Comparison of 4 Northern Hemisphere regions: Physical oceanographic responses to recent climate variability

1

2

3

4

56789

10

1 2

131415

1 6

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

Contents lists available at ScienceDirect

Progress in Oceanography

journal homepage: www.elsevier .com/ locate /pocean

RO

OFRecent climate forcing and physical oceanographic changes in Northern

Hemisphere regions: A review and comparison of four marine ecosystems

K.F. Drinkwater a,*, F. Mueter b, K. Friedland c, M. Taylor d, G.L. Hunt Jr. e, J. Hare c, W. Melle a

a Institute of Marine Research, P.O. Box 1870, Nordnes, 5817 Bergen, Norwayb University of Alaska Fairbanks, School of Fisheries and Ocean Sciences, Fisheries Division, 17101 Point Lena Loop Road, Juneau, AK 99801, USAc NOAA/NMFS/NEFSC, Narragansett Laboratory, 28 Tarzwell Drive, Narragansett, RI 02882, USAd NOAA/NMFS/NEFSC, 166 Water Street, Woods Hole, MA 02543-1026, USAe School of Aquatic and Fishery Sciences, Box 355020, University of Washington, Seattle, WA 98195, USA

a r t i c l e i n f o a b s t r a c t

1718

Article history:Available online xxxx

192021222324252627282930313233

0079-6611/$ - see front matter � 2009 Published bydoi:10.1016/j.pocean.2009.04.003

* Corresponding author. Tel.: +47 55236990; fax: +E-mail address: [email protected] (K.F. Drink

Please cite this article in press as: Drinkwater, Kreview and comparison of four marine ecosyste

EC

TED

PAs part of a project comparing the structure and function of four marine ecosystems off Norway and theUnited States, this paper examines the oceanographic responses to climate forcing, with emphasis onrecent changes. The four Northern Hemisphere ecosystems include two in the Pacific Ocean (BeringSea and Gulf of Alaska) and two in the Atlantic Ocean (Georges Bank/Gulf of Maine and the Barents/Nor-wegian Seas). Air temperatures, wind forcing and heat fluxes over the four regions are compared as wellas ocean hydrography and sea-ice conditions where seasonal sea ice is found. The long-term interannualvariability in air temperatures, winds and net heat fluxes show strong similarity between adjacent eco-systems and within subregions of an ecosystem, but no significant correlations between Pacific andAtlantic ecosystems and few across the Atlantic. In spite of the lack of correlation between climate forcingand ocean conditions between most of the ecosystems, recent years have seen record or near record highsin air and sea temperatures in all ecosystems. The apparent causes of the warming differ. In the Atlantic,they appear to be due to advection, while in the Pacific temperatures are more closely linked to air-seaheat exchanges. Advection is also responsible for the observed changes in salinity in the Atlantic ecosys-tems (generally increasing salinity in the Barents and Norwegian Seas and decreasing in the Gulf of Maineand Georges Bank) while salinity changes in the Gulf of Alaska are largely related to increased localrunoff.

� 2009 Published by Elsevier Ltd.

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

UN

CO

R1. Introduction

During the last couple of decades, temperatures throughoutmuch of the globe have been rising, a result largely attributed toanthropogenic effects (IPCC, 2007). Some of the largest changesare occurring in polar regions. The rise in temperature has resultedin significant decreases in the seasonal coverage and thickness ofsea ice in the Arctic (Stroeve et al., 2008; Haas et al., 2008), rapidmelting of the Greenland ice sheet (Wouters et al., 2008), and calv-ing of large ice shelves in both the Arctic (Copeland et al., 2007)and Antarctic (Scambos et al., 2003; Massom et al., 2008). Oceancirculation patterns have shifted due, in part, to changes in atmo-spheric pressure patterns (Polyakov et al., 2005). These observedchanges are having significant effects on the ecology of the polarand subpolar regions (Hunt et al., 2002, 2008; Grebmeier et al.,2006; Clarke et al., 2007). Even greater changes in climate areanticipated in the future (IPCC, 2007). However, our ability to pre-

68

69

70

71

Elsevier Ltd.

47 55238687.water).

.F., et al. Recent climate forcinms. Prog. Oceanogr. (2009), do

dict and prepare for the expected fluctuations in critical marine re-sources caused by such changes is limited. This lack of process-level predictive capability for systems that are important for com-mercial and subsistence harvests, as well as to non-harvested mar-ine life, is a powerful motivation to learn more about how climatechange will affect the ecosystems of these regions.

While studies on individual ecosystems can provide increasedknowledge and understanding, comparative studies offer theadded opportunity to determine what processes are fundamentaland what might be unique to a particular ecosystem. Examiningdifferent ecosystems increases the degrees of freedom in a statisti-cal sense when testing and determining relationships between cli-mate forcing, physical oceanography and biological responses. Forthese reasons, Norway and the United States undertook a project tocompare several of their high latitude regions: the eastern BeringSea, the coastal region of the northern Gulf of Alaska, the GeorgesBank/Gulf of Maine region and the Norwegian and Barents Seas offNorway (Fig. 1). The project was entitled Comparison of MarