salmon are in decline when compared to catches from the mid-1960s to the early 1990s.
403
404
17
405
406
407
Chapter 4
408
Sea Res. II, 65–70, doi: 10.1016/j.dsr2.2012.02.009, 72–83.
Ladd, C., and P.J. Stabeno (2012): Stratification on the Eastern Bering Sea Shelf revisited. Deep-
409
410
http://www.sciencedirect.com/science/article/pii/S0967064512000136
411
412
ABSTRACT: The timing and magnitude of stratification can have profound influences on the
413
marine ecosystem. On the Eastern Bering Sea shelf, in the absence of strong wind mixing,
414
stratification can be initiated by the melting of seasonal sea ice or by springtime warming of the
415
surface. Temperature and salinity both influence the stratification of the Eastern Bering Sea shelf
416
with their relative importance varying spatially and temporally. In the northern middle shelf
417
domain (north of ∼60°N), salinity stratification is often as important as temperature stratification.
418
On the southern middle shelf, while the influence of temperature on stratification dominates
419
during summer, the influence of salinity stratification plays a role in the interannual variability.
420
Mooring 2 (M2; 56.9°N, 164.1°W) has been deployed at ∼70 m depth in the southern middle
421
shelf domain since 1995. Data from this mooring show that stratification typically begins to set
422
up in May and to break down in September/October, but these dates can vary by >30 d. While no
423
trend is found in the timing of the spring setup, the fall stratification breakdown exhibited a trend
424
toward later breakdown (∼2 d later per year from 1996 to 2009). Results suggest that it may be
425
difficult to forecast stratification on the Eastern Bering Sea shelf from climate models as simple
426
indices of wind mixing or heat fluxes are not correlated with stratification. Contrary to intuition,
427
the strength of summer stratification is not correlated with depth averaged temperature. Warm
428
years such as 2000 and 2001 can have low stratification and cold years such as 2007 can have
429
very high stratification. This decoupling of stratification and temperature has implications for
430
forecasting the ecosystem in the face of climate change, as we cannot assume that projections of a
431
warmer climate simply imply higher stratification in the future.
432
433
18
434
Chapter 5
435
436
Miksis-Olds, J.L., P.J. Stabeno, J.M. Napp, A.I. Pinchuk, J.A. Nystuen, J.D. Warren, and S.L.
437
Denes (2013): Ecosystem response to a temporary sea ice retreat in the Bering Sea: Winter 2009.
438
Prog. Oceanogr., 111, doi: 10.1016/j.pocean.2012.10.010, 38–51.
439
440
http://www.sciencedirect.com/science/article/pii/S0079661112001590
441
442
ABSTRACT: Adding acoustic systems onto ocean moorings and observatories provides
443
additional data to more fully document ecosystem responses to environmental perturbations. A
444
passive acoustic recorder and three-frequency echosounder system were integrated into a
445
biophysical mooring on the central eastern Bering Sea continental shelf. An unexpected,
446
transient, mid-winter retreat of the seasonal sea ice was observed over the mooring for a 2-week
447
period in March 2009. Interpretation of the passive acoustic data provided information about sea
448
ice conditions and included the detection and identification of vocalizing marine mammals, while
449
the acoustic backscatter provided information on relative zooplankton and fish abundance before,
450
during, and after the retreat. Hydrographic data confirmed the acoustic signal was associated with
451
changing surface ice conditions, and the combined information from the biophysical mooring
452
sensors revealed changes in winter trophic level dynamics during the retreat, which would have
453
otherwise been undetected by traditional ship-based observations. Changes in the acoustic
454
environment, zooplankton dynamics, and acoustic detection of marine mammals were observed
455
amidst a physically stable and uniform water column with no indication of a phytoplankton
456
bloom. These data demonstrate the value of acoustic technologies to monitor changing
457
ecosystems dynamics in remote and hazardous locations.
458
459
19
460
Chapter 6
461
462
Mordy, C.W., E.D. Cokelet, C. Ladd, F.A. Menzia, P. Proctor, P.J. Stabeno, and E. Wisegarver
463
(2012): Net community production on the middle shelf of the Eastern Bering Sea. Deep-Sea Res.
464
II, 65–70, doi: 10.1016/j.dsr2.2012.02.012, 110–125.
465
466
http://www.sciencedirect.com/science/article/pii/S0967064512000264
467
468
ABSTRACT: To estimate temporal changes of nutrients and calculate the seasonal net
469
community production (NCP) on the eastern shelf of the Bering Sea, hydrographic sampling
470
along the 70-m isobath of the middle shelf was conducted in spring (2007–2009), summer (2008–
471
2009), and fall (2007). These were cold years, with sea ice covering much of the eastern Bering
472
Sea in April. Each spring, there was a region with relatively low nitrate in the middle portion
473
(59°–60°N) of the transect prior to the spring phytoplankton bloom. This water appeared to have
474
originated in the coastal domain and was advected offshore into the middle domain. Seasonal
475
NCP (mean±standard deviation) in this region was low (26±12 g C m−2), and may be indicative of
476
a portion of the middle shelf ecosystem that is chronically short of fixed carbon in spring. In
477
2007, the post-bloom cruise occurred during the fall transition when deep mixing,
478
remineralization, and denitrification/anammox compromised seasonal estimates of NCP. In other
479
years (2008–2009), the post-bloom cruise occurred in summer. On those cruises, the euphotic
480
zone, elevated chlorophyll fluorescence, and oxygen supersaturation were occasionally deeper
481
than the pycnocline, and there was a seasonal loss of nitrate and phosphate in the bottom layer. In
482
2008, preferential uptake of ammonium may have sustained sub-surface production in the north.
483
Therefore, seasonal estimates of NCP were not only evaluated in the upper mixed layer, but
484
throughout the water column. During summer, denitrification/anammox in bottom waters did not
485
appear to compromise seasonal estimates of NCP. Seasonal NCP averaged for 2008 and 2009
486
was slightly but significantly higher (p<0.0041) in the south (47±9 g C m−2, n=80) than in the
487
north (41±16 g C m−2, n=78). In the south, interannual variability of seasonal NCP was related to
488
the wind mixing in spring rather than the presence or absence of ice.
489
490
20
491
Chapter 7
492
493
Nystuen, J.A., S.E. Moore, and P.J. Stabeno (2010): A sound budget for the southeastern Bering
494
Sea: Measuring wind, rainfall, shipping, and other sources of underwater sound. J. Acoust. Soc.
495
Am., 128(1), doi: 10.1121/1.3436547, 58−65.
496
497
http://scitation.aip.org/content/asa/journal/jasa/128/1/10.1121/1.3436547
498
499
ABSTRACT: Ambient sound in the ocean contains quantifiable information about the marine
500
environment. A passive aquatic listener (PAL) was deployed at a long-term mooring site in the
501
southeastern Bering Sea from 27 April through 28 September 2004. This was a chain mooring
502
with lots of clanking. However, the sampling strategy of the PAL filtered through this noise and
503
allowed the background sound field to be quantified for natural signals. Distinctive signals
504
include the sound from wind, drizzle and rain. These sources dominate the sound budget and their
505
intensity can be used to quantify wind speed and rainfall rate. The wind speed measurement has
506
an accuracy of ±0.4 m s−1 when compared to a buoy-mounted anemometer. The rainfall rate
507
measurement is consistent with a land-based measurement in the Aleutian chain at Cold Bay, AK
508
(170 km south of the mooring location). Other identifiable sounds include ships and short
509
transient tones. The PAL was designed to reject transients in the range important for
510
quantification of wind speed and rainfall, but serendipitously recorded peaks in the sound
511
spectrum between 200 Hz and 3 kHz. Some of these tones are consistent with whale calls, but
512
most are apparently associated with mooring self-noise.
513
514
21
515
Chapter 8
516
517
Sigler, M.F., H.R. Harvey, C.J. Ashjian, M.W. Lomas, J.M. Napp, P.J. Stabeno, and T.I. Van Pelt
518
(2010): How does climate change affect the Bering Sea ecosystem? Eos Trans. AGU, 91(48), doi:
519
10.1029/2010EO480001, 457–458.
520
521
http://onlinelibrary.wiley.com/doi/10.1029/2010EO480001/abstract
522
ABSTRACT: The Bering Sea is one of the most productive marine ecosystems in the world,
523
sustaining nearly half of U.S. annual commercial fish catches and providing food and cultural
524
value to thousands of coastal and island residents. Fish and crab are abundant in the Bering Sea;
525
whales, seals, and seabirds migrate there every year. In winter, the topography, latitude,
526
atmosphere, and ocean circulation combine to produce a sea ice advance in the Bering Sea
527
unmatched elsewhere in the Northern Hemisphere, and in spring the retreating ice; longer
528
daylight hours; and nutrient-rich, deep-ocean waters forced up onto the broad continental shelf
529
result in intense marine productivity (Figure 1). This seasonal ice cover is a major driver of
530
Bering Sea ecology, making this ecosystem particularly sensitive to changes in climate. Predicted
531
changes in ice cover in the coming decades have intensified concern about the future of this
532
economically and culturally important region. In response, the North Pacific Research Board
533
(NPRB) and the U.S. National Science Foundation (NSF) entered into a partnership in 2007 to
534
support the Bering Sea Project, a comprehensive $52 million investigation to understand how
535
climate change is affecting the Bering Sea ecosystem, ranging from lower trophic levels (e.g.,
536
plankton) to fish, seabirds, marine mammals, and, ultimately, humans. The project integrates two
537
research programs, the NSF Bering Ecosystem Study (BEST) and the NPRB Bering Sea
538
Integrated Ecosystem Research Program (BSIERP), with substantial in-kind contributions from
539
the U.S. National Oceanic and Atmospheric Administration (NOAA) and the U.S. Fish and
540
Wildlife Service.
541
542
22
543
544
545
Chapter 9
546
and M.F. Sigler (2012): A comparison of the physics of the northern and southern shelves of the
547
eastern Bering Sea and some implications for the ecosystem. Deep-Sea Res. II, 65–70, doi:
548
10.1016/j.dsr2.2012.02.019, 14–30.
Stabeno, P.J., E. Farley, N. Kachel, S. Moore, C. Mordy, J.M. Napp, J.E. Overland, A.I. Pinchuk,
549
550
http://www.sciencedirect.com/science/article/pii/S0967064512000331
551
552
ABSTRACT: Sufficient oceanographic measurements have been made in recent years to
553
describe the latitudinal variation in the physics of the eastern Bering Sea shelf and the potential
554
impact of climate change on the species assemblages in the two ecosystems (north and south).
555
Many of the predicted ecosystem changes will result from alterations in the timing and extent of
556
sea ice. It is predicted that the sea ice in the northern Bering Sea will be less common in May, but
557
will continue to be extensive through April. In contrast, the southern shelf will have, on average,
558
much less sea ice than currently observed, but with large interannual and multiyear variability
559
until at least 2050. Thus, even under current climate warming scenarios, bottom temperatures on
560
the northern shelf will remain cold. Based on biophysical measurements, the southern and
561
northern ecosystems were divided by a North–South Transition at ∼60°N. The northern middle
562
shelf was characterized by a freshwater lens at the surface, cold bottom temperatures, and a
563
thicker pycnocline than found on the southern shelf. Subsurface phytoplankton blooms were
564
common. In contrast, the southern shelf stratification was largely determined by temperature
565
alone; the pycnocline was thin (often<3 m) and subsurface blooms were uncommon. Biological
566
responses to climate warming could include greater north–south differences in zooplankton
567
community structure, the transport of large Outer Shelf Domain crustacean zooplankton to the
568
middle shelf, and the disappearance of two principal prey taxa (Calanus spp. and Thysanoessa
569
spp.) of planktivorous fish, seabirds and whales. The response of commercially and ecologically
570
important fish species is predicted to vary. Some species of fish (e.g., juvenile sockeye salmon,
571
Oncorhynchus nerka) may expand their summer range into the northern Bering Sea; some (e.g.,
572
pink salmon, O. gorbuscha) may increase in abundance while still other species (e.g., walleye
573
pollock and arrowtooth flounder; Theragra chalcogramma and Atheresthes stomias, respectively)
574
are unlikely to become common in the north. The projected warming of the southern shelf will
575
limit the distribution of arctic species (e.g., snow crab, Chionoecetes opilio) to the northern shelf
576
and will likely permit expansion of subarctic species into the southern Bering Sea. The
23
577
distribution and abundance of baleen whales will respond to shifts in prey availability; for
578
instance, if prey are advected northward from the southeastern Bering Sea, an extension of range
579
and an increase in seasonally migratory baleen whale numbers is anticipated. Thus, alteration of
580
this ecosystem in response to climate change is expected to result in something other than a
581
simple northward shift in the distribution of all species.
582
583
24
584
Chapter 10
585
586
Stabeno, P.J., N.B. Kachel, S.E. Moore, J.M. Napp, M. Sigler, A. Yamaguchi, and A.N. Zerbini
587
(2012): Comparison of warm and cold years on the southeastern Bering Sea shelf and some
588
implications for the ecosystem. Deep-Sea Res. II, 65–70, doi: 10.1016/j.dsr2.2012.02.020, 31–45.
589
590
http://www.sciencedirect.com/science/article/pii/S0967064512000343
591
592
ABSTRACT: The southeastern, middle shelf of the Bering Sea has exhibited extreme variability
593
in sea ice extent, temperature, and the distribution and abundance of species at multiple trophic
594
levels over the past four decades. From 1972–2000, there was high interannual variability of areal
595
extent of sea ice during spring (March–April). In 2000, this shifted to a 5-year (2001–2005)
596
period of low ice extent during spring, which transitioned to a 4-year (2007–2010) period of
597
extensive sea ice. High (low) areal extent of sea ice in spring was associated with cold (warm)
598
water column temperatures for the following 6–7 months. The ocean currents also differed
599
between warm and cold years. During cold years, the monthly-mean currents over the shelf were
600
largely westward, while in warm years the direction of currents was more variable, with
601
northward flow during December–February and relatively weak flow during the remainder of the
602
year. The types and abundance of zooplankton differed sharply between warm and cold years.
603
This was especially true during the prolonged warm period (2001–2005) and cold period (2007–
604
2010), and was less evident during the years of high interannual variability. During the warm
605
period, there was a lack of large copepods and euphausiids over the shelf; however, their
606
populations rebounded during cold period. Small crustacean zooplankton taxa did not appear to
607
vary between and warm and cold years. For both walleye pollock and Pacific cod, year-class
608
strength (recruitment) was low during the prolonged warm period, but improved during the
609
following cold period. Year-class strength did not appear to vary as a function of warm and cold
610
years during the period of high year-to-year variability. Also, recruitment of arrowtooth flounder
611
(a predator of pollock and cod) did not appear influenced by the warm or cold years. Finally, the
612
distribution and relative abundance of fin whales appeared to differ in warm and cold years, with
613
fewer whales on the southeastern, middle shelf during warm years.
614
615
25
616
Chapter 11
617
Stafford, K.M., S.E. Moore, P.J. Stabeno, D.V. Holliday, J.M. Napp, and D.K. Mellinger (2010):
618
Biophysical ocean observation in the southeastern Bering Sea. Geophys. Res. Lett., 37, L02606,
619
doi: 10.1029/2009GL040724.
620
621
http://onlinelibrary.wiley.com/doi/10.1029/2009GL040724/full
622
623
ABSTRACT: Integrated ocean observation, from physical and atmospheric forcing mechanisms
624
to the distribution and abundance of top-level predators, is critical to the investigation of marine
625
ecosystems and the impact of climate change on them. We integrated data from a biophysical
626
mooring in the southeast Bering Sea to create a one-year snapshot of ocean dynamics in this
627
remote large marine ecosystem. Distinct patterns in production (chlorophyll), zooplankton
628
biovolume (copepods and euphausiids) and the occurrence of zooplankton predators (fin and
629
right whales) were defined and related to discrete features in the annual physical cycle. Peaks in
630
prey and predator cycles were linked to spikes in fluorescence that occurred at the onset of water
631
column stratification in late spring 2006 and the appearance of sea ice in late winter 2007. These
632
data illustrate the capability and potential of integrated ocean observing systems (IOOS) to
633
describe seasonal variability and linkages in a remote marine ecosystem.
634
635
636
637
26
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
Submitted manuscripts
27
655
Chapter 12
656
657
Cheng, W., E. Curchitser, C. Ladd, and P.J. Stabeno (2013): Ice–ocean interactions in the eastern
658
Bering Sea: NCAR CESM simulations and comparison with observations. Deep-Sea Res. II.
659
660
Abstract
661
We examine ice-ocean interactions in the Eastern Bering Sea (EBS) using output from the
662
National Center for Atmospheric Research-Community Earth System Model and satellite
663
observations, and draw comparisons with published results based on ocean in situ data. The
664
model simulation generally captures the observed seasonal advance and retreat of sea ice in the
665
EBS and spatial patterns. In the northern EBS, however, modeled sea ice annual maximum occurs
666
in April instead in March, as observed by satellite observations. Sea ice also retreats slower in
667
spring in the simulations than observed. Simulations indicate that early winter changes in ice
668
coverage in the north are due to a combination of freezing and advection processes, while in the
669
south changes are primarily due to advection from the north and melting. Via this seasonal transit,
670
sea ice cools the entire EBS middle shelf relatively uniformly, but its influence on ocean salinity
671
has greater spatial variability, as suggested by both observations and simulations. On inter-annual
672
time scales, high ice cover in the EBS leads to cold anomalies in the bottom water, especially on
673
the middle shelf, with the strongest anomalies in the southern domain. The corresponding salinity
674
anomalies are positive, with the strongest signal in the northern near-shore locations. The
675
associated current anomalies at 50m are generally southward on the shelf, and northwestward in
676
the slope region. Comparing years 1961-2005 versus years 2005-2050, the Probability
677
Distribution Function of ice time series on the EBS has shifted northward by ~ 2° latitude. The
678
implications of these results on marine ecosystem responses are discussed.
679
680
Keywords: Bering Sea, sea ice, numerical model, ocean temperature, salinity
681
28
682
1.
Introduction
683
The Bering Sea is a North Pacific
684
subarctic ocean bounded to the north by
685
Bering Strait, which connects the Bering Sea
686
with the Arctic Ocean, and to the south by the
687
Aleutian Island chain (Fig. 1). The Eastern
688
Bering Sea (EBS) shelf is wide (approximately
689
500 km) and flat (with depth less than 180m).
690
As early as November, sea ice begins forming
691
in the northern EBS, and under the prevailing
692
wind conditions, is transported southward 700 area. The black dots mark the ocean moorings in the
693
– 1000 km, often covering much of the EBS shelf. Maximum ice extent typically occurs in
694
March, but it can occur as early as February or as late as April (Stabeno et al., 2012a). During
695
melting season, sea ice retreats and by June the Bering Sea is usually ice free. Sea ice extent is
696
highly variable on interannual scales, but thus far, shows no significant trends since the beginning
697
of the satellite record in 1979 (Brown and Arrigo, 2012).
Figure 2. Geographical information of the study
Eastern Bering Sea middle shelf.
698
The seasonal changes of sea ice profoundly influence the physical oceanography of the
699
EBS and its ecosystem. During extensive ice years, with sea ice persisting on the southern shelf
700
after mid March, the water column over the southern shelf (south of 60° N) tends to be colder and
701
fresher (Stabeno et al., 2010, 2012b; Ladd and Stabeno, 2012; Sullivan et al., this issue). The
702
timing of ice retreat also influences the timing of the spring phytoplankton bloom (Sigler et al.,
703
this issue), zooplankton species composition, and fish recruitment. For instance, a period of
704
limited ice over the southern shelf (2001 – 2005) was characterized by fewer large crustacean
705
zooplankton the following summer (Stabeno et al., 2012a). As a result the young of the year
706
walleye pollock (Theragra chalcogramma) had less lipids and failed to survive the following
29
707
winter (Heintz, personal communication) resulting in very low recruitment (Stabeno et al.,
708
2012b).
709
The EBS shelf is divided into three domains (Coachman, 1986). The inner or coastal
710
domain (water depth < 50 m) is characterized as well mixed or weakly stratified during the
711
summer. The middle domain (50 – 100 m) has well defined two layers during the summer with a
712
wind mixed surface layer and tidally mixed bottom layer. The outer domain (100 – 180 m) is
713
more oceanic, with a well-mixed surface layer and bottom layer separated by an intermediate
714
layer. On the middle domain, the influence of the sea ice from the winter and spring often persists
715
through the following summer. As a part of a biophysical observing network on the EBS shelf,
716
four moorings on the middle domain (M2, M4, M5 and M8) (Fig. 1) provide long time series (9 –
717
18 years) of temperature and salinity data that have been published (e.g. Stabeno et al., 2010,
718
2012a,b; Sigler et al., this issue; Sullivan et al., this issue).
719
The purpose of this study is to examine the impacts of seasonal sea ice on the EBS shelf
720
water properties, using output from a global climate model and compare it with in situ ocean
721
(primarily data from moorings) and satellite observations. We will examine sea ice variability in
722
the EBS on seasonal and inter-annual time scales. Pan-Arctic sea ice is projected to decline in the
723
future under Greenhouse Gases (GHGs) forcing but there are large uncertainties (Wang and
724
Overland, 2009). For instance since 2007, winter/spring sea ice extent has been well above
725
average in the EBS even though the summer minimum ice extent in the Arctic has reached record
726
minimums (Stroeve et al. 2007; Stabeno et al., 2012 a,b; Brown and Arrigo 2012).
727
The paper is organized as the following. In section 2 we describe sea ice satellite
728
observational data and the numerical model output used in this study. In section 3 we present the
729
main results, including comparing model simulations with satellite and in situ observations. In
730
section 4 we draw conclusions and discuss the implications of the results on the EBS marine
731
ecosystem.
732
30
733
2.
Data and methods
734
2.1.
Satellite and sea ice data
735
Ice concentration images are derived using data from the Bootstrap algorithm files of
736
daily ice concentration we download from the National Snow and Ice Data Center (NSIDC). The
737
files use data from the Scanning Multichannel Microwave Radiometer (SMMR) aboard the
738
Nimbus-7 satellite or the Special Sensor Microwave/Imager (SSM/I) aboard Defense
739
Meteorological Satellite Program (DMSP) satellites, depending on the year. We used data from
740
years 1987 to 2005.
741
To examine how the ice cover varies along the EBS middle shelf, a 100 km by 100 km
742
box was defined centered on each of four biophysical moorings (M2, M4, M5, and M8) that have
743
been maintained by NOAA since 1990s. Percent of ice cover from satellite was averaged over
744
these 100 km by 100 km regions and monthly climatologies were calculated.
745
746
2.2.
Numerical model output
747
The global climate model we used is the NCAR-CESM (National Center for Atmospheric
748
Research-Community Earth System Model). The NCAR-CESM (and its predecessors) is a global
749
coupled ocean-sea ice-atmosphere-land model (Gent et al. 2012) participating in the IPCC
750
(Intergovernmental Panel on Climate Change) assessment reports. Specific formulations of the
751
sea ice model used in CESM are summarized in Holland et al. (2012). CCSM4 (the version of
752
CESM used in up-to-date publications) simulated Arctic ice extent seasonal cycle in the 20 th
753
century follows observations closely (Jahn et al., 2012). Arctic sea ice simulated by CCSM4 in
754
the 21st century has been investigated in a number of studies (e.g., Vavrus et al., 2012).
755
The particular configuration of the CESM used in this study has 0.9° latitude by 1.25°
756
longitude resolution for the atmosphere and land components, and nominal 1° horizontal
757
resolution for the ocean and sea-ice components. The external climate forcing factors (including
31
758
natural and anthropogenic components) vary with time. Because it is a coupled model, the
759
feedbacks between different components of the climate system, such as the ocean and sea ice, are
760
preserved. Monthly averaged ocean and sea ice states from the CESM “historical” (from 1850 to
761
2005) and “RCP” (Representative Concentration Pathways) (year 2005 onward) simulations were
762
downloaded from NCAR data mass storage. The “historical” simulations of the CESM were
763
initialized from a spun-up state of the climate system under pre-industrial GHGs forcings. The
764
“historical” simulations cover years 1850 to 2005 (we only used output from the late 20 th
765
century), while its “RCP” simulations were initialized from the climate state of the model at the
766
end of the “historical” simulations (Gent et al. 2012), and RCP simulations cover years 2006-
767
2100 and later. We use results from the RCP6.0 forcing (radiative forcing stabilizes at 6.0 W/m2
768
by year 2100), a medium emission scenario.
769
To study the inter-annual variability of sea ice and its impact on the EBS shelf, we extract
770
Empirical Orthogonal Functions (EOFs) of physical variables from multiple years in March.
771
Mean states of the shelf water property and circulation averaged over “high” and “low” ice years
772
are computed and compared against each other. The “high” and “low” ice years are defined based
773
on March ice concentration at M5: if a year’s value is above +1STD of the multi-year time series,
774
that year is a “high” ice year; if it is below -1STD, it’s “low” ice year.
775
To facilitate the comparison with observations, we carried out an additional CESM
776
RCP6.0 simulation with daily sea ice and ocean output, from December 1, 2016 to June 30, 2017.
777
This time period is used as a “representative” year with substantial ice cover in the EBS (shown
778
later). The comparison between this model output and observations are process oriented. Due to
779
the internal variability of the coupled system, simulated climate state at any given year is not
780
guaranteed to “match” with observations for that year.
781
When applicable, the model simulations are compared with existing figures in Sullivan et
782
al. (this issue) which summarizes the ocean in situ observations at the four moorings (M2, M4,
783
M5 and M8) (Fig. 1).
32
784
785
3.
Results
786
3.1.
Sea ice climatology
787
Monthly
mean
climatology
788
(defined over years 1987-2005) of sea ice
789
concentration was computed from the
790
CESM
791
compared with the satellite data (Fig. 2).
792
Satellite data show a gradual increase in
793
ice cover from December through March
794
(growth in December is small), with the 1987-2005) of sea ice concentration (percentage) at the
“historical”
simulation
and
Figure 3. Monthly mean climatology (averaged over years
mooring sites from CESM simulation (red) and SSM/I data
(green). Names of the moorings are marked on each panel.
Vertical bars show one standard deviation.
795
percent
796
maximum in March on the middle shelf
797
regardless of latitude. After March, ice cover decreases rapidly. At the southern two sites, ice
798
cover is reduced to zero by May. Farther north, an additional month is required for complete loss
799
of sea ice (June at M5 and M8). The magnitude of the standard deviations at each location
800
suggests that ice cover is more variable at M4 than at the other locations examined.
of
ice
cover
reaching
its
801
While the model simulations show same general patterns as the observations, there are
802
marked differences in timing and magnitude. The CESM captures the March maximum at the
803
southern two sites (Fig. 2, M2 and M4); at the northern sites (M5 and M8), CESM reaches annual
804
maximum ice cover in April, a month later than observed. The model simulation underestimates
805
ice cover at all four locations leading up to the March maximum and overestimates ice cover after
806
March. Complete loss of ice occurs one month later than observed in CESM simulations, except
807
at the southernmost site. The magnitude of the variability is well reproduced by CESM except at
808
the M4 location where the model significantly underestimates the interannual variability.
33
809
The modeled spatial pattern of ice
810
concentration in March, the month of maximum
811
annual ice extent, is similar to observations
812
(Fig. 3). The biggest discrepancy is that the
813
simulated ice concentration gradient is weaker
814
than observations, especially in the western part
815
of the domain. The model results show
816
significant ice cover over the deep basin in the
817
northern
ice
is
818
constrained to the shelf in observations.
In
819
addition, in the southeast corner of the EBS
820
shelf, the influence of warm inflow from the
821
Pacific is observed to result in less ice close to
822
the Alaskan Peninsula (Fig. 3 upper panel). Figure 4. Multi-year (1987-2005) mean March sea ice
823
Due to the low spatial resolution, the CESM
824
model cannot be expected to reproduce such regional features.
Bering
Sea,
while
sea
concentration (percentage) from SSM/I data (upper
panel) and CESM simulation (lower panel).
825
826
3.2.
Ice formation: dynamic versus thermodynamic processes
827
The mechanisms that result in changes in the areal extent of ice can be divided into two
828
broad categories: dynamic processes and thermodynamic processes. Dynamic processes include
829
advection and mechanical ridging/rafting, while thermodynamic processes are ice melting and
830
freezing. Not surprisingly, ice cover changes seasonally in different locations because of different
831
processes.
34
832
At
all
four
mooring
833
locations, dynamical processes (for
834
simplicity,
835
“advection”) contribute to positive
836
ice growth in January (Fig. 4). At
837
the southern two moorings (M2 and
838
M5),
839
concentration
840
advection of ice into the region
841
offset by local melting (Fig. 4b, c). Figure 5. Time integrals of ice area tendency
we
will
call
instantaneous
is
a
it
ice
result
of
t
A
 t dt,T
0
 Dec1
T0
842
The
843
advection and the local melt terms instantaneous ice area A (black) at the moorings (names of them are
844
are weaker at M5 than at M2.
845
Further north at M8, ice growth is dominated by advection into the region, with a small
846
contribution from local formation (Fig. 4a). Thus, ice melt contributes to water column properties
847
throughout the winter over a majority of the shelf. Ice formation is only important in the
848
northernmost region from January through April. After April, local melting contributes to the
849
decline in ice cover even in the north.
magnitudes
of
both
the (here A is the ice area and t is time) due to dynamics (green) and
thermodynamics (red), and their comparisons with the daily
marked on the panels). Ice area is defined as percentage coverage.
 they are similar to results at
Results at M4 are not shown because
M2.
850
Spatially, near Bering Strait and along the inner shelf, ice formation (Fig. 5a) is largely
851
offset by advection out of the region (Fig. 5b). This balance is reversed in the southern and
852
offshore locations, where advection into the region is offset by local ice melt. The pattern thus
853
emerged is that, sea ice is thermodynamically formed in the northern domain and along the coast,
854
and then transported southward and offshore, and subsequently melted in the warmer ambient
855
environment. This is the well-known “freshwater conveyor” effect of seasonal sea ice on the
856
Bering Sea shelf (Pease 1980; Sullivan et al, this issue). During this seasonal transit, the largest
857
ice-ocean exchanges in heat and freshwater occur at the ice edge (Figs. 5 c, d).
35
858
The above ice growth and
859
retreat processes have strong interannual
860
variabilities, which are influenced by
861
wind
862
empirical orthogonal functions (EOFs)
863
of March ice advective growth (ice area
864
tendency
865
formation/melt (ice area tendency due to
866
thermodynamics), and wind stress are
867
extracted
868
another (Fig. 6). The dominant mode of
869
March
870
variability (accounting for over 60% of
871
the variability) is an oscillation between
872
northeasterly and southwesterly winds
873
(Fig. 6c,d). The principal component
874
(PC) of this wind stress mode (black line
875
on Fig. 6d) is highly correlated with PCs
876
of
877
thermodynamics and dynamics (Fig. 6d,
878
red and green lines): the correlation to dynamics (a) and thermodynamics (b), and spatial pattern of
879
coefficients between the three PCs on denote the 50m, 100m, and 200m ocean topography. Contours of
forcing.
due
To
to
illustrate
dynamics),
this,
ice
Figure 6. Colors show the spatial patterns of ice area tendency
ice
and
wind
area
compared
stress
with
one (percent/day) due to thermodynamics (a) and dynamics (b), and
inter-annual
tendencies
due
ice to ocean heat (c, in W/m2) and freshwater (d, in cm/day)
fluxes. White areas are ice-free. Positive signs in c) and d) mean
into the ocean. Contours in c) and d) mark the 0.8, 0.6, 0.4, and
0.2 ice area (in fraction). All results are averaged over February
of the model year shown in Figure 4. Circles mark the mooring
locations.
to
Figure 6. Spatial patterns of EOF1 of the ice area tendency due
EOF1 of the surface wind stress (c). Thin black lines on a) and b)
a) is redrawn in c). d) Corresponding principal components
(PCs) of the spatial patterns shown in a)–c). CESM March data
is used for this analysis. Percentage of variance explained by
EOF1 is marked on a)–c).
880
Fig. 6d range from 0.80 to 0.97. When
881
the winds are from the northeast, more
882
sea ice is advected into the middle shelf (Fig. 6a), where it melts (Fig. 6b). At the same time, in
883
the northern and coastal regions, freezing (Fig. 6b) and ice removal by advection (Fig. 6a) are
36
884
enhanced under northeasterly winds (Fig. 6c). This strong coupling between ice advection and
885
melting is a characteristic of the EBS.
886
Next we examine relationships
887
between seasonal ice growth/decline
888
(via
889
formation/melting) and ice-ocean heat
890
and freshwater exchange (Fig. 7).
891
Advection of ice into the region begins
892
in December in the north (M8) and
893
January farther south (M2 and M5) and Figure 7a. Modeled daily ice area tendency due to
894
cooling of the ocean occurs soon after.
895
At the northern sites (M5 and M8), there
896
is a period in which ice area tendency and ice-ocean fluxes are relatively constant between
897
January, the beginning of the seasonal ice increase, and March or April, the beginning of seasonal
898
decrease (Fig. 7). Advection induced ice area tendency (red lines in Fig. 7) is generally smaller at
899
the northern mooring M8 than at moorings farther south (M5 and M2).
advection
and
local
thermodynamics (red) and dynamics (green) and their
comparisons with the simultaneous ice to ocean heat fluxes
(black). Positive (negative) heat flux means into (out of) the
ocean. Panels a)–c) correspond to M8, M5, and M2 respectively.
M4 results are similar to M2 and not shown.
900
In terms of ice-ocean heat exchange, at all sites and throughout the season, the ocean
901
loses heat. At M2 and M5, during the ice growth season (January or February), the heat loss
902
appears correlated with ice area tendency but lags the latter by a few days (Fig. 7a). The lag-0
903
correlation coefficients between black and red (green) lines on Fig. 7 are 0.97 (-0.52) and 0.74 (-
904
0.54) for M2 and M5, respectively, which are statistically significant with the 95% confidence
905
level. Heat loss in these locations is controlled by both latent and sensible heat exchange. During
906
ice formation, latent heat would result in a positive heat flux to the ocean. At M8, freezing (Fig.
907
7a; positive red line) is accompanied by negative heat flux to the ocean (Fig. 7a; black line),
908
implying that sensible heat flux dominates latent heat flux. This is consistent with model output
909
(Fig. 9) showing that the ocean temperature is above the freezing point in January. Because ice
37
910
bottom temperature is at the freezing point, there is a sensible heat exchange at the ice-ocean
911
interface.
912
The relationship between ice
913
tendency and ice-ocean freshwater
914
exchange (Fig. 7b) is straightforward.
915
At all locations, the water column
916
freshens (becomes saltier) whenever ice
917
melts (forms). Thus, the ocean gains
918
freshwater in the south throughout the
919
season while, in the north, the ocean Figure 7b. Same as Fig. 7a except the black lines are ice to oean
920
gains salt during ice formation (mainly of) the ocean.
921
in January and February) and freshwater during ice melting (in April and May). Ice advective
922
growth (green lines in Fig. 7) plays an implicit role in ice-ocean freshwater exchange through its
923
control on ice thermodynamic growth (Fig. 5).
freshwater fluxes (cm/day), positive (negative) means into (out
924
925
3.3.
926
temperature and salinity
927
Influences
on
The vertical
ocean
structure
and
928
temporal
929
temperature and salinity are related to
930
ice cover. At M5 and M8, the water
931
column in CESM is vertically well
932
mixed
933
temperature
934
beginning of the winter (December) as
935
a result of wind mixing (Fig. 8). The weak stratification at M8 (and to a lesser extent at M5) is
variability
or
weakly
and
of
ocean
stratified
salinity
at
in
the
Figure 8. Colors show modeled ocean temperature (left) and
sality (right) as a function of depth (0-100m) and time at the
moorings. The thick black lines on each panel show ice area
(percentage) at that mooring. M4 results are not shown because
they are similar to M2 results.
38
936
maintained through spring under significant ice cover, and at M8, there is a hint of brine rejection
937
into the bottom layer when ice is formed. This is not dissimilar to observations at the
938
moorings. The observed water column at M8 tends to be well mixed under the ice, with a winter
939
season increase in salinity of ~1 psu, which is a result of brine rejection or on-shelf advection
940
during the cold winter months (e.g., Fig. 12 in Sullivan et al., this issue). This observed increased
941
in salinity contrasts with a small (~0.2 psu) depth-averaged decrease in salinity in the model.
942
At M2, however, the differences between the modeled and observed water column
943
structures are greater. In December, the modeled water column is stratified near the bottom, albeit
944
weakly, especially in salinity (Fig. 8, bottom panels). In observations, during years with little or
945
no ice over the mooring, the water column is well mixed usually by November (Table 2 and Fig.
946
6 in Sullivan et al., this issue). During years in which ice is extensive over the southern domain,
947
the ice often arrives when the water column is well mixed, but above freezing. The ice cools and
948
freshens the surface waters, but the near bottom waters remain warmer and saltier for several
949
weeks before the water column becomes
950
well mixed.
951
cooling and freshening at M2 after ice
952
arrival, similar to observations, but
953
stratification persists in the model. In
954
CESM the ocean stratification begins to
955
increase during the melting season at all
956
mooring locations (Fig. 8). Stratification
957
measured at the moorings usually begins
958
after ice concentrations have decreased ocean temperature (red, units: °C) and salinity (green, units:
959
CESM also simulates
Figure 9. Modeled daily ice area (black,, units: percentage) and
(Sullivan et al., this issue).
psu) at 10m depth at moorings M2, M5 and M8. M4 result is
similar to M2 result and therefore not shown.
960
The near-surface ocean gradually cools starting from fall (not shown), but the strongest
961
cooling is concurrent with largest sea ice increase at all moorings (Fig. 9, red lines). This result is
39
962
consistent with observations (Sullivan et al., this issue, their Figs. 7-12). In the model, as in
963
observations, cooling at M8 and M5 continues even under significant ice coverage (Fig. 9ab, red
964
line, late January), until the temperature reaches the freezing point and remains constant
965
thereafter. The quasi-steady thermal ocean lasted a couple months at M8, but only briefly at
966
M5. In observations, warming generally begins in March, although if ice retreats, the water
967
column can warm through advection especially at the southern moorings M2 and M4 (Stabeno et
968
al., 2010, 2012a; Sullivan et al., this issue). The model shows a similar pattern, in that after the
969
sea ice diminishes (which occurs sooner in the lower latitudes), the near-surface ocean warms
970
rapidly (Fig. 9).
971
The modeled seasonal evolution of near-surface salinity is more variable and differs from
972
north to south (Fig. 9, green lines). Starting in December, modeled near-surface salinity at the two
973
northern mooring sites generally decreases. At M8, the linear trend continues throughout the
974
season, although it is also punctuated by episodes of salinity increase. At M5, salinity decreases
975
from December to early February along with ice increase, it then exhibits a period of weak
976
changes from late February to March, followed by a gradual increase. At M2, large synoptic scale
977
variability is observed in the near-surface salinity. In contrast, at each of the mooring sites,
978
observed salinity increases from December to March because of brine rejection and/or on-shelf
979
advection during the cold winter months, after which ice begins to melt, thus freshening the water
980
column (Sullivan et al., this issue).
981
982
983
984
3.4.
Contrast in ocean properties between “high” and “low” ice years
Change in bottom temperature, salinity, and density in high ice years relative to low ice
years indicates cooling anomalies over the entire bottom layer, with the strongest
40
985
anomaly occurring in the southern
986
domain (Fig. 10a); concurrently, a
987
large part of the inner and middle
988
shelf domain has positive salinity
989
anomalies (Fig. 10b). While the
990
gradient
991
anomalies is the strongest in the
992
north-south direction, the salinity
993
anomaly has the strongest gradient
994
in the onshore-offshore direction. As Figure 10. Shading indicates anomalies in March bottom
995
expected, in this subarctic ocean, the
996
bottom density anomaly (Fig. 10d) is
997
dominantly controlled by salinity.
998
The
in
the
temperature
temperature (a), salinity (b), and density (d) averaged over high ice
years minus that over low ice years. Thin black lines in a) and b)
mark the 50m, 100m, and 200m isobath. Vectors in d) show the
ocean ciruculation anomalies at 50m depth, dark blue ones are the
results on the shelf (depth< 200m), while the light blue ones are
results in the slope region (200m<depth<2000m). March bottom
density (shade) and 50m currents (vectors) averaged over high ice
years are shown in c) as a reference. CESM output from years
1961-2005 is used in this analysis.
corresponding upper
999
ocean circulation is anomalously southward on the shelf except in the Gulf of Anadyr (Fig. 10d,
1000
vectors). The anomalous circulation is close to the shelf-break between 59°N and 60°N with a
1001
slightly offshore direction, but is located further in-shore in the southern domain (south of 59°N),
1002
and shows a strong on-shore movement toward the southeast corner of the EBS when it reaches
1003
the Aleutian Island chain. However, the circulation anomalies next to the shelf-break in the slope
1004
region (200m < depth < 2000m) are opposite to circulation anomalies on the shelf, and move
1005
1006
northwestward.
41
1007
3.5.
1008
We compare March ice cover at the mooring sites between two periods: years 1961-2005 vs.
1009
years 2006-2050, since March is near
1010
the annual maximum ice extent (Fig.
1011
2). During the earlier period, ice
1012
cover at M2 and M4 is usually low (<
1013
50%) while ice cover at M8 is usually
1014
high (>50%) (Fig. 11, left panels). Ice
1015
cover at M5 fluctuates between low
1016
and high conditions on interannual
1017
time scales (Fig. 11, left panels). Ice Figure 11. Time series of March sea ice area (percentage) at the
1018
cover at all locations in the later
1019
mooring locations during years 1961-2005 (left panels) and years
2006-2050 (right panels) from CESM. Year 2005 is the end of
“historical” period, while year 2006 is the beginning of “RCP”
period.
period (Fig. 11, right panels) is
1020
smaller relative to the earlier period.
1021
M2 and M4 are mostly ice-free in
1022
the later period, only occasionally
1023
having significant ice cover. In the
1024
earlier
1025
distribution functions (PDFs) of ice
1026
cover at M2 and M4 show increased
1027
occurrence of low ice cover events
1028
while the PDF at M8 has more Figure 12. Probability distribution functions (PDFs) of March ice
1029
occurrence of high ice cover events
1030
(Fig. 12, left panels); PDF at M5 is approximately flat among all categories (Fig. 12, left panel,
1031
M5). In the later period, PDF at M5 has more occurrence of low ice cover events while PDF at
Decadal changes: “historical” versus “RCP” simulations
period,
the
probability
area (percentage) at the mooring locations. Left panels are results
from period 1 (years 1961-2005), right panels are from period 2
(years 2006-2050).
42
1032
M8 is more or less flat among all categories (Fig. 12, right panels). In this sense, PDF at M8 in
1033
the later period is similar to PDF at M5 in the earlier period.
1034
1035
4.
Conclusions and Discussion
1036
Seasonal cycle and spatial pattern of the annual maximum ice cover in the EBS modeled
1037
by CESM are reasonable compared to satellite observations (Figs. 2, 3). However, biggest
1038
discrepancies include delayed onset of the annual maxima (by a month) and slower summer
1039
melting in the northern Bering Sea in the simulation (Fig. 2).
1040
One possible cause for the slower summer melt could be that the simulation maintains
1041
weak stratification of the water column under the ice (Fig. 8), while observations show that the
1042
water column is well mixed throughout the winter. The EBS shelf loses heat and gains freshwater
1043
during the melting season. If the ocean layer receiving these fluxes is too shallow (due to
1044
stratification), this cooling and freshening effect on the surface layer would be too strong. A
1045
cooled ocean slows down further melting, causing a negative feedback on ice melting. An
1046
anomalously fresh surface layer further increases water column stability. In addition, an initial
1047
delay in ice retreat allows ice to remain into the period when spring winds are weakening, further
1048
delaying ice break-up and melt. These feedbacks may contribute to the slower melting in the
1049
model than in observations.
1050
Modeled seasonal ice advance in the EBS is consistent with the “freshwater conveyor”
1051
conceptual model suggested by observations (Pease 1980; Sullivan et al., this issue). Sea ice is
1052
both advected in and formed locally in the northern domain. In the southern domain, sea ice is
1053
advected from the north and melts locally. The regional differences in the partition between
1054
advection and formation/melt influence water column properties, particularly affecting local
1055
salinity budgets. The boundary between the northern domain where formation dominates and the
1056
southern domain where advection and melt dominate changes seasonally (Fig. 5) and
43
1057
interannually (Fig. 6). On interannual time scales, the division is strongly influenced by surface
1058
wind.
1059
At all mooring sites and throughout the season, modeled heat flux at the ice-ocean
1060
interface is from the ocean to the ice (Fig. 7a) while freshwater flux can be either into the ocean
1061
(during melting) or out of the ocean (brine rejection during freezing). The salt/freshwater flux in
1062
nature is influenced by island polynyas in the Bering Sea. St. Lawrence Island, to the north of
1063
M8, and to a lesser extent St. Matthews Island to the north of M5, is associated with important
1064
polynyas. CESM does not resolve either polynya. Thus, the model will underestimate ice
1065
formation due to the polynyas and the associated salinity flux into the ocean at these locations.
1066
Despite these limitations, in CESM, higher spring ice concentration in the northern EBS
1067
leads to positive salinity anomalies (relative to years with less ice concentration) on a large part of
1068
the bottom water (Fig. 10b), and the anomaly is the strongest in the northern domain near-shore
1069
locations. Weak negative salinity anomalies are seen in the southern end of the middle- and outer-
1070
shelf (Fig. 10b). Correspondingly, bottom water on the entire shelf has cold anomalies in high ice
1071
years relative to low ice years (Fig. 10a), and the anomalies are the strongest in the southern
1072
domain. Both temperature and salinity anomalies persists from spring into summer (not shown).
1073
According to the model, circulation on the middle shelf is anomalously southward during
1074
high ice years; however, next to the shelf-break in the slope region, circulation is anomalously
1075
northwestward during high ice years, opposite to situation on the shelf (Fig. 10 d). Broadly
1076
speaking, the southward circulation anomalies on the shelf are consistent with the southward
1077
wind anomalies during high ice years (Fig. 6). Such wind anomalies are also expected to cause
1078
the Anadyr current to flow stronger in the on-shelf direction toward Bering Strait (Danielson et al.
1079
2012). The reason for the northwestward anomalies in the slope region is unclear. The strength of
1080
the Bering Slope Current is correlated with the multivariate ENSO index and North Pacific Index
1081
on interannual time scales (Ladd, this issue), it is possible that these climate indices have
1082
manifestations in the EBS surface wind and sea ice variability.
44
1083
Over a 45-year time span (between period 1: 1961-2005 and period 2: 2006-2050), the
1084
EBS ice concentration PDF have shifted northward by 2 degrees. PDF of Ice concentration at M8
1085
in the later period is more or less flat, similar to PDF of ice concentration at M5 in the earlier
1086
period, while PDF at M5 in the later period becomes similar to PDF at M2 and M4 in the earlier
1087
period.
1088
Sea ice has profound effect on the EBS primary production (Hunt et al., 2002; Stabeno et
1089
al. 2010; Brown and Arrigo, 2012; Sigler et al., this issue), zooplankton structure (Hunt et al.
1090
2002; Coyle et al., 2011; Eisner et al., this issue), and upper trophic level species (e.g., Wyllie-
1091
Echeverria and Wooster, 1998). Timing of ice retreat determines spring bloom. Over the southern
1092
shelf, if ice is present after mid-March, an early ice-associated bloom occurs; otherwise the spring
1093
bloom usually occurs in May (Stabeno et al. 2001; Hunt et al. 2002). In this regard, CESM
1094
captures well the March maximum ice cover in the southern domain in a climatological mean
1095
sense.
1096
Lower water temperatures associated with more sea ice reduce energy required for
1097
zooplankton overwinter survival, resulting in increased spawning biomass and production (Sigler
1098
et al., this issue). In addition, the distribution of low bottom temperatures associated with ice
1099
distribution (the cold pool) determine available habitat for some fish species (Hollowed et al.
1100
2012). Conversely, warmer temperatures increase metabolic requirements, leading to food
1101
limitation. This could result in reduced net primary production (Lomas et al., 2012), leading to
1102
poor overwinter survival of zooplankton (Coyle et al., 2011).
1103
In addition to influences on temperature and salinity, ice melt results in input of ice algae
1104
to the water column, and ice algae provides an early sources of food for zooplankton growth and
1105
spawning (Sigler et al., this issue). Thus, the “freshwater conveyor” associated with seasonal sea
1106
ice in the EBS can also be thought of as conveyor belt for ice alge, and possibly zooplankton. The
1107
balances between ice advection and ice formation/melt are important to the ecosystem, and
1108
understanding what processes controls the seasonal to interannual variability in these balances
45
1109
will have important implications. Our results suggest surface winds in the EBS should be closely
1110
monitored because of its strong influences on ice advection and local formation.
1111
Global climate models are more often used to study global and basin scale processes,
1112
their applications on regional scales are relatively rare. Understanding how climate models work
1113
on all scales, however, is critically important to our ability to predict future changes in the
1114
physical environment and marine ecosystems, the latter tend to operate on regional scales. Our
1115
next step is to investigate the Bering Sea ice-ocean system in a coupled biophysical ocean-ice
1116
circulation model. This class of models can afford higher spatial resolutions, although they
1117
misrepresent atmosphere-ocean or atmosphere-sea ice interactions. Combining knowledge gained
1118
from different models and observations should ultimately improve our understanding of the
1119
system.
1120
1121
Acknowledgements
1122
We thank W. Floering, C. Dewitt and S. McKeever for maintaining, deploying and recovering the
1123
instruments on the mooring, and D. Kachel, and P. Proctor for data processing. The research was
1124
generously supported by grants from NPRB (grant numbers), and the NPRB-sponsored BSIERP
1125
and NOAA’s North Pacific Climate Regimes and Ecosystem Productivity (NPCREP) programs.
1126
The CESM project is supported by the National Science Foundation and the Office of Science
1127
(BER) of the U.S. Department of Energy. Computing resources were provided by the Climate
1128
Simulation Laboratory at NCAR's Computational and Information Systems Laboratory (CISL),
1129
sponsored by the National Science Foundation and other agencies. This research was enabled by
1130
CISL compute and storage resources. Bluefire, a 4,064-processor IBM Power6 resource with a
1131
peak of 77 TeraFLOPS provided more than 7.5 million computing hours, the GLADE high-speed
1132
disk resources provided 0.4 PetaBytes of dedicated disk and CISL's 12-PB HPSS archive
1133
provided over 1 PetaByte of storage in support of this research project. This is contribution
1134
EcoFOCI-0806 to the Ecosystems & Fisheries Oceanography Coordinated Investigations; 2093 to
46
1135
Joint Institute for the Study of Atmosphere and Ocean, University of Washington, and 3995 to
1136
NOAA/Pacific Marine Environmental Laboratory. This is the BEST_BSIERP Bering Sea Project
1137
publication number XX.
1138
1139
Figure Captions
1140
Figure 1. Geographical information of the study area. The black dots mark the ocean moorings in
1141
the Eastern Bering Sea middle shelf.
1142
Figure 2. Monthly mean climatology (averaged over years 1987-2005) of sea ice concentration
1143
(percentage) at the mooring sites from CESM simulation (red) and SSM/I data (green). Names of
1144
the moorings are marked on each panel. Vertical bars show one standard deviation.
1145
Figure 3. Multi-year (1987-2005) mean March sea ice concentration (percentage) from SSM/I
1146
data (upper panel) and CESM simulation (lower panel).
1147
Figure 4. Time integrals of ice area tendency
t
A
 t dt,T
0
 Dec1 (here A is the ice area and t is
T0
1148
time) due to dynamics (green) and thermodynamics (red), and their comparisons with the daily
1149
instantaneous ice area A (black) at 
the moorings (names of them are marked on the panels). Ice
1150
area is defined as percentage coverage. Results at M4 are not shown because they are similar to
1151
results at M2.
1152
Figure 5. Colors show the spatial patterns of ice area tendency (percent/day) due to
1153
thermodynamics (a) and dynamics (b), and ice to ocean heat (c, in W/m2) and freshwater (d, in
1154
cm/day) fluxes. White areas are ice-free. Positive signs in c) and d) mean into the ocean.
1155
Contours in c) and d) mark the 0.8, 0.6, 0.4, and 0.2 ice area (in fraction). All results are averaged
1156
over February of the model year shown in Figure 4. Circles mark the mooring locations.
1157
Figure 6. Spatial patterns of EOF1 of the ice area tendency due to dynamics (a) and
1158
thermodynamics (b), and spatial pattern of EOF1 of the surface wind stress (c). Thin black lines
1159
on a) and b) denote the 50m, 100m, and 200m ocean topography. Contours of a) is redrawn in c).
47
1160
d) Corresponding principal components (PCs) of the spatial patterns shown in a)–c). CESM
1161
March data is used for this analysis. Percentage of variance explained by EOF1 is marked on a)–
1162
c).
1163
Figure 7a. Modeled daily ice area tendency due to thermodynamics (red) and dynamics (green)
1164
and their comparisons with the simultaneous ice to ocean heat fluxes (black). Positive (negative)
1165
heat flux means into (out of) the ocean. Panels a)–c) correspond to M8, M5, and M2 respectively.
1166
M4 results are similar to M2 and not shown.
1167
Figure 7b. Same as Fig. 7a except the black lines are ice to oean freshwater fluxes (cm/day),
1168
positive (negative) means into (out of) the ocean.
1169
Figure 8. Colors show modeled ocean temperature (left) and sality (right) as a function of depth
1170
(0-100m) and time at the moorings. The thick black lines on each panel show ice area
1171
(percentage) at that mooring. M4 results are not shown because they are similar to M2 results.
1172
Figure 9. Modeled daily ice area (black,, units: percentage) and ocean temperature (red, units: °C)
1173
and salinity (green, units: psu) at 10m depth at moorings M2, M5 and M8. M4 result is similar to
1174
M2 result and therefore not shown.
1175
Figure 10. Shading indicates anomalies in March bottom temperature (a), salinity (b), and
1176
density (d) averaged over high ice years minus that over low ice years. Thin black lines in a) and
1177
b) mark the 50m, 100m, and 200m isobath. Vectors in d) show the ocean ciruculation anomalies
1178
at 50m depth, dark blue ones are the results on the shelf (depth< 200m), while the light blue ones
1179
are results in the slope region (200m<depth<2000m). March bottom density (shade) and 50m
1180
currents (vectors) averaged over high ice years are shown in c) as a reference. CESM output from
1181
years 1961-2005 is used in this analysis.
1182
Figure 11. Time series of March sea ice area (percentage) at the mooring locations during years
1183
1961-2005 (left panels) and years 2006-2050 (right panels) from CESM. Year 2005 is the end of
1184
“historical” period, while year 2006 is the beginning of “RCP” period.
48
1185
Figure 12. Probability distribution functions (PDFs) of March ice area (percentage) at the
1186
mooring locations. Left panels are results from period 1 (years 1961-2005), right panels are from
1187
period 2 (years 2006-2050).
1188
1189
References
1190
Brown, Z.W., Arrigo, K.R., 2012. Contrasting trends in sea ice and primary production in the
1191
Bering Sea and Arctic Ocean. ICES J. Mar. Sci. 69, 1180-1193, 10.1093/icesjms/fss113.
1192
Coachman, L.K., 1986. Circulation, water masses and fluxes on the southeastern Bering Sea
1193
shelf. Cont. Shelf Res. 5(1–2), 23–108.
1194
Coyle, K.O., Eisner, L.B., Mueter, F.J., Pinchuk, A.I., Janout, M.A., Cieciel, K.D., Farley, E.V.,
1195
Andrews, A.G., 2011. Climate change in the southeastern Bering Sea: impacts on pollock
1196
stocks and implications for the oscillating control hypothesis. Fish. Oceanogr. 20, 139-156,
1197
10.1111/j.1365-2419.2011.00574.x.
1198
Danielson, S., Hedstrom, K., Aagaard, K., Weingartner, T., Curchitser, E., 2012. Wind-induced
1199
reorganization
1200
http://dx.doi.org/10.1029/2012GL051231.
of
the
Bering
Shelf
circulation.
Geophys.
Res.
Lett.
39(8),
1201
Eisner, L.B., Napp, J.M., Mier, K.L., Pinchuk, A.I., Andrews, A.G., submitted. Climate-mediated
1202
changes in zooplankton community structure for the eastern Bering Sea. Deep-Sea Res. II.
1203
Gent, P. R., G. Danabasoglu, L. J. Donner, M. M. Holland, E. C. Hunke, S. R. Jayne, D. M.
1204
Lawrence, R. B. Neale, P. J. Rasch, M. Vertenstein, P. H. Worley, Z. -L. Yang, and M.
1205
Zhang, 2011: The Community Climate System Model version 4. J. Climate, 24, 4973-4991,
1206
doi: 10.1175/2011JCLI4083.1.
1207
Holland, M. M., D. A. Bailey, B. P. Briegleb, B. Light, and E. Hunke, 2012: Improved Sea Ice
1208
Shortwave Radiation Physics in CCSM4: The Impact of Melt Ponds and Aerosols on Arctic
1209
Sea Ice. J. Climate, 25, 1413-1430, doi: 10.1175/JCLI-D-11-00078.1.
49
1210
Hollowed, A.B., Barbeaux, S.J., Cokelet, E.D., Farley, E., Kotwicki, S., Ressler, P.H., Spital, C.,
1211
Wilson, C.D., 2012. Effects of climate variations on pelagic ocean habitats and their role in
1212
structuring forage fish distributions in the Bering Sea. Deep-Sea Res. II 65–70, 230-250,
1213
http://dx.doi.org/10.1016/j.dsr2.2012.02.008.
1214
Hunt, G.L., Jr., Stabeno, P., Walters, G., Sinclair, E., Brodeur, R.D., Napp, J.M., Bond, N.A.,
1215
2002. Climate change and control of the southeastern Bering Sea pelagic ecosystem. Deep-
1216
Sea Res. II 49, 5821-5853.
1217
Jahn, A., D.A. Bailey, C.M. Bitz, M.M. Holland, E.C. Hunke, J. E. Kay, W. H. Lipscomb, J.A.
1218
Maslanik, D. Pollak, K. Sterling, J. Strœve, 2012: Late 20th Century Simulation of Arctic Sea
1219
Ice and Ocean Properties in the CCSM4. J. Climate, 25, 1431-1452, doi: 10.1175/JCLI-D-11-
1220
00201.1.
1221
Vavrus, S. J., M. M. Holland, A. Jahn, D. A. Bailey, and B. A. Blazey, 2012: 21st-Century
1222
Arctic Climate Change in CCSM4. J. Climate, 25, 2696-2710, doi: 10.1175/JCLI-D-11-
1223
00220.1.
1224
1225
Ladd, C.A., Stabeno, P.J., 2012. Stratification on the Eastern Bering Sea shelf revisited. Deep-Sea
Res. II 65–70, 72–83.
1226
Lomas, M.W., Moran, S.B., Casey, J.R., Bell, D.W., Tiahlo, M., Whitefield, J., Kelly, R.P.,
1227
Mathis, J.T., Cokelet, E.D., 2012. Spatial and seasonal variability of primary production on
1228
the
1229
http://dx.doi.org/10.1016/j.dsr2.2012.02.010.
Eastern
Bering
Sea
shelf.
Deep-Sea
Res.
II
65–70,
126-140,
1230
Pease, C.H., 1980. Eastern Bering Sea ice processes. Mon. Weather Rev. 108(12), 2015–2023.
1231
Sigler, M.F., Stabeno, P.J., Eisner, L.B., Napp, J.M., Mueter, F.J., this issue. Spring and fall
1232
phytoplankton blooms in a productive subarctic ecosystem, the eastern Bering Sea, during
1233
1955-2011.
50
1234
Stabeno, P.J., Bond, N.A., Kachel, N.B., Salo, S.A., Schumacher, J.D., 2001. On the temporal
1235
variability of the physical environment over the south-eastern Bering Sea. Fish. Oceanogr.
1236
10, 81-98.
1237
Stabeno, P.J., Napp, J., Mordy, C., Whitledge, T., 2010. Factors influencing physical structure
1238
and lower trophic levels of the eastern Bering Sea shelf in 2005: Sea ice, tides and winds.
1239
Prog. Oceanogr. 85(3–4), 180–196.
1240
Stabeno, P.J., Farley Jr., E.V., Kachel, N.B., Moore, S., Mordy, C.W., Napp, J.M., Overland, J.E.,
1241
Pinchuk, A.I., Sigler, M.F., 2012a. A comparison of the physics of the northern and southern
1242
shelves of the eastern Bering Sea and some implications for the ecosystem, Deep-Sea Res. II
1243
65–70, 14–30.
1244
Stabeno, P.J., Kachel, N.B., Moore, S.E., Napp, J.M., Sigler, M., Yamaguchi, A., Zerbini, A.N.,
1245
2012b. Comparison of warm and cold years on the southeastern Bering Sea shelf and some
1246
implications for the ecosystem, Deep-Sea Res. II 65–70, 31–45.
1247
1248
1249
1250
1251
1252
1253
1254
Stroeve, J., M. M. Holland, W. Meier, T. Scambos, and M. Serreze, 2007: Arctic sea ice decline:
Faster than forecast. Geophys. Res. Lett., 34, L09501, doi:10.1029/2007GL029703.
Magaret Sullivan, M., C. Mordy, S. Salo, and P. Stabeno, this issue: The influence of sea ice on
seasonal water column changes on the eastern Bering Sea shelf.
Wang, M., and J. E. Overland, 2009: A sea ice free summer Arctic within 30 years? Geophys.
Res. Lett., 36, L07502, doi:10.1029/2009GL037820.
Wyllie-Echeverria, T., Wooster, W.S., 1998. Year-to-year variations in Bering Sea ice cover and
some consequences for fish distributions. Fish. Oceanogr. 7, 159-170.
1255
1256
51
1257
Chapter 13
1258
1259
Eisner, L. B., J. M. Napp, K. L. Mier, A. I. Pinchuk, A. G. Andrews III. (2013). Climate-
1260
mediated changes in zooplankton community structure for the eastern Bering Sea , Deep-Sea Res.
1261
II.
1262
1263
Climate-mediated changes in zooplankton community structure for the eastern Bering Sea
1264
1265
Lisa B. Eisnera*, Jeffrey M. Nappa, Kathryn L. Miera, Alexei I. Pinchukb, Alexander G. Andrews
1266
IIIc.
1267
1268
a
1269
98115, Lisa.Eisner@NOAA.gov, 001-206-526-4060; Jeff.Napp@NOAA.gov;
1270
Kathy.Mier@NOAA.gov
1271
b
1272
Loop Road, Juneau, Alaska 99801, aipinchuk@alaska.edu
1273
c
1274
Alaska 99801, Alex.Andrews@NOAA.gov
1275
1276
1277
1278
*Corresponding Author Address
1279
Zooplankton are critical to energy transfer between higher and lower trophic levels in the eastern
1280
Bering Sea ecosystem. Previous studies from the southeastern Bering Sea shelf documented
1281
substantial differences in zooplankton taxa in the Middle and Inner Shelf Domains between warm
1282
and cold years. Our investigation expands this analysis into the northern Bering Sea and the south
1283
Outer Domain, looking at zooplankton community structure during a period of climate-mediated,
1284
large-scale change. Elevated air temperatures in the early 2000s resulted in regional warming and
NOAA-Fisheries, Alaska Fisheries Science Center, 7600 Sand Point Way NE, Seattle, WA
University of Alaska Fairbanks, School of Fisheries and Ocean Sciences, 17101 Point Lena
NOAA-Fisheries, Alaska Fisheries Science Center, 17109 Point Lena Loop Road, Juneau,
Abstract:
52
1285
low sea ice extent in the southern shelf whereas the late 2000s were characterized by cold
1286
winters, extensive spring sea ice, and a well- developed pool of cold water over the entire Middle
1287
Domain. The abundance of large zooplankton taxa such as Calanus spp. (C. marshallae and C.
1288
glacialis), and Parasagitta elegans, increased from warm to cold periods, while the abundance of
1289
gelatinous zooplankton (Cnidaria) and small taxa decreased. Biomass followed the same trends as
1290
abundance, except that the biomass of small taxa in the southeastern Bering Sea remained
1291
constant due to changes in abundance of small copepod taxa (increases in Acartia spp. and
1292
Pseudocalanus spp. and decreases in Oithona spp.). Statistically significant changes in
1293
zooplankton community structure and individual species were greatest in the Middle Domain, but
1294
were evident in all shelf domains, and in both the northern and southern portions of the eastern
1295
shelf. Changes in community structure did not occur abruptly during the transition from warm to
1296
cold, but seemed to begin gradually and build as the influence of the sea ice and cold water
1297
temperatures persisted. The change occurred one year earlier in the northern than the southern
1298
Middle Shelf. These and previous observations demonstrate that lower trophic levels within the
1299
eastern Bering Sea respond to climate-mediated changes on a variety of time scales, including
1300
those shorter than the commonly accepted quasi-decadal time periods. This lack of resilience or
1301
inertia at the lowest trophic levels affects production at higher trophic levels and must be
1302
considered in management strategy evaluations of living marine resources.
1303
1304
Keywords: Zooplankton, community structure, climate change, eastern Bering Sea
1305
1306
1.0 Introduction:
1307
In recent years, climate change in the western Arctic has led to rapid changes in the eastern
1308
Bering Sea shelf resulting in variations in seasonal sea ice coverage and water column
1309
temperatures, and these variations have affected the entire ecosystem (e.g. Napp and Hunt, 2001;
1310
Stabeno et al., 2012b). Zooplankters are essential prey for many fish, seabirds, and marine
53
1311
mammals, therefore there is considerable interest in changes in zooplankton abundance and how
1312
these changes may propagate through the food web and impact higher trophic levels (Coyle et al.,
1313
2011; Hunt et al., 2011). Many planktivorous fishes, seabirds, and marine mammals inhabit both
1314
the northeastern and southeastern Bering Sea during summer (e.g. Piatt and Springer, 2003;
1315
Friday et al., 2012; Hollowed et al., 2012; Stevenson and Lauth, 2012). Historical research
1316
documented variations in single species abundance and cross shelf patterns of zooplankton in the
1317
southeastern Bering Sea (Cooney and Coyle, 1982; Coyle and Pinchuk, 2002; Napp et al., 2002;
1318
Stabeno et al., 2010; Coyle et al., 2011; Stabeno et al., 2012a & b). However, there are few
1319
modern descriptions of the broad-scale distribution of eastern Bering Sea zooplankton,
1320
particularly for the northern Bering Sea (Motoda and Minoda, 1974; Coyle et al., 1996 and
1321
references therein). To better understand how climate variability and change are affecting the
1322
recruitment and broad-scale distributions of planktivores, we need a more thorough understanding
1323
of how climate and other forcing affects the production and distribution of zooplankton across the
1324
entire region.
1325
1326
The eastern Bering Sea is a large marine ecosystem characterized by a broad continental shelf >
1327
500 km wide and > 1000 km long oriented in a north westerly direction from the Alaskan
1328
Peninsula in the south, to below St. Lawrence Island, approximately 63° latitude, in the north.
1329
Near the northern boundary, the system shifts from a pelagic to benthic dominated system (e.g.
1330
Grebmeier et al., 2006; Stevenson and Lauth, 2012; Sigler et al., this issue). The shelf can be
1331
divided into regions or domains, each with their characteristic hydrography, circulation, and
1332
fauna (Iverson et al., 1979; Cooney and Coyle, 1982; Coachman, 1986; Kachel et al., 2002).
1333
Cross shelf domains are most distinct during summer and fall in the southeastern Bering Sea
1334
where there are 3 domains approximated by water depth: Outer (100 – 180 m), Middle (50 – 100
1335
m) and Inner (< 50 m). The Inner Domain is well mixed, the Middle Domain is stratified into 2
1336
layers, and the Outer Domain is stratified, but with 3 layers. In addition to cross shelf variations,
54
1337
there are latitudinal variations in wind fields, water column properties, and most importantly, sea
1338
ice coverage (Stabeno et al., 2010, 2012a). The northeastern Bering Sea is ice covered every year;
1339
whereas coverage in the south historically varies by more than 100 km (Stabeno et al., 2012a).
1340
The cold pool (a layer of < 2 °C water that resides on the shelf bottom), is formed as a result of
1341
winter cooling and mixing that most often precedes sea ice formation), and typically persists
1342
through the following summer. Its extent and magnitude varies with climatic conditions affecting
1343
the distribution of many upper trophic level organisms in the region (e.g. Wyllie-Echeverria,
1344
1998; Hollowed et al., 2012, Kotwicki and Lauth, 2013).
1345
1346
The goal of this manuscript is to describe the broad-scale spatial variations in large and small
1347
zooplankton community composition in the north and southeastern Bering Sea during warm and
1348
cold climate states. Additionally, we address several important questions regarding zooplankton
1349
ecology affected by changing climate conditions (see Table 1). The data for these analyses come
1350
from broad-scale surveys of zooplankton conducted by the Bering-Aleutian Salmon International
1351
Survey (BASIS) program at NOAA, and provide the opportunity to describe the spatial cross-
1352
shelf and latitudinal distribution of zooplankton for the entire eastern Bering Sea shelf. These
1353
surveys began in 1999 before the Bering Sea Project (Bering Sea Ecosystem Study [BEST] and
1354
Bering Sea Integrated Ecosystem Research Program [BSIERP]) was launched in 2007 (Wiese et
1355
al., 2012) and have continued into the present after the field components of BEST and BSIERP
1356
ended. During the 2000s the region experienced different multi-year climate regimes (Stabeno et
1357
al., 2012b) above average sea water temperatures and very low sea ice coverage (2000 – 2005), a
1358
single year of average sea water temperatures and sea ice extent (2006), and cold years with
1359
extensive sea ice (2007 – 2009). Thus, in contrast to most regional scientific programs in the last
1360
50 years whose timing or duration allowed limited sampling in a single ecosystem state, our
1361
dataset allows us to evaluate changes in the eastern Bering Sea shelf pelagic ecosystem during
1362
multiple thermal phases.
55
1363
2.0 Methods:
1364
2.1 Survey station locations and
1365
oceanographic sample collection
1366
From mid August to early October,
1367
2003 – 2009, samples were collected in
1368
the eastern Bering Sea, at stations
1369
located from 54.5 – 63.0 N and 159.0
1370
– 174.0 W, and spaced approximately
1371
60 km apart (Cieciel et al., 2009;
1372
Farley and Moss, 2009; Fig 1). Stations
1373
were divided into shelf regions for
1374
analyses using the Bering Sea
1375
Integrated Ecosystem Program (BSIERP) designations based on oceanographic/hydrographic,
1376
fisheries and ecosystem characterizations (Ortiz et al., 2012). We further grouped these regions
1377
into five major domains using approximately 60°N to split the shelf into south and north (Stabeno
1378
et al. 2010) and approximating cross-shelf domains defined by Coachman (1986) yielding: S
1379
Inner (< 50 m bathymetry, BSIERP regions 2 and 7), S Middle (50 – 100 m, regions 3 and 6), S
1380
Outer (100 – 200 m, region 4), N Inner (<~40 m, region11), and N Middle (~40 – 100 m, regions
1381
9 and 10).
Figure 7. Bering Sea study area showing stations used for
analyses in each oceanographic shelf domain for: A) large
zooplankton and B) small zooplankton. Black lines indicate
spatial “blocks” which are thiessen polygons around points.
Blocks containing colored points (reduced data set) were
visited at least two years during warm (2003 – 2005) and two
years during cold (2006 – 2009) regimes; colors signify
domain.
1382
1383
Vertical profiles of temperature, salinity, and chlorophyll a (Chla) fluorescence were collected at
1384
each station with a Sea-Bird 1Model 25 or Model 9plus CTD. A rosette sampler was used to
1385
obtain discrete water samples for surface and bottom nutrients, total Chla (Whatman GF/F) and
1386
large size-fractionated Chla (> 10µm, Millipore Isopore polycarbonate membrane filters); all
1
Use of trade names does not signify an endorsement by the U.S. National Oceanic and Atmospheric
Administration.
56
1387
samples were stored frozen (-80 C) for 6 months maximum and analyzed using standard
1388
procedures at shore-based facilities (Parsons et al., 1984; Gordon et al., 1993). Vertical profiles of
1389
Chla concentrations were derived from regressions between in vivo fluorescence and discrete
1390
Chla samples, mean r2 = 0.65). Water column stability (energy required to mix the water column
1391
to 70 m, J m-3) was calculated at each station from CTD temperature and salinity data (Simpson et
1392
al., 1977). Ice coverage (number of days of ice at each station during the prior winter) and timing
1393
of retreat (the last day of ice in spring) for a 60 km2 square box centered on each station was
1394
obtained from the Advanced Microwave Scanning Radiometer (AMSR) on MODIS Aqua from
1395
the National Snow and Ice Data Center (NSIDC) (S. Salo, NOAA PMEL, personal comm.). We
1396
assumed no ice when concentrations dropped below 15%. Wind mixing data (wind velocity3, u*3)
1397
for August were obtained from National Centers for Environmental Prediction (NCEP) reanalysis
1398
data set (N. Bond, NOAA PMEL, personal comm.). Wind mixing data were averaged over a 2°
1399
latitude by 5° longitude box centered on a NOAA PMEL mooring at site M4 (57. 9°N, 168. 9°W;
1400
Stabeno et al., 2010). We assumed M4 winds approximated wind fields over our entire survey
1401
area, since winds at all four 70 m moorings in the northeast and southeast Bering Sea (M2, M4,
1402
M5 and M8) have shown strong coherence during recent years (Stabeno et al., 2010). Winter
1403
wind fields for the prior period (October to April) based on shelf-wide model results were used to
1404
estimate potential onshore and offshore flow, assuming that southeasterly winds promote stronger
1405
onshore flows than northwesterly winds (Danielson et al., 2012).
1406
1407
2.2 Zooplankton sample collection and lab analysis
1408
Zooplankton samples were collected and analyzed using methods described in Coyle et al.
1409
(2011). Briefly, small zooplankton assemblages were sampled at every other station with a 0.1 m2
1410
Juday net with 168 µm mesh (Volkov, 1984; Volkov et al., 2007). The net was towed vertically
1411
from within 5 – 10 m of the bottom (or maximum depth of 200 m) to the surface at about 1 m s-1.
1412
Juday net samples were sieved into size fractions prior to counting on board ship (Volkov et al.
57
1413
2007); however, data from all size fractions combined were used to quantify small zooplankton
1414
taxa. Large zooplankton assemblages were collected at every station with a 60 cm MARMAP-
1415
style bongo frame with a 505 µm mesh net. Oblique tows were conducted from the surface to
1416
within 5–10 m of the bottom, and volume filtered was measured with calibrated General Oceanics
1417
flowmeters. All samples were preserved in 5% formalin: buffered with seawater. Zooplankton
1418
collections from 2003 – 2004 were sorted at the Polish Plankton Sorting and Identification Center
1419
(Szczecin, Poland); collections from 2005 – 2009 were processed at the University of Alaska
1420
(Coyle et al., 2008). The lowest taxonomic level of sorting varied between labs, so we used the
1421
lowest taxonomic stage available across all years. Zooplankton tows were collected primarily
1422
during the daytime. Many euphausiids on the Bering Sea shelf stay within 1 – 2 m of the bottom
1423
during the day (Coyle and Pinchuk, 2002) and those in the water column are difficult to
1424
quantitatively capture with nets (e.g. Sameoto et al., 1993; Wiebe et al., 2004); therefore
1425
euphausiids were not included in analyses. It is also possible that under representation of other
1426
diel migrants (e.g. Metridia spp.) occurred. The small volume filtered by Juday and bongo nets
1427
compared to larger nets (e.g. Multinet or MOCNESS) could have resulted in under sampling of
1428
relatively large and rare taxa. Accordingly, very large medusae such as Chrysaora melanaster
1429
were excluded since they were not quantitatively sampled by our bongo nets. Unidentified
1430
copepod nauplii were also excluded, although they were fairly numerous in the Juday net samples
1431
(~ 11% of total abundance). A total of 424 Juday and 717 bongo net samples were collected for
1432
all years combined.
1433
1434
2.3 Statistical analyses
1435
Separate analyses were conducted for large (bongo samples) and small (Juday samples)
1436
zooplankton taxa. Zooplankton data were 4th root transformed (unless otherwise indicated), to
1437
down weight the contribution of dominant taxa to similarity/dissimilarity patterns (Clarke and
1438
Warwick, 2001). Initial analyses were conducted using both abundance (no. m-3) and integrated
58
1439
abundance (no. m-2) for comparisons among regions with large changes in bathymetry (Questions
1440
1 and 2, Table 1). We found no substantial differences in the outcome when using number
1441
normalized by volume or area; therefore we chose to present abundances normalized by volume
1442
filtered (no. m-3) to allow for comparisons with most prior studies on the eastern Bering Sea shelf
1443
which reported their abundance data normalized by volume filtered. Zooplankton biomass wet
1444
weights (g m-3, Coyle et al., 2011) were used in Question 2 analyses. The warm regime was
1445
designated as 2003 – 2005 and the cold regime as 2006 – 2009 based on late summer /fall mean
1446
water column temperature anomalies for 2003 – 2009, estimated from CTD data collected during
1447
our surveys (cf. Stabeno et al., 2012b, which categorizes 2006 as a year of average temperatures
1448
and sea ice extent). For all analyses, excluding those for Question 1, it was necessary to account
1449
for uneven sampling across years. Therefore, a reduced data set was used that included only
1450
station locations within equally spaced geographically partitioned regions (Blocks) that were
1451
sampled a minimum of two warm and two cold years (Fig. 1). Blocks were created in
1452
ArcMap/ArcInfo, 10.1 (ESRI, 2012) using thiessen polygons, polygons constructed around
1453
stations by calculating the midline between each pair of adjacent stations (Thiessen and Alter,
1454
1911; Gold, 1991). Blocks were approximately 60 x 60 km for the large taxa (equivalent to the
1455
spacing of stations) and approximately 80 x 80 km for the small (equivalent to the diagonal
1456
spacing of every other station). The small taxa required larger blocks due to sparser sampling.
1457
Separate analysis were conducted for each of the five major domains (S Middle, N Middle, S
1458
Inner, N Inner, S Outer) for Questions 3and 4 analyses (Table 1).
1459
1460
Statistical analyses included cluster analysis, similarity percentage contribution of individual taxa
1461
(SIMPER), and permutational, multivariate analysis of variance (PERMANOVA) in the software
1462
package, PRIMER-E,version 6.1.15 with PERMANOVA+ version 1.0.5 (Clarke and Warwick,
1463
2001; Clarke and Gorley, 2006; Anderson et al., 2008). All procedures used Bray-Curtis
1464
dissimilarities, a statistic commonly used to quantify the dissimilarity in composition between
59
1465
samples (Bray and Curtis 1957). In our study, these procedures were used to evaluate variations
1466
in community composition (the relative abundances of zooplankton taxa). SIMPER calculates the
1467
average contribution of each taxon to the overall Bray-Curtis dissimilarity in community
1468
composition. PERMANOVA allows the use of distance measures, in our case, Bray-Curtis
1469
dissimilarity in community composition, to test for the significance of specific effects, similar to a
1470
multivariate ANOVA. The advantage of using a permutation test is that it does not require the
1471
data to follow a particular distribution and is therefore more robust than parametric statistics. This
1472
procedure also allows for the partitioning of the multivariate variation in abundance according to
1473
nested sampling designs and 2-way designs with interactions, as well as the addition of
1474
covariates, such as environmental variables (Anderson et al. 2008). Statistical procedures used for
1475
each of our research questions are detailed below.
1476
1477
Q1) The taxonomic composition and distribution of zooplankton were evaluated qualitatively
1478
using abundance data from all stations and years. For grouping stations, we applied a hierarchical
1479
cluster analysis using group average linkage of Bray-Curtis dissimilarities among stations. We
1480
then used SIMPER analysis to determine the average percent contribution of each taxa in each
1481
station cluster. For grouping taxa, we applied a separate cluster analysis of Bray-Curtis
1482
dissimilarities using mean abundance (standardized untransformed data) of each taxon across
1483
station clusters. Standardization rather than transformation was necessary for grouping species
1484
since we were more interested in relative similarity/dissimilarity patterns rather than down
1485
weighting abundant species, which was necessary when grouping stations (Clarke and Warwick,
1486
2001).Clusters were identified by drawing a line across branches on each dendrogram (one for
1487
station and one for taxa) determined by at least 50% similarity, length of branches (indicating
1488
stability of groups), a permutation procedure that objectively tests for grouping a priori
1489
(SIMPROF, Clarke and Warwick, 2001; Clarke and Gorley, 2006), and subjective biological
60
1490
interpretation of clusters. For the station cluster analysis, we designated station clusters as outliers
1491
if they contained less than 1% of the sampled stations.
1492
1493
Q2) The total abundance and total biomass of all large and all small zooplankton in the north and
1494
south eastern Bering Sea for each year were estimated by summing the abundance or biomass
1495
over all taxa within each block and dividing by number of stations. The average abundance or
1496
biomass was computed over all blocks within each region, i.e. the south (< 60 °N, Inner, Middle
1497
and Outer Domains) and north (60 – 63 °N, Inner and Middle Domains). Then, we estimated the
1498
mean and standard error (SE) of the 4th root transformed data, then back-transformed the data to
1499
estimate geometric means and SE of untransformed total abundance and biomass.
1500
1501
Q3a) To test for significant differences in zooplankton community composition between Regimes
1502
(warm vs. cold), a PERMANOVA was applied to all domains combined and then to each domain
1503
separately. In addition to the Regime effect, a geographic blocking factor (Block) was included as
1504
a random effect to account for the uneven sampling across years (defined above). Year was also
1505
included in the model as a random nested effect within Regime. Interactions among these effects
1506
were also included. A permutational dispersion (PERMDISP) analysis was applied to test for any
1507
significant dispersion effects between Regimes (Anderson et al. 2008). This is analogous to a
1508
univariate Levene test for homogeneity of variances in ANOVA and tests for differences in
1509
distances between the yearly centroids (center of points in multivariate space).
1510
1511
Q3b) A 2-way (Regime and Block) SIMPER analysis was run to determine which zooplankton
1512
taxa contributed most to the differences between regimes for domains where significant
1513
differences were found between Regimes.
1514
61
1515
Q3c) A PERMANOVA was applied to each domain to determine if zooplankton communities
1516
were significantly different for one or more years. In this case, Year was considered a fixed
1517
effect, and Block was considered a random effect accounting for spatial variability and the
1518
uneven sampling across years. To determine which consecutive years had significantly different
1519
community composition, pairwise multiple comparison tests were applied to consecutive years
1520
only. No Bonferroni correction was applied here as PERMANOVA provides exact tests, and a
1521
Bonferroni correction would be overly conservative in this case (Anderson et al., 2008).
1522
1523
Q3d) The mean untransformed zooplankton abundances for the most common taxa (ones that
1524
jointly contributed to ≥ 97% of the total untransformed abundance for all years combined) were
1525
calculated for the north and southeastern Bering Sea to determine which taxa contributed to the
1526
interannual variations in community structure and total abundance.
1527
1528
Q4) A set of local environmental variables measured concurrently at the same stations as
1529
zooplankton abundance and regional variables measured yearly over the eastern Bering Sea were
1530
used to evaluate the conditions most often associated with variations in community composition
1531
(dependent variable), for warm vs. cold Regimes within each shelf domain. Variables were
1532
removed if they were correlated with other variables (R > 0.8) or were determined to be of lesser
1533
biological importance. The variables of interest were included as covariates (independent
1534
variables) in PERMANOVA tests for domains where significant differences were found between
1535
warm and cold years (based on results of Question 3a). The local covariates included temperature,
1536
Chla (a measure of phytoplankton biomass), nitrate, and ammonium measured at the surface (5
1537
m) or above the mixed layer depth; stability over the top 70 m; integrated Chla and mean large
1538
size-fractionated Chla/total Chla over the top 50 m; day of ice retreat in spring prior to our late
1539
summer/ fall sampling; and temperature, salinity, and ammonium measured at the bottom or
1540
below the mixed layer depth. The regional covariate tested was the yearly winter (October-April)
62
1541
wind direction (% of time winds were from the SE and NW; Danielson et al., 2012). We also
1542
tested latitude and longitude to see if there were significant interannual geographic gradients
1543
within oceanographic domains.
1544
1545
We first removed the within year
1546
spatial variability by including a
-------------------------Middle Shelf ------------------------------------Inner Shelf-----------------Outer Shelf ---
Large Zooplankton Clusters of Stations
Eukrohnia hamata
Calanus pacificus
Neocalanus spp.
Eucalanus bungii
Metridia pacifica
Ctenophora
Clione limacina
Caridea
Epilabidocera
amphitrites
Anomura
Mysida
Limacina helicina
Brachyura
Hyperiidae
Gammaridae
Cumacea
Cnidaria
Parasagitta elegans
Calanus spp.
Appendicularia
predominant regime
L1
L2
L3
L4
L5
L6
L7
L8
W
W/C
W
C
W/C
C
W/C
C
blocking factor (described
1548
previously). Then, each covariate
1549
was sequentially added using
1550
Type I sum of squares which
1551
tests for the significance of the
1552
addition of that variable with all
1553
prior covariates accounted for.
1554
The best fit model was chosen
1555
by including the most significant
1556
covariates in a particular order
1557
that gave the lowest mean square
1558
error. The Regime effect and
1559
interactions were included as the
1560
last covariates in the model. Whenever a Regime effect (added last) became insignificant (P >
1561
0.05) compared to the Regime effect before adding environmental covariates (P < 0.05), we
1562
concluded that the environmental covariates were explaining at least some of the variability in
1563
community composition between warm and cold years. To examine yearly changes in
1564
temperature and sea ice variations we plotted surface and bottom temperature, day of sea ice
1565
retreat and number of days of ice coverage the prior winter, averaged over stations in the N and S
1566
Middle Domains.
Large Zooplankton Clusters of Taxa
1547
Abundance (no. m¯³)0.25
=0
0.1 - 0.5
0.5 - 1.0
1.0 - 2.0
2.0 - 3.0
>3
Figure 8. Heat map of large zooplankton showing clusters of
stations (x axis) by taxa (y axis) using all stations sampled. Shade of
each cell indicates the 4th root transformed abundance of each
taxon. Approximate domain designations are shown above clusters.
Designations at bottom indicate climate regime (W = warm; C =
cold; W/C = both) when each cluster was commonly observed.
Horizontal lines separate clusters of species (indicating taxa which
tended to co-occur over all years and domains combined).
63
1567
1568
3.0 Results:
1569
3.1 Spatial and temporal patterns of large and small zooplankton (Q1).
1570
The clusters of stations (station clusters) show which stations had similar taxonomic
1571
compositions over the eastern Bering Sea shelf for all years combined. For large
1572
1573
1574
1575
Figure 3. Spatial variation of large zooplankton clusters for each year using 4th root
transformed abundance data. Warm years highlighted with red and cold with blue. Refer
to Fig. 2 to see taxa abundances for each cluster. Intended for color reproduction on the
Web and in print
64
1576
1577
zooplankton, the multivariate analyses yielded eight station clusters (Fig. 2, 3). The taxa that
1578
made up approximately 80% of the total abundance are listed from high to low for the clusters
1579
described below. One pattern that stands out is the replacement of Middle Domain clusters during
1580
the transition from warm to cold years (Fig. 3). Cluster L3 characterized by Cnidaria (gelatinous
1581
zooplankton dominated by Aglantha digitale in our study), Parasagitta elegans (Chaetognatha)
1582
and Appendicularia (Larvacea) extended over a broad area in the warm years, particularly 2003 –
1583
2004, and was replaced in cold years by Cluster L4 characterized by Calanus spp.2 (Copepoda),
1584
Cnidaria and P. elegans. In addition, in warm years, there was a cluster confined primarily to the
1585
Inner Domain, Cluster L1, characterized by P. elegans, Cnidaria, Epilabidocera amphitrites
1586
(Copepoda) and Caridea (shrimp), but in cold years this group was observed less often.
1587
Latitudinal variations in the timing of changes were also seen, with increased coverage for
1588
Cluster L4 in the northeastern Bering Sea in 2004 – 2005, and expansion southward and inshore
1589
into the Inner Domain in the following cold years. In contrast, in the Outer Domain, Cluster L8
1590
characterized by Limacina helicina (Pteropoda), Calanus spp., Eucalanus bungii (Copepoda), and
1591
Metridia pacifica (Copepoda), was observed during all years.
1592
1593
For small zooplankton, the analyses yielded nine station clusters (Fig. 4 and 5). Similar to large
1594
taxa, in the Middle Domain there was one cluster with broad spatial coverage in warm years,
1595
Cluster S4, characterized by Oithona spp.(Copepoda), Pseudocalanus spp.(Copepoda),
1596
Echinodermata larvae, Polychaeta (annelid worms), molluscan Bivalvia larvae and Cirripedia
1597
(barnacle larvae), which was replaced in cold years by Cluster S7, characterized by
1598
2
Calanus spp. referred to in this work is most likely a mixture of C. marshallae and C. glacialis based on
recent genetic analyses by Campbell et al., personal comm. and J. Nelson (2009).
65
Pseudocalanus spp., Oithona spp.,
1600
Acartia spp. (Copepoda) and
1601
Polychaeta (Fig. 5), with fewer taxa
1602
making up the bulk of abundance in
1603
cold years. In the Inner Domain, a
1604
cluster with broad coverage in the
1605
southeastern Bering Sea, Cluster
1606
S1, characterized by Oithona spp.,
----------- Middle Shelf ---------------------------------- Inner Shelf -------------------- Outer Shelf
Small Zooplankton Clusters of Taxa
1599
Metridia spp. copepodites
Globigerina sp.
Tortanus discaudatus
Eurytemora herdmani
Podon sp.
Evadne sp.
Harpacticoida
Cirripedia
Echinodermata
Microcalanus pygmaeus
Centropages abdominalis
Bivalvia
Pseudocalanus spp.
Acartia spp.
Oithona spp.
Polychaeta
Fritillaria sp.
Scolecithricella sp.
predominant regime
S1
Small Zooplankton Clusters of Stations
S2
S3
S4
S5
S6
S7
S8
W
W
C
W
W
C
C
S9
W/C W/C
Abundance (no. m¯³)0.25
=0
0-1
1-2
2-4
4-6
>6
1607
Pseudocalanus spp., larval
1608
Bivalvia, Centropages abdominalis
1609
(Copepoda), Polychaeta, Acartia
1610
spp., Echinodermata larvae and Podon sp. (Cladocera) was partially replaced in the coldest years,
1611
2008 – 2009, by Cluster S6, characterized by Pseudocalanus spp. and Acartia spp. and C.
1612
abdominalis and by Cluster S7. Again, similar to the large taxa, in the Outer Domain there was
1613
one cluster common during all years, Cluster S9, characterized by Oithona spp., Pseudocalanus
1614
spp., Metridia spp. copepodites, Acartia spp., Microcalanus spp. (Copepoda), and Bivalvia
1615
larvae. Small taxa clusters did not show large north – south variations.
Figure 4. Heat map of small zooplankton clusters of stations
(x axis) by taxa (y axis). Refer to Fig. 2 for details.
1616
66
1617
Clusters of co-occurring taxa (taxa clusters)
1618
show how taxa grouped together over all
1619
years and spatial areas combined (Fig. 2, 4,
1620
y-axis). Large zooplankton taxa that tended
1621
to co-occur included Neocalanus spp., E.
1622
bungii, M. pacifica; these are oceanic
1623
copepods resident in the Outer Domain or
1624
brought in from beyond the shelf break
1625
(Gibson et al. 2013). Other taxa that co-
1626
occurred were P. elegans, Calanus spp., and
1627
Appendicularia; all are common taxa with
1628
relatively high abundances in the Middle
1629
Domain (Coyle et al. 2008, 2011). Small taxa
1630
that tended to co-occur included Evadne sp.
1631
(Cladocera), Podon sp., Eurytemora
1632
herdmani (Copepoda), and Tortanus
1633
discaudatus (Copepoda), taxa typically
1634
neritic and common nearshore (Gieskes 1971, Johnson 1934, Kos 1977); and Oithona spp.,
1635
Acartia spp., Pseudocalanus spp., Polychaeta, Bivalvia larvae, and C. abdominalis, taxa that
1636
often reside in upper layers of the water column (Marlowe and Miller, 1975).
Figure 5. Spatial variation of small zooplankton clusters by
year using 4th root transformed abundance data. Warm
years highlighted with red and cold with blue. Refer to Fig. 4
to see taxa abundances for each cluster. Intended for color
reproduction on the Web and in print.
1637
1638
3.2 Total abundance and total biomass of large and small zooplankton during warm and cold
1639
regimes (Q2).
1640
Trends in total abundance and biomass varied for large and small zooplankton and between the
1641
north and south eastern Bering Sea (Fig. 6). In the south, large zooplankton increased in
1642
abundance starting in 2007, peaked in 2008 and declined in 2009; in contrast, small zooplankton
67
1643
steadily decreased in abundance
1644
from 2003 – 2009 (Fig. 6A). In the
1645
north, total abundances of large
1646
zooplankton increased between 2004
1647
and 2005 and decreased between
1648
2007 and 2009; small zooplankton
1649
abundances did not show clear
1650
trends with time, and were only
1651
noticeably higher in 2004 than in
1652
other years (Fig. 6B). The overall
1653
magnitude of large zooplankton abundances was similar in the north and south, however, for
1654
small zooplankton, abundances were lower in the north compared to the south. Trends in biomass
1655
mirrored those in abundance except for small taxa in the south where despite the observed
1656
declines in abundance from 2003 to 2009, the biomass remained constant (Fig. 6C, D). It is also
1657
worth noting that in spite of high abundance of small taxa during the warm years, their biomass
1658
was similar to that of large taxa during the warm period in the south and during all years in the
1659
north (Fig. 6).
Figure 6. Small and large zooplankton total mean
abundance (no. m-3) in the eastern Bering Sea in the A)
south (< ~60 N) and B) north (~60-63 N), and total mean
biomass (g m-3) in the C) south and D) north. Geometric
means and standard errors are shown.
1660
1661
3.3 Statistical differences among assemblages, shelf domains, and years (Q3)
1662
3.3.1 Zooplankton abundances between regimes by domain (Q3a)
1663
We next evaluated if the observed differences in zooplankton assemblages between regimes were
1664
statistically significant for the five major shelf domains. For all domains combined,
1665
PERMANOVAs showed significant interactions between Regime and Domain (P = 0.012, P =
1666
0.001, for large and small taxa, respectively), indicating that community structure varied by
1667
Regime depending on Domain. Therefore, each domain was evaluated separately for all further
1668
analyses. There were significant differences in the community structure between warm and cold
68
1669
regimes for all domains except the N Inner for large taxa and N Inner and S Outer for small taxa,
1670
after accounting for significant spatial block effects (Table 2). The strongest differences were
1671
observed in the S Middle and S Inner Domains.
1672
1673
3.3.2 Zooplankton taxa differences between warm and cold regimes (Q3b)
1674
Large zooplankton taxa that substantially decreased in abundance (> 2-fold) during the transition
1675
from warm to cold regimes include Cnidaria, Gammaridae amphipods, Mysida, the neritic
1676
copepod Epilabidocera amphitrites, and the temperate copepod Calanus pacificus (Table 3).
1677
Taxa that substantially increased in abundance from warm and cold regimes include Anomuran
1678
and Brachyuran crab larvae, Calanus spp., Parasagitta elegans, the oceanic copepods
1679
Neocalanus spp. Eucalanus bungii, Hyperiidea amphipods and Cumacea. Taxa that showed both
1680
increases and decreases depending on domain include Appendicularia, Caridea decapods and
1681
Limacina helicina. Calanus spp. were the top contributor to changes in community composition
1682
between regimes (13% – 23% of total dissimilarity), particularly in the Middle Domain (both
1683
north and south). High contributions to dissimilarity (> 10%) were also seen for Caridea decapods
1684
in the S Inner, P. elegans and Appendicularia in the S Middle; Limacina helicina in the S Outer,
1685
and Appendicularia and Cnidaria in the N Middle Domains.
1686
1687
Most small zooplankton taxa decreased in abundance in the S Inner, S Middle and N Middle
1688
Domain from warm to cold years (Table 3). The only small taxa showing increases in cold years
1689
were Acartia spp. in all three domains, Metridia spp. copepodites and Pseudocalanus spp. in the
1690
S Middle Domain, and Fritillaria spp. (Appendicularia) in the N Middle Domain. Oithona spp.
1691
and larval bivalves had high contributions to dissimilarity in community composition between
1692
regimes for all three domains, with Acartia spp., Polychaeta, and Echinodermata larvae also
1693
important, particularly in the Middle Domain.
1694
69
1695
3.3.3 Zooplankton abundances between consecutive years (Q3c and d)
1696
Next, we evaluated which consecutive years showed the greatest changes in community
1697
composition in each domain to determine if the differences between regimes occurred during our
1698
designated break point between warm to cold regimes (i.e. between 2005 and 2006) or if the
1699
changes in community composition
1700
took place over a period of years.
1701
We also evaluated which taxa
1702
showed the largest interannual
1703
variations in the south and north
1704
eastern Bering Sea.
1705
1706
For large and small zooplankton
1707
communities, most regions with the
1708
exception of the S Outer showed
1709
significant differences (based on
1710
pairwise multiple comparisons)
1711
between pairs of consecutive years
1712
(Table 4) indicating that interannual
1713
changes in community composition
1714
were observed over our entire study
1715
period, not just for the years when the temperature regime changed most rapidly from warm to
1716
cold (eg. 2005-2007 in south). For large zooplankton, the greatest interannual variations, in the
1717
south were seen by Calanus spp., which increased in cold years starting in 2007 (Fig. 7); while in
1718
the north, increases were due to Cnidaria in 2005 and 2007, and Calanus spp. starting in 2006.
1719
The small taxa that decreased in abundance from warm to cold years in the south were Oithona
1720
spp. and Bivalvia larvae, while in the north, Oithona spp. decreased only in 2005, 2007 and 2009
Figure 7. Untransformed mean zooplankton abundances in
the eastern Bering Sea for large taxa in the A) south (< ~60
south and D) north. Intended for color reproduction on the
Web and in print.
70
1721
(Fig. 7). Small zooplankton composition in the north was most different in 2009, with higher
1722
levels of polychaetes and Acartia
1723
spp.; no data exist for 2008 to
1724
determine if this change was
1725
gradual. Overall, L. helicina, M.
1726
pacifica, Pseudocalanus spp., and
1727
Bivalvia larvae appear to be more
1728
abundant in the south, and Cnidaria
1729
and Polychaeta more abundant in
1730
the north. Some of these north –
1731
south variations can be explained by
1732
the lack of sampling in the Outer
1733
Domain in the northeastern Bering
1734
Sea.
1735
1736
3.4 Environmental factors
1737
contributing to observed differences
1738
in zooplankton community
1739
composition (Q4)
1740
The interannual variations in
1741
zooplankton abundance and
1742
community composition appear to relate to changes in temperature and sea ice retreat in the
1743
eastern Bering Sea. Temperature above and below the mixed layer depth in the N and S Middle
1744
Domains showed a declining trend from 2003 – 2009, concurrent with increases in the date of ice
1745
retreat and duration of ice coverage (Fig. 8). The greatest difference in timing of ice retreat and
1746
ice coverage was found in the S Middle Domain, with increases seen from 2005 – 2008, which
Figure 8. A) Mean temperature above (surface) and below
(deep) the mixed layer depth during our surveys, B) mean
day of year for sea ice retreat during the prior spring, and
C) mean number of days of sea ice coverage during the
prior winter, in the south (S) and north (N) Middle
Domains. Standard errors are shown.
71
1747
overlaps the increase in Calanus spp. (Fig. 7, 8). Decreases in surface and bottom temperatures
1748
coincided with decreases in small taxa such as Oithona spp. and increases in Pseudocalanus spp.
1749
and Acartia spp. in the S Inner and S Middle Domains. Temperatures and small taxa abundance
1750
were lower in the north than the south (Fig. 7, 8). For large zooplankton taxa, models with
1751
environmental co-variates were constructed for the S Inner, S Middle, S Outer and N Middle
1752
Domains (where significant Regime effect was found). Environmental covariates were able to
1753
account for at least some of the variability in zooplankton community composition between
1754
regimes for all but the S Middle Domain based on P values for Regime with and without
1755
environmental covariates (Table 5 vs. 2). Physical and geographic covariates explained the
1756
greatest amount of variability in large zooplankton community composition (Table 5).
1757
Temperature below the mixed layer depth was important in the N and S Middle Domains
1758
although temperature above the mixed layer depth, salinity below the mixed layer depth, ice
1759
retreat timing, latitude and longitude were also highly significant (P < 0.01), depending on
1760
domain (Table 5). For small taxa, environmental covariates explained some of the variability in
1761
community composition between regimes in the S Inner, S Middle and N Middle Domains (Table
1762
2 vs. 5). Similar to large taxa, physical covariates were highly significant, with temperature above
1763
and below the mixed layer depth important for the S Inner and S Middle Domains as well as
1764
stability for the S Middle Domain (Table 5). Overall, temperature explained the greatest amount
1765
of temporal variability in community structure, and was significant in models for all domains
1766
except the S Outer for large and the N Middle for small zooplankton. Most of the biologically
1767
related covariates, nutrients (ammonium and nitrate) at surface and bottom, surface Chla,
1768
percentage of large phytoplankton, and August wind mixing, were not significant.
1769
1770
1771
1772
72
1773
4.0 Discussion:
1774
4. 1 Overview
1775
This study provides a rare look into how stanzas of warm and cold years affect the zooplankton
1776
community structure in both the north and south eastern Bering Sea and the environmental
1777
variables associated with these changes. Our analyses demonstrate the considerable cross-shelf
1778
variability in zooplankton communities as seen in prior studies (e.g. Cooney and Coyle, 1982),
1779
but also shows variability between assemblages in the north and south both in timing of changes
1780
in abundance of key taxa and total abundance and biomass. Water temperature showed the
1781
strongest relationship with large and small zooplankton community structure. We hypothesize
1782
that large taxa abundance has a direct or indirect relationship to bottom temperature (depending
1783
on taxa) and a direct relationship to the winter/spring southerly extent of sea ice; while small taxa
1784
abundance has a direct relationship to surface and bottom temperature. Ice algae and sea-ice
1785
related phytoplankton blooms in cold years may provide an early food source for large
1786
zooplankton on the southern shelf (Siger et al. this issue). In contrast, lower temperatures may
1787
reduce temperature-dependent growth (e.g. Hirst and Lampitt 1998), negatively impacting
1788
reproduction rates of small zooplankton with fewer cohorts produced during the growing season.
1789
However, temperature may also affect zooplankton survival, so multiple generations do not
1790
guarantee higher abundances. The observed changes in community composition were not abrupt,
1791
but showed gradual changes over a period of years, concordant with changes in sea ice and water
1792
temperature. Species-specific responses in cold years include both increases in omnivorous
1793
Calanus spp., Neocalanus spp., and in predators such as P. elegans, and decreases (with large
1794
interannual variability) in predatory Cnidaria (A.digitale).
1795
1796
4.2 Temporal and spatial patterns in community structure between regimes
1797
Assemblages often were clustered together based on domain with separate groups of oceanic (e.g.
1798
Neocalanus spp., Metridia pacifica, Eucalanus bungii) and neritic taxa (Podon sp., Evadne sp.,
73
1799
Polychaeta, Bivalvia larvae). These differences are attributed to, and maintained by, the presence
1800
of hydrographic fronts that separate the domains during summer and the consistency of
1801
hydrographic and biological processes that are inherent to the different domains (Iverson et al.,
1802
1979). The extension of a Middle Domain cluster into the Inner Domain in cold years for large
1803
zooplankton , suggests that the Inner Front may not be as strong (or is more leaky) in cold years,
1804
possibly resulting in increased transport across the Inner Shelf. Alternatively, the Inner Front may
1805
be established later in the season in cold than in warm years allowing more time for Middle
1806
Domain taxa to become established in the areas that eventually become the Inner Domain.
1807
Previous studies suggest the Inner Front was established later in the season during years with low
1808
sea surface temperatures (SSTs) (1998 and 1999), due to storminess and late ice retreat, than
1809
during a year with high SSTs (1997) (Kachel et al., 2002). Additionally, winter wind direction
1810
was a significant co-variate for large zooplankton in the S Inner Domain, suggesting the
1811
differences in wind fields may be related to onshore transport or timing of set up for the Inner
1812
Front (Kachel, et al. 2002, Danielson et al., 2012). Inshore – offshore changes in centers of
1813
distribution within domains are also suggested by the importance of latitude or longitude in our
1814
best-fit models for small taxa within the N and S Middle Domains and large taxa within the N
1815
Middle and S Outer Domains.
1816
1817
The recent regime shifts influenced large and small zooplankton community structure over most
1818
shelf domains, but variations were greatest in the S Middle Domain, where the largest physical
1819
changes (e.g. sea ice retreat timing, temperature) were observed (Stabeno et al., 2012a). The high
1820
significance of surface temperature (and higher abundances) for small taxa in the south but not in
1821
the north may be because in the north, the Middle Domain summer surface temperature oscillated
1822
around 8 ºC, whereas in the south, it declined from almost 12 ºC in 2004 to below 8 ºC in 2009.
1823
The surface layer may be where small taxa predominate, as found in the Gulf of Alaska (Marlowe
1824
and Miller, 1975).
74
1825
1826
North – south variations in Middle Domain zooplankton assemblages were also lower in cold
1827
years compared to a single warm year (2005) (Stabeno et al., 2012a). We found a similar result
1828
for 2005, but the other warm years (2003, 2004) showed much lower latitudinal variations in
1829
zooplankton assemblages. These differences may relate to interannual variability in the N-S
1830
strength of the temperature gradients. After the recent period of extended warming, even the north
1831
received heat so that in 2003 and 2004 the surface temperature gradients between the north and
1832
south were diminished. However, the north in 2005 was colder than in 2003 – 2004, while the
1833
south in 2005 was still warm (Danielson et al., 2011).
1834
1835
1836
4.3 Interannual variations in total abundance and biomass
1837
Total abundance and total biomass of zooplankton also varied by N – S region and regime. In the
1838
south, total abundances of large taxa were highest in cold years, and similarly total biomass for
1839
large zooplankton taxa increased 5 – 10 fold between 2005 – 2007 and 2008 – 2009. This is
1840
relevant to the feeding of small age-0 fishes, such as Walleye Pollock (Gadus chalcogramma,
1841
herein referred to as pollock). The improvement in age-0 pollock condition in cold years as
1842
evidenced by increased total energy content (kJ gm-1 wet weight) is attributed to the higher
1843
availability of large, lipid bearing zooplankton prey such as Calanus spp. and Thysanoessa spp.
1844
(Euphausiacea) in cold years (Coyle et al. 2011, Hunt et al., 2011; Heintz et al., 2013). Calanus
1845
spp., and Neocalanus spp. are also important prey for other forage fish such as juvenile Pacific
1846
salmon (Oncorhynchus spp.), Pacific Herring (Clupea pallasi) and Capelin (Mallotus villosus),
1847
planktivorous sea birds and endangered baleen whales such as the North Pacific Right Whale
1848
(Eubalaena japonica) (Russell et al., 1999; Springer et al., 2008; Sheffield-Guy et al., 2009;
1849
Coyle et al., 2011; Wade et al. 2011; Heintz et al., 2013). Therefore decreases in large lipid-rich
1850
crustacean zooplankton in warm years may be detrimental to a variety of predators, not just
75
1851
pollock. In contrast to large taxa, biomass for small zooplankton in the south eastern Bering Sea
1852
was not different between regimes, even though abundances were lower in cold than warm years.
1853
This indicates that reductions in abundances of very small taxa such as Oithona spp. were
1854
compensated for by increases in abundances of larger taxa such as Pseudocalanus spp. and
1855
Acartia spp. Future work is needed to determine its impact on trophic transfer and production at
1856
higher trophic levels.
1857
1858
The north eastern Bering Sea did not follow the same pattern as the south. In the north, the
1859
abundance of large taxa increased 4-fold between 2004 and 2005, and in 2007 reached the same
1860
levels as observed in the south a year later (2008). Increases in the north were due to both
1861
Cnidaria and Calanus spp., while in the south they were due primarily to Calanus spp. Unlike in
1862
the south, total biomass for large zooplankton remained fairly constant in the north, even though
1863
taxonomic composition changed. The abundance and biomass of small taxa in the north declined
1864
from 2004 to 2005, coincident with a drop in water temperature, and then remained at that level
1865
for the remainder of our time series.
1866
1867
Total zooplankton biomass (sum of the small and large taxa) was highest in the coldest years
1868
(2008 – 2009) in the south due to increases in large zooplankton biomass, but remained constant
1869
across all years in the north. Large and small taxa contributed equally to total biomass except in
1870
the south in 2008-2009, where the contribution to total biomass was 2 – 4 times higher for large
1871
compared to small taxa. This result for the southeastern Bering Sea was previously shown by
1872
Coyle et al. (2011), however biomass levels for the north and their relation to the south in warm
1873
and cold years is new. This suggests that as the climate warms, the total biomass available for
1874
higher trophic levels will decrease in the south, but may not change in the northeastern Bering
1875
Sea.
1876
76
1877
4.4 Proposed mechanisms driving variations in abundances of key taxa
1878
Finally, to synthesize the information on changes in abundances and the environmental factors
1879
responsible for these changes between regimes and north – south regions, we describe potential
1880
mechanisms for some of the key herbivorous and predatory zooplankton taxa: Calanus spp.,
1881
Neocalanus spp., P. elegans and A. digitale (Cnidaria).
1882
1883
Calanus spp. had the greatest impact on large zooplankton community structure between warm
1884
and cold regimes. Increases in Calanus spp. in cold years have been documented for the
1885
southeastern Bering Sea (Hunt et al., 2008; Coyle et al. 2011; Hunt et al. 2011, Stabeno et al.,
1886
2008, 2010, 2012b), but data were not available for the north. One proposed mechanism driving
1887
increases in cold years in the south is that when ice algae are available there is an early source of
1888
food for Calanus spp. reproduction (Hunt et al., 2002; Baier and Napp, 2003, Sigler et al., this
1889
issue). An early season ice-associated phytoplankton bloom can also provide a source of energy
1890
for metamorphosis and growth (Baier and Napp, 2003; Sigler et al., this issue). Lower winter
1891
water temperatures will also lower metabolic rates which allow Calanus spp. to reduce the energy
1892
required for survival. Therefore, a series of cold years may facilitate increases in spawning
1893
biomass and reproductive output resulting in a stronger year class (Sigler et al., this issue). A
1894
corollary of this is that warmer summer temperatures lead to increased food limitation due to
1895
increased metabolic requirements and reductions in net primary production (Lomas et al., 2012),
1896
which result in reductions in zooplankton lipid (energy) stores, poor overwinter survival, and
1897
fewer copepodites produced the following spring (Coyle et al., 2011).
1898
1899
In the southeastern Bering Sea, late summer abundance of Calanus spp. increased in 2007, the
1900
first year the majority of the S Middle Domain was ice covered, so the availability of an early
1901
spring algal food source, may have been important in this region. In the north, increases in ice
1902
coverage and retreat were seen from 2003 – 2006, but had a smaller range of variability than seen
77
1903
in the south. We do not yet understand what processes promoted increases in Calanus spp. in the
1904
northeastern Bering Sea starting in 2006; because this region was at least partially ice covered
1905
until mid-April in all years and presumably ice algae were available all years. Finally, because the
1906
Calanus spp. we sampled were a combination of Calanus marshallae and Calanus glacialis,
1907
changes in advection in the north and south of these different taxa may also be at play, and these
1908
two taxa may respond to changes in climate and temperature in different ways. C. marshallae was
1909
previously thought to be dominant in the southeastern Bering Sea, while C. glacialis dominated
1910
further northward in the Bering, Chukchi and Beaufort Seas (Frost, 1974). The highest
1911
abundances of Calanus spp. were observed in 2007 in the north and in 2008 in the south, a year
1912
after Calanus spp. began to increase in each region. This suggests that sequential years of cold
1913
conditions are advantageous to build large biomass of the species. Decreases in Calanus spp. in
1914
2009 in the current study could be due to increases in predation from Themisto libellula, an arctic
1915
hyperid amphipod (Pinchuk et al., 2013).
1916
1917
Parasagitta elegans also increased in cold years with the largest abundances observed in 2007. P.
1918
elegans is a voracious predator of the early developmental stages of large and small copepods in
1919
the eastern Bering Sea (Baier and Terazaki, 2005), so this increase may be a response to increases
1920
in prey availability. On the south eastern shelf for 1995 – 1999, the most frequently consumed
1921
prey of P. elegans was Calanus nauplii in cold icy years when Calanus spp. were in higher
1922
abundance, with Pseudocalanus spp. predominantly consumed in relatively warmer years (Baier
1923
and Terazaki, 2005).
1924
1925
Neocalanus spp. overwinter off the shelf at depths of 200-2000 m (e.g. Miller et al., 1984). The
1926
newly hatched nauplii and copepodites ascend to the upper layer in late winter to early spring,
1927
where they complete their development by early summer (Vidal and Smith, 1986), and start to
1928
descend to deeper layers. Recent modeling suggests that wind is a primary factor controlling
78
1929
inter-annual variability in on-shelf transport of Neocalanus spp. (Gibson et al, 2013).
1930
Southeasterly winds, which prevailed during warm years, enhanced the on-shelf transport of
1931
oceanic zooplankton over the southern shelf, while northwesterly winds, more common in cold
1932
years, reduced onshore transport. This suggests that Neocalanus spp. abundance would be lower
1933
during cold years since it would be more difficult for this taxa to move onto the southern shelf in
1934
spring (Danielson et al 2012, Gibson et al. 2013). However, Neocalanus spp. abundance
1935
increased in cold years over the S Middle Domain in our data set and there was no significant
1936
difference in its abundance between warm and cold years over the Outer shelf. Thus, the
1937
explanation for variations in Neocalanus spp. abundance over the shelf is likely more complicated
1938
with multiple factors important (e.g. transport, food availability, predation).
1939
1940
Aglantha digitale, a small (10-40 mm long) hydromedusan jellyfish, was the most numerous
1941
Cnidarian collected in our bongo nets from 2005-2009, making up 93% (ranged 87 – 97% by
1942
year) of the total Cnidarian abundance. A digitale is an ambush predator capable of retaining
1943
small, motile prey (Costello and Colin, 2002). This taxon has been shown to ingest adults and
1944
larvae of small copepods (Oithona spp., Acartia spp., Pseudocalanus spp., Temora longicornis),
1945
Evadne sp., Oikopleura spp., zooplankton eggs, veligers, tintinnids and dinoflagellates (Matsakis
1946
and Conover, 1991; Pages et al., 1996; Costello and Colin, 2002) with an estimated maximum
1947
prey size of 1.5 mm (Prudkovsky, 2006). Matsakis and Conover (1991) found that medusan
1948
carbon weight and total zooplankton carbon biomass varied reciprocally in studies in Bedford
1949
Basin, Nova Scotia. For the current study, in the northeastern Bering Sea with the exception of
1950
2009, small copepod abundances tended to be higher when Cnidaria (A. digitale primarily)
1951
abundances were lower and vice versa (Fig. 7). Thus, predation by A. digitale on small copepods
1952
may be one factor to help explain these interannual changes in abundances. Higher numbers of
1953
Cnidarians in the northeastern Bering Sea may also partially account for lower numbers of small
1954
taxa in the north compared to the south. Average abundances of A. digitale and total copepods,
79
1955
respectively, were 42.5 m-3 and 3800 m-3 in the north and 4.3 m-3 and 6600 m-3 in the south, for
1956
2005 – 2009 for the Inner and Middle Domains combined. If A. digitale consumes an average of
1957
13.9 copepod prey per day (based on feeding experiments at 4°C , Matsakis and Conover, 1991),
1958
then for the north and south respectively, they can consume 15% and 1% of the total copepod
1959
standing stock per day. In other studies, gelatinous zooplankton often consume < 10% of prey
1960
resources, but can consume over 50% during episodic events (Matsakis and Conover, 1991 and
1961
references there in; Pages et al., 1996). We also found that A. digitale abundances were 63% and
1962
72% higher in a warm year (2005) than in cold years (2006 – 2009) for the north and southeastern
1963
Bering Sea, respectively.
1964
1965
5.0 Conclusions
1966
Changes in eastern Bering Sea zooplankton community composition in response to climate
1967
change will vary by oceanographic domain and by latitude, with larger effects predicted for the
1968
south compared to the northeastern Bering Sea. This impact will be stronger on large, than on
1969
small zooplankton community structure. Reductions in sea ice coverage and changes in ice retreat
1970
timing and subsequent decreases in food resources in the spring, increases in temperature and
1971
basal metabolic requirements, and potential changes in advection may be detrimental for large
1972
crustacean taxa such as Calanus spp. and Neocalanus spp. The negative impact of warming on
1973
large, lipid-rich zooplankton taxa is expected to propagate through the food web affecting
1974
commercial fisheries, seabirds, and protected species. In the eastern Bering Sea, statistical
1975
relationships between water temperature and recruitment of pollock have been used in
1976
combination with accepted climate change scenarios to project a future decrease in one of the
1977
world’s largest fisheries (Ianelli et al., 2011 and Mueter et al., 2011). Our work helps to elucidate
1978
the mechanisms behind that statistical relationship.
1979
80
1980
Alterations in the composition of small zooplankton assemblages from climate change may also
1981
contribute to anticipated effects at higher trophic levels. In the eastern Bering Sea, Oithona spp.,
1982
dominant during the warm regime, decreased in abundance in the cold regime, while
1983
Pseudocalanus and Acartia spp. increased. This species replacement had no net effect on the total
1984
biomass, but we speculate that it did affect trophic pathways, as Pseudocalanus and Acartia are
1985
considered more favorable as food for other planktivores (e.g. Oithona spp. nauplii are under-
1986
represented in larval pollock stomachs). Accordingly under future warming scenarios, global food
1987
webs, which include humans, (sensu Rice et al., 2011) may be impacted to a higher degree in
1988
regions that become ice-free compared to those which retain seasonal ice coverage.
1989
1990
Acknowledgements:
1991
We thank the captain and crews of the NOAA ship Oscar Dyson, and charter vessels, Sea Storm,
1992
NW Explorer, and Epic Explorer, for their years of hard work and diligence on our surveys. We
1993
are grateful for the assistance in field sampling, data processing and analysis from NOAA
1994
scientific staff and volunteers. In particular we’d like to thank K. Cieciel, J. Gann and F. Van
1995
Tulder for sampling, processing and database maintenance of oceanographic data, M. Auburn-
1996
Cook, M. Courtney, A. Feldman, E. Farley, J. Murphy and B. Wing for the many years of survey
1997
participation and assistance with zooplankton collection and processing. We are deeply
1998
appreciative of the TINRO scientists, in particular N. Kusnetsova and A.Volkov, who conducted
1999
the onboard taxonomic analysis of Juday samples. We thank K. Coyle and his lab for assistance
2000
and guidance regarding zooplankton sampling and analysis, as well as the Polish Plankton Sorting
2001
and Identification Center for their contributions to the early part of this time series. K. Coyle also
2002
provided insightful comments on an earlier version of this manuscript. We greatly appreciate the
2003
helpful comments from the three anonymous reviewers and the journal editor, C. Ashjion.
2004
Funding was provided by the North Pacific Research Board, Bering Sea Fisherman’s Association,
2005
Arctic-Yukon-Kuskokwim-Sustainable-Salmon-Initiative, and NOAA National Marine Fisheries
81
2006
Service including the Fisheries and the Environment (FATE) and the North Pacific Climate
2007
Regimes and Ecosystem Productivity (NPCREP) programs. This is contribution EcoFOCI-0807
2008
to NOAA’s Ecosystems Fisheries Oceanography Coordinated Investigations, contribution #####
2009
from BEST/BSIERP.
2010
References:
2011
Anderson, M.J., Gorley, R.N., Clarke, K.R., 2008. PERMANOVA+ for PRIMER: guide to
2012
software and statistical methods. PRIMER-E, Plymouth.
2013
2014
2015
Baier, C.T. Napp, J.M., 2003. Climate-induced variability in Calanus marshallae populations. J.
Plankton Res. 25, 771-782.
2016
2017
Baier, C. T.,Terazaki, M., 2005. Interannual variability in a predator-prey interaction: climate,
2018
chaetognaths and copepods in the southeastern Bering Sea. J. Plankton Res. 27(11),
2019
1113-1125.
2020
2021
2022
Bray J.R., Curtis J.T., 1957. An ordination of the upland forest communities of Southern
Wisconsin. Ecological Monographies 27:325-349
2023
2024
Cieciel, K., Farley, E.V. Jr, Eisner, L.B., 2009. Jellyfish and juvenile salmon associations with
2025
oceanographic characteristics during warm and cool years in the eastern Bering Sea.
2026
North Pacific Anadromous Fish Commission Bulletin 5, 209–224.
2027
2028
Clarke, K.R., Gorley, R.N., 2006. PRIMER v6: user manual/tutorial. PRIMER-E, Plymouth.
2029
2030
2031
Clarke, K.R., Warwick, R.M., 2001. Change in marine communities: an approach to statistical
analysis and interpretation, 2nd edn. PRIMER-E, Plymouth.
2032
2033
2034
Coachman, L.K., 1986. Circulation, water masses and fluxes on the southeastern Bering Sea
shelf. Cont. Shelf Res. 5, 23–108.
2035
2036
2037
Cooney, R.T., Coyle, K.O., 1982. Trophic implications of cross-shelf copepod distributions in the
southeastern Bering Sea. Mar. Biol. 70(2), 187-196.
82
2038
2039
2040
Costello, J.H., Colin, S.P., 2002. Prey resource use by coexistent hydromedusae from Friday
Harbor, Washington. Limnol. Oceanogr. 47(4), 934-942.
2041
2042
Coyle, K.O., Chavtur, V.G., Pinchuk, A.I., 1996. Zooplankton of the Bering Sea: A review of
2043
Russian-language literature, pp. 97-133. In: Mathisen, O. and Coyle, K.O. (Eds.),
2044
Ecology of the Bering Sea: A review of Russian Literature, Alaska Sea Grant Program
2045
Report No. 96-01, University of Alaska, Fairbanks.
2046
2047
Coyle, K.O., Eisner, L.B., Mueter, F.J., Pinchuk, A.I., Janout, M.A., Cieciel, K.D., Farley, E.V.,
2048
Andrews, A.G., 2011. Climate change in the southeastern Bering Sea: impacts on pollock
2049
stocks and implications for the oscillating control hypothesis. Fish. Oceanogr. 20, 139–
2050
156.
2051
2052
Coyle, K.O., Pinchuk, A.I., 2002. Climate-related differences in zooplankton density and growth
2053
on the inner shelf of the southeastern Bering Sea. Prog. Oceanogr. 55, 177–194.
2054
2055
Coyle, K.O., Pinchuk, A.I., Eisner, L.B., Napp, J.M., 2008. Zooplankton species composition,
2056
abundance and biomass on the eastern Bering Sea shelf during summer: the potential role
2057
of water column stability and nutrients in structuring the zooplankton community. Deep
2058
Sea Res. II 55, 1775–1791.
2059
2060
Danielson, S., Eisner, L., Weingartner, T. and Aagaard, K., 2011. Thermal and haline variability
2061
over the central Bering Sea shelf: Seasonal and interannual perspectives. Cont. Shelf Res.
2062
31, 539-554.
2063
2064
Danielson, S., Hedstrom, K., Aagaard, K., Weingartner, T., Curchitser, E., 2012. Wind-induced
2065
reorganization of the Bering shelf circulation. Geophys. Res. Lett. 39, L08601,
2066
doi:10.1029/2012GL051231
2067
2068
2069
ESRI, 2012. ArcGIS Desktop: Release 10.1. Redlands, CA: Environmental Systems Research
Institute
2070
83
2071
Farley, E.V. ,Moss, J.H., 2009. Growth rate potential of juvenile chum salmon on the eastern
2072
Bering Sea shelf: an assessment of salmon carrying capacity. North Pacific Anadromous
2073
Fish Commission Bulletin 5, 265–277.
2074
2075
Friday, N.A., Waite, J.M., Moore, S.E., Zerbini, A.N., 2012. Cetacean distribution and abundance
2076
in relation to oceanographic domains on the eastern Bering Sea shelf: 1999–2004. Deep
2077
Sea Res. II 65–70, 260–272.
2078
2079
2080
Frost, B.W., 1974. Calanus marshallae, a new species of Calanoid copepod closely allied to the
sibling species C. finmarchicus and C. glacialis. Marine Biology 26, 77-99.
2081
2082
Gibson, G. A., Coyle, K.O., Hedstrom, K, Curchitser, E. N. 2013. A modeling study to explore
2083
on-shelf transport of oceanic zooplankton in the Eastern Bering Sea. Journal of Marine
2084
Systems. 121, 47-64
2085
2086
2087
Gieskes, W.W.C., 1971. Ecology of the cladocera of the North Atlantic and the North Sea.
Netherlands Journal of Sea Research. 5(3), 342-376.
2088
2089
2090
Gold, C.M., 1991. Problems with handling spatial data--the Voronoi approach. CISM Journal 45,
65-80
2091
2092
Gordon, L.I., Jennings, J.C. Jr., Ross, A.A., Krest, J.M., 1993. A suggested protocol for
2093
continuous flow automated analysis of seawater nutrients (phosphate, nitrate, nitrite and
2094
silicic acid) in the WOCE Hydrographic Program and the Joint Global Ocean Fluxes
2095
Study. WOCE Hydrographic Program Office, Methods Manual WHPO 91-1, Oregon
2096
State University, Corvallis.
2097
2098
Grebmeier, J.M., Overland, J.E., Moore, S.E., Farley, E.V., Carmack, E.C., Cooper, L.W., Frey,
2099
K.E., Helle, J.H., McLaughlin, F.A., McNutt, S.L., 2006. A major ecosystem shift in the
2100
northern Bering Sea. 311(5766), 1461-4.
2101
2102
Heintz, R.A., Siddon, E.C., Farley, E.V. Jr., Napp, J.M., 2013. Climate-related changes in the
2103
nutritional condition of young-of-the-year walleye pollock (Theragra chalcogramma)
2104
from the eastern Bering Sea. Deep Sea Res. II BEST-BSIERP Special Issue 2.
84
2105
2106
2107
Hirst, A.G., Lampitt, R.S., 1998. Towards a global model of in situ weight specific growth in
marine planktonic copepods. Mar. Biol. 132, 247-257.
2108
2109
Hollowed, A.B., Barbeaux, S.J., Cokelet, E.D., Farley, E., Kotwicki, S., Ressler, P.H., Spital, C.,
2110
Wilson, C.D., 2012. Effects of climate variations on pelagic ocean habitats and their role
2111
in structuring forage fish distributions in the Bering Sea. Deep Sea Res. II, 65-70, 230-
2112
250.
2113
2114
Hunt, G.L. Jr., Coyle, K.O., Eisner, L.B., Farley, E.V., Heintz, R., Mueter, F., Napp, J.M.,
2115
Overland, J.E., Ressler,P.H., Salo, S., Stabeno, P.J., 2011. Climate impacts on eastern
2116
Bering Sea food webs: A synthesis of new data and an assessment of the Oscillating
2117
Control Hypothesis. ICES J. Mar. Sci. 68, 1230-1243.
2118
2119
Hunt, G.L., Jr., Stabeno, P.J., Walters, G., Sinclair, E., Brodeur, R.D., Napp, J.M., Bond, N.A.,
2120
2002. Climate change and control of the southeastern Bering Sea pelagic ecosystem.
2121
Deep Sea Res. II 49, 5821-5853.
2122
2123
Hunt, G.L., Stabeno, P.J., Strom, S., and Napp, J.M., 2008. Patterns of spatial and temporal
2124
variation in the marine ecosystem of the southeastern Bering Sea, with special reference
2125
to the Pribilof Domain. Deep Sea Res. II, 55, 1919–1944.
2126
2127
Ianelli, J.N, Hollowed, A.B., Haynie A.C., Mueter, F.J., Bond, N.A., 2011. Evaluating
2128
management strategies for eastern Bering Sea walleye pollock (Theragra chalcogramma)
2129
in a changing environment. ICES J. Mar. Sci. 68(6), 1297-1304.
2130
2131
Iverson, R.L., Coachman, L.K., Cooney, R.T., English, T.S., Goering, J.J., Hunt, Jr., G.L.,
2132
McCauley, M.C., McRoy, C.P., Reeburg, W.S., Whitledge, T.E., 1979. Ecological
2133
significance of fronts in the southeastern Bering Sea, pp. 437-466. In: Ecological
2134
Processes in Coastal and Marine Systems, R.J. Livingston (Ed.), Plenum Press.
2135
2136
2137
Johnson, M., 1934. The life history of the copepod Tortanus discaudatus (Thomspon and Scott).
The Biological Bulletin, Woods Hole 67, 182-200.
2138
85
2139
Kachel, N.B., Hunt Jr., G.L., Salo, S.A., Schumacher, J.D., Stabeno, P.J., Whitledge, T.E., 2002.
2140
Characteristics and variability of the inner front of the southeastern Bering Sea. Deep Sea
2141
Res. II 49 (26), 5889–5909.
2142
2143
Kos, M.S., 1977. Species of the genus Eurytemora (Copepoda, Calanoida) in the northern part of
2144
The Pacific Ocean: systematics, distribution, variability. Explorations of the Fauna of the
2145
Seas 20, 20-53.
2146
2147
Kotwicki, S. Lauth,R.R., 2013. Detecting temporal trends and environmentally-driven changes in
2148
the spatial distribution of groundfishes and crabs on the eastern Bering Sea shelf. Deep
2149
Sea Res. II.
2150
2151
Lomas, M.W., Moran, S.B., Casey, J.R., Bell, D.W., Tiahlo, M., Whitefield, J., Kelly, R.P.,
2152
Mathis, J.T., Cokelet, E.D., 2012. Spatial and seasonal variability of primary production
2153
on the Eastern Bering Sea shelf. Deep Sea Res II 65-70, 126-140.
2154
2155
2156
Marlowe, C.J., Miller, C.B., 1975. Patterns of vertical distribution and migration of zooplankton
at ocean Station “P”. Limnol. Oceanogr. 20(5), 824-844.
2157
2158
Matsakis, S., Conover, R. J., 1991. Abundance and feeding of medusae and their potential impact
2159
as predators on other zooplankton in Bedford Basin (Nova Scotia, Canada) during spring.
2160
Can. J. Fish. Aquat. Sci. 48(8), 1419-1430.
2161
2162
Miller, C. B., Frost, B. W., Batchelder, H. P., Clemons, M. J., Conway, R. E., 1984. Life histories
2163
of large, grazing copepods in a subarctic ocean gyre: Neocalanus plumchrus, Neocalanus
2164
cristatus and Eucalanus bungii in the northeast Pacific. Prog. Oceanogr. 13, 201-243.
2165
2166
Motoda, S., Minoda, T., 1974. Plankton of the Bering Sea, pp. 207-241. In: D.W. Hood and E.J.
2167
Kelley, (Eds), Oceanography of the Bering Sea with Emphasis on Renewable Resources.
2168
University of Alaska Press.
2169
2170
Mueter, F. J., Bond, N.A., Ianelli, J.N., Hollowed, A.B., 2011. Expected declines in recruitment
2171
of walleye pollock (Theragra chalcogramma) in the eastern Bering Sea under future
2172
climate change. ICES J. Mar. Sci. 68(6), 1284-1296.
86
2173
2174
Napp, J.M., Baier, C.T., Brodeur, R.D., Coyle, K.O., Shiga, N., Mier, K., 2002. Interannual and
2175
decadal variability in zooplankton communities of the south- east Bering Sea shelf. Deep
2176
Sea Res. II 49, 5991–6008.
2177
2178
2179
Napp, J.M., Hunt Jr., G.L., 2001. Anomalous conditions in the southeastern Bering Sea, 1997:
linkages among climate, weather, ocean, and biology. Fish. Oceanogr. 10, 61–68.
2180
2181
Ortiz, I,. Weise, F., Greig, A., 2012. Marine Regions Boundary Data for the Bering Sea Shelf and
2182
Slope. UCAR/NCAR - Earth Observing Laboratory/Computing, Data, and Software
2183
Facility. Dataset. doi:10.5065/D6DF6P6C
2184
2185
Pages, F., Gonzalez, H.E., Gonzalez, S.R., 1996. Diet of the gelatinous zooplankton in
2186
Hardangerfjord (Norway) and potential predatory impact by Aglantha digitale
2187
(trachymedusae). Mar. Ecol. Prog. Ser. 139(1-3), 69-77.
2188
2189
2190
Parsons, T.R., Maita, Y., Lalli, C.M., 1984. A Manual of Chemical and Biological Methods for
Seawater Analysis. Pergamon Press, Oxford, England, 173 pp.
2191
2192
2193
Piatt, J.F., Springer, A.M., 2003. Advection, pelagic food webs and the biogeography of seabirds
in Beringia. Mar. Ornithol. 31, 141–154.
2194
2195
Pinchuk, A.I., Coyle, K.O., Farley, E.V., Renner, H.M., 2013. Emergence of the Arctic Themisto
2196
libellula (Amphipoda: Hyperiidae) on the southeastern Bering Sea shelf as a result of the
2197
recent cooling and their potential impact on pelagic food web. ICES J. Mar. Sci.
2198
DOI######
2199
2200
2201
Prudkovsky, A.A., 2006. Food selectivity in Aglantha digitale (Hydrozoa, Trachymedusae) in the
White Sea. Zoologicheskii Zhurnal 85(11), 1294-1302.
2202
2203
Rice, J.C., Garcia, S.M., 2011. Fisheries, food security, climate change, and biodiversity:
2204
characteristics of the sector and perspectives on emerging issues. ICES J. Mar. Sci. 68,
2205
1343-1353.
2206
87
2207
Russell, R.W., Harrison, N.M., Hunt, G. J., 1999. Foraging at a front: Hydrography, zooplankton,
2208
and avian planktivory in the northern Bering Sea. Mar. Ecol. Prog. Ser.182, 77-93.
2209
2210
Sameoto, D., Cochrane, N. and Herman, A., 1993. Convergence of acoustic, optical, and net-
2211
catch estimates of euphausiid abundance: Use of artificial light to reduce net avoidance.
2212
Can. J. Fish. Aquat. Sci. 50, 334-346.
2213
2214
Sheffield-Guy, L., Roby, D.D., Gall, A.E., Irons, D. B., Rose, I. C., 2009. The influence of diet
2215
and ocean conditions on productivity of auklets on St Lawrence Island, Alaska. Mar.
2216
Ornithol. 37(3), 227-236.
2217
2218
Sigler, M.F. Stabeno, P., In review, this issue. Spring and fall phytoplankton blooms in a
2219
productive subarctic ecosystem, the eastern Bering Sea, during 1995-2011. Deep-Sea
2220
Res. II
2221
2222
2223
Simpson, J.H., Hughes, D.G., Morris, N.C.G., 1977. The relation of seasonal stratification to tidal
mixing on the continental shelf. Deep Sea Res. 24(Suppl.), 327–340.
2224
2225
2226
Springer, A.M., Byrd, G.V., Iverson, S.J., 2008. Hot oceanography: Planktivorous seabirds reveal
ecosystem responses to warming of the Bering Sea. Mar. Ecol. Prog. Ser. 352, 289-297.
2227
2228
Stabeno, P.J., Farley, E.V., Kachel, N., Moore, S., Mordy, C., Napp, J.M., Overland, J.E.,
2229
Pinchuk, A.I., Sigler, M., 2012a. A comparison of the physics, of the northern and
2230
southern shelves of the eastern Bering Sea and some implications to the ecosystem. Deep
2231
Sea Res. II 65-70:14-30.
2232
2233
Stabeno, P.J., Kachel, N.B., Moore, S.E., Napp, J.M., Sigler, M., Yamaguchi, A., Zerbini, A.N.,
2234
2012b. Comparison of warm and cold years on the southeastern Bering Sea shelf and
2235
some implications for the ecosystem. Deep Sea Res. II 65-70:31-45.
2236
2237
Stabeno, P. J., Kachel, N., Mordy, C., Righi, D., Dalo, S., 2008. An examination of the physical
2238
variability around the Pribilof Islands in 2004. Deep Sea Res. II 55(16-17), 1701-1716.
2239
88
2240
Stabeno, P.J., Napp, J., Mordy, C., Whitledge, T., 2010. Factors influencing physical structure
2241
and lower trophic levels of the eastern Bering Sea shelf in 2005: sea ice, tides and winds.
2242
Prog. Oceanogr. 85 (3-4), 180–196.
2243
2244
Stevenson, D.E., Lauth,R.R., 2012. Latitudinal trends and temporal shifts in the catch
2245
composition of bottom trawls conducted on the eastern Bering Sea shelf. Deep Sea Res.
2246
II, 65-70, 251-259.
2247
2248
2249
Thiessen, A. J., Alter, J.C., 1911. Precipitation Averages for Large Areas. Monthly Weather Rev.
39, 1082-84.
2250
2251
Vidal, J., Smith, S.L., 1986. Biomass, growth, and development of populations of herbivorous
2252
zooplankton in the southeastern Bering Sea during spring. Deep Sea Res. 33, 523-556.
2253
2254
2255
Volkov, A.F., 1984. Recommendations for Expedited Zooplankton Sample Processing at Sea.
Vladivostok: TINRO, 31 p (in Russian).
2256
2257
Volkov, A.F., Efimkin, A.Y. Kuznetseva, N.A., 2007. Characteristics of the plankton community
2258
of the Bering Sea and several regions of the North Pacific in 2002 – 2006. Izvestiya
2259
TINRO 151, 338–364 (in Russian).
2260
2261
Wade, P. R., De Robertis, A., Hough, K.,. Booth, R , Kennedy, A., LeDuc, R., Munger, L., Napp,
2262
J., Shelden, K.E.W., Rankin, S., Vasquez, O., Wilson., C., 2011. Rare detections of North
2263
Pacific right whales in the Gulf of Alaska, with observations of their potential prey.
2264
Endang. Species Res. 13:99-109.
2265
2266
Wiebe, P.H., Ashjian, C.J., Gallager, S.M., Davis, C.S., Lawson, G.L., and Copley, N.J., 2004.
2267
Using a high-powered strobe light to increase the catch of Antarctic krill. Mar. Biol. 144,
2268
493-502.
2269
2270
2271
Wiese, F.K., Wiseman, W.J., van Pelt, T.I., 2012. Bering Sea linkages. Deep Sea Res. II 65-70, 25.
2272
89
2273
2274
Wyllie-Echevarria, T., Wooster, W.S., 1998. Year-to-year variations in Bering Sea ice cover and
some consequences for fish distributions. Fish. Oceanogr. 7, 159–170.
2275
90
Table 1. Questions addressed in manuscript.
Q1
What are the spatial and temporal patterns of large and small zooplankton taxa in
the eastern Bering Sea for 2003–2009?
Q2
How do the total abundances and biomasses of large and small zooplankton vary
between warm and cold years for the southeastern and northeastern Bering Sea?
Q3a
Are the community compositions of large and small taxa different between warm
and cold regimes within each domain?
b
When these communities differed between warm and cold regimes which taxa
contributed most to the differences?
c
Did the observed differences in zooplankton community composition between
regimes occur exactly during the break between low and high ice years?
d
Which taxa showed the largest variations among years?
Q4
For each domain, which environmental factors explained the greatest amount of
variability in zooplankton community composition overall and between warm
and cold regimes?
91
Table 2. PERMANOVA results (P-values from permutation test) for differences in community
structure of large and small zooplankton taxa (abundance data) in the warm (2003 – 2005) versus
cold (2006 – 2009) years by domain (Fig. 1). Other effects in model are not shown (Block, Year
(Regime), and interactions). *Significant at P < 0.05.
Domain
S Inner
S Middle
S Outer
N Inner
N Middle
Large
P(perm)
0.002*
0.001*
0.008*
0.152
0.024*
Small
P(perm)
0.005*
0.001*
0.731
0.438
0.029*
92
93
Table 4. PERMANOVA results (P-values) showing differences in community composition for
large and small zooplankton taxa (abundance data) by domain. P-values for at least one difference
among years (All years) are followed by pairwise multiple comparisons for consecutive years. * P
< 0.05.
Large
zooplankton:
S Inner
S Middle
S Outer
N Inner
N Middle
All years
< 0.001*
< 0.001*
0.015*
< 0.001*
< 0.001*
2003
2004
0.005*
0.007*
0.382
0.290
0.157
2004
2005
< 0.001*
< 0.001*
0.084
0.001*
0.014*
2005
2006
< 0.001*
< 0.001*
0.364
0.026*
0.007*
2006
2007
0.001*
< 0.001*
0.764
0.023*
0.002*
2007
2008
0.108
< 0.001*
2008
2009
0.043*
zooplankton:
S Inner
S Middle
S Outer
N Inner
N Middle
All years
< 0.001*
< 0.001*
0.011*
< 0.001*
< 0.001*
0.043*
Small
2003
2004
0.003*
0.296
0.743
0.028*
0.154
2004
2005
0.001*
0.045*
0.044*
0.026*
0.003*
2005
2006
0.160
0.031*
0.149
0.715
0.020*
2006
2007
0.438
0.084
0.243
0.734
0.133
2007
2008
0.038*
0.320
0.770
2008
2009
0.087
0.057
0.204
94
Table 5. Results of best-fit PERMANOVA models using Type I (sequential) sum of squares
showing P-values of environmental covariates for large and small zooplankton community
composition. Order of inclusion in the model is shown in parentheses. An insignificant (P > 0.05)
Regime (factor) indicates covariates explained some of the variability between regimes.
*Significant at P < 0.05.
Large zooplankton
Variable
longitude
latitude
Integrated Chla
Winter wind
direction
S below MLD
T below MLD
T above MLD
Ice retreat timing
Regime (factor)
Small Zooplankton
Variable
longitude
latitude
T below MLD
T above MLD
Ice retreat timing
stability
Regime (factor)
S Inner
0.012 (1)*
0.014 (2)*
S Middle
-
S Outer
0.006 (1)*
-
N Middle
0.006 (1)*
-
0.030 (3)*
0.079
0.001 (1)*
0.001 (2)*
0.001 (3)*
0.001*
0.006 (2)*
0.383
0.001 (2)*
0.082 (3)*
0.335
S Inner
0.001 (1)*
0.001 (2)*
0.393
S Middle
0.025 (4)*
0.003 (3)*
0.001 (2)*
0.001 (1)*
0.061
N Middle
0.020 (1)*
0.042 (2)*
0.202 (3)
0.252
95
2276
Figure Captions:
2277
Figure 1. Bering Sea study area showing stations used for analyses in each oceanographic shelf
2278
domain for: A) large zooplankton and B) small zooplankton. Black lines indicate spatial “blocks”
2279
which are thiessen polygons around points. Blocks containing colored points (reduced data set)
2280
were visited at least two years during warm (2003 – 2005) and two years during cold (2006 –
2281
2009) regimes; colors signify domain.
2282
2283
Figure 2. Heat map of large zooplankton showing clusters of stations (x axis) by taxa (y axis)
2284
using all stations sampled. Shade of each cell indicates the 4th root transformed abundance of each
2285
taxon. Approximate domain designations are shown above clusters. Designations at bottom
2286
indicate climate regime (W = warm; C = cold; W/C = both) when each cluster was commonly
2287
observed. Horizontal lines separate clusters of species (indicating taxa which tended to co-occur
2288
over all years and domains combined).
2289
2290
Figure 3. Spatial variation of large zooplankton clusters for each year using 4th root transformed
2291
abundance data. Warm years highlighted with red and cold with blue. Refer to Fig. 2 to see taxa
2292
abundances for each cluster. Intended for color reproduction on the Web and in print
2293
2294
Figure 4. Heat map of small zooplankton clusters of stations (x axis) by taxa (y axis). Refer to
2295
Fig. 2 for details.
2296
2297
Figure 5. Spatial variation of small zooplankton clusters by year using 4th root transformed
2298
abundance data. Warm years highlighted with red and cold with blue. Refer to Fig. 4 to see taxa
2299
abundances for each cluster. Intended for color reproduction on the Web and in print.
2300
96
2301
Figure 6. Small and large zooplankton total mean abundance (no. m-3) in the eastern Bering Sea
2302
in the A) south (< ~60N) and B) north (~60-63 N), and total mean biomass (g m-3) in the C)
2303
south and D) north. Geometric means and standard errors are shown.
2304
2305
Figure 7. Untransformed mean zooplankton abundances in the eastern Bering Sea for large taxa
2306
in the A) south (< ~60 N) and B) north (~60-63 N) and for small taxa in the C) south and D)
2307
north. Intended for color reproduction on the Web and in print.
2308
2309
Figure 8. A) Mean temperature above (surface) and below (deep) the mixed layer depth during
2310
our surveys, B) mean day of year for sea ice retreat during the prior spring, and C) mean number
2311
of days of sea ice coverage during the prior winter, in the south (S) and north (N) Middle
2312
Domains. Standard errors are shown.
97
2313
Chapter 14
2314
2315
Sigler, M.F., P.J. Stabeno, L.B. Eisner, and J.M. Napp (2012): Spring and fall phytoplankton
2316
blooms in a productive subarctic ecosystem, the eastern Bering Sea, during 1995-2011. Deep-Sea
2317
Res. II.
2318
2319
Spring and fall phytoplankton blooms in a productive subarctic ecosystem, the eastern Bering
2320
Sea, during 1995-2011
2321
Michael F. Sigler1, Phyllis J. Stabeno2, Lisa B. Eisner1, Jeffrey M. Napp3, Franz J. Mueter4
2322
2323
1
2324
Atmospheric Administration, 17109 Pt. Lena Loop Rd., Juneau, AK 99801 USA
2325
(907) 789-6037; mike.sigler@noaa.gov and (907) 789-6602; lisa.eisner@noaa.gov
Alaska Fisheries Science Center, National Marine Fisheries Service, National Oceanic and
2326
2327
2
2328
and Atmospheric Administration, 7600 Sand Point Way NE, Seattle, WA 98115-0070 USA
2329
(206) 526-6453; phyllis.stabeno@noaa.gov
Pacific Marine Environmental Laboratory, Oceans and Atmospheric Research, National Oceanic
2330
2331
3
2332
Atmospheric Administration, 7600 Sand Point Way NE, Seattle, WA 98115-0070 USA
2333
Phone: (206) 526-4148; jeff.napp@noaa.gov
Alaska Fisheries Science Center, National Marine Fisheries Service, National Oceanic and
2334
2335
3
2336
Rd., Juneau, AK 99801 USA
2337
Phone: (907) 790-5448; fmueter@alaska.edu
University of Alaska Fairbanks, Juneau Center for Fisheries and Oceans, 17309 Pt. Lena Loop
2338
2339
2340
Abstract
The timing and magnitude of phytoplankton blooms in subarctic ecosystems often
2341 strongly influence the amount of energy that is transferred through each trophic pathway. In the
2342 eastern Bering Sea, spring bloom timing has been linked to ice retreat timing and production of
2343 zooplankton and fish. A large part of the eastern Bering Sea shelf (~500 km wide) is ice-covered
2344 during winter and spring. Four oceanographic moorings have been deployed along the 70-m
2345 depth contour of the eastern Bering Sea shelf with the southern location occupied annually since
2346 1995, the two northern locations since 2004 and the remaining location since 2001. Chlorophyll
98
2347 a fluorescence data from the four moorings provide 37 realizations of a spring bloom and 33
2348 realizations of a fall bloom. We found that in the eastern Bering Sea: if ice was present after
2349 mid-March, spring bloom timing was related to ice retreat timing (p < 0.001, df = 1,24); if ice
2350 was absent or retreats before mid-March, a spring bloom usually occurred in May or early June
2351 (average day 148, SE = 3.5, n = 11). A fall bloom also commonly occurred, usually in late
2352 September (average day 274, SE = 4.2, n = 33), and its timing was not significantly related to
2353 the timing of storms (p = 0.88, df = 1,27) or fall water column overturn (p = 0.49, df = 1,27).
2354 The magnitudes of the spring and fall blooms were correlated (p = 0.011, df = 28). The interval
2355 between the spring and fall blooms varied between four to six months depending on year and
2356 location. We present a hypothesis to explain how the large crustacean zooplankton taxa Calanus
2357 spp. likely responds to variation in the interval between blooms (spring to fall and fall to spring).
2358
2359 1. Introduction
2360
The timing and magnitude of phytoplankton blooms in subarctic ecosystems often
2361
strongly influence the amount of energy transferred through each trophic pathway. In the
2362
southeastern Bering Sea, spring bloom timing has been linked to production of large crustacean
2363
zooplankton and walleye pollock (Theragra chalcogramma) (Hunt et al., 2002, 2011; Coyle et
2364
al., 2011); if ice is present after mid-March, an early ice-associated bloom occurs there; otherwise
2365
a spring bloom usually occurs in May (Hunt et al., 2002; Stabeno et al., 2001). Although spring
2366
bloom timing is well-characterized in the southeastern part of the shelf (Brown and Arrigo, 2011,
2367
2013; Hunt et al., 2002, 2011; Rho and Whitledge 2007; Stabeno et al., 2001), less is known
2368
about the spring bloom elsewhere in the eastern Bering Sea (Brown and Arrigo 2013), as well as
2369
the characteristics of the fall bloom (Rho and Whitledge 2007).
2370
The eastern Bering Sea is dominated by a broad continental shelf (~500 km wide), a large
2371
part of which is ice-covered during winter, with the maximum extent varying >100 km among
2372
years. In ice-covered areas, the seasonal cycle of primary production begins with ice algae
2373
(primarily large diatoms), which begin to grow in the spring when light level becomes adequate.
2374
Ice algae are adapted to lower light levels than pelagic phytoplankton (Kirst and Wiencke, 1995)
2375
and grow within the ice and at the ice-water interface depending on the amount of overlying snow
2376
cover. Ice algae begin to grow in mid-February in the Bering Sea (R. Gradinger, University of
2377
Alaska, Fairbanks, pers. comm.) and may provide an early concentrated food source (chlorophyll
2378
a maximum ~300 µg l-1; Mock and Gradinger, 2000; Niemi et al., 2011) for zooplankton (e.g.
2379
O’Brien, 1987; Runge and Ingram, 1991). Sea ice algae that are not grazed can seed the spring
2380
phytoplankton bloom or may aggregate and sink out of the upper water column (Tremblay et al.,
99
2381
1989, Riebesell et al., 1991; Haecky et al., 1998; Fortier et al., 2002) with seeding more likely if
2382
the release from ice is not followed by a strong mixing event that breaks down stratification (Jin
2383
et al., 2007). Sedimentation of ice algae occurs during and immediately after the bloom starts
2384
over the Bering Sea middle shelf (R. Gradinger, University of Alaska, Fairbanks, pers. comm.).
2385
Ice algae increases within the ice typically are terminated during ice melt in late spring or early
2386
summer.
2387
Phytoplankton in the Bering Sea begin to bloom in the spring once the water becomes
2388
stratified and the day length becomes long enough (sensu Sverdrup 1953). Prior to this,
2389
phytoplankton are considered to be light-limited, but have adequate nutrients due to the advection
2390
of nutrient rich slope water onto the shelf during the previous winter, which is mixed throughout
2391
the water column (Niebauer et al., 1995); nutrient recycling on the shelf also is important
2392
(Granger et al., 2011). The phytoplankton spring bloom typically ends when the surface nutrient
2393
supply is exhausted and phytoplankton growth becomes nutrient limited (typically below 1 µM
2394
nitrate). Grazing pressure from mesozooplankton and microzooplankton also increases as the
2395
spring progresses, which lowers phytoplankton standing stocks.
2396
In the summer, phytoplankton concentration in the surface mixed layer is typically low
2397
due to nutrient limitation and continued grazing pressure. Episodic wind events can break down
2398
stratification and mix nutrients and viable phytoplankton cells to the surface during this period
2399
(Sambrotto 1986; Stabeno et al., 2010). During fall, increased storminess and overall cooling of
2400
the water column reduces stratification and deepens the mixed layer so that nutrients are mixed to
2401
the surface to fuel fall phytoplankton blooms. The fall bloom ends when phytoplankton become
2402
light limited, due to decreased day length and deepening of the mixed layer, and when
2403
zooplankton grazing reduces standing stocks. Pelagic-benthic coupling (Grebmeier et al., 2006)
2404
and luxury consumption by diatoms of nutrients near the sediment may also be important (Droop
2405
1973), since the Bering Sea shelf is a shallow shelf system with 1% light levels located 10-20 m
2406
below the pycnocline (Mordy et al., 2012). Phytoplankton present in the water column during late
2407
fall, when ice begins to form, can be incorporated into the ice with large diatom cells
2408
preferentially selected (Gradinger and Ikavalko1998, Niemi et al., 2011). Ice algal cells may grow
2409
slowly during winter when light levels are very low (Melnikov 1998) and then begin to grow
2410
faster when light increases in spring.
2411
In this paper we focus on the middle domain of the eastern Bering Sea shelf where four
2412
oceanographic moorings have been located. The measurements on the moorings include
2413
temperature and chlorophyll a fluorescence. In summer, the middle domain is strongly stratified
2414
into two layers, with a wind-mixed upper layer and tidally-mixed lower layer. The middle domain
100
2415
typically extends from the 50 m isobath to the 100 m isobath, and is bounded by oceanic fronts or
2416
transition zones (Iverson et al., 1979). In winter, the middle domain is usually well mixed and
2417
cold, with a large part (>50%) ice-covered. These four oceanographic moorings provide the
2418
longest, daily record of in situ oceanographic measurements in the eastern Bering Sea. This paper
2419
is the first examination of the chlorophyll a fluorescence data, excepting previous analyses of the
2420
spring bloom at the southernmost mooring (Hunt et al., 2002, 2011; Stabeno et al., 2001). In this
2421
paper our objectives are to: characterize spring and fall blooms over the eastern Bering Sea
2422
middle shelf; relate their timing and strength to physical characteristics including spring ice
2423
retreat and fall overturn; and discuss some implications of these results for one of the large
2424
crustacean zooplankton taxa characteristic of that domain (Calanus spp.).
2425
2426
2. Data and methods
2427
2.1. Moorings
2428
Four oceanographic moorings have been
2429
deployed along the 70-m depth contour of the eastern
2430
Bering Sea shelf with two southern locations sampled
2431
almost continually since 1995 (M2) and 1999 (M4), and
2432
two northern locations since 2004 (M8) and 2005 (M5)
2433
(ure 1). Prior to 2005, moorings were recovered and
2434
redeployed twice a year, once in the spring (April/May)
2435
and again in the late summer or early fall (September /
2436
October). Since 2006, there has been extensive ice on
2437
the northern shelf in spring and M8 (and sometimes M5)
2438
has only been recovered and redeployed once a year in
2439
August or September. Data collected by instruments on
2440
the moorings included temperature (miniature temperature recorders, SeaBird3 SBE-37 and SBE-
2441
39) and chlorophyll a fluorescence (WET Labs DLSB ECO Fluorometer). A transition to
2442
fluorometer sensors with wipers that sharply reduced fouling occurred during 2001-2004. Data
2443
were collected at least hourly and instruments were calibrated according to manufacturer’s
2444
specification prior to deployment. During autumn, winter, and early spring, the shallowest
2445
instrument was at 11 m at M2 and M4, at 15 m at M5 and at 20 m at M8. During late spring to
2446
early autumn (the ice-free period), the mooring at M2 included a surface toroid measuring
3
Figure 9. Study area and mooring
locations.
Use of trade names does not constitute an endorsement by NOAA
101
2447
temperature at a depth of 1 m, and the upper instrument at M4 was at 11 m, as was the upper
2448
instrument at M5 and M8 if a summer mooring was deployed. For consistency, our analyses
2449
focus on data recorded at 11 m (or the shallowest instrument at M5 and M8 during autumn,
2450
winter and early spring). This standardization is reasonable because for parallel data sets collected
2451
at both 1 and 11 m at M2 and M4, the data recorded at 11 m captured 96% of variability recorded
2452
at 1 m (Stabeno et al., 2007). At these mooring locations, the water column is well-mixed during
2453
winter and develops a 15-35 m wind-mixed layer once stratification develops during spring
2454
(Stabeno et al., 2007). Farther north, at M8 and possibly M5, the spring bloom can occupy a
2455
narrower depth range and slowly sink, so that identified bloom strengths and timings are
2456
approximate.
2457
Additional data were collected when the moorings were deployed and recovered and
2458
were used for quality control of the chlorophyll a fluorescence data collected by mooring
2459
instruments. The additional data included chlorophyll a fluorescence measurements made with a
2460
Seabird SBE 911 plus system with chlorophyll a fluorescence sensors (WET Labs WETStar).
2461
Data were recorded during the downcast, with a descent rate of usually 15 m min-1 to a depth of
2462
35 m, and 30 m min-1 below that. The additional data also included water samples for extracted
2463
chlorophyll a collected during CTD casts. The water samples were filtered through glass fiber
2464
filters (nominal pore size 0.7 µm), then frozen at -80 °C until analysis. Frozen chlorophyll a
2465
samples were analyzed with a calibrated benchtop Turner TD-700 fluorometer using standard
2466
acidification methods (Parsons et al., 1984).
2467
Conversion of in vivo chlorophyll a fluorescence (volts) to chlorophyll a concentration
2468
(µg l-1) was performed for both the moored and CTD chlorophyll a fluorescence sensors using
2469
relationships provided by the manufacturer for each instrument during annual service. We did not
2470
attempt to create a relation between mooring chlorophyll a fluorescence and extracted chlorophyll
2471
a from water samples due to the limited number of water samples at deployment and recovery of
2472
the moorings and the difficulty of matching the timing and location of the water samples and
2473
mooring measurements. The chlorophyll a estimates based on the fluorescence sensors were
2474
compared to the chlorophyll a water samples collected during CTD casts for quality control of the
2475
mooring-based measurements; when high (>40 µg l-1) values of chlorophyll a occurred or isolated
2476
spikes occurred, these measurements were excluded from the analysis. In addition, the mooring–
2477
based measurements also were reviewed for measurement drift due to fouling and when drift
2478
occurred, these measurements were excluded from the analysis. Typically drift did not occur
2479
during winter and was more common during summer, which was mostly eliminated for
2480
instruments with wipers.
102
2481
2482
2.2. Sea Ice
2483
Two sources of sea-ice data were used. The first source was the National Ice Center
2484
(NIC), with data available from 1972 to 2005; the second source was the Advanced Microwave
2485
Scanning Radiometer EOS (AMSR), with data available from 2002 to 2012 (Spreen et al., 2008).
2486
These two data sets provide data over the entire period (1972–2012) for which high-quality data
2487
on sea-ice extent and areal concentration are available. We downloaded data available from the
2488
NIC website (http://www.natice.noaa.gov) for our study period (1995-2005) which were
2489
interpolated to a 0.25 degree grid. NIC data are derived from a variety of sources including the
2490
Advanced Very High Resolution Radiometer (AVHRR) aboard the Polar Orbiting Environmental
2491
Satellites (POES). AMSR data consist of daily ice concentration at 12.5 km resolution, which are
2492
available from the National Snow and Ice Data Center (NSIDC; http://nsidc.org). To examine
2493
how the ice cover varies along the 70-m isobath, a 100 km by 100 km box was defined around
2494
each of our biophysical moorings (M2, M4, M5, and M8) maintained by NOAA. AMSR and NIC
2495
data overlap during the four-year period 2002-2005, during which time they have very similar
2496
values (M2, Stabeno et al., 2012b). To span the period 1972-2012, we used both NIC and AMSR
2497
data, using the average value in the overlap years to derive the annual cycle of percent ice cover
2498
for each mooring location.
2499
2500
2.3. Wind
2501
Winds were estimated using daily data from the National Centers for Environmental
2502
Prediction (NCEP)/National Center for Atmospheric Research (NCAR) Reanalysis (Kalnay et al.,
2503
1996). We follow the procedure used in Bond and Adams (2002) to specify particular elements of
2504
the atmospheric forcing on a daily basis for selected periods and interpolated wind velocity to the
2505
locations of four moorings (Fig. 1). The daily winds from the reanalysis are reliable in this region
2506
based on a comparison to independent buoy measurements from 1995 to 2000 (Ladd and Bond,
2507
2002).
2508
2509
2510
2.4. Data analysis
Ice cover for each mooring and year was examined to determine if ice was present at any
2511
time that winter or spring and if present, when the ice retreated for the last time. Ice retreat was
2512
considered to have occurred when ice cover fell below 15% for the last time during that spring.
2513
The temperature records for each mooring and year were examined to determine: 1) when the
2514
ocean began warming following ice retreat; and 2) when fall overturn occurred. When ice was
103
2515
present the temperature was approximately -1.7 °C. Warming was considered to have started
2516
when the near surface temperature rose above -1 °C for the last time that spring. Fall overturn
2517
was considered to have occurred once temperature (at 11 m) fell 2 °C below the summer
2518
maximum.
2519
We also considered basing timing
2520
of the fall overturn on timing of mixed
2521
layer depth deepening. Unfortunately, the
2522
mooring data at M5 and M8 were
2523
inadequate to do this. We tested whether
2524
timing of mixed layer depth deepening and
2525
fall temperature drop were related, which
2526
would support using fall temperature drop
2527
as a measure of stratification breakdown
2528
and the resultant nutrient replenishment
2529
that could support a fall bloom. We found
2530
that they are related (Fig. 2) (Pearson’s
2531
product-moment correlation = 0.61, df = 20, p = 0.002). The relationship was stronger for M4
2532
(Pearson’s product-moment correlation = 0.95, df = 6, p < 0.001) than M2 (Pearson’s product-
2533
moment correlation = 0.50, df = 12, p = 0.07), but the latter relationship was strongly influenced
2534
by one data point. In 1997, the mixed layer depth deepened about 40 days after the temperature
2535
dropped, which is later than usual and may have occurred because winds were weak that late
2536
summer and early fall. If this data point is excluded, the relationship for M2 is much better
2537
(Pearson’s product-moment correlation = 0.76, df = 11, p = 0.003).
Figure 10. The relationship of mixed layer depth
deepening (day) and temperature drop (day) for
moorings M2 and M4. The mooring data at M5 and M8
were inadequate to compute mixed layer deepening.
2538
The average wind speed records by calendar day and location were examined to
2539
determine when the first stormy period occurred after day 210 (ca. July 27). The wind data were
2540
divided into 10-day periods; if daily average wind was greater than 11 m s-1 for half or more of
2541
the ten-day period, then this period was defined as stormy (Sullivan et al., this issue).
2542
Fluorescence for each mooring and year was examined to determine the time and
2543
magnitude of the maximum value in spring and fall. These records were plotted and the times and
2544
magnitudes of the spring and fall blooms were assigned (Appendix Figure 1). Each year, the
2545
spring bloom was assigned to the maximum fluorescence value before day 180 (ca. June 27) and
2546
the fall bloom was assigned to the maximum fluorescence value after day 210 (ca. July 27). In
2547
some years the fluorescence record was discontinuous and the maximum value could not be
2548
determined (e.g., spring bloom at M2 in 1996. The values are tabled in Appendix Table 1.
104
2549
2550
2. 5. Comparison to shelf-wide patterns
2551
for fluorescence
2552
Satellite ocean color data were
2553
examined to understand the spatial
2554
pattern of chlorophyll a over the middle
2555
shelf. In particular, we wanted to find
2556
out whether it was reasonable to assume
2557
that the mooring fluorescence data were
2558
representative of fluorescence values
2559
over a much larger area. To do so, we
2560
retrieved eight-day composite Level-3
2561
SeaWiFS and MODIS-Aqua
2562
chlorophyll a data at 9-km spatial
2563
resolution using the Giovanni online data system, developed and maintained by the NASA
2564
Goddard Earth Sciences Data and Information Services Center (Acker and Leptoukh 2007). Data
2565
for a rectangular region (54 – 66°N and 157 – 180°W) that encompasses the eastern Bering Sea
2566
shelf were extracted from the global coverage
2567
datasets (units: mg m-3). Open ocean values
2568
(offshore of the 500 m isobaths) and nearshore
2569
values (inshore of 10 m isobaths) were excluded.
2570
We compared eight full years of overlap between
2571
the two data sets (2003-2010) based on correlations
2572
and absolute differences (bias) of the
2573
logarithmically transformed data on a pixel-by-
2574
pixel basis. Estimates from the two satellites were
2575
strongly correlated (r = 0.65) and bias was found to
2576
be negligible (median bias < 0.01, 88% of 1.4
2577
million paired observations had absolute bias <
2578
0.05). The SeaWiFS data from 1998-2002 were
2579
combined with the MODIS-Aqua data from 2003-
2580
2011 to produce the longest possible continuous
Figure 11. A boxplot of ice retreat (day) by mooring. The
box extends from the first quartile to the third quartile,
the heavy line dividing the box is the median and the
whiskers are the smallest and largest values. The number
of years with ice present were 12 (M2), 15 (M4), 17 (M5)
and 17 (M8).
Figure 4. Scatterplots of maximum spring
bloom strength (ug Chl l-1), spring bloom
timing (day), minimum winter-spring
temperature (day), day temperature rose
above -1 ºC and day ice cover fell below 15%.
If ice was absent that year, then the ice
retreat day is zero.
105
2581
time series of remotely-sensed chlorophyll a data. Finally, we correlated eight-day composites of
2582
the fluorescence data with 8-day remotely-sensed chlorophyll a data for all non-missing pairs on
2583
a pixel-by-pixel basis to produce correlation maps for each of the four moorings.
2584
2585
2.6. Zooplankton Metabolism
2586
We examined some implications of these results for a large crustacean zooplankton taxa
2587
characteristic of that domain (Calanus spp.).The effect of temperature on the respiration of late
2588
stage Calanus spp. was approximated from literature values. We used a Q10 of 2.77 and nitrogen-
2589
specific respiration of Calanus finmarchicus (130 µmol O2 gN-1 h-1; Saumweber and Durbin,
2590
2006), average bottom temperatures in cold and warm years (-1.8 and 2°C, respectively; Stabeno
2591
et al., 2012a), wet weight measurements on preserved Calanus spp. C5 obtained from the eastern
2592
Bering Sea during August and September (1.82 x 10-3 g), and literature values for the relationship
2593
of dry to wet weight (15%) and nitrogen to dry weight (8%) for copepods to estimate respiration
2594
rates.
2595
2596
3. Results
2597
3.1. Spring
2598
Ice was present for about two
2599
out of three years at M2 (12 of 17),
2600
nearly all years at M4 (15 of 17), and all
2601
years at M5 and M8 (17 of 17). When
2602
ice was present, retreat was significantly
2603
(ANOVA, p < 0.001, df = 3, 57) earlier
2604
in the south (mean M2: Day 88; M4: 99;
2605
M5: 130; M8: 138) (Fig. 3). In some
2606
years when ice was present, retreat
2607
occurred before mid-March (M2 during
2608
1998, 2000, 2002, 2006 and M4 during
2609
2000 and 2002) (Appendix Table 1). Ice
2610
was absent at M2 during 1996, 2001 and
2611
2003-2005 and at M4 during 2001 and
2612
2005. Ice presence reduced ocean
2613
temperature below -1ºC in all cases, but
2614
one (ice was present at M2 in 2002, but for only 11 days, and minimum temperature was -0.1°C).
Figure 5. Scatterplots of spring bloom maximum (day)
versus ice retreat (day) by mooring. If ice was absent
that year, then the ice retreat date is zero. The diagonal
dashed line is the 1:1 line to compare timings of spring
bloom and ice retreat. The vertical dashed line is March
15.
106
2615
The timing of warming (near-
2616
surface temperature > -1°C) was
2617
strongly correlated with the time of ice
2618
retreat (Pearson’s r = 0.96, df = 40, p <
2619
0.001). The interval between ice retreat
2620
and warming was longer in the north
2621
(M2 warming averaged Day 100; M4:
2622
110; M5: 153; M8: 181, corresponding
2623
to intervals for M2 of 12 days; M4: 11;
2624
M5: 23; M8: 43).
2625
The spring bloom usually
Figure 6. Scatterplot of the observed (number) and fitted
(line, based on simple linear regression, y-intercept = 53,
slope = 0.66, df = 1, 24, p < 0.001) values of spring bloom
maximum (day) versus ice retreat (day) for all moorings
when ice was present after March 15. The number indicates
mooring number.
2626
reached only one peak (34 of 37 cases)
2627
(Appendix Figure 1), so that the
2628
maximum value and when the
2629
maximum value occurred are reasonable measures of the strength and timing of the spring bloom.
2630
Double rather than single spring peaks occurred at M2 in 2003 and 2011 and at M4 in 1997.
2631
During years when ice was absent (or was present, but retreated before mid-March), the
2632
spring bloom maximum occurred in late May to early June (average day 148, SE = 3.5, n = 11)
2633
(Fig. 4, panel row 2, column 5). This pattern occurred for M2 and M4 but not M5 or M8 where
2634
ice always was present after mid-March
2635
(Fig. 5). There was no statistically
2636
significant difference in timing of the spring
2637
bloom maximum in years when ice was
2638
absent (average day 141) and when ice was
2639
present, but retreated before mid-March
2640
(average day 153) (two-way t-test, df = 9, p
2641
= 0.10). During years when ice was absent
2642
or retreated before mid-March, spring bloom
2643
maximum averaged 19 µg Chl l-1 (log-
2644
transformed, SE = 1.2, n = 11) (Fig. 4, panel
2645
row 1, column 5). There was no statistically
2646
significant difference in spring bloom
2647
maximum in years when ice was absent (average 25 ug Chl l-1) and when ice retreated before
2648
mid-March (average 15 ug Chl l-1) (log-transformed, two-way t-test, df = 9, p = 0.14).
Figure 7. A boxplot of fall overturn (day) by mooring.
The box extends from the first quartile to the third
quartile, the heavy line dividing the box is the
median and the whiskers are the smallest and largest
values. The numbers of years when fall overturn
could be determined were 15 (M2), 10 (M4), 16 (M5)
and 7 (M8).
107
2649
In contrast to the late May to
2650
early June timing of the spring bloom
2651
in years when ice was absent or
2652
retreated early, if ice was present after
2653
mid-March (day 75), an ice-associated
2654
bloom occurred between early April
2655
and mid-June and the bloom timing
2656
was related to ice retreat timing (Fig. 4,
2657
panel row 2, column 5), regardless of
2658
mooring (Fig. 5). Later blooms
2659
occurred when ice retreat was later
2660
(linear regression, y-intercept = 53,
2661
slope = 0.66, df = 1, 24, p < 0.001)
2662
(Fig. 6). This relationship implies that
2663
bloom day is 119 when ice retreat
2664
occurs on day 100, 152 when ice
2665
retreat occurs on day 150; and 172 when ice retreat occurs on day 180. There was no statistically
2666
significant difference in spring bloom strength whether ice was absent or present, but retreated
2667
before mid-March (average 19 µg Chl l-1) or ice was present after mid-March (average 17 µg Chl
2668
l-1) (log-transformed, two-way t-test, df = 35, p = 0.70); the overall average spring bloom strength
2669
was 17 µg Chl l-1 (log-transformed, SE = 1.2, n = 37).
Figure 8. Scatterplots of maximum fall bloom strength (ug
Chl l-1), fall bloom timing (day), maximum summer-fall
temperature (ºC), fall overturn (day) and first storm
(day).
2670
2671
3.2. Fall
2672
Fall overturn timing was similar among moorings on average (M2 mean was Day 259;
2673
M4: 261; M5: 259; M8: 268 (ANOVA, df = 3, 34, p = 0.49)), but was more variable at M2 and
2674
M4 (Fig. 7). Sometimes an early major storm prompted early fall overturn; on five occasions
2675
when overturn was earlier than usual (approximately day 240 versus a range of 250-280), the day
2676
of the first major storm was earlier than usual (about day 240 versus a range of 240-320) (Figure
2677
8, panel row 4, column 5).
108
2678
The fall bloom usually reached
2679
only one peak (30 of 33 cases) (Appendix
2680
Figure 1), so that the maximum value and
2681
when the maximum value occurred are
2682
reasonable measures of the strength and
2683
timing of the fall bloom. Double rather
2684
than single fall blooms occurred at M2
2685
once (1998) and M4 twice (2004, 2007).
2686
In six cases (1996 M4, 2002 M2, 2005
2687
M4, 2007 M2, 2009 M5, 2010 M8), only
2688
the descending limb of the fall
2689
fluorescence data is available due to
2690
instrument failure, hence the fall bloom
2691
peak is uncertain and no values for
2692
magnitude or timing were assigned.
2693
There was no significant
Figure 9. Scatterplots of fall bloom maximum (day)
versus fall overturn (day) by mooring. The diagonal
dashed line is the 1:1 line to compare timings of fall
bloom maximum and fall overturn. The horizontal
dashed line is August 15.
2694
relationship between fall overturn timing and fall bloom timing (linear regression, df = 1, 27, p =
2695
0.88) (Fig. 8, panel row 2, column 4) or fall bloom strength (linear regression, log-transform, df =
2696
1, 27, p = 0.49) (Fig. 8, panel row 1, column 4).
2697
Fall bloom timing was similar among moorings (M2 mean: Day 276; M4: 277; M5: 258;
2698
M8: 281) (ANOVA, p = 0.32, df = 3, 29) (Fig. 9). Fall bloom strength also was similar among
2699
moorings (ANOVA, log-transform, p = 0.93, df = 3, 30) (Fig. 9). On average, the fall bloom
2700
occurred on day 274 (late September) (SE = 4.2, n = 33) with strength 8 µg Chl l-1 (SE = 1.2, n =
2701
34).
2702
The strengths of the spring and fall blooms were correlated (Pearson’s r = 0.46, df = 28, p
2703
= 0.011, log-transformed values) (Fig. 10). The interval of time between the spring and fall
2704
blooms ranged from about 4-6 months, with longer intervals occurring for earlier spring blooms
2705
(linear regression, intercept = 294, slope = -1.16, df = 1, 27, p < 0.001) (Fig. 11).
2706
2707
3.3. Comparison to shelf-wide patterns
109
2708
Eight-day composites of chlorophyll a values at both M2 and M4 were moderately to
2709
strongly correlated with remotely-sensed chlorophyll a values over broad regions of the eastern
2710
shelf (Appendix Figure 2), including the middle shelf in the southeast and most of the northern
2711
Bering Sea except Norton Sound. In contrast,
2712
chlorophyll a values at M5 were positively
2713
correlated with remotely-sensed values over
2714
much of the shelf north of about 60°N, but
2715
negatively correlated with remotely-sensed
2716
values to the south of M5. Chlorophyll a
2717
values at M8 were positively correlated with
2718
remotely-sensed values over the northwestern
2719
part of the shelf and much of the outer shelf
2720
and negatively correlated with remotely-
2721
sensed values on much of the inner shelf.
2722
These negative correlations are likely due to
2723
differences in bloom timing.
Figure 10. Scatterplot of maximum spring bloom
strength (ug Chl l-1) and maximum fall bloom
strength. The number indicates mooring number.
2724
2725
4. Discussion
2726
4.1. Spring
2727
We found that in the eastern Bering
2728
Sea: if ice was present after mid-March,
2729
spring bloom timing was related to ice
2730
retreat timing; if ice was absent or retreats
2731
before mid-March, a spring bloom usually
2732
occurred in May or early June. Our findings
2733
generally match Brown and Arrigo (2013)
2734
based on their satellite-based observations
2735
(see exception below), as well as those of
2736
Hunt et al. (2002, 2011) and Stabeno et al.
2737
(2001). In general, ice-associated
2738
phytoplankton blooms are observed near the
2739
retreating ice edge on shipboard surveys and in ocean color data, due to melting ice increasing the
2740
stability of the water column (Alexander and Niebauer, 1981; Niebauer et al., 1995; Hunt et al.,
2741
2010; Brown et al., 2011). While the spring bloom usually moves northward as the eastern Bering
Figure 11. Scatterplot of observed (number) and fitted
(line, based on linear regression, intercept = 294,
slope = -1.16, df = 1, 27, p < 0.001) values for all
moorings. The y=axis is the interval between the
spring and fall bloom maximum (days). The x-axis is
the spring bloom maximum (day). The number
indicates mooring number.
110
2742
Sea becomes ice free, sometimes ice melts in the north before disappearing farther south and a
2743
spring bloom occurs in the northern ice-free area before it occurs in the south if light level is
2744
sufficient. For example in 2010, the northern Bering Sea melted earlier than in the south leaving
2745
an area of persistent ice between M2 and M4. Percent ice cover also influences whether or not
2746
ice-associated blooms occur by modulating both light and stratification (Sullivan et al., this
2747
issue).
2748
The relationship between the timing of the spring bloom and ice retreat when ice retreat
2749
occurred after mid-March was statistically significant for the eastern Bering Sea (pooled data
2750
from the north and south; Figure 6), which contradicts Brown and Arrigo (2013), who concluded
2751
that ice retreat timing was a good predictor of spring bloom timing for only the northern Bering
2752
Sea. The different conclusions for the southeastern Bering Sea likely resulted from how the two
2753
data sets were analyzed. Brown and Arrigo (2013) classified their data into four regions
2754
corresponding to M2, M4, M5 and M8 (i.e., similar geographic grouping to our approach). In one
2755
analysis, Brown and Arrigo (2013) regressed spring bloom timing versus ice retreat timing and
2756
found significant relationships for the northern, but not southeastern Bering Sea. In a second
2757
analysis, they classified the southeastern Bering Sea data based on whether or not ice retreat
2758
occurred before (early) or after (late) the mean ice retreat date for the two regions (M2 and M4)
2759
and whether the bloom occurred within (ice-edge bloom) or after (open-water bloom). Open-
2760
water blooms were defined as those occurring 20 days or more after ice retreat. They found no
2761
significant difference in mean timing between open-water blooms of early retreat (or ice free)
2762
years and ice-edge blooms of late retreat years.
2763
Like their first analysis, we regressed spring bloom timing versus ice retreat timing, but
2764
instead classified the data into two groups: ice absent or ice retreat before March 15; and ice
2765
retreat after March 15 following Hunt et al., (2002, 2011). Mid-March was chosen as the time
2766
when sufficient light becomes available for a phytoplankton bloom in the southeastern Bering Sea
2767
(Hunt et al. 2002). We pooled all mooring data for years when ice was present after mid-March
2768
(Figure 6) to test whether ice retreat and the spring bloom move northward as fson progresses.
2769
We suspect that if the Brown and Arrigo (2013) data set were similarly analyzed (regress bloom
2770
timing versus ice retreat timing when ice retreat occurred after mid-March), our results would
2771
match and both data sets would support the conclusion that the spring bloom generally moves
2772
northward as the eastern Bering Sea becomes ice free, and that the bloom occurs later when the
2773
ice retreats later.
2774
2775
The present work corrects and clarifies the relationship of spring bloom timing and ice
retreat timing reported by Hunt et al. (2002) for the southeastern Bering Sea. Their conclusion of
111
2776
late ice retreats with early ice-related blooms oversimplifies the data; while supported by early
2777
ice-related blooms for 1995, 1997 and 1999, later ice-related blooms occurred for 1975 and 1976
2778
(Hunt et al., 2002, their Figure 7). There is some evidence that there were early blooms under the
2779
ice in these two years, but the data are not very reliable. With more years of data, Hunt et al.
2780
(2011) observed that because ice can remain in the southeastern Bering Sea until May or rarely
2781
June, the calendar date of an ice-associated bloom could be later than that of an open-water
2782
bloom (i.e., if the ice retreat is late, the spring bloom also is late).
2783
In addition to ice retreat timing, wind patterns can affect spring bloom timing. When ice
2784
was absent or retreated before mid-March, spring bloom timing averaged day 148 (May 28) and
2785
varied from day 132 (May 12) to 166 (June 15). Short-lived surface pulses of chlorophyll a can
2786
occur with early ice retreat during a temporary relaxation of winds without significant
2787
stratification (Brown and Arrigo, 2013; Stabeno et al., 2010) and precede the larger primary
2788
bloom in May or early June. When ice retreat occurs after mid-March in the southeastern Bering
2789
Sea, a very early bloom can occur when strong winds advect ice which melts in place during a
2790
temporary relaxation of winds (Brown and Arrigo, 2013; Stockwell et al., 2001). This scenario
2791
matched the two earliest spring blooms we observed in the southeastern Bering Sea (1997 at M2;
2792
2003 at M4, Appendix Table 1). An open-water bloom subsequent to the ice-related spring bloom
2793
also can occur after a strong mixing event. This occurred in 1997, with a strong ice associated
2794
bloom in April and then a very strong storm in late May that resulted in an open water bloom.
2795
We summarize these findings on spring bloom timing and ice retreat timing (ours; Brown
2796
and Arrigo 2013; Hunt et al., 2002, 2011) as follows. If ice retreats after mid-March, an ice-
2797
associated bloom occurs; this pattern applies throughout the eastern Bering Sea. If the ice retreats
2798
before mid-March, there is no ice-associated bloom because sunlight is insufficient to initiate an
2799
ice-associated bloom. To date, early ice retreat occurs only in the southeastern Bering Sea and not
2800
in the northern Bering Sea; this pattern of persistent ice in the northern Bering Sea is expected to
2801
continue into the foreseeable future (Stabeno et al., 2012a). An open-water bloom occurs if ice
2802
retreats before March 15 or ice is absent; this pattern applies only in the southeastern Bering Sea.
2803
The strengths of spring blooms were similar for years when ice retreated after mid-
2804
March, and for years when ice retreated before mid-March or was absent. We used the maximum
2805
chlorophyll a value to measure spring bloom strength; its timing usually coincides with that of ice
2806
retreat, indicating that these values represent both ice algae and phytoplankton. We anticipated
2807
spring blooms for late retreat years would be weaker due to consumption of nutrients by ice algae
2808
prior to the spring bloom maximum; ice algae growth and the spring phytoplankton bloom
2809
together can create a prolonged period of production or two spring blooms (Niebauer et al., 1995,
112
2810
Stabeno et al., 2001). We found some evidence that ice algae production reduces the subsequent
2811
value of maximum chlorophyll a. The spring bloom maximum averaged 25 (µg chlorophyll a l-l)
2812
when ice was absent, 15 when ice retreated before mid-March and 17 when ice retreated after
2813
mid-March. Although not significant, these differences imply that phytoplankton bloom strength
2814
is diminished if preceded by ice cover and ice algal growth. In contrast to our findings, satellite-
2815
based estimates of open-water blooms of early retreat years were >70% more productive than the
2816
ice-edge blooms of late retreat years, but under-ice production was not quantified (Brown and
2817
Arrigo, 2013). An analysis of measurements of nutrient draw-down would help to resolve this
2818
question.
2819
2820
2821
4.2. Fall
A fall bloom commonly occurred in both the northern and southeastern Bering Sea, on
2822
average in late September. Winds at M2, M4, M5, and M8 are significantly correlated (Stabeno et
2823
al., 2010), which will tend to synchronize fall blooms, as was observed. However bloom timing
2824
was not significantly related to either storm or fall overturn timing. A timing effect for the fall
2825
bloom may be difficult to detect because the month-long interval when blooms usually occurred
2826
is short and because the timing is affected by several interacting factors including wind strength,
2827
stratification, fall cooling and light level. For example, fall overturn requires strong winds, but
2828
once cooling begins in late September and early October, less wind energy is necessary to
2829
overturn the ocean. A fall bloom also depends on sufficient light, introduction of nutrients from
2830
below the pycnocline and short periods of stabilization that allow phytoplankton to remain in the
2831
sunlit waters and grow.
2832
The timing of the fall overturns at M2 and M4 were much more variable than at M5 and
2833
M8 which may be due to north-south differences in stratification. In the south, stratification is
2834
almost completely controlled by temperature, while in the north it is equally controlled by
2835
temperature and salinity so variations in temperature have more effect on stratification in the
2836
south than in the north. This contrast implies that the southeastern Bering Sea is more sensitive to
2837
the timing of fall cooling than the northern Bering Sea.
2838
Our mooring data and previous middle-shelf observations (Rho and Whitledge, 2007;
2839
Brown et al., 2011) indicated that the strengths of fall blooms were weaker than spring blooms on
2840
average. It may be that nutrients are limiting and storm mixing can’t provide sufficient
2841
replenishment of nutrients for a strong fall bloom or the bloom is limited by grazing. For instance,
2842
if the lower layer has 20 μM l-1 of nitrate and the surface mixed layer is 20 m deep and depleted
2843
of nitrate, then deepening the mixed layer by 5 m will result in only 4 μM l-1 of nitrate, compared
113
2844
to 18 μM l-1 available for consumption in the top 20 m of water column in spring. In addition,
2845
variations in chlorophyll a reflect the result of multiple processes including phytoplankton
2846
production, grazing, sinking, and advection. Chlorophyll a will not increase until cell growth
2847
exceeds losses by grazing and other factors. While grazing impact has been measured for spring
2848
(Cooney and Coyle 1982; Sherr et al., 2013), grazing impact has not been measured for fall, so
2849
comparing grazing pressure on spring and fall blooms is not possible. Finally, phytoplankton
2850
physiological status will impact the bloom intensity, with healthier cells growing faster, so
2851
phytoplankton populations in spring may grow faster than those in fall.
2852
Broad-scale sampling of the eastern Bering Sea shelf found that late summer and early
2853
fall chlorophyll a values were higher during warm years compared to cold years (Eisner et al.,
2854
2012). Ocean color data for 1998-2007 in the Bering Sea also show that annual mean chlorophyll
2855
a and total primary production were higher in warm years than cold years (Brown et al., 2011,
2856
Mueter et al., 2009). We used the warm-average-cold year classification of Stabeno et al. (2012)
2857
to classify our fall bloom data. Likewise we found that fall bloom strength was stronger in warm
2858
(average = 12.3 µg Chl l-1, log-transform) versus cold (6.3) and average (8.6) years, but the
2859
differences were not statistically significant (ANOVA, log-transform, df = 2, 31, p = 0.32).
2860
2861
4.3 Spring/fall comparisons
2862
Spring and fall bloom strengths were related, implying that a common factor influences
2863
spring and fall primary production (e.g., overwinter replenishment of nutrients). In addition, this
2864
relationship likely amplifies secondary production during good years (both spring and fall blooms
2865
tend to be strong) and vice versa, bad years (both spring and fall blooms tend to be weak). An
2866
analysis of nutrient information, comparing spring-fall differences by year, would help us to
2867
understand the mechanism for this relationship, but has not yet been published.
2868
The fall bloom occurred in late September on average, and the timing was less variable
2869
than for the spring bloom (varies over ~60 day compared to ~120 day period) regardless of
2870
location, so the spring-fall interval largely depends on spring bloom timing. In the northern
2871
Bering Sea, where ice is present every year and usually retreats in late May (Fig. 5), the interval
2872
typically lasted four months (Fig. 11). The interval also typically lasted four months in the
2873
southeastern Bering Sea in years when ice was absent or retreated before March 15. In contrast,
2874
the interval lasted up to six months in the southeastern Bering Sea in years when ice was present
2875
after March 15, but retreated soon thereafter.
2876
2877
4.4. The representativeness of the fluorescence data
114
2878
Distinct spring and fall blooms occurred at these moorings located in the middle domain
2879
with the spring bloom stronger than the fall bloom. Likewise, a bimodal spring-fall pattern was
2880
found for net primary production for the middle domain with spring stronger than fall for both
2881
satellite-based (Brown et al., 2011) and temporally and spatially scattered in situ measurements
2882
(Rho and Whitledge 2007). During the summer interval between the spring and fall blooms, we
2883
found that chlorophyll a fluorescence was low in the middle domain, similar to Goes et al. (this
2884
issue). We also found that the mooring observations of chlorophyll a fluorescence were
2885
consistently correlated with satellite observations of chlorophyll a fluorescence in the middle
2886
domain at the same approximate latitude (Appendix Figure 2). This pattern implies that the
2887
mooring information represents conditions of the middle domain over a range of ca. 150-200 km.
2888
The bimodal pattern was replicated for the inner domain except that spring and fall bloom
2889
strengths appear similar for satellite-based measurements (Brown et al., 2011). In contrast, the
2890
fall bloom was less distinct for the outer domain (Rho and Whitledge 2007; Brown et al., 2012).
2891
These patterns imply that bloom patterns differ somewhat among the inner, middle and outer
2892
domains, and that the mooring information do not well represent the oceanographic conditions of
2893
the inner (influence of the coastal current, river runoff, and tidal and wind mixing ) or outer
2894
domains (influence of slope, episodic on-shelf flow and upwelling). The broad-scale data
2895
collected during the late summer BASIS sampling also indicates that chlorophyll a concentrations
2896
and phytoplankton particle size (large >= 10 um and small < 10 um) varied among shelf domains
2897
(Eisner et al., 2012). The relatively high concentrations found in the outer domain were primarily
2898
comprised of small taxa (< 25% large); chlorophyll a concentrations on average were lower in the
2899
middle domain during these surveys, but when increases in chlorophyll a were observed (usually
2900
near M4), there was a higher proportion of large taxa (40-70% large). Besides phytoplankton,
2901
mixotrophic ciliates (a taxonomic group of microzooplankton that also photosynthesize) also can
2902
contribute to total chlorophyll a measurements in summer (Stoecker et al., this issue).
2903
2904
4.5 Biological Implications
2905
4.5.1 Lower trophic levels
2906
Spring bloom timing predictably varied between early April and mid-June depending on
2907
ice presence/absence and ice retreat timing. In contrast, we found no statistically significant
2908
difference in spring bloom maximum among the independent factors examined including ice
2909
presence/absence and ice retreat timing. Thus phytoplankton biomass is the same, but the
2910
secondary community (zooplankton species) that benefits will depend on how close timing of
2911
their spring energy-intensive needs such as reproduction and awakening from winter diapause
115
2912
matches the timing of the spring bloom. Species that require an early pulse of energy will benefit
2913
from years when ice is present, but retreats in late March (conditions that tend to result in early
2914
April phytoplankton bloom). In contrast, species with a phenology timed for a late energy pulse
2915
will benefit from years with no ice (conditions that tend to result in a late May to early June
2916
bloom).
2917
These observations also imply that climate, through its connection to the production,
2918
transport, and dissipation of sea ice has the potential to affect the success of zooplankton
2919
populations and the strength of coupling between primary production and higher trophic levels.
2920
For example, the large crustacean zooplankton taxa Calanus spp. may benefit in years when ice is
2921
present after March 15, but retreats relatively early. Baier and Napp (2003) observed that spring
2922
Calanus concentrations in the southeastern Bering Sea were higher in cold years than in warm
2923
years, and that these individuals likely metamorphosed from naupliar to copepodite stages during
2924
the early spring bloom. They hypothesized that metamorphosis was a recruitment bottleneck and
2925
that an early spring bloom benefitted copepodite recruitment. The spring bloom occurs during
2926
April in years when ice is present, but retreats relatively early, which may promote strong
2927
recruitment of copepodites. In addition, colder winters with ice present likely reduces metabolism
2928
and lipid utilization by Calanus spp. and thus may promote winter survival (Coyle et al., 2011).
2929
Recruitment of Calanus spp. on the eastern Bering Sea shelf is complicated, involving
2930
many different processes. Life history strategies of large crustacean zooplankton (LCZ) in the
2931
eastern Bering Sea may be able to take advantage of the oscillations between warm and cold
2932
periods, regardless of whether or not there are stanzas of multiple years of the same conditions
2933
(Overland et al., 2012). The timing of reproductive events for the Calanus congener Calanus
2934
glacialis varies between different physical and biological environments of the Arctic (Daase et al.
2935
2013). The following scenarios were developed as a working hypothesis to explain some of the
2936
present observations and to serve as a guide for future research. In these scenarios, we simplified
2937
Calanus life history into four major steps: spawning, metamorphosis, accumulation of depot
2938
lipids and overwintering. These scenarios make the following assumptions: 1) the timing of
2939
spawning by Calanus spp. is protracted (from February to May) (Baier and Napp 2003 and
2940
references therein) and the longer that conditions are suitable, the more total eggs each individual
2941
will produce (E. Durbin, U. Rhode Island, pers. comm.); 2) early-spawned individuals enter
2942
diapause before late-spawned individuals; and 3) overwinter respiration is a function of
2943
temperature (Saumweber and Durbin 2006). The working hypothesis does not address the role of
2944
advection of individuals into or out of the area.
116
2945
Case A. Cold/Ice-Replete Years
2946
with Early Ice Retreat (Fig. 12). We
2947
hypothesize that access to under ice
2948
algae (beginning as early as mid
2949
February) increases egg production rates
2950
of early spawning Calanus (Runge and
2951
Ingram, 1991). Metamorphosis to the
2952
copepodite stage benefits from an early
2953
ice retreat and spring bloom production
2954
(Baier and Napp, 2003). Summer
2955
accumulation of depot lipids is
2956
dependent on summer primary and
2957
microzooplankton production
2958
(Sambrotto et al., 1986; Sherr et al.,
2959
2013; Stoecker et al., this volume).
2960
Overwintering for individuals spawned
2961
in late winter begins relatively early and
2962
these individuals may not be able to take
2963
advantage of the fall bloom. The ability
2964
of individuals (C5 copepodites) to
2965
survive the following winter benefits
2966
from low respiration rates (RC5 copepodite = 5.66 µm 02 d-1 at -1.8 °C (see methods)) resulting from
2967
cold bottom temperatures. Reproductive output during the following year(s) is maximized when
2968
there are subsequent ice-replete years and under ice algae.
2969
Figure 12. Timing of Calanus reproduction relative to
the presence of ice and ice algae, and the spring and fall
blooms. The Calanus life history is simplified and the
timings are approximate. Black is the period of ice
cover; gray is the 50th percentiles for the timing of ice
retreat. White rectangles are the 50th percentile dates
for bloom maximum in the spring and fall. Ice algae
blooms during ice cover (solid gray line) and ice retreat
(dashed gray line) and begin as early as mid-February
(R. Gradinger, personal comm.). A – a cold icy year, with
ice retreat shortly after March 15; B – a cold icy year
with ice retreat after April 14; C - no ice after March 15.
Case B. Cold/Ice-Replete Years with Late Ice Retreat. In this scenario, both early and late
2970
spawning by Calanus spp. females benefits, as reproduction timing coincides with either the ice
2971
algae or the spring bloom. The nauplii and copepodites from late spawning would likely develop
2972
post-spring bloom (i.e., metamorphosis is not coincident with the spring bloom), and would
2973
depend upon summer primary and microzooplankton production to grow and accumulate depot
2974
lipids. The onset of overwintering occurs latest for the late-spawned progeny which may take
2975
advantage of the fall bloom to accumulate additional lipids before entering diapause. Ability to
2976
survive the overwintering period benefits from the cold bottom temperatures present in late fall
2977
through early spring.
117
2978
Case C. Warm/Ice-Deplete Years. Egg production by early spawning of Calanus is
2979
reduced by the absence or short duration of sea ice and under ice algae. The timing of the spring
2980
bloom is late; it benefits late spawning of Calanus spp. females. The late spring bloom may
2981
coincide with metamorphosis of the early-spawned nauplii to copepodites, but is mismatched
2982
with the metamorphosis from the late spawners. Growth and development are dependent upon
2983
summer primary and microzooplankton production. Diapause that is delayed until after the fall
2984
bloom benefits late spawners. Warm bottom temperatures result in higher respiration rates -- 47%
2985
higher daily respiration than during cold years (RC5 copepodite = 8.34 µm 02 d-1 at 2.0 °C). The impact
2986
on lipid reserves is exacerbated if the period without food (beginning of diapause to the following
2987
spring bloom) is longer in warm years relative to cold. Thus we hypothesize that survival of
2988
Calanus spp. over a warm winter is low relative to a cold winter. Predation losses would also
2989
likely be higher in warm winters at a given predator abundance due to higher metabolism and
2990
food requirements of fishes. Absence of sea ice for a second year results in low reproductive
2991
output by the early spawners, and a decrease in the population.
2992
2993
2994
4.4.2. Connections to Higher Trophic Levels
The six to eight month variation in the fall to spring bloom interval also may contribute to
2995
the peak of the dome-shaped relationship between juvenile pollock production and sea surface
2996
temperature observed by Mueter et al. (2006, 2011) and Coyle et al. (2011). Pollock recruitment
2997
was generally high when ice retreated early and adult biomass was low (Mueter et al., 2006); we
2998
found that a six-month fall to spring bloom interval occurs in years with an early ice retreat. The
2999
likely mechanism is the relationship between early ice retreat, delayed spring bloom, and the
3000
production of large crustacean zooplankton as pollock recruitment was generally high when large
3001
crustacean zooplankton were abundant (Hunt et al., 2011). The mechanism linking secondary
3002
production and pollock recruitment has been attributed to enhanced energy content and
3003
overwinter survival of age-0 pollock during cold years when the lipid-rich, large crustacean
3004
zooplankton biomass is highest (Heintz et al., accepted).
3005
3006
3007
5. Conclusions
● In the eastern Bering Sea: if ice is present after mid-March, spring bloom timing is
3008
related to ice retreat timing; if ice is absent or retreats before mid-March, a spring bloom
3009
usually occurs in May or early June.
3010
3011
● Spring and fall bloom strengths are related, implying that a common factor influences
spring and fall primary production.
118
3012
● We hypothesize that large crustacean zooplankton such as Calanus spp. benefit from
3013
cold, icy winters in the southeastern Bering Sea because ice algae or ice associated
3014
phytoplankton blooms provide an early spring food source and respiration rates are lower
3015
during cold winters.
3016
3017
Acknowledgements
3018
We thank D. Kachel for data analysis, S. Salo for providing the ice data and W. Floering and C.
3019
Dewitt for deploying and recovering the moorings. We thank the officers and crews of the NOAA
3020
ships Miller Freeman and Oscar Dyson for their invaluable assistance in deploying and
3021
recovering the moorings. We thank Zach Brown, Rolf Gradinger and one anonymous reviewers
3022
whose comments improved our manuscript. The research was generously supported by the North
3023
Pacific Research Board and NOAA’s North Pacific Climate Regimes and Ecosystem Productivity
3024
(NPCREP) and Fisheries Oceanography Coordinated Investigations programs. This is BEST-
3025
BSIERP publication no. XXX and NPRB publication no. YYY. This is contribution FOCI-B***
3026
to the Ecosystems & Fisheries Oceanography Coordinated Investigations; and 3953 to Pacific
3027
Marine Environmental Laboratory. The findings and conclusions in this paper are those of the
3028
authors and do not necessarily represent the views of NOAA’s National Marine Fisheries Service
3029
or Oceans and Atmospheric Research.
3030
3031
References
3032
Acker, J.G., Leptoukh, G. 2007. Online analysis enhances use of NASA Earth science data. Eos,
3033
Transactions of the American Geophysical Union 88(2): 14.
3034
3035
Alexander V., Niebauer, H.J.1981. Oceanography of the eastern Bering Sea ice edge zone in
3036
spring. Limnology & Oceanography. 26: 1111-1125.
3037
3038
Baier, C.T., Napp, J.M. 2003. Climate-induced variability in Calanus marshallae populations. J.
3039
Plankton Res. 25: 771–782.
3040
3041
Baier, C.T., Terazaki, M. 2005. Interannual variability in a predator-prey interaction: climate,
3042
chaetognaths and copepods in the southeastern Bering Sea. J. Plankton Res., 27: 1113–1125.
3043
3044
Bond, N.A., Adams, J.M. 2002. Atmospheric forcing of the southeast Bering Sea Shelf during
3045
1995–99 in the context of a 40-year historical record. Deep-Sea Research II 49: 5869–5887.
119
3046
3047
Brown, Z.W., Arrigo, K.R. 2013. Sea ice impacts on spring bloom dynamics and net primary
3048
production in the eastern Bering Sea. J. Geophys. Res. 118, 1-20, doi: 10.1029/2012JC008034.
3049
3050
Brown, Z.W., van Dijken, G.L., Arrigo, K.R. 2011. A reassessment of primary production and
3051
environmental change in the Bering Sea. J. Geophys. Res. 116, C08014, doi:
3052
10.1029/2010JC006766.
3053
3054
Cooney, R.T., Coyle, K. O. 1982. Trophic implications of cross-shelf copepod distributions in the
3055
southeastern Bering Sea. Mar. Biol. 70: 187-196.
3056
3057
Coyle, K.O., Eisner, L.B., Mueter, F.J., Pinchuk, A.I., Janout, M.A., Cieciel, K.D., Farley, E.V.,
3058
Andrews, A. G. 2011. Climate change in the southeastern Bering Sea: impacts on pollock stocks
3059
and implications for the Oscillating Control Hypothesis. Fish. Oceanogr. 20: 139–156.
3060
3061
Daase, M., Falk-Petersen, S., Varpe, Ø., Darnis, G., Søreide, J.E., Wold, A., Leu, E., Berge, J.,
3062
Philippe, B., Fortier, L. 2013. Timing of reproductive events in the marine copepod Calanus
3063
glacialis: a pan-Arctic perspective. Can. J. Fish. Aquat. Sci. 70: 1-14.
3064
3065
Droop, M. R. 1973. Some thoughts on nutrient limitation in algae. J. Phycol. 9(3): 264-272.
3066
3067
Eisner, L.B., Cieciel, K., Gann, J. 2012. Phytoplankton biomass and size structure during late
3068
summer to early fall in the eastern Bering Sea. In Zador, S.[Ed.], Ecosystem Considerations 2012,
3069
Stock Assessment and Fishery Evaluation Report, North Pacific Fisheries Management Council,
3070
605 W 4th Ave, Suite 306, Anchorage, AK.
3071
http://www.afsc.noaa.gov/REFM/docs/2012/ecosystem.pdf
3072
3073
Fortier, M., Fortier, L., Michel, C., and Legendre, L. 2002. Climatic and biological forcing of the
3074
vertical flux of biogenic particles under seasonal Arctic sea ice. Mar. Ecol. Prog. Ser. 225:1-16.
3075
3076
Goes, J.I., Gomes, H.R., Haugen, E., McKee, K., D’Sa, E., Chekalyuk, A.M., Stoecker, D.,
3077
Stabeno, P., Saitoh, S., Sambrotto, R. This issue. Fluorescence, pigment and microscopic
3078
characterization of Bering Sea phytoplankton community structure and photosynthetic
3079
competency in the presence of a cold pool during summer. Deep-Sea Res. II.
120
3080
3081
Gradinger, R., Ikavalko, J. 1998. Organism incorporation into newly forming Arctic sea ice in the
3082
Greenland Sea. J. Plankton Res. 20: 871-886.
3083
3084
Granger, J., Prokopenko, M.G., Sigman, D.M., Mordy, C.W., Morse, Z.M., Morales, L.V.,
3085
Sambrotto, R.N., Plessen, B. 2011. Coupled nitrification-denitrification in sediment of the eastern
3086
Bering Sea shelf leads to 15N enrichment of fixed N in shelf waters. J. Geophys. Res.: Oceans
3087
(1978–2012) 116(C11) doi: 10.1029/2010JC006751
3088
3089
Grebmeier, J.M., Cooper, L.W., Feder, H.M., & Sirenko, B.I. 2006. Ecosystem dynamics of the
3090
Pacific-influenced northern Bering and Chukchi Seas in the Amerasian Arctic. Prog.
3091
Oceanogr. 71: 331-361.
3092
3093
Haecky, P., Jonsson, S., Andersson, A. 1998. Influence of sea ice on the composition of the
3094
spring phytoplankton bloom in the northern Baltic Sea. Polar Biol. 20: 1-8.
3095
3096
Heintz, R.A., Siddon, E.C., Farley, E.V., Napp, J.M. Accepted. Climate related changes in the
3097
nutritional condition of young-of-the-year pollock (Theragra chalcogramma) from the eastern
3098
Bering Sea. Deep-Sea Res. II
3099
3100
Hunt, G.L, Allen, B.M., Angliss, R.P., Baker, T., Bond,N., Buck, G. and others. 2010. Status and
3101
trends of the Bering Sea region, 2003-2008, pp. 196-267 In Mckinnell, S.M. and Dagg M.J.
3102
[Eds.], Marine Ecosystems of the North Pacific Ocean, 2003-2008. PICES Special Publication 4,
3103
393 p.
3104
3105
Hunt, G.L., Coyle, K.O., Eisner, L.B., Farley, E.V., Heintz, R.A., Mueter, F.J., Napp, J.M.,
3106
Overland, J.E., Ressler, P.H., Salo, S., Stabeno, P.J. 2011. Climate impacts on eastern Bering Sea
3107
foodwebs: a synthesis of new data and an assessment of the Oscillating Control Hypothesis. ICES
3108
J. Mar. Sci. 68: 1230–1243.
3109
3110
Hunt, G.L., Stabeno, P.J., Walters, G., Sinclair, E., Brodeur, R.D., Napp, J.M., Bond, N.A. 2002.
3111
Climate change and control of the southeastern Bering Sea pelagic ecosystem. Deep-Sea Res. II.
3112
49: 5821–5853.
3113
121
3114
Iverson, R.L., Coachman, L.K., Cooney, R.T., English, T.S., Goering, J.J., Hunt, Jr., G.L.,
3115
McCauley, M.C., McRoy, C.P., Reeburg, W.S., Whitledge, T.E. 1979. Ecological significance of
3116
fronts in the southeastern Bering Sea, pp. 437-466. In: Ecological Processes in Coastal and
3117
Marine Systems, R.J. Livingston (Ed.), Plenum Press.
3118
3119
Jin, M., Deal, C., Wang, J., Alexander, V., Gradinger, R., Saitoh, S., Iida, T., Wan, Z., Stabeno,
3120
P. 2007. Ice-associated phytoplankton blooms in the southeastern Bering Sea, Geophys. Res.
3121
Lett., 34, L06612, doi:10.1029/2006GL028849.
3122
3123
Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha, S.,
3124
White, G., Woollen, J., Zhu, Y., Kirst G. and Wienchke C. 1995. Ecophysiology of polar algae. J.
3125
Phycol. 31: 181-199.
3126
3127
Kirst, G.O., Wiencke, C.1995. Ecophysiology of polar algae. Journal of Phycology, 31: 181–199.
3128
doi: 10.1111/j.0022-3646.1995.00181.x
3129
3130
Ladd, C., Bond, N.A., 2002. Evaluation of the NCEP-NCAR Reanalysis in the northeast Pacific
3131
and Bering Sea. J. Geophys. Res. 107, 3158. doi:10.1029/2001JC001157.
3132
3133
Melnikov, I.A. 1998. Winter production of sea ice algae in the western Weddell Sea. J. Mar. Sys.
3134
17: 195-205.
3135
3136
Mock T., and Gradinger R., 2000. Changes in photosynthetic carbon allocation in algal
3137
assemblages of Arctic sea ice with decreasing nutrient concentrations and irradiance. Mar. Ecol.
3138
Prog. Ser. 202: 1-11.
3139
3140
Mueter, F.J., Bond, N.A., Ianelli, J.N., Hollowed, A.B. 2011. Expected declines in recruitment of
3141
walleye pollock (Theragra chalcogramma) in the eastern Bering Sea under future climate change.
3142
ICES J. Mar. Sci. 68: 1284-1296.
3143
3144
Mueter, F.J., Broms, C., Drinkwater, K.F., Friedland, K.D., Hare, J.A., Hunt, G.L., Melle, W.,
3145
Taylor, M. 2009. Ecosystem responses to recent oceanographic variability in high-latitude
3146
Northern Hemisphere ecosystems. Prog. Oceanography 81: 93-110.
3147
122
3148
Mueter, F.J., Ladd, C., Palmer, M.C., Norcross, B.L. 2006. Bottom-up and top-down controls of
3149
walleye pollock (Theragra chalcogramma) on the Eastern Bering Sea shelf. Prog. Oceanography
3150
68: 152-183.
3151
3152
Niebauer, H.J., Alexander, V., Henrichs, S.M. 1995. A time-series study of the spring bloom at
3153
the Bering Sea ice edge I. Physical processes, chlorophyll and nutrient chemistry. Cont. Shelf
3154
Res. 15: 1859-1877.
3155
3156
Niemi, A., Michel, C., Hille, K., Poulin, M. 2011 Protist assemblages in winter sea ice: setting the
3157
stage for the spring ice algal bloom. Polar Biol. 34: 1803-1817.
3158
3159
O'Brien, D.P. 1987. Direct observations of the behavior of Euphausia superba and Euphausia
3160
crystallorophias (Crustacea: Euphausiacea) under pack ice during the Antarctic spring of 1985. J.
3161
Crustacean Biol. 7: 437-448.
3162
3163
Overland, J.E., Wang, M., Wood, K.R., Percival, D.B., Bond, N.A. 2012. Recent Bering Sea
3164
warm and cold events in a 95-year context. Deep-Sea Res. II 65-70: 6-13.
3165
3166
Parsons, T.R., Maita, Y., Lalli, C.M. 1984. A manual of biological and chemical methods for
3167
seawater analysis. Pergamon Press, Oxford, 173pp.
3168
3169
Rho, T., Whitledge, T.E. 2007. Characteristics of seasonal and spatial variations of primary
3170
production over the southeastern Bering Sea shelf. Continental Shelf Research 27: 2556-2569.
3171
3172
Riebesell U., Schloss I., Smetacek V. 1991. Aggregation of algae form melting ice: implications
3173
for seeding and sedimentation. Polar Biol. 11: 239-248.
3174
3175
Runge, J.A., Ingram, R.G. 1991. Under-ice feeding and diel migration by the planktonic
3176
copepods Calanus glacialis and Pseudocalanus minutus in relation to the ice algal production
3177
cycle in southeastern Hudson Bay, Canada. Mar. Biol. 108: 217-225
3178
3179
Sambrotto R.N., H.J. Niebauer, J.J. Goering, and R.L. Iverson, 1986. Relationships among
3180
vertical mixing nitrate uptake and phytoplankton growth during the spring bloom in the southeast
3181
Bering Sea middle shelf. Cont. Shelf Res. 5: 161–198.
123
3182
3183
Saumweber, W.J., Durbin, E.G., 2006. Estimating potential diapause duration in Calanus
3184
finmarchicus. Deep-Sea Res. II 53: 2597-2617.
3185
3186
Sherr, E.B., Sherr, B.F., Ross, C. 2013. Microzooplankton grazing impact in the Bering Sea
3187
during spring sea ice conditions. Deep-Sea Res. II http://dx.doi.org/10.1016/j.dsr2.2013.03.019
3188
3189
Spreen, G., Kaleschke, L., Heygster, G. 2008. Sea ice remote sensing using AMSR-E 89 GHz
3190
channels. J. Geophys. Res. 113(C2), C02.
3191
3192
Stabeno, P.J., Bond, N.A., Kachel, N.B., Salo, S.A., 2007. On the recent warming of the
3193
Southeastern Bering Sea Shelf. Deep-Sea Res. II 54: 2599–2618.
3194
3195
Stabeno, P.J., Bond, N.A., Kachel, N.B., Salo, S.A., Schumacher, J. D. 2001. On the temporal
3196
variability of the physical environment over the southeastern Bering Sea. Fish. Oceanogr. 10: 81–
3197
98.
3198
3199
Stabeno, P.J., Farley, E., Kachel, N.B., Moore, S.E., Mordy, C.W., Napp, J.M., Overland, J.E.,
3200
Pinchuk, A.I., Sigler, M.F. 2012a. A comparison of the physics of the northern and southern
3201
shelves of the eastern Bering Sea and some implications to the ecosystem. Deep-Sea Res. II.
3202
doi:10.1016/j.dsr2.2012.02.019
3203
3204
Stabeno, P.J., Kachel, N.B., Moore, S.E., Napp, J.M., Sigler, M.F., Yamaguchi, A., Zerbini, A.N.
3205
2012b. Comparison of warm and cold years on the southeastern Bering Sea shelf and some
3206
implications to the ecosystem. Deep-Sea Res. II. doi:10.1016/j.dsr2.2012.02.020
3207
3208
Stabeno, P.J., Napp, J., Mordy, C., Whitledge, T., 2010. Factors influencing physical structure
3209
and lower trophic levels of the eastern Bering Sea shelf in 2005: Sea ice, tides and winds. Prog.
3210
Oceanogr. 85(3–4), 180–196.
3211
3212
Stoecker, D.K., Weigel, A.C., Stockwell, D.A., Lomas, M.W. This issue. Microzooplankton:
3213
Abundance, biomass and contribution to chlorophyll in the eastern Bering Sea in summer. Deep-
3214
Sea Res. II.
3215
124
3216
Sullivan, M.E., Mordy, C.W., Salo, S.A., Stabeno, P.J. This issue. The influence of sea ice on
3217
seasonal water column changes on the southeastern Bering Sea Shelf. Deep-Sea Res. II.
3218
3219
Sverdrup, H. U. 1953. On conditions for the vernal blooming of phytoplankton. J. Conseil perm.
3220
int. Explor. Mer. 18: 287-295.
3221
3222
Tremblay, C., Runge, J., Legendre, L. 1989. Grazing and sedimentation of ice algae during and
3223
immediately after a bloom at the ice-water interface. Mar. Ecol. Prog. Ser. 56: 291-300.
3224
3225
Figure Caption
3226
3227
Figure 1. Study area and mooring locations.
3228
3229
Figure 2. The relationship of mixed layer depth deepening (day) and temperature drop (day) for
3230
moorings M2 and M4. The mooring data at M5 and M8 were inadequate to compute mixed layer
3231
deepening.
3232
3233
Figure 3. A boxplot of ice retreat (day) by mooring. The box extends from the first quartile to the
3234
third quartile, the heavy line dividing the box is the median and the whiskers are the smallest and
3235
largest values. The number of years with ice present were 12 (M2), 15 (M4), 17 (M5) and 17
3236
(M8).
3237
3238
Figure 4. Scatterplots of maximum spring bloom strength (ug Chl l-1), spring bloom timing (day),
3239
minimum winter-spring temperature (day), day temperature rose above -1 ºC and day ice cover
3240
fell below 15%. If ice was absent that year, then the ice retreat day is zero.
3241
3242
Figure 5. Scatterplots of spring bloom maximum (day) versus ice retreat (day) by mooring. If ice
3243
was absent that year, then the ice retreat date is zero. The diagonal dashed line is the 1:1 line to
3244
compare timings of spring bloom and ice retreat. The vertical dashed line is March 15.
3245
3246
Figure 6. Scatterplot of the observed (number) and fitted (line, based on simple linear regression, y-
3247
intercept = 53, slope = 0.66, df = 1, 24, p < 0.001) values of spring bloom maximum (day) versus ice
3248
retreat (day) for all moorings when ice was present after March 15. The number indicates mooring
3249
number.
125
3250
3251
Figure 7. A boxplot of fall overturn (day) by mooring. The box extends from the first quartile to
3252
the third quartile, the heavy line dividing the box is the median and the whiskers are the smallest
3253
and largest values. The numbers of years when fall overturn could be determined were 15 (M2),
3254
10 (M4), 16 (M5) and 7 (M8).
3255
3256
Figure 8. Scatterplots of maximum fall bloom strength (ug Chl l-1), fall bloom timing (day),
3257
maximum summer-fall temperature (ºC), fall overturn (day) and first storm (day).
3258
3259
Figure 9. Scatterplots of fall bloom maximum (day) versus fall overturn (day) by mooring. The
3260
diagonal dashed line is the 1:1 line to compare timings of fall bloom maximum and fall overturn.
3261
The horizontal dashed line is August 15.
3262
3263
Figure 10. Scatterplot of maximum spring bloom strength (ug Chl l-1) and maximum fall bloom
3264
strength. The number indicates mooring number.
3265
3266
Figure 11. Scatterplot of observed (number) and fitted (line, based on linear regression, intercept
3267
= 294, slope = -1.16, df = 1, 27, p < 0.001) values for all moorings. The y=axis is the interval
3268
between the spring and fall bloom maximum (days). The x-axis is the spring bloom maximum
3269
(day). The number indicates mooring number.
3270
3271
Figure 12. Timing of Calanus reproduction relative to the presence of ice and ice algae, and the
3272
spring and fall blooms. The Calanus life history is simplified and the timings are approximate.
3273
Black is the period of ice cover; gray is the 50th percentiles for the timing of ice retreat. White
3274
rectangles are the 50th percentile dates for bloom maximum in the spring and fall. Ice algae
3275
blooms during ice cover (solid gray line) and ice retreat (dashed gray line) and begin as early as
3276
mid-February (R. Gradinger, personal comm.). A – a cold icy year, with ice retreat shortly after
3277
March 15; B – a cold icy year with ice retreat after April 14; C - no ice after March 15.
126
3278
Appendix Table 1. Annual values compiled for this manuscript; methods for compiling these
3279
values are described in the methods section. In all cases “Day” refers to Julian day. Day ice
3280
retreat: the day that percent ice cover fell below 15% (a zero value indicates that ice cover never
3281
fell below 15%); Day temp >-1 ºC: the day that ocean temperature rose above -1 ºC (a zero value
3282
indicates that ocean temperature never fell below -1 ºC); Min WS temp: Minimum temperature
3283
during winter and spring; Day max spring bloom: Day that the chlorophyll a (µg l-1) values
3284
reached their maximum (NA indicates missing data); Max spring bloom: Maximum chlorophyll a
3285
value during winter and spring; Max fall temp: Maximum temperature during summer and fall;
3286
Day first storm: first day a storm occurred during summer and fall as defined in the methods; Day
3287
overturn: first day that temperature drop indicated overturn was underway; Max SF temp:
3288
Maximum temperature during summer and fall; Day max fall bloom: Day that the chlorophyll a
3289
values reached their maximum; Max fall bloom: Maximum chlorophyll a value during summer
3290
and fall.
127
Moor Year
M2
1995
M2
1996
M2
1997
M2
1998
M2
1999
M2
2000
M2
2001
M2
2002
M2
2003
M2
2004
M2
2005
M2
2006
M2
2007
M2
2008
M2
2009
M2
2010
M2
2011
M4
1995
M4
1996
M4
1997
M4
1998
M4
1999
M4
2000
M4
2001
M4
2002
M4
2003
M4
2004
M4
2005
Day
ice
retreat
104
0
99
60
129
35
0
53
0
0
0
36
98
110
119
103
114
112
44
100
72
128
40
0
60
87
101
0
Day
temp
>-1C
110
0
92
56
134
58
0
0
0
0
0
36
111
128
125
127
120
NA
NA
120
NA
NA
NA
0
65
87
100
0
Min
WS
temp
-1.60
-0.90
-1.56
-1.68
-1.73
-1.20
1.67
-0.14
1.63
1.31
2.12
-1.61
-1.67
-1.69
-1.73
-1.38
-1.38
NA
NA
-1.69
NA
NA
NA
0.86
-1.68
-1.24
-1.28
1.34
Day
max
Max
spring spring
bloom bloom
NA
NA
NA
NA
106
39.6
146
24.5
113
94.8
132
4.2
144
26.7
153
13.1
138
22.4
143
23.8
132
30.8
165
28.1
115
9.7
122
20.0
155
9.0
126
16.1
158
18.5
NA
NA
NA
NA
127
27.2
NA
NA
124
6.0
166
12.5
149
21.9
155
22.0
94
76.2
144
17.6
NA
NA
Max
fall
temp
10.5
9.6
13.6
11.8
9.7
10.3
11.6
12.5
12.1
13.6
13.0
10.4
12.2
9.8
10.1
8.4
10.4
NA
7.6
NA
NA
8.6
10.3
10.3
11.4
11.2
11.7
12.0
Day
first
Day
storm overturn
295
NA
245
242
245
277
235
228
295
263
285
240
295
246
265
247
295
262
275
265
255
235
285
262
255
268
325
284
315
NA
265
251
265
263
325
NA
245
NA
245
NA
235
NA
215
259
245
242
305
267
295
268
275
274
285
266
255
239
Max
SF
temp
239
207
220
217
195
195
229
242
251
224
220
235
246
254
237
213
231
NA
266
NA
NA
200
232
232
231
258
241
218
Day
max
Max
fall
fall
bloom bloom
NA
NA
306
6.9
293
16.6
275
9.1
238
22.6
248
7.6
NA
NA
NA
NA
294
59.3
271
19.8
261
15.5
268
4.6
NA
NA
300
3.2
NA
NA
301
4.6
261
6.3
NA
NA
NA
NA
NA
NA
NA
NA
249
8.4
263
5.4
272
14.4
288
59.6
274
9.7
286
8.3
NA
NA
128
M4
M4
M4
M4
M4
M4
2006
2007
2008
2009
2010
2011
135
100
135
117
138
121
120
123
127
125
NA
124
-1.70
-1.70
-1.74
-1.03
NA
-1.62
134
104
129
145
NA
129
32.1
10.6
52.3
2.9
NA
2.8
9.9
10.8
9.6
9.7
8.1
NA
255
255
285
315
265
265
258
254
279
223
NA
NA
242
223
236
230
268
NA
270
298
282
NA
291
NA
24.6
9.1
5.4
0.1
8.9
NA
129
Moor Year
M5
1995
M5
1996
M5
1997
M5
1998
M5
1999
M5
2000
M5
2001
M5
2002
M5
2003
M5
2004
M5
2005
M5
2006
M5
2007
M5
2008
M5
2009
M5
2010
M5
2011
M8
1995
M8
1996
M8
1997
M8
1998
M8
1999
M8
2000
M8
2001
M8
2002
M8
2003
M8
2004
M8
2005
M8
2006
Day
ice
retreat
132
120
129
124
156
127
93
109
102
108
128
153
151
166
137
136
136
127
131
138
142
158
134
141
136
117
128
122
144
Day
temp
>-1C
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
130
177
166
158
160
139
141
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
131
176
Min
WS
temp
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
-1.43
-1.74
-1.74
-1.75
-1.74
-1.71
-1.71
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
-1.76
-1.79
Day
max
Max
spring spring
bloom bloom
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
162
6.0
141
14.5
156
35.3
NA
NA
146
6.3
144
2.4
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
133
17.0
144
10.3
Max
fall
temp
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
11.3
9.3
11.0
9.3
9.4
9.5
7.7
NA
NA
NA
NA
NA
NA
NA
NA
NA
11.2
10.0
9.1
Day
first
Day
storm overturn
255
NA
245
NA
265
NA
315
NA
295
NA
315
NA
295
NA
285
NA
275
NA
285
NA
255
NA
255
NA
265
NA
325
NA
335
NA
265
NA
215
NA
255
NA
245
NA
265
NA
295
NA
265
NA
285
NA
275
NA
295
NA
285
NA
285
NA
315
NA
335
NA
Max
SF
temp
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
217
242
223
247
232
223
215
NA
NA
NA
NA
NA
NA
NA
NA
NA
225
239
241
Day
max
Max
fall
fall
bloom bloom
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
281
6.4
298
10.7
279
14.5
246
14.8
NA
NA
218
10.0
223
4.1
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
290
2.8
259
4.8
315
10.3
130
M8
M8
M8
M8
M8
2007
2008
2009
2010
2011
144
152
155
135
141
186
187
185
185
217
-1.76
-1.80
-1.78
-1.77
-1.76
161
160
166
NA
NA
35.4
49.6
49.3
NA
NA
9.9
8.6
7.9
9.7
1.8
295
275
265
305
265
NA
NA
NA
NA
NA
260
257
249
254
NA
244
298
NA
NA
NA
15.3
5.7
NA
NA
NA
131
3282
3283
Appendix Figure 1. Plots of chlorophyll a fluorescence by mooring, year and Julian day. The x-
3284
axis is Julian day and the y-axis is chlorophyll a (µg l-1). Each plot is labeled with mooring and
3285
year. The open symbols indicate the assigned spring (circle) and fall (triangle) blooms.
3286
132
3287
133
3288
134
3289
135
3290
136
3291
137
3292
138
3293
139
3294
3295
3296
140
3297
Appendix Figure 2. Maps of Pearson's product moment correlations between 8-day averages of
3298
log-transformed fluorescence at four mooring sites and 8-day averages of log-transformed
3299
chlorophyll a as estimated from satellite data (SeaWiFS 1998-2002, Modis Aqua 2003-2011).
3300
Black squares denote reference mooring locations for each map. Only correlations individually
3301
significant at approximately 95% are shown.
3302
3303
141
3304
Chapter 15
3305
3306
Sullivan, M., C. Mordy, S. Salo, and P.J. Stabeno (2013): The influence of sea ice on seasonal
3307
water column changes on the eastern Bering Sea shelf. Deep-Sea Res. II.
3308
Sea ice and water column structure on the eastern Bering Sea shelf
Margaret E. Sullivanab*, Nancy B. Kachel ab, Calvin W. Mordyab. Sigrid A. Salob,
and Phyllis J. Stabenob
a
Joint Institute for the Study of the Atmosphere and Ocean, University of Washington, 3737
Brooklyn Ave NE, Box 355672, Seattle, WA 98105-5672 USA
b
Pacific Marine Environmental Laboratory, NOAA, 7600 Sand Point Way NE, Seattle, WA
98115 USA
*Corresponding Author
7600 Sand Point Way NE,
Seattle, WA 98115 USA
msulliva@u.washington.edu,
Phone: (206) 526-6185, Fax: (206) 526-6485
Revised: September 5, 2013
For submission to Deep-Sea Research Part II: Topical Studies in Oceanography
3309
142
3310
Abstract
3311
Seasonal sea ice is a defining characteristic of the eastern Bering Sea shelf, and plays a critical
3312
role in determining the vertical structure of temperature and salinity over this shelf. This paper
3313
examines ice movement relative to local winds, and the impact of the arrival and retreat of ice on
3314
the water column at four mooring sites on the middle shelf. Ice forms primarily in coastal regions
3315
and is advected over the southern and outer shelves. Ice drift was ~2% of local wind speed and
3316
oriented 44 ° to the right of the winds (R-squared = 0.25). Measurements from 30 ice cores
3317
collected in 2006 -2009 gave an average salinity of ~6, and an average nitrate concentration of ~1
3318
µM. Time series data collected at the biophysical moorings in the Bering Sea (1995 – 2012) were
3319
used to explore the evolution of stratification under the ice. At mooring M8 in the north, the
3320
water column mixed and cooled prior to the arrival of ice, hence, little melt occurred when the ice
3321
arrived. At the other three moorings, the ocean temperature was 2 – 4 °C above freezing, resulting
3322
in extensive ice melt. The melting ice freshened and cooled the upper part of the water
3323
column. The water column remained stratified for 10 to 25 days. Mixing was a slower process
3324
under the ice than in open water. An estimated 1.3 m of ice melted with the first arrival of ice at
3325
three southern moorings and the latent heat of cooling of this ice accounted for approximately
3326
half the observed cooling. During ice retreat there appeared to be little ice melt around the
3327
southern two moorings, but an estimated 0.8 m at M5 and M8. The extent of ice melt sets up the
3328
water column for the following summer.
3329
3330
Keywords: Bering Sea, sea ice, ocean temperature, salinity, currents, stratification, ice melt
3331
3332
1. Introduction
3333
Situated between the North Pacific and the Arctic Ocean, the eastern Bering Sea is a
3334
marginal sea ice zone. During summer it is ice free, with ice beginning to form along the coast in
3335
the northern Bering Sea as early as November (Pease, 1980). During subsequent months, the sea
3336
ice is advected southward with maximum ice extent typically occurring in March or April
3337
(Stabeno et al., 2012b). In extensive ice years, sea ice can advance more than 1000 km, which,
3338
before the extensive melting now occurring in the Arctic Ocean, was the largest ice advance in
3339
any arctic or subarctic region (Niebauer, 1998). Sea ice typically begins retreat in late winter or
3340
early spring, and by late June the Bering Sea is ice free.
3341
Pease (1980) describes the expansion of ice into the southern Bering Sea as a conveyor belt
3342
in which ice forms in polynyas in the northern Bering Sea and is advected southward by winds.
3343
Polynyas are most common off southern coastlines due to prevailing northerly winds in winter,
143
3344
but they also can form when winds are southerly off northern and western coastlines (Niebauer
3345
and Schell, 1993, Niebauer et al., 1999). The St. Lawrence polynya occurs along the southern
3346
coast of St. Lawrence Island, and is a dominant ice-producing feature (Schumacher et al., 1983;
3347
Stringer and Groves, 1991; Danielson et al., 2006; Drucker et al., 2003), with ice being advected
3348
mostly south and southwestward (McNutt, 1981). The leading edge of the ice melts as it
3349
encounters warmer ocean temperatures in the south or along the shelf break (McNutt, 1981;
3350
Zhang et al., 2010).
3351
The eastern Bering Sea continental shelf is broad (>500 km), less than 180 meters deep,
3352
and extends more than 1000 km south from Bering Strait to the Aleutian Islands. The shelf is
3353
divided into inner (water depth < 50 m), middle (~50-100 m), and outer (~100-180 m) domains
3354
(Coachman, 1986), which are separated by transition zones or fronts (Schumacher, et al., 1979;
3355
Stabeno, et al., 2001; Kachel, et al., 2002). Ice typically appears first in the coastal domain, and
3356
latest in the outer domain. On average, sea ice covers the middle shelf north of 58 °N for five
3357
months each year, with ice persisting, on average, for less than one month on the southern shelf
3358
(south of 57 °N) (Stabeno et al., 2012a).
3359
These north-south differences in timing of ice retreat and advance, together with a north
3360
(weaker) – south (stronger) gradient in tidal energy, result in distinct differences in the water
3361
column structure on the northern and southern shelves during the summer (Stabeno et al., 2012a).
3362
The northern shelf with its more extensive and persistent ice cover remains cold until solar input
3363
warms the ocean, initiates ice melt and sets up summer stratification. This results in a warmer,
3364
fresher surface layer overlaying a colder, saltier bottom layer separated by a ~10 m interface. The
3365
cold bottom layer (< 2 °C) over the middle shelf is known as the cold pool, and remains largely
3366
well mixed due to the tides. Ice retreat in the south typically occurs in March or April with
3367
southerly winds, when winds are usually still strong enough to mix the water column (Stabeno et
3368
al, 2010). The water column becomes thermally stratified when winds weaken and solar
3369
insolation increases (Stabeno et al, 2007). In years with extensive ice in March or April, the
3370
bottom temperatures in ice-covered areas in the south are < 2 °C, and part of the cold pool
3371
(Wyllie-Echeverria, et al., 1998; Stabeno et al., 2012b). During years when the sea ice does not
3372
extend onto the southern shelf, the bottom temperatures remain above 2 °C, and the cold pool
3373
exists only over the northern shelf (Stabeno et al., 2012a,b).
3374
The presence or absence of sea ice impacts spatial and temporal variability of biological
3375
production across the shelf and slope (Stabeno et al., 2012a,b; Sigler, et al., this volume).
3376
Constituents of sea ice, including iron and phytoplankton, are advected southward, and influence
3377
the amount and fate of spring production (Aguilar-Islas et al., 2008; Schandelmeier and
144
3378
Alexander, 1981). In addition, sea ice impacts the timing of the spring phytoplankton bloom,
3379
with a bloom occurring under the ice if ice is present after mid-March, and an open-water bloom
3380
occurring in May or June if no ice is present after mid-March (Brown and Arrigo, 2013; Stabeno
3381
et al., 2012a; Sigler et al., this volume). Vertical stratification of the water column is important to
3382
3383
3384
3385
3386
3387
Figure 12. a) A MODIS True-Color image is combined with MODIS sea surface temperature in the eastern
Bering Sea for March 11 - 13, 2007. The large, red letters indicate the following: A polynyas; B heavy ice;
and C cloud cover. b) A shelf map showing distribution of ice sampling sites per year (2006-2009), and
the bathymetry. The four mooring locations (M2, M4, M5, M8) and geographic names are indicated in
both panels.
3388
Support under-ice phytoplankton blooms (Alexander and Niebauer, 1981; Mundy et al., 2009).
3389
The southern extent of the cold pool and the timing of spring primary production influence the
3390
food chain and higher trophic level predator-prey relationships for the following 6 - 8 months
3391
(Sigler et al., this volume), and over longer time scales (Mueter and Litzow, 2008).
3392
The purpose of this paper is to examine the impact of sea ice on the vertical structure of the
3393
water column. Measurements from satellites, ice cores and four long-term biophysical moorings
3394
distributed along the 70-m isobath from 56.9 °N to 62 °N were used to determine the rate and
3395
direction of ice movement, the arrival of ice at moorings, and the extent of ice melt and resulting
3396
stratification at each of the mooring sites (Fig. 1). First, the rate and direction of ice movement in
145
3397
the region were calculated using satellite images. Next, the characteristics of 30 ice cores
3398
collected from 2006 – 2009 are presented. Using time series of temperature, salinity and currents
3399
from the biophysical moorings (M2, M4, M5 and M8; Fig. 1), the impact of sea ice on the water
3400
column structure is examined. Using selected time series from the ~42 yearly records, the
3401
thickness of the ice melted at the mooring sites is estimated.
3402
3403
2. Methods
3404
2.1. Satellite Data
3405
Satellite measurements of ocean color and sea surface temperature (SST) were obtained
3406
using MODIS (Moderate Resolution Imaging Spectroradiometer) data files from the ocean color
3407
website at NASA (http://oceancolor.gsfc.nasa.gov, assessed 2012). SeaDAS, a suite of programs
3408
created and maintained by NASA, was used to map true-color and SST images. Individual floes
3409
were identified in a series of true-color images and tracked from one image to another, until they
3410
were no longer recognizable (Holt, et al, 1992, Leberl, et al, 1983). Ocean color and SST
3411
observations cannot be made through clouds, severely limiting the number of useful MODIS
3412
images from the Bering Sea. In addition, increased cloud cover is associated with southerly
3413
winds, so the set of floes that were tracked were biased against times when southerly winds were
3414
prevalent. This limited the analysis to years with a sufficient number of clear days to follow a floe
3415
over a period of time, and to larger floes in regions where horizontal shear was weak enough not
3416
to deform the floes.
3417
For 2002–2011, ice concentration data from the AMSR_E (Advanced Microwave Scanning
3418
Radiometer-Earth Observing System) sensor aboard the MODIS Aqua satellite were downloaded
3419
from the National Snow and Ice Data Center website (http://nsidc.org/data/amsre/, assessed 2012;
3420
Cavalieri et al., 2003). Ice concentrations from the NASA Special Sensor Microwave Imager
3421
(http://nsidc.org/data/docs/daac/ssmi_instrument.gd.html; Maslanik and Stroeve, 1999) were used
3422
for fall 1994–2002 to extend the time span prior to the start date of AMSR-E data. Daily-mean ice
3423
concentration values were used in this paper. The data were used to calculate mean daily ice
3424
concentrations within boxes centered on the four moorings (M2, M4, M5, and M8 in Fig. 1). The
3425
grid size of the ice data is approximately 12.5 km × 12.5 km. Following Stabeno et al. (2012a) we
3426
used 100 km × 100 km boxes.
3427
3428
2.2 Wind stress and winds
146
3429
National Center for Environmental Prediction (NCEP) – Department of Energy Reanalysis
3430
2 uses a state-of-the-art analysis/forecast system to perform data assimilation with data ranging
3431
from January 1979 to August 2012 (http://www.cpc.ncep.noaa.gov/products/wesley/reanalysis2/;
3432
Kanamitsu et al., 2002), and is an update to the NCEP/NCAR (National Center for Atmospheric
3433
Research) reanalysis. Six-hourly wind and wind-stress data were extracted from the NCEP
3434
Reanalysis 2 at the grid points nearest to the four mooring locations. NCEP Reanalysis 2 data
3435
were obtained from the website maintained by NOAA Earth System Research Laboratory,
3436
Physical Sciences Division in Boulder, Colorado, USA (http://www.esrl.noaa.gov/psd/). NCEP
3437
winds are well correlated with the observed winds in the Bering Sea (Ladd and Bond, 2002).
3438
QuikSCAT wind data were downloaded from the Jet Propulsion Laboratory, Physical
3439
Oceanography Distributed Active Archive Center (http://podaac.jpl.nasa.gov). Scatterometer
3440
winds are derived from ripple patterns on the surface of the ocean, and are therefore limited to
3441
ice-free areas. Although the instrument sees through clouds, measurements are affected by heavy
3442
rain. The records identify pixels where rain was suspected, and these records were excluded from
3443
our analysis. Grid spacing for QuikSCAT winds is 12.5 km (0.25 ° longitude), but the data were
3444
binned into 2° bins for regional plots.
3445
3446
3447
2.3. Ice core measurements
During four springs (2006-2009), a total of 30 ice cores were collected by NOAA scientists
3448
from floes distributed over the eastern Bering Sea shelf (Fig. 1b). In 2006, three floes were
3449
sampled on the R/V Thomas G. Thompson (Cruise TN193). The Thompson had limited
3450
maneuverability in the ice, so sampling was restricted to floes at the ice edge. Cruises in 2007-
3451
2009 were conducted on the icebreaker, U.S. Coast Guard Cutter Healy as part of the Bering Sea
3452
Project (http://bsierp.nprb.org/). The Healy is capable of breaking 1.4 meters of ice at 3 knots
3453
continuously, and 2.4 meters by backing and ramming. This permitted sampling deep into the
3454
pack-ice in water depths between 30 and 100 meters.
3455
Floes were chosen for accessibility and for ease of data collection. Sampling was biased
3456
against ice that was too thin to walk on or highly rafted, and to times when air temperature was
3457
above -15 °C. All samples were taken in spring when both melting and refreezing were prevalent.
3458
In 2006, the three ice stations were located along the ice edge in the outer domain, while the
3459
seven stations in 2007, the 11 in 2008, and the nine in 2009 were mostly located deeper within the
3460
ice field in the middle domain north of 58 °N (Fig. 1b; Table 1).
3461
3462
At most sites, air temperature, freeboard, ice thickness, and snow cover were recorded;
coordinates were recorded at the beginning and end of each site visit using a GPS unit. All cores
147
3463
were obtained using a Kovacs4 Mark-II, 9-cm diameter, 1.15-meter-long ice corer driven by an
3464
attachable gas engine. Cores were positioned on a black plastic surface, then measured and
3465
photographed. Top and bottom conditions and banding were noted. Small holes were drilled and
3466
temperatures were recorded every 10 cm along the core using either an Oakton Acorn or an
3467
Omega thermometer. All cores were sliced into 10-cm sections and double-bagged for later
3468
analysis. On the ship, sections from each core were thawed in the dark. Volume of the melt water
3469
was measured and then sub-sampled to measure salinity, and nutrient concentrations. Salinities
3470
were measured at sea using a salinometer referenced against IAPSO standard water. Ice density
3471
was determined from the dimensions of each ice core, and the volume and density of the ice melt.
3472
Samples for nutrient analysis were filtered using syringe with 0.45 μm cellulose acetate
3473
membranes, and collected in 30 ml acid washed, high-density polyethelene bottles after being
3474
rinsed three times with sample water. Samples were analyzed shipboard within 1 - 12 hrs of
3475
collection. Nitrate and nitrite concentrations were determined using a combination of analytical
3476
components from Alpkem, Perstorp and Technicon. WOCE-JGOFS standardization and analysis
3477
procedures specified by Gordon et al. (1994) were followed, including reagent preparation,
3478
calibration of labware, preparation of primary and secondary standards, and corrections for blanks
3479
and refractive index.
3480
For determination of chlorophyll and phaeopigment content a second core was taken at the
3481
sites, and sections of ice were melted in ~1.5 times the ice volume in filtered seawater to prevent
3482
cell lysis. Sub-samples (~50 ml) were filtered though Whatman GF/C filters, and the filters were
3483
frozen at -80 °C, remaining frozen until analysis several months later. Pigments were extracted in
3484
cold 90% acetone in the dark for 24 hrs, and determined using a calibrated fluorometer according
3485
to Strickland and Parsons (1968).
3486
3487
2.4. Moorings
3488
Four biophysical mooring sites (M2, M4, M5, and M8) are the cardinal locations of the
3489
Bering Sea biophysical observing network (Fig. 1). When possible, the moorings are recovered
3490
and redeployed twice a year, once in the spring (April/May) and again in the late summer or early
3491
fall (September/October). During years with extensive ice, M8 and sometimes M5 were only
3492
recovered/deployed in September. Moorings at the M2 site have been maintained almost
3493
continually since 1995, while moorings have been deployed at M4 since 1996 (continuously since
3494
2000), at M5 since 2005 and at M8 since 2004. At the writing of this manuscript, these moorings
3495
provided 42 fall-through-spring sets of time series describing water column conditions.
4
Use of trade names does not constitute an endorsement by NOAA
148
3496
Data collected by instruments on the moorings included temperature (miniature
3497
temperature recorders, SeaBird SBE-37 and SBE-39) and salinity (SBE-37). Currents were
3498
measured using an upward-looking, bottom-mounted, 300 or 600 kHz Teledyne RD Instruments
3499
acoustic Doppler current profiler (ADCP) deployed within 500 m of the main mooring. All
3500
instruments were calibrated prior to deployment, and the data were processed according to
3501
manufacturers’ specifications. The main mooring “line” at each site was chain and the
3502
instruments were mounted in-line in metal frames. Such construction is necessary to reduce the
3503
chances of losing the mooring due to sea ice and heavy fishing pressure in the region.
3504
Sampling intervals varied among the different instruments and ranged from 10 to 60
3505
minutes. Small data gaps (< 1 day) were filled using
3506
linear interpolation or spectral methods. Unless
3507
otherwise noted, data presented here have been filtered
3508
using a 35-hour, cosine squared, tapered Lanczos filter to
3509
remove semidiurnal and diurnal tidal signals, and re-
3510
sampled to 6-hour intervals.
3511
The top instrument at each mooring was at a depth
3512
between 1 and 20 m. When the mixed layer was deeper
3513
than the depth of the top instrument, the data from the
3514
top instrument could be reliably extrapolated to the
3515
surface (Stabeno et al., 2007). However, when the top
3516
instrument was below ~20 m and the ice had retreated
3517
from the mooring in the spring, the mixed layer was
3518
usually shallower than the top instrument and there was
3519
no reliable way of estimating localized near-surface
3520
temperatures.
3521
3522
2.5. Calculation of volume of ice melt
3523
Estimates of magnitude of ice melt can be made
3524
at the mooring sites from the change in salinity in the
3525
water column using the following equation:
3526
hw Si + (H-hw) Sw = H Sf
3527
or
(1)
Figure 13. Change in ice
concentration (color filled)
associated with 2-day mean
QuikSCAT winds (vectors) are
shown for three time periods: a)
February 21 - 23, 2003; b) March 2 3, 2003; and c) March 12 - 13, 2003.
149
3528
hw = H (Sw - Sf) / (Sw – Si)
(2)
3529
where hw is the thickness of water resulting from ice melt, Si the salinity of the ice, H the depth of
3530
the upper mixed layer (which may be the whole water column), Sw the initial salinity of the water
3531
column and Sf the final salinity of the water column. This is an approximation, since an
3532
introduced piece of ice does not melt completely, but rather the ice is advected over the water,
3533
with only a portion of it melting at that location. When the water reaches the freezing point, there
3534
is no further net melting of the ice.
3535
The thickness of the ice melted, hi, was calculated using hi = hw
3536
i
w
i
w
is the
is the density of ice. Using the value of hw calculated from salinity, the
3537
change in temperature of the water column per unit area due to latent heat can be calculated using
3538
the following equation
3539
3540
i
hi
(3)
3541
where L is the latent heat of fusion of sea ice and T is
3542
temperature.
3543
3544
3. Results and Discussion
3545
3.1. Satellite floe-tracking and winds
3546
Ice in the Bering Sea can be broadly
3547
characterized using true-color satellite imagery from
3548
which polynyas (dark shadows leeward of the
3549
coastline), thin ice (gray, often seaward of the open
3550
water of the polynya), and thicker ice (white) can be
3551
identified (Fig. 1a). Ice conditions, including the
3552
location of polynyas and the direction and extent of
3553
ice advance, can change rapidly and are largely forced
3554
by winds (Fig. 2). For example, in late February 2003,
3555
northeasterly winds near Siberia and easterly winds
3556
off the Alaskan coast reduced the ice concentration
3557
near Siberia, St. Lawrence Island, and the Alaskan
3558
coast (Fig. 2a). Nine days later, southerly winds over
3559
the eastern shelf resulted in ice edge retreat in the east
Figure 14. Ice floe trajectories tracked
using MODIS True Color images are
shown for two years: a) February 19 –
April 15, 2003; and b) March 1 – April
20, 2007. The colors indicate the dates
of observations. See color bar at right
of plots for dates. Mooring locations
are indicated.
150
3560
and compression of the ice along those coastlines (Fig. 2b). In mid-March, northerly winds
3561
opened polynyas in the north, advanced the ice edge, and increased the ice concentration at the
3562
ice edge (Fig. 2c).
3563
Floe trajectories obtained from true-color satellite images are shown for 2003 and 2007
3564
(Fig. 3), a warm and a cold year, respectively (Stabeno et al., 2012b). These trajectories are
3565
primarily from advancing ice. While many floes could only be tracked for a few days, several
3566
were tracked for 1–2 weeks while they moved southward more than 100 km. The maps from both
3567
years show a corridor in which ice floes moved from the shelf north and east of St. Lawrence
3568
Island through Shpanberg Strait, with more westward ice motion in 2003 than 2007 as floes
3569
moved beyond Shpanberg Strait. These observations are similar to the conveyor belt corridor
3570
observations in Pease (1980). With southerly winds, ice produced from a polynya north of
3571
Nunivak Island could also contribute to the conveyor belt corridor (Figs. 2 and 3a). South of the
3572
Chukotka Peninsula, floes moved eastward with prevailing currents (Danielson, 2000), and those
3573
nearest the coast continued north through Anadyr Strait. A few floes originating southwest of St.
3574
Lawrence were tracked to M8 (Fig. 3).
3575
These floe trajectories were used to infer the origin of ice in each of the four mooring areas
3576
(Figs. 1 and 3). The origin of the ice at M8 was a combination of ice from the conveyor belt and
3577
from the polynya south of St. Lawrence Island (McNutt, 1981; Schumacher et al., 1983), and
3578
occasionally Anadyr Strait (Schumacher, et al., 1983). M4 and M5 locations were directly
3579
influenced by ice from the conveyor belt corridor (Muench and Ahlnas,1976; Pease, 1980). The
3580
area around M2 was outside the dominant influence of the conveyor belt corridor, and subject to
3581
ice formation from polynyas south of Nunivak Island and along the northern Bristol Bay coast
3582
during northerly or northeasterly
3583
winds.
3584
Tracked floes, which had
3585
typical speeds of 0.1 to 0.3 m s-1,
3586
were influenced by both winds and
3587
currents. These findings are
3588
comparable to earlier observations
3589
(Shapiro and Burns, 1975; Muench
3590
and Ahlnas, 1976; McNutt, 1981;
3591
Weeks and Weller, 1984), as well as
3592
results from a recent ice model
3593
(Zhang et al, 2010), but slower than
Figure 15. Wind vector plots, from 6-hourly NCEP-2
Reanalysis data centered at mooring site M4, for the cold
seasons of a) 2005 – 2006 and b) 2006 – 2007. Northward is
upward. Ice extent in the 100-km x 100-km around M4 is
included as a blue line.
151
3594
the 0.5 m s-1 average reported by Pease (1980) and the 0.45 m s-1 average reported by Macklin et
3595
al. (1984). Ice drift was compared to NCEP2 winds interpolated in space and time, and found to
3596
be ~2% of local wind speed (n = 440), and oriented 44 ° to the right of the wind direction with an
3597
R-squared value of 0.25, which is higher than the findings of Macklin et al. (1984).
3598
While wind largely controls the movement of sea ice (Figs. 2, 3), the timing of storms is
3599
also important. The influence of wind events on ice formation varied dependent upon the season.
3600
Northerly winds during active freeze-up (January into March) created more ice than similar winds
3601
would have generated during warming (mid April to May). Nonetheless, northerly winds later in
3602
the season could prolong ice cover by continuing to transport ice southward. For example, in the
3603
winter of 2005-2006, northerly winds in December and January (Fig. 4a) pushed the ice to its
3604
most southern extent in late January. This was in spite of the fact that water column was ~2.8 °C.
3605
The next episode of sustained northerly winds did not occur until mid-April (initial water
3606
temperature of 0.5 °C), and although the ice advanced at this time, the ice edge did not regain its
3607
earlier position, because the increase solar insolation. At the beginning of April the water column
3608
warmed from -1 °C to >0 °C. In contrast, during the winter of 2006 – 2007, there were several
3609
storms in December and January (Fig. 4b), but the longest period of northerly winds was during
3610
March and April, and sea ice reached its maximum extent in late March (Rodionov et al., 2007).
3611
The water column cooled in early March (from 1.1 C to -1.7 °C) and did not begin warming until
3612
the ice began retreating.
3613
3614
3.2. Sea-ice cores
3615
Thirty ice stations were
3616
sampled on the eastern Bering Sea
3617
shelf in April and early May on four
3618
cruises between 2006 and 2009 (Fig.
3619
1b). Ice thickness ranged from 37 to
3620
127 cm with a single outlier of >329
3621
cm of rafted ice at an inner domain
3622
station (Table 1). Mean ice thickness
3623
was calculated excluding this outlier.
3624
Surface water from nearby
3625
hydrographic casts had mean salinity of 31.9, mean water temperature of -1.7 °C, and mean
Figure 16. a) Vertical profiles of temperature, salinity and
nitrate from an ice core collected at Ice Station 2 in 2008
(Table 1). b) Ice core photograph from same station. c)
Selected vertical profiles of salinity from 2009 ice cores.
3626
alinities in the ice were low (mean ~6, Table 1), reflecting brine rejection
3627
during ice formation and brine drainage. There was considerable variability in the shapes of the
152
3628
salinity profiles (Fig. 5c). Some salinity profiles showed the well-defined C-shape (e.g., green
3629
line, Fig. 5c) of marginal ice zones and young ice (Malmgren, 1927, Eicken, 2003). Another
3630
common pattern was a double C-shape (e.g., red line, Fig. 5c), often a result of rafting. Mean ice
3631
density was 0.89 ± 0.03 x 103 kg m-3, which is in agreement with the Table 1 summary in Timco
3632
and Frederking (1996). Nutrient concentrations were low in the ice cores (mean nitrate ~1 µM),
3633
suggesting that ice was not a conveyor of macro nutrients in the Bering Sea. Melting sea ice,
3634
however, does provide a significant contribution of iron to the water column (Aguilar-Islas et al.,
3635
2008).
3636
Horizontal banding was evident in the ice cores and often, dark coloration occurred at the
3637
ice-water interface (Fig. 5). Some banding was correlated with elevated chlorophyll-a and
3638
phaeopigments, while in other cores the banding appeared to be sediment. Highest chlorophyll
3639
concentrations were most commonly found in the bottom section of each ice core. For example,
3640
in 2009, bottom chlorophyll was as high as 209 µg l-1, with an average of ~80 µg l-1 which was
3641
approximately 20 times greater than the average concentration (4 µg l-1) found in the rest of the
3642
core slices. Mean snow depth on floes was variable, but generally less than 10 cm (Table 1).
3643
3644
3645
3.3. Influence of sea ice on the water column
Typically, the depth-averaged water column temperature reaches a maximum in September
3646
(Stabeno et al., 2012a, b), when the mixed layer begins to deepen. In examining the 18 years of
3647
data at M2, the depth-averaged temperature is almost as warm in August as in September, and
3648
usually the water column has not mixed. Temperature data from August are used to represent
3649
maximum summer near-surface temperature. During August, the mixed layer is usually 20-30 m
3650
deep. The highest mean August near-surface temperature was 12.5 °C (M2, 2002, a warm year)
3651
and the lowest monthly mean bottom temperature was -1.7 °C (M8, 2009, a cold year) (Table 2).
3652
The differences between monthly mean surface and bottom temperatures in August range from
3653
5.2 to 10.8 °C.
3654
3655
153
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
3677
Figure 6. Time series from M2 showing a) wind stress, b) temperature at 12
depths distributed through the water column, and c) salinity at four depths.
The breaks in the time series (b and c) occurred when the winter mooring
was replace by the summer mooring. The gray shading indicates times when
the water column was well mixed. There was no ice cover at M2 during this
period.
3678
3679
3680
154
3681
The time at which the water
3682
column became well mixed varied
3683
from north to south (Table 2), with M8
3684
remaining stratified the longest of the
3685
four mooring areas. At M8 during
3686
summer, both temperature and salinity
3687
contribute equally to stratification,
3688
while at the southern moorings
3689
stratification is controlled by
3690
temperature. Since the vertical
3691
differences in temperature are similar
3692
north-to-south, the stratification is
3693
almost twice as strong at M8 than at
3694
the other mooring sites (Stabeno et al.,
3695
2010, 2012a; Ladd and Stabeno, 2012),
3696
thus delaying mixing.
3697
While there is considerable
Figure 6. Time series from M2 showing a) wind stress, b)
temperature at 12 depths distributed through the water
column, and c) salinity at four depths. The breaks in the
time series (b and c) occurred when the winter mooring
was replace by the summer mooring. The gray shading
indicates times when the water column was well mixed.
There was no ice cover at M2 during this period.
3698
information on the timing and
3699
characteristics of ice retreat and melting (Alexander and Niebauer, 1981; Hunt and Stabeno,
3700
2002; Stabeno et al., 1998; Stabeno et al., 2012b), less is known about the short-term response of
3701
the water column to melting sea ice. Instruments at each of the four biophysical moorings in the
3702
Bering Sea have measured temperature, salinity and currents in the water column as ice was
3703
advected over the moorings. In total, there were 42 sets of data showing the water column
3704
response to the ice. Several patterns were evident in the time series at the various moorings. To
3705
explore these patterns, seven examples were selected (two from M2, three from M4, one each
3706
from M5 and M8).
3707
3708
3709
3.3.1. No ice (Example: M2, autumn 2004 – spring 2005)
From 1995 to the present, there was little (areal coverage of < 20% concentration) or no ice
3710
present for five years of observations at M2 and two years at M4. Data from these seven
3711
examples show how the water column mixes, cools and becomes more saline without the
3712
influence of ice. In October 2004, the water column at M2 was cooling and mixing (Fig. 6).
155
3713
From a mean August surface temperature of 12.5 °C and mean bottom temperature of 3.2 °C (not
3714
shown), the water column became well mixed at 5.7 °C at the end of October. The mean wind
3715
stress during the fall of 2004 (mean 0.23 N m-2) was not significantly stronger than the long-term
3716
(1995 – 2012) mean (0.20 N m-2). The water column remained well-mixed through the winter
3717
until April 8, with the bottom temperature reaching a minimum of ~2.0 °C in mid April.
3718
Minimum water-column temperature in 2004 was anomalously high, and was surpassed by only
3719
two other warm-years: January, 2001 (3.0 °C) and January, 2003 (3.4 °C). All three instances are
3720
from an unusually warm period, 2001 to 2005 (Stabeno et al., 2012b). Mean minimum winter
3721
temperature from 1995 to 2011 at M2 was -0.6 °C, while the mean for the nine years excluding
3722
this warm window (2001 – 2005) was -1.7 °C (Table 2). Salinity increased from ~31.7 in early
3723
October to ~32.3 in mid February most likely from advection, of more saline water from the outer
3724
shelf and slope (Sullivan et al., 2008).
3725
3726
3727
3.3.2. A single ice event (Example: M2, fall 2007 – spring 2008)
During the winter and spring 2007
3728
ice was present near M2 for
3729
approximately one month (Fig. 7).
3730
During this period, winds were moderate,
3731
with isolated storms interspersed with
3732
periods of weak winds. The water
3733
column became well mixed (defined as
3734
when the vertical temperature difference
3735
was < 0.04 °C, and the salinity difference
3736
was < 0.02) at 3.6 °C in mid November,
3737
2006 (not shown). Ice arrived in the
3738
vicinity of the mooring on March 2, and
3739
four days later, as the ice coverage
3740
reached ~20%, the water column began
3741
to stratify with colder, fresher water
3742
overlying warmer, saltier water. The top
3743
39 m of the water column cooled to the
3744
freezing point (-1.7 °C) by mid March
3745
and ice melt freshened the surface salinity by ~0.5. The water column remained stratified for ~20
3746
days. There were four moderate storms during the ice period (Fig. 7, top, 1 – 4). The first storm
Figure 7. Time series from M2, February 27 – April 17,
2007, showing a) wind stress, b) temperature at 12
depths distributed through the water column and the
percent areal ice cover in the 100-km x 100-km box
around M2, and c) salinity at three depths. The gray
shading indicates times when the water column was
well mixed.
156
3747
(March 6-9) coincided with the arrival of the ice; the second (March 18-22) occurred during a
3748
period of >90% ice cover; and the third (March 25 – 27) occurred during a period of decreasing
3749
ice. The second storm coincided with a slight mixing at 39 – 44 m and the third event resulted in
3750
the total mixing of the water column. The primary cause of the greater mixing associated with
3751
the third storm, which was only slightly stronger than the second, was likely reduced ice cover.
3752
Less areal ice cover permits the more rapid movement of ice floes and also greater direct energy
3753
transfer between the atmosphere and the water.
3754
The duration and ice concentration at M2 in 2007 (Fig. 7) typified ice coverage in the
3755
south. Ice cover at M2 is more variable, has lower areal concentration and persists for shorter
3756
periods of time than at the other three mooring areas (Stabeno et al., 2012a). M2 is outside the
3757
direct path of the ice conveyor belt corridor and it is likely that ice near M2 arrived from Nunivak
3758
Island and the coastal polynya to the east. Over 17 year period (late winter 1995 to spring, 2011),
3759
single winter ice events occurred 6 times at M2 (lasting 20 - 83 days) and 7 times at M4 (lasting
3760
11 - 121 days).
3761
3762
3763
3.3.3. Multiple ice events with incomplete mixing (Example: M4, fall 2010 – spring 2011)
During some years, ice can be intermittent, with ice advancing and retreating several
3764
times during the cold season (December – May). The winter and spring of 2011 was such a
3765
season at M4. The water column became well mixed at ~1.7 °C in late November 2010 (not
3766
shown) and was still well mixed when northerly winds forced the ice over the mooring on
3767
January 20 (Fig. 8). Strong winds kept the water column well mixed as the ice cooled and
3768
freshened the water. The ice persisted for approximately one week and then retreated with
3769
southeasterly winds. Following this initial ice advance and retreat, the winds remained variable,
3770
forcing four major ice events over the next three months. As expected, periods of northerly or
3771
northeasterly winds resulted in ice advancing, while periods of southerly winds resulted in ice
3772
retreat. Maximum areal ice concentration remained below 70%, except for the final ice event.
3773
While variable, salinity (especially at 31 m) generally increased from January into late
3774
March in spite of the melting ice evidenced by the cooling of the surface waters. Large
3775
freshening events (> 0.3) of the complete water column occurred (e.g., early March, early April),
3776
but freshening of just the surface waters (from ice melt) occurred more often. The greatest
3777
decrease in salinity (0.6) occurred in late April when ice concentrations approached 100%. The
3778
157
3779
Figure 8. Time series from M4, January 19 – May 1, 2011, showing a) 6-hourly wind stress,
b) 12-hourly wind velocity, c) temperature at 16 depths distributed through the water
column and the percent areal ice cover in the 100-km x 100-km box around M4, and d)
salinity at three depths. The gray shading indicates times when the water column was well
mixed. Asterisks (*) indicate periods of increased temperature and salinity, likely a result
of advection. Two storms are indicated in each panel by a red bar.
158
3780
3781
temperature time series showed recurring, albeit weak, stratification (< 1.5 °C difference between
3782
top and bottom), interspersed with well-mixed periods (Fig. 8, gray shading). The cooling and
3783
freshening of the upper 20 - 27 m of water with the advent of ice was similar to the event
3784
described at M2 in 2007 (Fig. 7). The
3785
water column began mixing after each
3786
ice retreat and was usually well mixed
3787
before the next arrival of ice, only to
3788
become stratified again. The two wind
3789
events (A and B in Fig. 8 a) were similar
3790
in magnitude. During the first storm the
3791
water column was uniformly stratified
3792
(Fig. 8c; ~1 °C in temperature, 0.3 in
3793
salinity). The winds during the first
3794
storm mixed the water column, but not
3795
completely. During the second storm,
3796
the water column mixed completely.
3797
Wind mixing became most effective in
3798
both cases when ice concentration fell
3799
below 30%.
3800
The bottom temperatures often
3801
warmed after mixing episodes (Fig. 8b),
3802
an indication of advection. Similarly the
3803
0.7 increase in salinity from mid January
3804
through late March was also indicative of
3805
advection. Mean winter currents at M4
3806
are southward (Stabeno et al., 2010).
3807
Currents, which flow north along the
3808
100-m isobath, can circle around St. Paul Island (Sullivan et al., 2008) and advect warmer, more
3809
saline water onto the middle-shelf near M4.
3810
Figure 9. Time series from M4, February 2 – April 24,
2008, showing a) wind stress, b) temperature at 12
depths distributed through the water column and the
percent areal ice cover in the 100-km x 100-km box
around M4, and c) salinity at three depths. Low-pass
filtered currents are shown at d) 11 m and e) 55 m. The
gray shading indicates times when the water column
was well mixed. The red bar at the bottom of each
panel represents the time period shown in Fig. 10.
Multiple instances of ice advance and retreat (>2 events per year) were not uncommon at
3811
the southern moorings: 4 of 17 years at M2; 1 of 17 at M4; and 3 of 17 at M5. The example
3812
discussed here (Fig. 8) had the most ice events. This example, with a persistent series of
159
3813
advances and retreats, serves as a good example of both wind and advection influencing ice melt
3814
and winter stratification events.
3815
3816
3.3.4. A cold year with persistent ice (Example: M4, fall 2007 – spring 2008)
3817
During 2007-2008 (both cold years, Stabeno et al., 2012b), the water column became well
3818
mixed on November 15 at 2.8 °C and cooled to -0.4 °C by January 13 (not shown). Ice reached
3819
the area in early February coincident with strong winds (Fig. 9). The water column was well
3820
mixed for 3 - 4 days in early February, and then became stratified as winds weakened and ice
3821
concentration increased to > 20%. The ice-covered water column became well mixed again in
3822
late February and remained well mixed or weakly stratified until late March. Once again, there
3823
are strong indications that advection played a role in stratifying the water column. In both early
3824
February and late March there was a marked increase in temperature and salinity below 40 m,
3825
while the temperature and salinity in the upper water column remained relatively constant. Low-
3826
pass filtered, near-bottom currents during the February event were > 15 cm s-1 from the south.
3827
During the March event, currents were variable and relatively weak (< 10 cm s-1) (Fig. 9 d, e).
3828
While the daily mean currents were relatively weak in mid-March (Fig. 9), the tidal
3829
currents were significantly stronger,
3830
regularly exceeding 40 cm s-1 (Fig. 10).
3831
The first and second increase in bottom
3832
temperature and salinity occurred when
3833
the currents began to turn
3834
eastward. Once the boundary (front)
3835
between the warmer, more saline bottom
3836
water and the cooler, fresher bottom
3837
water passed the mooring, the
3838
relationship between the events and the
3839
direction of the current became complex.
3840
Bottom frontal structures are often not
3841
straight lines, but rather are crooked,
3842
sometimes following the bottom
3843
bathymetry. The low-pass filtered data
3844
show a relatively steady increase in
3845
stratification in late March (Fig. 9b, c),
3846
but the unfiltered data show that the
Figure 10. Hourly time series at M4 of a) temperature at 12
depths, b) hourly salinity at three depths, and hourly
current velocity at c) 11 m and d) 55 m. The 2008 time
period shown is the same as that indicated by the red bars
in Fig. 9. The blue shading indicates periods when the
bottom currents have a northward component. Northward
is upward.
160
3847
increase in stratification occurred as a series of events related to the tidal currents. That the
3848
surface temperatures remain near freezing is not surprising since there was ~100% ice cover in
3849
the 100-km x 100-km box around M4 keeping the upper water column cold and fresh.
3850
3851
3852
3.3.5. Two ice events, one stratified and one well mixed (Example: M5, 2005 – 2006)
During the cold season of 2005 – 2006 extensive ice arrived at M5 in December and finally
3853
departed in early June. Four periods of strong ice retreat occurred over this time (Fig. 11). The
3854
water column became fully mixed in early November at ~3.1 °C (not shown), and remained so
3855
until mid December when ice cover reached ~30%. Two distinct periods of > 90% ice
3856
concentration occurred. During the first ice event, late December through late February, the
3857
winds were relatively weak for the first 6 weeks and strengthened for the last 2 weeks. In
3858
contrast, during the second ice event, the winds were stronger during the first 5 weeks and weaker
3859
during the last 2 – 3 weeks. Ice, combined with low wind stress during the first ice event, cooled
3860
and freshened only the upper water column (upper
3861
24 – 42 m) without greatly impacting the deeper
3862
water. The sharp increase in bottom salinity in early
3863
February, and the accompanying increase in
3864
temperature was likely due to advection. Areal ice
3865
concentration fell to <30% in late February, and
3866
under the influence of relatively strong winds, the
3867
water mixed. The storms on February 12 with
3868
>50% ice cover had little impact mixing the water
3869
column, while a week later after the ice began
3870
retreating, winds of similar magnitude readily
3871
mixed the water column. A second large influx of
3872
ice occurred in mid-March. In sharp contrast to the
3873
first event, the water column remained largely well
3874
mixed through two months of 60-90% ice
3875
concentration. During this time the temperature (~ -
3876
1.8 °C) and the salinity (~ 31.8) remained relatively
3877
constant. That the water column temperature was at
3878
the freezing point with constant salinity supports the conclusion that there was little net ice melt
3879
during the second ice event.
Figure 11. Time series from M5, December 14,
2005 – May 31, 2006, showing a) wind stress, b)
temperature at 11 depths distributed through
the water column and the percent areal ice
cover in the 100-km x 100-km box around M5,
and c) salinity at three depths. The gray shading
indicates times when the water column was well
mixed.
161
3880
M5, similar to M4, is solidly within the ice conveyor belt (Fig. 1). Ice cover was often 80-
3881
100%, either steadily retaining a high concentration of ice all season, or partially retreating for a
3882
periods of days to weeks and returning to a high concentration. Areal ice concentration data at
3883
M4 show five years of the single ice event, and 10 years of multiple ice events for the period fall
3884
1994 – spring 2011. As such, the cold season of 2005 – 2006 was fairly typical at M5.
3885
3886
3887
3888
3.3.6. Months of ~100% ice cover (Example: M8, 2008 – 2009)
M8 is the farthest north of the four moorings (Fig. 1), and representative of the northern
3889
shelf (Stabeno et al., 2012a). The water column at M8 from 2005-2011 mixed notably later (early
3890
December to late January) and at colder temperatures (average -1.0 °C) than at M2, M4, or M5
3891
(Table 2). The average minimum winter temperature over those years was -1.8 °C, which was
3892
colder than at the southern moorings as a result of higher salinity. In 2008 the water column
3893
became well mixed at -0.1 °C in early December (Fig. 12). Sea ice arrived in late December and
3894
stratified the water column for a few days until the water column cooled to < -1.7 °C. The high
3895
areal ice cover (>90%) and a well-mixed or weakly stratified water column persisted for ~165
3896
days. Surface water began to warm
3897
and stratify ahead of May ice melt,
3898
and ice concentration was < 20% by
3899
the end of May, 2009.
3900
The patterns of strong pre-
3901
ice cooling, minimal ice melt,
3902
minimal decrease in water
3903
temperature, and areal ice
3904
concentrations >80% for more than 4
3905
months were typical at M8. The
3906
duration of ice cover (Fig. 12) was
3907
typical for the period (1995 – 2011)
3908
for the northern shelf. Over the 17-
3909
year period, ice was present an
3910
average of 127 days, with the longest
3911
single, near-100% ice concentration
3912
event lasting 189 days in 2005 –
3913
2006. Twelve years of the 17 years
Figure 12. Time series from M8, November 15, 2008 – June
30, 2009, showing a) wind stress, b) temperature at 11
depths distributed through the water column and the
percent areal ice cover in the 100-km x 100-km box around
M8, and c) salinity at three depths. The gray shading
indicates times when the water column was well mixed.
162
3914
examined had only one ice event (80 – 100% coverage) during the cold season, and the remaining
3915
five had two.
3916
3917
3918
3.4 Estimating the magnitude of ice melt and its direct influence on temperature
The thickness of the total ice melt can be estimated using the time series from the moorings
3919
and Equations (2) and (3). The largest problem with selecting data for this calculation was
3920
advection. Care was made in selecting examples where freshening was well defined and
3921
advection did not play an important role in modifying the water column. Of the 42 sets of winter
3922
and spring data (1995 – 2012), 15 sets were suitable for this calculation (Table 3). For each of
3923
these data sets the water column was well mixed at the time of the ice’s arrival; the water column
3924
(or at least the upper layer) cooled to the freezing point within a month of ice arrival; and there
3925
was measurable cooling and freshening of the water column. Advection did not appear to
3926
influence the system in these selected instances.
3927
Among the 14 annual examples (the two sets from M2, 2009 were combined together to
3928
provide total cooling for that year) most were from the southern moorings with only three from
3929
M5 and none from M8. The lack of examples from M8 was primarily because the water column
3930
was already at or near the freezing point when ice first appeared at that mooring site and any
3931
stratification resulting from the arrival of ice was minimal and short lived.
3932
From each of the examples, the average initial salinity (Sw) and initial ocean temperature (Tw)
3933
were determined (Table 3). After the arrival of ice, the water column often became stratified and
3934
then re-mixed to a depth of H. Usually it mixed to the bottom, but in certain instances only the
3935
surface was mixed before the retreat of ice. The change in salinity and temperature of this mixed
3936
layer was calculated. Using Equation (2) and an ice salinity of 6 (Table 1), hw was calculated.
3937
Using the mean density (0.89 x103 kg m-3) from Table 1 the thickness (hi) of ice melted was
3938
estimated for each data set, giving a mean thickness of 1.3 m. At -1.7 °C and a salinity of 6, the
3939
latent heat of fusion of sea ice (L) is 65 cal g-1 (Weeks and Weller, 1984). Using Eq. 3, the
3940
L)
was calculated for each example
3941
(Table 3). The latent heat of fusion accounted for an estimated 26-100 % of total observed
3942
cooling. On average almost half of the cooling of the water column was due solely to ice melt; the
3943
remainder was from air-sea interaction either directly or through the ice.
3944
A similar estimate can be made when ice melted in place in the spring (Table 4). At the
3945
southern moorings most ice retreated through advection. At the northern two moorings, however,
3946
often there was a sharp decrease in salinity associated with loss of sea ice. This decrease in
3947
salinity resulted in a fresh surface layer with a mean depth of 28 m (Table 4). This fresh water
163
3948
layer persists at M8 throughout the summer season (Stabeno et al., 2012a). The estimated amount
3949
of ice melted ranged from 0.31 m to 1.51 m, with an average of 0.84 m. The balance between
3950
atmospheric warming and cooling due to ice melt (latent heat) resulted in the upper-layer
3951
temperature remaining cold.
3952
In the winter, the arrival of ice freshened and cooled the southern shelf, while in spring the
3953
retreat of ice freshened the northern shelf. These estimates of ice melt are probably
3954
underestimates, because ice melt resulting from the advection of warmer water was not included
3955
and numerous minimal changes in temperature/salinity were also neglected. The estimation that
3956
>1.3 m of ice melted over the middle domain of the southern shelf could not be used to calculate
3957
the total ice melt over the southern shelf. Once the water column reached the freezing point, ice
3958
continued to be transported southward without modifying the water column and melted at the
3959
leading edge often over the slope.
3960
3961
3962
4. Summary and Conclusions
While variability in the winds resulted in sudden changes in the distribution and areal
3963
concentration of the ice over the cold season, the winds were primarily from the north resulting in
3964
the creation of polynyas along the leeward coasts and a south- or southwestward transport of sea
3965
ice over the Bering Sea shelf. The wind-driven conveyor belt of Pease (1980) originated north-
3966
northeast of St. Lawrence Island fanning south and westward over the shelf. This main corridor
3967
provided ice primarily to the areas around moorings M4 and M5. The St. Lawrence polynya and
3968
the region around Anadyr contributed ice to the area around M8. Finally, the Alaskan coastal
3969
polynyas in Bristol Bay and south of Nunivak Island were the primary suppliers to sea ice in the
3970
vicinity of M2.
3971
The moorings extend along the 70-m isobath of the middle shelf, with M8 on the northern
3972
shelf, M5 in the transition zone between north and south, and M2 and M4 on the southern shelf.
3973
Thus, examining data from these moorings gives insight on how the middle shelf responded
3974
spatially and temporally to ice melt. Over the southern part of the middle shelf (south of ~ 60
3975
°N; Stabeno et al., 2012a), the primary ice melt occurred when the ice was advected southward.
3976
Approximately 1.3 m of ice melted, freshening the water column and reducing the concentration
3977
of nutrients. Over the northern part of the middle shelf, as represented by M8, there was little ice
3978
melt with the first arrival of ice. Most of the melt at this site occurred when ice retreated, and ice
3979
melt during retreat left behind a fresh surface layer, which persisted at M8 through the summer.
3980
This increased the strength of stratification and delayed the fall mixing of the water column on
164
3981
the northern shelf by ~1.3 months relative to M2 (Table 2). The transition zone at M5 had
3982
substantial ice melt during both ice advance and retreat (a total ~2.1 m; Tables 3 and 4).
3983
Typically when ice first arrived, the water column (or the surface mixed layer) cooled
3984
and freshened, and remained well mixed. Once the areal ice cover exceeded 20 – 25 %, the
3985
melting ice stratified the water column, with a fresher, colder layer overlaying warmer, more
3986
saline water. This surface layer was usually > 20 m deep. The vertical stratification usually was
3987
t
= ~ 0.18 (using average temperature and salinity values from Table 3 of -1.7
3988
°C and 31.7, respectively, for surface layer, and 1.3 °C and 32.2 for bottom layer), but persisted
3989
for 10 - 25 days. Storms quickly re-mixed the water column if ice retreated (e.g. Fig. 8); storms
3990
had less influence under extensive ice, but the water column would eventually become well
3991
mixed (e.g. Figs. 7, 9 and 10). Advection plays a role in the stratification of the water column,
3992
particularly in the vicinity of the ice edge front.
3993
Without sea ice, the water column at M2 and M4 usually cools by ~4 °C to ~2 °C, while
3994
in years with extensive ice the water column cools on average more than 6 °C to -1.7 °C (Stabeno
3995
et al., 2012b). Multiple consecutive cold (warm) years modify this, making the depth-averaged
3996
summer temperatures colder (warmer). Average cooling in a series of cold years was ~5.5 °C,
3997
while average cooling in a series of warm years was ~4.5 °C (from Fig. 4A in Stabeno et al.,
3998
2012b). For comparison, more cooling (~ 7.5 °C) occurred between warm and cold years. The
3999
greater cooling during cold years rather than warm years comes from two sources. First, the air
4000
temperature during years with extensive ice is colder than years with little or no ice; thus, cooling
4001
due to sensible heat flux is greater. The second and probably more important source of cooling is
4002
the latent heat flux due to ice melt. The cooling from melting 1.3 m of ice (probably an
4003
underestimate) accounts for approximately half the observed cooling of the water column. In
4004
sharp contrast, on the northern shelf near M8, the water column cooled to ~ -1 °C before ice
4005
arrived (Table 2). Ice arrival cooled water temperatures to ~ -1.8 °C with little melt because water
4006
temperatures were already low.
4007
This discussion of sea-ice melt applies to the middle shelf domain, and not to the coastal
4008
or outer shelf domains. In the coastal domain, ice is formed and shore fast ice and rafting are
4009
common. The water column is typically well mixed by tides which are stronger on the
4010
southeastern shelf than in the north. In the outer shelf domain advection dominates, and intrusions
4011
of warmer, more saline slope water are common (e.g. Stabeno and van Meurs, 1999; Sullivan et
4012
al., 2008; Coachman, 1986). The water column is deeper (> 100 m) and not well mixed even in
4013
the winter (Stabeno et al., 1998). A few hypotheses, however, can be made about how the coastal
4014
and outer domains may respond to ice melt. Probably little ice is melted in the coastal domain
165
4015
until ice retreat at the end of the cold season. The coastal domain, with its weak alongshore flow,
4016
retains the signature of lower salinity through the summer as a result of ice melt (Danielson et al.,
4017
2012; Kachel et al., 2002). With the breakdown of the Inner Front in late summer, this freshwater
4018
signature spreads out over the shelf (Ladd and Stabeno, 2012). The signature of ice melt on the
4019
outer domain is ephemeral; it does not persist through the summer because of the persistent
4020
alongshore currents, and onshore intrusions of more saline, nutrient rich water (Stabeno et al.,
4021
2010).
4022
It is becoming evident that the presence or absence of sea ice in spring is the single most
4023
important component determining the physical and biological structure of the eastern Bering Sea
4024
shelf ecosystem, not only in the winter and spring when it is present, but also during the warm
4025
season (Stabeno et al., 2010, 2012a, b; Hunt et al., 2002, 2008, 2010). The advance of sea ice
4026
early in the season transports fresh water southward, which then melts at the ice edge reducing
4027
the salinity of the middle shelf by ~0.4 (Table 3). In contrast late in the cold season, the retreat of
4028
ice transports fresh water northward, where it melts providing approximately half of the summer
4029
stratification on the northern shelf (Stabeno et al., 2012a; Ladd and Stabeno, 2012). Sea ice had
4030
low concentrations of nitrate and, thus, could not contribute to the water column as the ice
4031
melted. There are significant amounts of iron, however, in the sea ice (Aguilar-Islas, 2008),
4032
which contributes to the production along the slope. Significant concentrations of chlorophyll-a
4033
were found in bottom sections of the ice cores. This ice algae, located at the ice-seawater
4034
interface, may provide an early concentrated food source (Mock and Gradinger, 2000; Niemi et
4035
al., 2011) for zooplankton (e.g. O’Brien 1987, Runge and Ingram 1991). While chlorophyll-a is
4036
transported by the ice toward the ice edge, it is not known if the ice also transports large amounts
4037
of zooplankton which feed on the ice algae. The absence of ice over the southern shelf during
4038
consecutive warm years (2001-2005) likely contributed, directly or indirectly, to a drop in
4039
populations of large crustacean zooplankton (e.g. Calanus marshallae, Thysanoessa raschii) and
4040
a resultant reduction in the pollock fishery (Hunt et al., 2008, 2010; Coyle et al., 2008, Stabeno et
4041
al., 2012b).
4042
4043
Acknowledgements
4044
We thank the officers and crews of the NOAA ships Miller Freeman and Oscar Dyson, U. S.
4045
Coast Guard Cutter Healy, and R/V Thomas G. Thompson for the invaluable assistance in making
4046
these oceanographic measurements. Special thanks go to the personnel who collected the ice
4047
cores including E. Cokelet, D. Kachel, P. Proctor, J. Cross, and members of the National Marine
4048
Mammal Lab during the 2006 ice cruise lead by M. Cameron and J. Bengston. We thank W.
166
4049
Floering, C. Dewitt and S. McKeever for maintaining, deploying and recovering the instruments
4050
on the mooring, D. Kachel and P. Proctor for processing the mooring and ship board
4051
measurements, P. Proctor for nutrient analysis on some samples, and A. Hermann for comparison
4052
of ice drift data to interpolated NCEP winds. Graphics were provided by K. Birchfield. This
4053
research was funded by grants from The North Pacific Research Board (grant B52) and the
4054
National Science Foundation (grants 0732430, 1107250, and 0813985), and support from
4055
NOAA’s North Pacific Climate Regimes and Ecosystem Productivity (NPCREP) program. This
4056
publication was partially funded by the University of Washington Joint Institute for the Study of
4057
the Atmosphere and Ocean (JISAO) under NOAA Cooperative Agreements NA17RJ1232 and
4058
NA10OAR4320148, and is contribution EcoFOCI-0802 to NOAA’s Ecosystems and Fisheries-
4059
Oceanography Coordinated Investigations, contribution 2157 to JISAO, contribution 3975 to
4060
NOAA’s Pacific Marine Environmental Laboratory, and BEST-BSIERP publication number XX.
4061
The findings and conclusions in this paper are those of the authors and do not necessarily
4062
represent the views of NOAA’s Oceans and Atmospheric Research.
4063
4064
References
4065
Aguilar-Islas, A.M., Rember, R.D, Mordy, C.W., Wu, J., 2008. Sea ice-derived dissolved iron
4066
and its potential influence on the spring algal bloom in the Bering Sea. Geophys. Res. Lett.
4067
35, http://dx.doi.org/10.1029/2008GL035736.
4068
4069
4070
Alexander, V., Niebauer, H.J., 1981. Oceanography of the eastern Bering Sea ice-edge zone in
spring, Limnol. Oceanogr. 26 (6), 1111–1125.
Brown, Z. W., Arrigo, K. R., 2013, Sea ice impacts on spring bloom dynamics and net primary
4071
production in the Eastern Bering Sea, J. Geophys. Res. J. Oceans, 118, 43–62,
4072
doi:10.1029/2012JC008034.
4073
Cavalieri, D.J., Markus, T., Comiso, J.C., 2003. AMSR-E/Aqua Daily L3 12.5 km Brightness
4074
Temperature, Sea Ice Concentration, & Snow Depth Polar Grids. Version 2. 2002-2011.
4075
Boulder, Colorado USA: NASA DAAC at the National Snow and Ice Data Center,
4076
http://nsidc.org/data/amsre/.
4077
4078
Coachman, L.K., 1986. Circulation, water masses and fluxes on the southeastern Bering Sea
shelf. Cont. Shelf Res. 5 (1–2), 23–108.
4079
Coyle, K. O., Pinchuk, A. I., Eisner, L. B., Napp, J. M., 2008. Zooplankton species composition,
4080
abundance and biomass on the eastern Bering Sea shelf during summer: the potential role of
4081
water column stability and nutrients in structuring the zooplankton community. Deep Sea
4082
Research II, 55: 1755–1791.
167
4083
Danielson, S., Aagaard, K., Weingartner, T., Martin, S., Winsor, P., Gawarkiewicz, G.,
4084
Quadfasel, D., 2006. The St. Lawrence polynya and the Bering shelf circulation: New
4085
observations that test the models, J. Geophys. Res., J111, C09023,
4086
doi:10.1029/2005JC003268.
4087
Danielson, S., Hedstrom, K., Aagaard, K., Weingartner, T., Curchitser, E., 2012. Wind-induced
4088
reorganization of the Bering Shelf circulation. Geophys. Res. Lett. 39 (8),
4089
http://dx.doi.org/10.1029/2012GL051231.
4090
Drucker, R., Martin, S., Moritz, R., 2003. Observations of ice thickness and frazil ice in the St.
4091
Lawrence Island polynya from satellite imagery, upward looking sonar, and
4092
salinity/temperature moorings, J. Geophys. Res., 108(C5), 3149, doi:10.1029/2001JC001213.
4093
Eicken, H., 2003. From the microscopic to the macroscopic to the regional scale: Growth,
4094
microstructure and properties of sea ice, in: Thomas, D.N., Dieckmann, G.S. (Eds.), Sea
4095
Ice—An Introduction to its Physics, Biology, Chemistry and Geology. Blackwells Scientific
4096
Ltd., London, pp. 22–81.
4097
Gordon, L.I., Jennings Jr, J.C., Ross, A.A., Krest, J.M., 1994. A suggested protocol for
4098
continuous flow automated analysis of seawater nutrients (phosphate, nitrate, nitrite and
4099
silicic acid) in the WOCE Hydrographic Program and the Joint Global Ocean Fluxes Study.
4100
WHP Operations and Methods. WOCE Hydrographic Program Office, Methods Manual 91-
4101
1. November.
4102
Holt, B., Rothrock, D.A., Kwok, R., 1992. Determination of sea ice motion from satellite images,
4103
in: Carsey, F.D. (Ed.), Microwave Remote Sensing of Sea Ice, Geophys. Monogr. Ser., Vol.
4104
68, AGU, Washington, D.C., doi:10.1029/GM068p0343, pp. 343–354.
4105
Hunt, Jr., G.L., B.M. Allen, R.P. Angliss, T. Baker, N. Bond, G. Buck, G.V. Byrd, K.O. Coyle,
4106
A. Devol, D.M. Eggers, L. Eisner, R.A. Feely, S. Fitzgerald, L.W. Fritz, E.V. Gritsay, C.
4107
Ladd, W. Lewis, J. Mathis, C.W. Mordy, F. Mueter, J. Napp, E. Sherr, D. Shull, P. Stabeno,
4108
M.A. Stepanenko, S. Strom, T.E. Whitledge (2010): Status and trends of the Bering Sea
4109
region, 2003-2008. In Marine Ecosystems of the North Pacific Ocean, 2003-2008, S. M.
4110
McKinnell and M. J. Dagg (eds.), PICES Special Publication 4, 196-267.
4111
4112
4113
Hunt Jr., G.L., Stabeno, P.J., 2002. Climate change and the control of energy flow in the
southeastern Bering Sea. Prog. Oceanogr. 55 (1–2), 5–22.
Kachel, N.B., Hunt Jr., G.L., Salo, S.A., Schumacher, J.D., Stabeno, P.J., Whitledge, T.E., 2002.
4114
Characteristics and variability of the inner front of the southeastern Bering Sea. Deep-Sea
4115
Res. II 49 (26), 5889–5909.
168
4116
Kanamitsu, M., Ebisuzaki, W., Woollen, J., Yang, S.-K., Hnilo, J.J., Fiorino, M., Potter, G.L.,
4117
2002. NCEP-DOE AMIP-II Reanalysis (R-2). Bull. Amer. Meteor. Soc. 83, 1631–1643.
4118
Ladd, C., Bond, N.A., 2002. Evaluation of the NCEP/NCAR reanalysis in the NE Pacific and the
4119
4120
4121
4122
Bering Sea. J. Geophys. Res., 107 (C10), 3158, doi: 10.1029/2001JC001157.
Ladd, C.A., Stabeno, P.J., 2012. Stratification on the Eastern Bering Sea shelf revisited. Deep-Sea
Res. II 65–70, 72–83.
Leberl, F., Raggam, J., Elachi, C., Campbell, W.J., 1983. Sea ice motion measurements from
4123
SEASAT SAR images, J. Geophys. Res. 88 (C3), 1915–1928,
4124
doi:10.1029/JC088iC03p01915.
4125
Leu, E., Søreide, J. E., Hessen, D. O., Falk-Petersen, S., Berge, J., 2011, Consequences of
4126
changing sea-ice cover for primary and secondary producers in the European Arctic shelf
4127
seas: timing, quantity, and quality. Prog. Oceanogr. 90(1), 18-32.
4128
4129
4130
4131
Macklin, S.A., Pease, C.H., Reynolds, R.M., 1984. Bering Air-Sea-Ice Study (BASICS),
February and March 1981. NOAA Tech Memo ERL PMEL-52, 104 pp.
Malmgren, F., 1927. On the Properties of Sea Ice, Norwegian North Polar Expedition with the
“Maud” 1918–1925: Scientific Results, Vol. 1, No. 5, 67 pp.
4132
Maslanik, J., Stroeve, J., 1999. Near-Real-Time DMSP SSM/I-SSMIS Daily Polar Gridded Sea
4133
Ice Concentrations. 1994-2002. (Updated daily.) Boulder, Colorado USA: NASA DAAC at
4134
the National Snow and Ice Data Center,
4135
http://nsidc.org/data/docs/daac/ssmi_instrument.gd.html.
4136
McNutt, S.L., 1981. Remote sensing analysis of ice growth and distribution in the Eastern Bering
4137
Sea. Chapter 10 in Eastern Bering Sea Shelf: Oceanography and Resources, D.W. Hood and
4138
J.A. Calder (eds.), Vol. 1, USDOC/NOAA/OMPA, 141–165.
4139
Michel, C., Legendre, L., Ingram, R. G., Gosselin, M., Levasseur, M., 1996. Carbon budget of
4140
sea-ice algae in spring: evidence of a significant transfer to zooplankton grazers. J. Geophys.
4141
Res. 101(C8), 18345-18
4142
Mock T., Gradinger R., 2000. Changes in photosynthetic carbon allocation in algal assemblages
4143
of Arctic sea ice with decreasing nutrient concentrations and irradiance. Mar. Ecol. Prog. Ser.
4144
202: 1-11.
4145
4146
4147
4148
Muench, R.D., Ahlnas, K., 1976. Ice movement and distribution in the Bering Sea from March to
June 1974. J. Geophys. Res. 81, 4469–4476.
Muench, R.D., Schumacher, J.D., Salo, S. A., 1988: Winter currents and hydrographic conditions
on the northern central Bering Sea shelf. J. Geophys. Res., 93 (C1), 516-526.
169
4149
4150
4151
Mueter, F.J., Litzow, M.A., 2008. Sea ice retreat alters the biogeography of the Bering Sea
continental shelf. Ecol. Appl. 18, 309–320, doi:10.1890/07-0564.1.
Mundy, C.J., et al., 2009. Contribution of under-ice primary production to an ice-edge upwelling
4152
phytoplankton bloom in the Canadian Beaufort Sea. Geophys. Res. Lett. 36, L17601,
4153
doi:10.1029/2009GL038837.
4154
Niebauer, H.J., 1998. Variability in Bering Sea ice cover as affected by a regime shift in the
4155
North Pacific in the period 1947–1996. J. Geophys. Res. 103(C12), 27,717–27,737.
4156
Niebauer, H.J., Bond, N.A., Yakunin, L.P., Plotnikov, V.V., 1999. An update on the climatology
4157
and sea ice of the Bering Sea, in: Loughlin, T.R., Ohtani, K. (Eds.), The Bering Sea: A
4158
Summary of Physical, Chemical and Biological Characteristics and a Synopsis of Research.
4159
North Pacific Marine Science Organization, PICES, Alaska Sea Grant, pp. 29–60.
4160
Niebauer, H.J., Schell, D.M., 1993. Physical environment of the Bering Sea population, in: Burns,
4161
J.J., Montague, J.J., Cowles, C.J. (Eds.), The Bowhead Whale. The Society for Marine
4162
Mammalogy Special Pub. No. 2, Seattle, Washington, pp. 23–43.
4163
4164
Niemi, A., Michel, C., Hille, K., Poulin, M. 2011 Protist assemblages in winter sea ice: setting the
stage for the spring ice algal bloom. Polar Biol. 34: 1803-1817.
4165
O'Brien, D.P. 1987. Direct observations of the behavior of Euphausia superba and Euphausia
4166
crystallorophias (Crustacea: Euphausiacea) under pack ice during the Antarctic spring of
4167
1985. J. Crustacean Biol. 7: 437-448.
4168
Pease, C.H., 1980. Eastern Bering Sea ice processes. Mon. Weather Rev. 108 (12), 2015–2023.
4169
Rodionov, S.N., Bond, N.A., Overland, J.E., 2007. The Aleutian Low, storm tracks, and winter
4170
4171
climate variability in the Bering Sea. Deep-Sea Res. II 54, 2560–2577.
Runge, J.A., Ingram, R.G. 1991. Under-ice feeding and diel migration by the planktonic
4172
copepods Calanus glacialis and Pseudocalanus minutus in relation to the ice algal production
4173
cycle in southeastern Hudson Bay, Canada. Mar. Biol. 108: 217-225
4174
Schandelmeier, L., Alexander, V., 1981. An analysis of the influence of ice on spring
4175
phytoplankton population structure in the southeast Bering Sea. Limnol. Oceanogr. 26 (5),
4176
935–943.
4177
Schumacher, J.D., Aagaard, K., Pease, C. H., Tripp, R.B., 1983. Effects of a shelf polynya on
4178
flow and water properties in the Northern Bering Sea. J. Geophys. Res., 88, 2723-2732;
4179
Schumacher, J.D., Kinder, T.H., Pashinski, D.J., Charnell, R.L., 1979. A structural front over the
4180
4181
4182
continental shelf of the eastern Bering Sea. J. Phys. Oceanogr. 9 (1), 79–87.
Shapiro, L.H., Burns, J.J., 1975. Satellite observations of sea ice movement in the Bering Strait
region, in: Weller, G., Bowling, S.A. (Eds.), Climate of the Arctic. Twenty-fourth Alaska
170
4183
Science Conference, Fairbanks, Alaska, August 15 to 17, 1973. Geophysical Institute,
4184
University of Alaska, Fairbanks, pp. 379–386.
4185
Sigler, M.F., Stabeno, P.J., Eisner, L.B., Napp, J.M., Mueter, F.J., this issue. Spring and fall
4186
phytoplankton blooms in a productive subarctic ecosystem, the eastern Bering Sea, during
4187
1955–2011.
4188
4189
Stabeno, P.J., Bond, N.A., Kachel, N.B., Salo, S.A., 2007. On the recent warming of the
southeastern Bering Sea shelf. Deep-Sea Res. II 54, 2599–2618.
4190
Stabeno, P.J., Bond, N.A., Kachel, N.B., Salo, S.A., Schumacher, J.D., 2001. On the temporal
4191
variability of the physical environment over the south-eastern Bering Sea. Fish. Oceanogr.
4192
10(1), 81–98.
4193
Stabeno, P.J., Farley Jr., E.V., Kachel, N.B., Moore, S., Mordy, C.W., Napp, J.M., Overland, J.E.,
4194
Pinchuk, A.I., Sigler, M.F., 2012a. A comparison of the physics of the northern and southern
4195
shelves of the eastern Bering Sea and some implications for the ecosystem, Deep-Sea Res. II
4196
65–70, 14–30.
4197
Stabeno, P.J., Kachel, N., Mordy, C., Righi, D., Salo, S., 2008. An examination of the physical
4198
variability around the Pribilof Islands in 2004. Deep-Sea Res. II 55 (16–17), doi:
4199
10.1016/j.dsr2.2008.03.006, 1701–1716.
4200
Stabeno, P.J., Kachel, N.B., Moore, S.E., Napp, J.M., Sigler, M., Yamaguchi, A., Zerbini, A.N.,
4201
2012b. Comparison of warm and cold years on the southeastern Bering Sea shelf and some
4202
implications for the ecosystem, Deep-Sea Res. II 65–70, 31–45.
4203
Stabeno, P.J., Napp, J., Mordy, C., Whitledge, T., 2010. Factors influencing physical structure
4204
and lower trophic levels of the eastern Bering Sea shelf in 2005: Sea ice, tides and winds.
4205
Prog. Oceanogr. 85 (3–4), 180–196.
4206
Stabeno, P.J., Schumacher, J.D., Davis, R.F., Napp, J.M., 1998. Under-ice observations of water
4207
column temperature, salinity and spring phytoplankton dynamics: Eastern Bering Sea shelf. J.
4208
Mar. Res. 56 (1), 239–255.
4209
4210
4211
4212
4213
4214
Stabeno, P.J., van Meurs, P., 1999. Evidence of episodic on-shelf flow in the southeastern Bering
Sea. J. Geophys. Res., 104 (C12), doi: 10.1029/1999JC900242, 29,715–29,720.
Strickland, J.D.H., Parsons, T.R., 1968. A Practical Handbook of Seawater Analysis, 1st Ed. Bull.
167, Fish. Res. Bd Can., Ottawa, 293 pp.
Stringer, W. J., Groves, J. E., 1991. Location and areal extent of polynyas in the Bering and
Chukchi Seas, Arctic, 44, 164-171.
171
4215
Sullivan, M.E., Kachel, N.B., Mordy, C.W., Stabeno, P.J., 2008. The Pribilof Islands:
4216
Temperature, salinity and nitrate during summer 2004. Deep-Sea Res. II 55 (16–17),
4217
http://dx.doi.org/10.1016/j.dsr2.2008.03.004, 1729–1737.
4218
4219
Timco, G.W., Frederking, R.M.W., 1996. A review of sea ice density. Cold Regions Science and
Technology 24 (1), 1–6.
4220
Weeks, W.F., Weller, G., 1984. Offshore oil in the Alaskan Arctic. Science 225, 371–378.
4221
Wyllie-Echeverria, T., Wooster, W.S., 1998. Year to year variations in Bering Sea ice cover and
4222
4223
4224
some consequences for fish distributions. Fisheries Oceanography 7, 159–170.
Zhang, J., Woodgate, R., Moritz, R., 2010. Sea ice response to atmospheric and oceanic forcing
in the Bering Sea. J. Phys. Oceanogr. 40 (8), 1729–1747.
4225
172
4226
4227
Tables
4228
4229
Table 1. Ice stations, 2006-2009, with average data from the core. April 19, 2007 Ice Core 2 ice
4230
thickness value was excluded from the mean thickness.
4231
Date
Ice
Snow
station
depth
(cm)
Ice
Air
Mean
thickness temp. salinity
(cm)
(°C)
Mean
Mean
Mean
temp.
ice
nitrate
(°C)
density
(µM)
(103 kg
m-3)
29-Apr-06
01 (D3)
2
50
-5.5
4.86
-1.72
0.89
0.4
30-Apr-06
02 (D4)
4
37
-5.9
7.23
-2.48
0.85
0.5
3-May-06
03 (E3)
27
100
-2.6
6.00
-1.58
0.97
2.2
17-Apr-07
1
1
51
-9.8
5.83
-3.75
0.91
0.3
19-Apr-07
2
--
4.05
-4.14
0.89
1.1
21-Apr-07
3
2
30
-1.0
--
-2.00
--
0.2
24-Apr-07
4
3
73
-1.2
5.79
-1.46
0.91
1.2
30-Apr-07
5
2
64
--
4.74
-1.39
0.89
0.4
6-May-07
6
2
62
-1.0
5.26
-1.43
0.90
0.1
7-May-07
7
0
69
2.6
4.00
-0.10
0.82
3.1
3-Apr-08
1
10
44
0.9
6.57
-2.23
0.93
1.2
4-Apr-08
2
43
92
-10.0
5.64
-3.44
0.89
1.1
6-Apr-08
4
8
55
-5.3
5.77
-2.56
0.89
1.1
8-Apr-08
5
8
41
-13.8
5.86
-3.20
0.92
3.5
8-Apr-08
5a
6
40
--
5.44
-3.25
0.91
1.5
12-Apr-08
6
4
76
-8.0
7.24
-2.86
0.83
1.0
15-Apr-08
8
17
85
--
--
-2.73
--
0.3
16-Apr-08
9
11
59
-1.1
--
-2.38
--
0.8
28-Apr-08
10
5
41
-2.5
6.34
-2.10
0.88
0.2
29-Apr-08
11
7
69
-4.0
6.04
-2.19
0.85
0.7
--
>329
173
30-Apr-08
12
5
71
-3.0
5.29
-2.13
0.88
2.1
5-Apr-09
1
4
73
1.5
5.70
-2.97
0.92
1.5
7-Apr-09
2
17
87
-0.2
5.05
-3.09
0.89
0.3
14-Apr-09
3
5
93
-10.6
6.39
-5.46
0.91
0.5
16-Apr-09
4
9
43
-4.2
7.21
-4.35
0.92
1.2
18-Apr-09
5
28
87
-5.3
4.59
-1.60
0.92
0.7
20-Apr-09
6
30
86
-3.6
5.63
-2.56
0.88
0.4
1-May-09
8
16
64
--
--
-1.76
--
1.4
2-May-09
9
4
42
-0.2
4.70
-0.60
0.84
0.3
3-May-09
10
21
127
-0.6
5.04
-1.78
0.86
0.3
MEAN
10
66
-3.8
5.62
-2.44
0.89
0.99
Std.
10
23
4.1
0.88
1.11
0.03
0.83
Deviation
4232
4233
4234
4235
4236
174
4237
Table 2. Bering Sea shelf moorings: statistics and information. The location of the moorings and
4238
the years during which they were deployed are given on the left. Mixing information includes the
4239
range of time, and average date, at which the water column became well mixed with a
4240
temperature difference of < 0.02 °C. The years during which areal sea ice concentration was <
4241
20% for the entire cold season were 1996, 2001, 2003 – 2005 at M2, and 2001 and 2005 at M4.
4242
Mooring sites M5 and M8 always had ice present during the cold season.
4243
Mooring
M2
56.9 °N, 164.1
°W
1995 – 2011
M4
57.9 °N,
168.9°W
1996, 19992011
M5
59.9 °N, 171.7
°W
2005 – 2011
M8
62.2 °N, 174.7
°W
2005 – 2011
Mean August Temp
2005 – 2011
(all years)
Fall Water Column Mixing Minimum
date and temperature
Temp.
2005 –
2011 (all
Date Range Temperature years)
(mean)
2005 – 2011 (°C)
(all years)
(°C)
Surface
(°C)
Bottom
(°C)
10.1
(10.4)
1.4
(2.0)
10/21 – 12/4
(Nov. 9)
3.6
(4.1)
-1.7
(-0.6)
9.4
(9.7)
1.2
(2.2)
10/18 –
11/24
(Nov. 1)
3.4
(4.4)
-1.7
(-1.1)
8.9
-0.8
10/16 –
12/21
(Nov. 20)
1.7
-1.7
5.0
-1.5
12/4 – 1/25
(Dec. 19)
-1.0
-1.8
4244
4245
4246
4247
4248
4249
4250
175
4251
4252
Table 3. Observations and calculations of the impact of melting ice on the water column. Tw (Sw)
4253
is the depth-
4254
the average change in temperature (salinity) in the part of the water column that was modified by
4255
melting ice. H is depth of the water column cooled (usually the entire water column), h i is the
4256
thickness of ice melted as calculated from changes in salinity, and hw is the equivalent depth of
4257
L is
the change in temperature due to latent heat of fusion and the final column is
4258
the fraction of the observed cooling that is caused by the latent heat of fusion. Note that 2009 at
4259
M2 had two distinct melting events (a and b) which were combined together (2009t).
4260
Mooring
Tw
Sw
H
hw
hi
– year
(°C)
(°C)
(m)
(m)
(m)
(°C)
M2 – 1998
2.0
3.4
31.95
0.45
25
0.43
0.51
1.17
0.34
2000
1.1
2.9
31.94
0.55
70
1.48
1.75
1.42
0.49
2006
2.6
4.3
32.12
0.71
27
0.73
0.86
1.82
0.43
2007
1.2
2.8
32.30
0.50
70
1.33
1.57
1.27
0.45
0.35
70
0.95
1.11
0.91
0.29
3.1
2009t
L
L
a
2.2
2.0
32.00
0.20
70
0.54
0.63
0.52
0.26
b
-0.6
1.1
31.92
0.15
70
0.41
0.48
0.39
0.35
2012
0.8
2.5
31.94
0.32
70
0.86
1.02
0.83
0.33
M4 – 2004
1.2
2.2
32.24
0.34
70
0.91
1.07
0.87
0.40
2006
2.5
3.8
32.38
0.65
70
1.72
2.03
1.65
0.43
2007
0.7
2.4
32.18
0.46
70
1.23
1.45
1.18
0.49
2009
1.6
3.5
32.11
0.45
70
1.21
1.42
1.15
0.33
2012
0.2
1.9
31.96
0.34
70
0.92
1.08
0.88
0.46
0.9
2.6
32.20
0.92
32
1.12
1.32
2.35
0.90
2007
2.0
3.6
32.20
0.50
70
1.34
1.57
1.28
0.36
2010
-0.1
1.8
31.55
0.70
70
1.92
2.25
1.83
1.00
1.15
1.36
1.33
0.48
M5 – 2006
Mean
2.9
0.52
4261
4262
4263
176
4264
Table 4. Statistics for the amount of ice that melts at moorings sites when the ice retreats.
4265
Column headings are the same as in Table 3.
4266
4267
Mooring - year
H
Sw
hw
hi
M5 - 2007
0.7
25
31.5
0.69
0.81
2008
0.7
25
31.4
0.69
0.81
2010
1.3
25
31.3
1.28
1.51
2011
0.3
30
31.4
0.35
0.42
2012
0.9
25
31.4
0.89
1.04
M8 - 2006
0.2
35
32.4
0.27
0.31
2008
0.7
32
32.8
0.84
0.98
Mean
0.7
28
31.7
0.71
0.84
4268
4269
4270
4271
4272
177
4273
4274
4275
4276
4277
4278
4279
4280
4281
4282
4283
4284
4285
4286
4287
4288
4289
Papers in preparation
178
4290
4291
4292
Chapter 16
Acoustically-Determined Patterns of Summer Zooplankton Biomass and Vertical Migration
4293
Behavior from the Bering Sea Middle Shelf during a Cold Regime
4294
4295
4296
Jeffrey M. Nappa, Adam H. Spearb, Phyllis .J. Stabenoc
4297
Benjamin Blossc and David Strauszc
4298
4299
4300
a
4301
7600 Sand Point Way NE, Seattle, WA 98115, USA; Jeff.Napp@NOAA.gov; Tel:+1-206-526-
4302
4148; Fax: +1-206-526-6723
Corresponding author: NOAA-Fisheries, Alaska Fisheries Science Center, RACE Division,
4303
4304
b
4305
Seattle, WA 98115, USA; Adam.Spear@NOAA.gov
NOAA-Fisheries, Alaska Fisheries Science Center, RACE Division, 7600 Sand Point Way NE,
4306
4307
c
4308
WA 98115, USA; Phyllis.Stabeno@NOAA.gov
4309
4310
4311
4312
Abstract
4313
variations in climate, and influence recruitment success of planktivorous fishes (e.g walleye
4314
pollock, Gadus chalcogramma). The broad continental shelf in this region is remote and not
4315
easily sampled by ships, particularly during cold winters and springs when sea ice covers much of
4316
the region. Two active acoustic sensors were moored over the eastern Bering Sea shelf during the
4317
summers of 2007, 2009, and 2010 and were used to gain insight into the temporal variability of
4318
summer zooplankton on this highly productive marine ecosystem. The first sensor, a 300 kHz
4319
Acoustic Doppler Current Profiler (ADCP) was intended for current measurements, sampled the
4320
entire water column, but was not calibrated for acoustic detection of small scatterers. The second
4321
sensor, a Tracor Acoustic Profiling System (TAPS) was calibrated, contained 8 unique
4322
frequencies between 100 and 3000 kHZ, but only sampled a very small volume. Volume
4323
backscatter data from the two samplers each provided indices of zooplankton biomass and
4324
distribution through different processing techniques with differing sets of assumptions. The
4325
temporal patterns of scattering by small particles varied both seasonally and among years, even
NOAA-Research, Pacific Marine Environmental Laboratory, 7600 Sand Point Way NE, Seattle,
Zooplankton biomass and distribution in the eastern Bering Sea are strongly connected to
179
4326
though all three years had similar environmental forcing. The backscatter volume from the ADCP
4327
and the estimated zooplankton biovolume from the TAPS were correlated. Acoustically-derived
4328
estimates of large copepod biovolume was highest in 2009 and 2010 and lowest in 2007 in
4329
agreement with other data. The diel signal dominated temporal variability and we were unable to
4330
detect additional modes of high variability at frequencies below 0.5 day-1. The TAPS-8 estimated
4331
sizes of the organisms responsible for diel vertical migration roughly matched our expectation of
4332
the identity of those organisms from nets and the literature. Length of time migrators spent in the
4333
surface waters was inversely related to the number of daylight hours and increased vulnerability
4334
to predation by age-0 fishes. This latter observation, while not new, provides an advantage to
4335
planktivores like pollock which must accumulate sufficient depot lipids by the end of summer to
4336
survive their first winter.
4337
4338
Key Words: Bering Sea, zooplankton, bioacoustics, diel vertical migration, walleye pollock,
4339
climate regimes
4340
4341
4342
4343
4344
4345
Figure 4. Eastern Bering Sea and M2 study site where the acoustic instruments collected
4346
zooplankton volume backscatter data.
4347
4348
180
4349
4350
4351
Figure 5. ADCP-derived zooplankton volume backscatter. Top – 2007, Middle – 2009, Bottom
4352
– 2010.
4353
4354
4355
181
4356
4357
4358
Chapter 17
Patterns of flow on the eastern Bering Sea shelf: 20 years of observations
4359
4360
P. J. Stabenoa and N. B. Kachelb
4361
4362
a
4363
Point Way NE, Seattle, WA 98115, USA; Phyllis.Stabeno@NOAA.gov
Corresponding author: cNOAA-Research, Pacific Marine Environmental Laboratory, 7600 Sand
4364
4365
b
4366
WA 98115, USA
JISAO/UW and Pacific Marine Environmental Laboratory, 7600 Sand Point Way NE, Seattle,
4367
4368
Abstract
4369
Currents are examined on the eastern Bering Sea shelf using data from moorings,
4370
satellite-tracked drifters and hydrographic surveys. The major source of water on the
4371
Bering Sea shelf is Unimak Pass and the canyons that intersect the shelf break. Absolute
4372
geostrophic transport through Unimak Pass varies from an average of 0.25 x 106 m3 s-1
4373
in the warm months to 0.43 x 106 m3 s-1 in the cold months. Flow along the 50-m isobath
4374
is weak, with a transport of < 0.1 x 106 m3 s-1 in both summer and fall. The transport
4375
along the 100-m isobath is more than twice that along 50-m isobath: 0.2 x 106 m3 s-1 in
4376
the summer at the Pribilof Islands; and a northward geostrophic transport (referenced to
4377
the bottom) of 0.31 x 106 m3 s-1 during spring and summer at 60°N. Northward transport
4378
along the 100-m and 50-m isobaths accounts for approximately half of the northward
4379
flow through Bering Strait. The transit time north from Unimak Pass to Bering Strait is
4380
greater than 13 months. Thus, the source of most of the heat transported into the Arctic
4381
through Bering Strait is a result of air-sea interactions in the Bering Sea. Examination of
4382
the currents and water properties on the southern shelf indicates that ~50% of the shelf
4383
water is exchanged with slope in October – January each year, replenishing >50% of the
4384
nutrients consumed the previous growing season.
4385
4386
Key Words: Bering Sea, currents, middle shelf, climate regimes
4387
182
Fig. 6. Mean currents
during summer (May –
September) derived from
satellite-tracked drifters
drogued at 45 m. There
are at least 10
independent estimates of
velocity in each grid cell.
4388
4389
4390
Fig. 7. Bottom currents at M2
divided into octents determined
by wind directions (e.g. for 0 –
45 °T the vector is the average
velocity when winds are toward
that direction). Note currents are
not 45 – 90 ° to the right of the
winds (lower left). Number of
points in each bin shown in
lower right.
4391
183
4392
4393
4394
Fig. 8. Bottom currents at M5 divided into octents determined by wind directions (e.g. for 0 – 45
4395
°T the vector is the average velocity when winds are toward that direction). Note currents are not
4396
45 – 90 ° to the right of the winds (middle panel). Number of points in each bin shown in
4397
bottom.
4398
184
4399
4400
4401
4402
4403
Figure 9. Relationship between winds and currents at M8. In the top panels, vectors
4404
indicate the mean direction of the flow when the winds are toward the octant indicated
4405
(e.g. 270-315). The bottom panels show current direction as a function of wind direction
4406
(black line) and the number of points in each direction octant (shaded). The period
4407
examined is winter when there is no ice for a) near-surface (~12-17 m), and b) the
4408
difference between near-surface and near-bottom flow.
185
4409
4410
Conclusions
Major results from moorings, accompanying CTD and the analyses that resulted from these
4411
findings.
4412
1.
Almost 18 years of data from M2 were used to identify warm and cold years over the
4413
southeastern Bering Sea. There is close relationship between ice extent in March and
4414
April and depth averaged temperature at M2, which permitted the extension of the
4415
definition of warm and cold back to 1972.
4416
2.
Prior to 2000, the amount of ice in March and April was highly variable, but beginning
4417
in spring 2000, the SE Bering Sea enter 5.5 year window of low ice extent and warm
4418
conditions. This was followed by 8+ year period (2007 – present) more extensive ice in
4419
March and April and very cold conditions on the southern shelf.
4420
3.
The types and abundance of zooplankton differed sharply between warm and cold years.
4421
This was especially true during the prolonged warm period (2001–2005) and cold period
4422
(2007–2010), and was less evident during the years of high interannual variability.
4423
During the warm period, there was a lack of large copepods and euphausiids over the
4424
shelf, however, their populations rebounded during cold period. Small crustacean
4425
zooplankton taxa did not appear to vary between and warm and cold years.
4426
4.
Contrary to historical views, there is a sharp division between the northern and southern
4427
Bering Sea shelves which occurs at ~60 °N. These differences include tidal strength,
4428
sharpness of the interface, cause of vertical stratification, presence of subsurface
4429
chlorphyll maxima, and influence of ice melt. These are discussed below.
4430
5.
The ocean currents differed between warm and cold years. During cold years, the
4431
monthly-mean currents over the southeastern shelf were largely westward, while in
4432
warm years the direction of currents was more variable, with northward flow during
4433
December–February and relatively weak flow during the remainder of the year.
4434
6.
Unlike the more coastal domain on the northern shelf where winds dominate the flow,
4435
the flow on the southern middle shelf is more complex. During the winter, the bottom
4436
currents at M2 are not always couple to the wind,. This could have important
4437
implication for transport of plankton, such as snow crabs larvae.
4438
7.
Temperature and salinity both influence the stratification of the Eastern Bering Sea
4439
shelf: temperature dominates stratification on the southern shelf, and temperature and
4440
salinity are equally important on the northern shelf. The strength of stratification does
4441
not vary as a function of warm or cold years.
4442
8.
Approximately, 1 m of ice melts over the southern shelf during the ice advance,
186
4443
modifying the salinity. A similar amount melts over the northern shelf with ice retreat.
4444
Ice cover inhibits the vertical mixing of the water column. Approximately 50% of the
4445
cooling of the southern shelf is caused by ice melt, with the remainder a result of air-sea
4446
interaction.
4447
9.
Chlorophyll a fluorescence data from the four biophysical moorings provided 37
4448
realizations of a spring bloom and 33 realizations of a fall bloom. Basic findings: If
4449
ice was present after mid-March, spring bloom timing was related to ice retreat
4450
timing; if ice was absent or retreated before mid-March, a spring bloom usually
4451
occurred in May or early June. A fall bloom also commonly occurred, usually in late
4452
September, and its timing was not significantly related to the timing of storms or fall
4453
water column overturn. The magnitudes of the spring and fall blooms were correlated.
4454
10.
The breadth of the mooring instrumentation from thermistors, conductivity cells,
4455
current meters and ADCPs, nitrate sensors, fluorometers, TAPS-8 to measure
4456
zooplankton biovolume, and acoustic systems to listen for marine mammals. Passive
4457
acoustic data provided information about sea ice conditions and detected and
4458
identified of vocalizing marine mammals; the acoustic backscatter provided
4459
information on relative zooplankton and fish abundance before, during, and after the
4460
retreat; and the hydrographic data confirmed the acoustic signal was associated with
4461
changing surface ice conditions. The combined information from the biophysical
4462
mooring sensors revealed changes in winter trophic level dynamics during the ice
4463
retreat, which would have otherwise been undetected by traditional ship-based
4464
observations.
4465
11.
The nitrate sensors and nutrient measurements around the moorings provided
4466
information of the seasonal draw down of nutrients associated with primary production,
4467
and winter replenishment. Nutrients are typically replenished on the shelf from
4468
November – January. Summer storms result in mixing of nutrients into the photic zone
4469
and a phytoplankton bloom 5 – 7 days later.
4470
12.
Subsurface blooms in the region around M8 are common during the summer.
4471
13.
The biophysical mooring data were frequently used to calibrate biophysical models of
4472
the Bering Sea. These included side-by-side model/data comparisons of the vertical
4473
profiles of currents, temperature, salinity, and chlorophyll-a at multiple locations, from
4474
hourly to interdecadal time scales.
4475
4476
BSIERP and Bering Sea Project Connections
187
4477
We found this to be a highly successful program. Collaborations among groups
4478
was extensive and the sharing of data and resources strongly to contributed to the
4479
success of this program. By sharing ship time each research group gained a first-
4480
hand understanding of the goals, needs and problems of other groups. The PI
4481
meetings were comprehensive and productive. The results of these collaborations
4482
are seen in the papers published and the fact that individuals from variety of groups
4483
co-author manuscripts. The success in the program is evident in the series of
4484
special issues that have come out of it.
4485
Ship time is always a problem in high latitudes. Participation in this program
4486
provided us the opportunity to collect data during the winter/spring when the
4487
Bering Sea is covered in ice. In turn, data and samples were collected on our late
4488
summer cruises, which extended the sampling period into late summer. This
4489
benefited the other scientists of this program.
4490
EcoFOCI had extensive physical, chemical and biological data (moorings and
4491
shipboard measurements) collected during the warm period. These data provided
4492
the ability to make critical comparisons between warm period (2001 – 2005) and
4493
the cold period sampled by Bering Sea Project. In addition, the long time series at
4494
the moorings (M2, M4, M5 and M8) provided context for many of the other
4495
measurements made as part of the Bering Sea Project.
4496
This is the best large integrated program that we have participated in.
4497
4498
188
4499
Management or policy implications
4500
Our long-term biophysical mooring array provides critical and highly cited measurements of the
4501
physical drivers of the southeastern Bering Sea ecosystem. These measurements are used by the
4502
North Pacific Fishery Management Council (Fig. 10) to inform stock assessments and guide total
4503
allowable catch for some of the most economically important fisheries in the U.S. For example,
4504
during a recent warm phase in the Bering Sea (2000 – 2005) pollock recruitment was extremely
4505
poor and assessments of pollock stocks and environmental conditions warned of possible
4506
continued declines.
4507
ecosystem considerations advice by changing existing management practices – reducing pollock
4508
harvest levels. This occurred during a period when the pollock population was facing a variety of
4509
environmental stressors, and potentially averted a collapse of the fish stock and the fisheries,
4510
communities and economies that depend on it.
Resource managers responded to the combined stock assessment and
4511
4512
4513
4514
4515
Figure 10. Schematic of the pathway for management action for walleye pollock.
4516
4517
189
4518
4519
4520
4521
Publications
4522
Jeffrey M. Napp, J. M., A. H. Spear, P. .J. Stabeno, B. Bloss and D. Strausz, Acoustically-
In Preparation
4523
Determined Patterns of Summer Zooplankton Biomass and Vertical Migration Behavior
4524
from the Bering Sea Middle Shelf during a Cold Regime, For submission to the fourth
4525
special issue.
4526
4527
4528
Stabeno, P. J. and N. B. Kachel, Patterns of flow on the eastern Bering Sea shelf: 20 years of
data. For submission to the fourth special issue.
4529
4530
Submitted/in press
4531
4532
Cheng, W., E. Curchitser, C. Ladd, and P.J. Stabeno (2013): Ice–ocean interactions in the eastern
4533
Bering Sea: NCAR CESM simulations and comparison with observations. Deep-Sea Res. II.
4534
4535
Cross, J.N., J.T. Mathis, M.W. Lomas, S.B. Moran, J. Grebmeier, D. Shull, C.W. Mordy, P.J.
4536
Stabeno, R. Sambrotto, R. Gradinger, L. Juranek, M. Prokopenko, M.S. Baumann, and M.
4537
Ostendorf (2013): Integrated assessment of the carbon budget in the Southeastern Bering
4538
Sea. Deep-Sea Res. II.
4539
4540
Goes, J.I., H. Rosario Gomes, E. D'Sa, D. Stoecker, C.W. Mordy, R. Sambrotto, and P.J. Stabeno
4541
(2013): Distribution of phytoplankton communities in the Bering Sea during the summer of
4542
2008 that followed a cold spring. Deep-Sea Res. II.
4543
4544
Napp, J.M., R.D. Brodeur, D. Demer, R. Hewitt, P.J. Stabeno, and G.L. Hunt, Jr. (2008):
4545
Acoustic determination of nekton and zooplankton distributions at tidally generated shelf
4546
fronts in the eastern Bering Sea. Mar. Ecol. Prog. Ser.
4547
4548
Prokopenko, M.G., J. Granger, M. Long, C. Mordy, E. Barkan, N. Cassar, E. Cokelet, P. DiFiore,
4549
C. Ladd, R. Sambrotto, and B. Moran (2013): Mixed layer depth controls the instantaneous
4550
Net to Gross Production ratios and potential export efficiency in spring blooms on the
4551
eastern Bering Sea shelf. Deep-Sea Res. II.
4552
190
4553
Sigler, M.F., P.J. Stabeno, L.B. Eisner, and J.M. Napp (2012): Spring and fall phytoplankton
4554
blooms in a productive subarctic ecosystem, the eastern Bering Sea, during 1995-2011.
4555
Deep-Sea Res. II.
4556
4557
Stauffer, B., J.I. Goes, H. Rosario Gomes, K. McGee, and P.J. Stabeno (2013): Interannual
4558
variability of spring-time phytoplankton community composition across the continental shelf
4559
of the southeastern Bering Sea. Deep-Sea Res. II.
4560
4561
4562
Sullivan, M., C. Mordy, S. Salo, and P.J. Stabeno (2013): The influence of sea ice on seasonal
water column changes on the eastern Bering Sea shelf. Deep-Sea Res. II.
4563
4564
Published
4565
4566
Cooper, L.W., M. Sexson, J.M. Grebmeier, R. Gradinger, C.W. Mordy, and J.R. Lovvorn (2013):
4567
Linkages between sea-ice coverage, pelagic-benthic coupling, and the distribution of
4568
spectacled eiders: Observations in March 2008, 2009 and 2010, Northern Bering Sea. Deep-
4569
Sea Res. II, doi: 10.1016/j.dsr2.2013.03.009.
4570
4571
Coyle, K.O., A. I. Pinchuk, L. B. Eisner, J. M. Napp (2008). Zooplankton species composition,
4572
abundance and biomass on the eastern Bering Sea shelf during summer: The potential role of
4573
water-column stability and nutrients in structuring the zooplankton community. Deep-Sea
4574
Res. II. doi:10.1016/j.dsr2.2008.04.029.
4575
4576
Danielson, S., E. Curchitser, K. Hedstrom, T. Weingartner, and P. Stabeno (2011): On ocean and
4577
sea ice modes of variability in the Bering Sea. J. Geophys. Res., 116(C12), C12034, doi:
4578
10.1029/2011JC007389.
4579
4580
Granger, J, M.G. Prokopenko, C.W. Mordy, and D.M. Sigman (2013): The proportion of
4581
remineralized nitrate on the ice-covered eastern Bering Sea shelf evidenced from the oxygen
4582
isotope ratio of nitrate. Global Biogeochem. Cycles, doi: 10.1002/gbc.20075.
4583
4584
Hermann, A.J., G.A. Gibson, N.A. Bond, E.N. Curchitser, K. Hedstrom, W. Cheng, M. Wang,
4585
P.J. Stabeno, L. Eisner, and K.D. Cieciel (2013): A multivariate analysis of observed and
191
4586
modeled biophysical variability on the Bering Sea shelf: Multidecadal hindcasts (1970-
4587
2009) and forecasts (2010-2040). Deep-Sea Res. II, doi: 10.1016/j.dsr2.2013.04.007.
4588
4589
Hunt, Jr., G.L., B.M. Allen, R.P. Angliss, T. Baker, N. Bond, G. Buck, G.V. Byrd, K.O. Coyle,
4590
A. Devol, D.M. Eggers, L. Eisner, R.A. Feely, S. Fitzgerald, L.W. Fritz, E.V. Gritsay, C.
4591
Ladd, W. Lewis, J. Mathis, C.W. Mordy, F. Mueter, J. Napp, E. Sherr, D. Shull, P. Stabeno,
4592
M.A. Stepanenko, S. Strom, and T.E. Whitledge (2010): Status and trends of the Bering Sea
4593
region, 2003-2008. In Marine Ecosystems of the North Pacific Ocean, 2003-2008, S. M.
4594
McKinnell and M. J. Dagg (eds.), PICES Special Publication 4, 196-267.
4595
4596
Hunt, Jr., G. L., K.O. Coyle, L. Eisner, E.V. Farley, R. Heintz, F. Mueter, J.M. Napp, J. E.
4597
Overland, P.H. Ressler, S. Salo, and P. J. Stabeno (2011): Climate impacts on eastern Bering
4598
Sea foodwebs: A synthesis of new data and an assessment of the Oscillating Control
4599
Hypothesis. ICES J. Mar. Sci., 68(6), doi: 10.1093/icesjms/fsr036, 1230–1243.
4600
4601
Hunt, Jr., G.L., P.J. Stabeno, S. Strom, and J.M. Napp (2008): Patterns of spatial and temporal
4602
variation in the marine ecosystem of the southeastern Bering Sea, with special reference to
4603
the Pribilof Domain. Deep-Sea Res. II, 55(16–17), doi: 10.1016/j.dsr2.2008.04.032, 1919–
4604
1944.
4605
4606
4607
Ladd, C., and P.J. Stabeno (2009): Freshwater transport from the Pacific to the Bering Sea
through Amukta Pass. Geophys. Res. Lett., 36, L14608, doi: 10.1029/2009GL039095.
4608
4609
4610
Ladd, C., and P.J. Stabeno (2012): Stratification on the Eastern Bering Sea Shelf revisited. DeepSea Res. II, 65–70, doi: 10.1016/j.dsr2.2012.02.009, 72–83.
4611
4612
4613
Ladd, C., P.J. Stabeno, and J.E. O'Hern (2012): Observations of a Pribilof eddy. Deep-Sea Res. I,
66, doi: 10.1016/j.dsr.2012.04.003, 67–76.
4614
4615
Mathis, J.T., J.N. Cross, N.R. Bates, S.B. Moran, M.W. Lomas, C.W. Mordy, and P.J. Stabeno
4616
(2010): Seasonal distribution of dissolved inorganic carbon and net community production
4617
on the Bering Sea shelf. Biogeosciences, 7, doi: 10.5194/bg-7-1769-2010, 1769–1787.
4618
192
4619
Mathis, J.T., J.N. Cross, N. Monacci, R.A. Feely, and P.J. Stabeno (2013): Evidence of prolonged
4620
aragonite undersaturations in the bottom waters of the southern Bering Sea shelf from
4621
autonomous sensors. Deep-Sea Res. II, doi: 10.1016/j.dsr2.2013.07.019.
4622
4623
Miksis-Olds, J.L., P.J. Stabeno, J.M. Napp, A.I. Pinchuk, J.A. Nystuen, J.D. Warren, and S.L.
4624
Denes (2013): Ecosystem response to a temporary sea ice retreat in the Bering Sea: Winter
4625
2009. Prog. Oceanogr., 111, doi: 10.1016/j.pocean.2012.10.010, 38–51.
4626
4627
Mordy, C.W., E.D. Cokelet, C. Ladd, F.A. Menzia, P. Proctor, P.J. Stabeno, and E. Wisegarver
4628
(2012): Net community production on the middle shelf of the Eastern Bering Sea. Deep-Sea
4629
Res. II, 65–70, doi: 10.1016/j.dsr2.2012.02.012, 110–125.
4630
4631
Mordy, C.W., L. Eisner, P. Proctor, P.J. Stabeno, A. Devol, D.H. Shull, J.M. Napp, and T.
4632
Whitledge (2010): Temporary uncoupling of the marine nitrogen cycle: Accumulation of
4633
nitrite on the Bering Sea Shelf. Mar. Chem., 121, doi: 10.1016/j.marchem.2010.04.004,
4634
157–166.
4635
4636
Nystuen, J.A., S.E. Moore, and P.J. Stabeno (2010): A sound budget for the southeastern Bering
4637
Sea: Measuring wind, rainfall, shipping, and other sources of underwater sound. J. Acoust.
4638
Soc. Am., 128(1), doi: 10.1121/1.3436547, 58−65.
4639
4640
Sigler, M.F., H.R. Harvey, C.J. Ashjian, M.W. Lomas, J.M. Napp, P.J. Stabeno, and T.I. Van Pelt
4641
(2010): How does climate change affect the Bering Sea ecosystem? Eos Trans. AGU,
4642
91(48), doi: 10.1029/2010EO480001, 457–458.
4643
4644
Stabeno, P.J., E. Farley, N. Kachel, S. Moore, C. Mordy, J.M. Napp, J.E. Overland, A.I. Pinchuk,
4645
and M.F. Sigler (2012): A comparison of the physics of the northern and southern shelves of
4646
the eastern Bering Sea and some implications for the ecosystem. Deep-Sea Res. II, 65–70,
4647
doi: 10.1016/j.dsr2.2012.02.019, 14–30.
4648
4649
Stabeno, P.J., N.B. Kachel, S.E. Moore, J.M. Napp, M. Sigler, A. Yamaguchi, and A.N. Zerbini
4650
(2012): Comparison of warm and cold years on the southeastern Bering Sea shelf and some
4651
implications for the ecosystem. Deep-Sea Res. II, 65–70, doi: 10.1016/j.dsr2.2012.02.020,
4652
31–45.
193
4653
4654
Stabeno, P.J., J. Napp, C. Mordy, and T. Whitledge (2010): Factors influencing physical structure
4655
and lower trophic levels of the eastern Bering Sea shelf in 2005: Sea ice, tides and winds.
4656
Prog. Oceanogr., 85(3–4), doi: 10.1016/j.pocean.2010.02.010, 180–196.
4657
4658
Stafford, K.M., S.E. Moore, P.J. Stabeno, D.V. Holliday, J.M. Napp, and D.K. Mellinger (2010):
4659
Biophysical ocean observation in the southeastern Bering Sea. Geophys. Res. Lett., 37,
4660
L02606, doi: 10.1029/2009GL040724.
4661
4662
4663
Outreach
4664
Web-page
4665
EcoFOCI maintains a suite of NPRB related web sites that can found at
4666
http://www.pmel.noaa.gov/foci/.
4667
4668
Schools, mentoring, radio and press
4669
Carol Ladd made a presentation to four 5thgrade classrooms at Sunset Elementary, Issaquah
4670
4671
4672
4673
School District, about the Bering Sea Ice Expedition 2008. May 22, 2008.
Ned Cokelet contributed to a public presentation after the spring BEST cruise in May 2009 and
discussed with the public goals and findings of the research.
Members of PMEL/Eco-FOCI collaborated with Aleut Community of St. Paul Island in
4674
maintaining their temperature and salinity measurements in their harbor and on St. George
4675
Island from 2009-2012.
4676
4677
4678
4679
4680
4681
4682
4683
4684
4685
Nancy and Dave Kachel gave Science Day Presentations about BEST/BSIERP at Jane Adams K8 School, Seattle, WA. December 8, 2010.
Nancy and Dave Kachel gave Science Day Presentations about BEST/BSIERP at Jane Adams K8 School, Seattle, November 30, 2011.
Lisa Guy, Calvin Mordy and Scott McKeever. Satellite-Tracked Drifters, Polar Science Weekend
at the Pacific Science Center, March 1-2, 2012. Seattle, WA.
Scott McKeever. Volunteered as a judge at the Orca Bowl, School of Oceanography, University
of Washington, March, 2012.
Nancy and Dave Kachel participated as Science Fair Judges at Jane Adams K-8 School, Seattle,
WA. April 2012.
194
4686
4687
4688
Margaret Sullivan, Speed Networking, Young Leadership Program/NOAA Science Camp,
Seattle, WA. July 2012.
Guy, Lisa. Press Release titled "New Bering Sea research reveals how changing ecosystem
4689
impacts America's most valuable fisheries" Link to article:
4690
http://www.noaanews.noaa.gov/stories2012/20120530_beringsea.html
4691
4692
Guy,Lisa N.Kachel and Scott McKeever provided outreach to Sea Grant Fellow Kristen
Jabanoski via in-person discussion, slide presentation and a tour on 29 January 2013.
4693
Nancy and Dave Kachel gave Science Day Presentations about BEST/BSIERP at Jane Adams K-
4694
8 School, Seattle, Kachel, N. Science Day Presentations at Jane Adams K-8 School, Seattle,
4695
WA. Feb 6, 2013.
4696
Margaret Sullivan provided EcoFOCI animations to Danielle Dickson (NPRB) for a presentation.
4697
The animation shows Bering Sea satellite ice (Modis true-color) with ice-floe tracking (dots)
4698
for 2003 (warm year, less ice) and 2007 (cold year, more ice, 20 Feb 2013.
4699
4700
4701
4702
4703
4704
4705
Scott McKeever. Volunteered as a judge at the Orca Bowl, School of Oceanography, University
of Washington, February 23, 2013.
Nancy and Dave Kachel participated as Science Fair Judges at Jane Adams K-8 School, Seattle,
WA. April 2013.
Nancy Kachel and Jeff Napp made presentations and Scott McKeever gave a facilities tour for
undergraduate Seattle University Marine Science class on 22 May 2013.
Carol Ladd, is a mentor for the MPOWIR program, providing mentoring to physical
4706
oceanographers from late graduate school through their early careers. The overall goal of
4707
MPOWIR is to make mentoring opportunities for junior physical oceanographers universally
4708
available and of higher quality by expanding the reach beyond individual home
4709
institutions. The aim is to reduce the barriers to career development for all junior scientists
4710
in the field, with a particular focus on improving the retention of junior women.
4711
Carol Ladd, Hosted a high school senior (Zach Hall) to discuss a career in science (6/28/2012).
4712
Jeff Napp and Phyllis Stabeno both served on the Science Advisory Board.
4713
4714
Conference Presentations
4715
Stabeno, P., J. Napp, C. Mordy and T. Whitledge, Southeast Bering Sea shelf: a contrast between
4716
a warm year and a cool year. Alaska Marine Science Symposium, Anchorage Alaska.
4717
January 2008.
4718
4719
Ladd, C. North Pacific meets the Bering Sea: transport through Amukta Pass. Alaska Marine
Science Symposium, Anchorage Alaska. January 2008.
195
4720
4721
Stabeno, P. and J. Napp. EcoFOCI: The Bering Sea ecosystem. PMEL Site Review. Seattle,
Washington. August 2008.
4722
Stabeno, P.J., J.M.Napp, T.E.Whitledge, and C. Mordy. The influence of seasonal sea ice on the
4723
eastern Bering Sea shelf ecosystem. AGU / ASLO Ocean Sciences Meeting, Orlando,
4724
Florida, February 2008.
4725
Napp, J. and P. Stabeno. Bottom-up control of eastern Bering Sea ecosystem: implications for
4726
integrated ecosystem assessment during a period of climate change. Library Lunch Seminar.
4727
Washington DC.
4728
4729
4730
4731
4732
Napp, J. The Bering Sea: Current Status and Recent Events. PICES Press, 16 (1), January 2008.
p. 31- 33.
Napp, J. The Bering Sea: Current Status and Recent Events. PICES Press, 16 (2), July 2008. p.
30-31.
Stabeno, P. J., N. Kachel, C. Mordy, D. Moore and J. Napp. Bering Sea Observing Network:
4733
Physics to Whales. Poster presented at OceanObs09, Venice, Italy, September 2009.
4734
Buckley, T.W., TenBrink, T., Napp, J. and Aydin, K. “Comparing zooplankton consumed by
4735
walleye pollock and caught by nets in the eastern Bering Sea.” AGU/ASLO Ocean Sciences
4736
Meeting, Portland, OR, February 2010.
4737
Cross, J. N.; J. T. Mathis; N. R. Bates; B. Moran; M. W. Lomas; P. J. Stabeno Seasonal
4738
Distribution and Controls on Dissolved Inorganic Carbon and Net Community Production
4739
on the Eastern Bering Sea Shelf. 2010 Ocean Science Meeting Portland OR
4740
Hermann, A. J.; K. Aydin; N. A. Bond; W. Cheng; E. N. Curchitser; G. A. Gibson; K. Hedstrom;
4741
I. Ortiz; M. Wang; P. J. Stabeno The Bering Sea Integrated Ecosystem Research Program:
4742
downscaling climate change to a subarctic region with coupled biophysical models. 2010
4743
Ocean Science Meeting Portland OR
4744
4745
Kachel, N. B.; C. W. Mordy; P. J. Stabeno Comparison of four decades of hydrographic structure
over the SE Bering Sea Shelf. 2010 Ocean Science Meeting Portland OR
4746
Mordy, C. W.; C. Ladd; M. W. Lomas; R. Sambrotto; P. J. Stabeno. The Annual Nutrient Cycle
4747
on the Bering Sea Shelf: Seasonal Drawdown vs. Winter Replenishment. 2010 Ocean
4748
Science Meeting, Portland OR
4749
Napp, J.M., Eisner, L.B., Farley, E.W., Stabeno, P.J., Hunt, Jr., G.L. Recent changes in
4750
zooplankton abundance and biomass in the eastern Bering Sea, Alaska Marine Science
4751
Symposium, Anchorage, AK, January 2010.
4752
4753
Napp, J.M., Sigler, M.F., and Stabeno, P.J. Understanding ecosystem processes in the Bering Sea.
NOAA Headquarters Lunch Seminar, Silver Spring, MD, March 2010.
196
4754
4755
Salo, S. A and P. J. Stabeno, Spatial Scales of Correlation on the Bering Sea Middle Shelf, 2010
Ocean Science Meeting Portland OR
4756
Stabeno, P. J.; N. B. Kachel; C. Mordy; J. M. Napp; M. C. Spillane, The north - south structure of
4757
eastern Bering Sea shelf ? Implications for a changing ecosystem. 2010 Ocean Science
4758
Meeting Portland OR
4759
Stabeno, P.J. N. A. Bond, C. M. Mordy, and J. M. Napp. Comparison between a cold year (2009)
4760
and a warm year (2005) and implications for the ecosystem on response to changes in
4761
climate, Alaska Marine Science Symposium, Cook Hotel, Anchorage, Alaska.
4762
4763
4764
K. Stafford; S. E. Moore; P. J. Stabeno; V. Holliday; J. M. Napp. Biophysical Ocean Observation
in the Southeastern Bering Sea. 2010 Ocean Science Meeting Portland OR
Sullivan, M. E.; C. W. Mordy; P. J. Stabeno; N. B. Kachel; D. G. Kachel; S. A. Salo A Brief
4765
History of an Ecosystem Spring Setup as Seen in Annual Ice Cover. 2010 Ocean Science
4766
Meeting Portland OR.
4767
Hunt, G.L. Jr., Eisner, L.B., Farley, E.W., Moss, J.H., Napp, J.M., and Stabeno, P.J. “The
4768
oscillating control hypothesis: Reassessment in view of new information from the eastern
4769
Bering Sea.” PICES/ICES Climate Change Effects on Fish and Fisheries, Sendai Japan,
4770
April, 2010.
4771
Napp, J.M., Eisner, L.B., Farley, E.W., Stabeno, P.J., Hunt, Jr., G.L. “Recent changes in
4772
zooplankton abundance and biomass in the eastern Bering Sea.” Alaska Marine Science
4773
Symposium, Anchorage, AK, January 2010.
4774
4775
4776
Napp, J.M., Sigler, M., and Stabeno, P.J. ”Understanding ecosystem processes in the Bering Sea.”
NOAA Headquarters, Silver Spring, MD, March 2010.
Napp, J.M., Hunt, Jr., G.L., Eisner, L.B., Farley, E.W., Stabeno, P.J., Andrews, A., and
4777
Yamaguchi, A. The response of eastern Bering Sea zooplankton communities to climate
4778
fluctuations: Community structure, biodiversity, and energy flow to higher trophic levels.
4779
PICES/ICES Climate Change Effects on Fish and Fisheries, Sendai Japan, April, 2010.
4780
Stabeno, P.J., Bond, N.A., Napp, J.M., and Mordy, C.W. “Comparison between a cold year
4781
(2009) and a warm year (2005) and implications for the ecosystem on response to changes in
4782
climate. Alaska Marine Science Symposium, Anchorage, AK, January 2010.
4783
Bond, N.A. “Climate and modeling the Bering Sea Ecosystem”. Bering Sea Ecosystem
4784
4785
4786
4787
Professional Development Workshop, Anchorage, Alaska. 11-15 October 2010.
Bond, N.A., Stabeno, P.J., Hermann, A.J., and M. Wang. “What controls the extent of Bering Sea
Ice in spring?” PICES Annual Meeting, Portland, Oregon. 25-29 October 2010.
Hermann, A.J., Aydin, K., Bond, N.A., Cheng, W., Curchitser, E.N., Gibson, G.A., Hedstrom, K.,
197
4788
Ortiz, I., Wang, M., and Stabeno, P.J. “Modes of biophysical variability on the Bering Sea
4789
shelf”. PICES Annual Meeting, Portland, Oregon. 25-29 October 2010.
4790
Hinckley, S., Burgos, J., Szuwalski, C., Parada, C., Orensanz, J.M., Punt, A., Armstrong, D.,
4791
Ernst, B., Hermann, A., Megrey, B., and J. Napp. Can we use a coupled biophysical IBM
4792
model experiment to predict recruitment of snow crab in the Bering Sea – The Gate
4793
Hypothesis. Alaska Marine Science Symposium, Anchorage, AK, January 2011.
4794
Hunt, G.L. Jr., Coyle, K.O., Eisner, L., Farley, E.V., Heintz, R., Mueter, F., Napp, J.M.,
4795
Overland, J.E., Ressler, P.H., Salo, S., and P.J. Stabeno. “Climate impacts on eastern Bering
4796
Sea food webs: A synthesis of new data and an assessment of the Oscillating Control
4797
Hypothesis. PICES Annual Meeting, Portland, Oregon. 25-29 October 2010.
4798
4799
4800
4801
4802
Ladd, C., Stabeno, P., and O’Hern, J. “The Pribilof eddy in the eastern Bering Sea”. PICES
Annual Meeting, Portland, Oregon. 25-29 October 2010.
Ladd, C., Stabeno, P.J., and J. O’Hern. “The Pribilof eddy in the eastern Bering Sea”. Alaska
Marine Science Symposium, Anchorage, AK, January 2011.
Mordy, C., Ladd, C., Menzia, F.A., Nabor, D., Proctor, P., Stabeno, P., Whitledge, T., and E.
4803
Wisegarver. Temporal and spatial variability of primary production on the eastern Bering
4804
Sea Shelf. Alaska Marine Science Symposium, Anchorage, AK, January 2011.
4805
Napp, J.M., Ashjian, C., Harvey, R., Lomas, M., Sigler, M., and Stabeno, P.J. ”Understanding
4806
ecosystem processes in the Bering Sea.” PICES Annual Meeting, Portland, Oregon. 25-29
4807
October 2010.
4808
Stabeno, P.J., Bond, N.A., and J.M. Napp. “Eastern Bering Sea shelf: Comparison between a cold
4809
period (2007-2010) and a warm period (2001-2005). PICES Annual Meeting, Portland,
4810
Oregon. 25-29 October 2010.
4811
4812
4813
Stabeno, P.J. and J.M. Napp. A continuation of cold conditions in the Bering Sea: 2010. Alaska
Marine Science Symposium, Anchorage, AK, January 2011.
Wang, M., Overland, J.E., and P. Stabeno. Examples of using global climate models for Bering
4814
Sea marine ecosystem projection. Alaska Marine Science Symposium, Anchorage, AK,
4815
January 2011.
4816
Napp, J.M., Eisner, L.B., Farley, E.V., Mier, K.L., Pinchuk, A.I., and Stabeno, P.J. ”North-south
4817
variation in eastern Bering Sea shelf spring and late summer zooplankton assemblage.”
4818
ESSAS Annual Science Meeting, Seattle, WA, May 2011
4819
4820
4821
Napp, J.M., Stabeno, P.J., and Sigler, M.F. “From climate to whales: Components of an
Integrated Ecosystem Assessment.” NOAA Headquarters, Silver Spring, MD, June 2011
Stabeno, P.J., Napp, J.M., and Sigler, M.F. “North Pacific Climate Regimes and Ecosystem
198
4822
Productivity (NPCREP): Ecosystem-based management in the Bering Sea.” Adaptations
4823
2011 Workshop, Invited Speakers, Washington, D.C., June 2011.
4824
4825
4826
4827
Napp, J.M., and the BEST/BSIERP PIs. “Integrated ecosystem research in the Bering Sea.” AFS
Annual Meeting, Seattle, WA, September 2011.
Stabeno, P., J. Napp, and N. Bond. The Status Of The Bering Sea In 2011: Cool. Alaska Marine
Science Symposium. January 16 – 19, Anchorage, AK.
4828
Hermann, A., Gibson, Bond, Cuchitser, Hedstrom, Chen, Want, Stabeno, Eisner, Janout, Aydin,
4829
A multivariate analysis of observed and modeled biophysical variability on the Bering Sea
4830
shelf: multidecadal hindcasts (1969-2009) and forecasts (2010-2040), Alaska Marine
4831
Science Symposium. January 16 – 19, Anchorage, AK.
4832
P. Stabeno and P. Ressler. Advection. Bering Sea Synthesis. February 6 -9, 2012, Bermuda.
4833
Sullivan, P., S. Salo, P. Stabeno, C. Mordy. Spring Ice and Salt Flux in the Eastern Bering Sea
4834
Marginal Ice Zone. 2012 Ocean Sciences Meeting, February, 2012, Salt Lake City, UT.
4835
P. Stabeno. Physics, Bering Sea PI Meeting. March 28 -30, 2012, Anchorage, AK.
4836
Stabeno, Danielson. Physics of the Bering Sea. Bering Sea Project PI meeting. Anchorage, AK.
4837
March, 2012. Oral Presentation
4838
Stabeno. EcoFOCI. Presentation to head of OAR. Seattle, WA, April 2012. Oral Presentation
4839
Stabeno and Napp.
4840
4841
4842
4843
The Bering Sea.
EcoFOCI Retreat.
Seattle WA, June 2012. Oral
Presentation
Stabeno and Overland. The New Normal in the U. S. Arctic. NOAA Senior Research Council.
Seattle, WA, July, 2012. Oral Presentation
Stabeno, Farley, Kachel, Moore, Mordy, Napp, Overland, Pinchuk and Sigler. Climate-mediated
4844
processes on the northern and southern shelves of the eastern Bering Sea and some
4845
implications for the ecosystem. Yeosu, Korea, May 2012. Oral Presentation
4846
4847
Reports
4848
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 20
4849
Number 2. July 2012.
4850
4851
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 20
4852
Number 1. January 2012.
4853
4854
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 19
4855
Number 2. July 2011.
199
4856
4857
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 18
4858
Number 2. July 2010.
4859
4860
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 18
4861
Number 1. January 2010.
4862
4863
Napp, Jeffrey. Current Status of the Bering Sea Ecosystem. PICES Press Volume 17 Number 2.
4864
July 2009.
4865
4866
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 17
4867
Number 1. January 2009.
4868
4869
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 16
4870
Number 2. July 2008.
4871
4872
Napp, Jeffrey. The Bering Sea: Current Status and Recent Events. PICES Press Volume 16
4873
Number 1. January 2008
4874
4875
Acknowledgements
4876
Special thanks goes to the Program Manager for the Bering Sea Project, Thomas Van Pelt, for his
4877
hard work, insight and support to this program. We thank all the scientists from EcoFOCI at both
4878
Alaska Fisheries Science Center and the Pacific Marine Environmental Laboratory for their hard
4879
work at sea, in processing data and in assisting in analysis. We thank the officers and crews of the
4880
NOAA ships Miller Freeman and Oscar Dyson, R/V Thomas G. Thompson, R/V Knorr and
4881
USCG Healy for invaluable assistance in deploying and recovery of the moorings, and collecting
4882
oceanographic measurements. Special thanks goes to NOAA’s Alaska Fisheries Science Center’s
4883
Resource Assessment and Conservation Engineering Division Groundfish and Shellfish
4884
Assessment Program who shared shiptime during particularly icy years to deploy moorings at M5
4885
and M8 on several occasions. NOAA’s North Pacific Climate Regimes and Ecosystem
4886
Productivity and Fisheries Oceanography Coordinated Investigations provided support including
4887
equipment, salaries and shiptime that helps to maintain these long-term moorings.
4888
4889
200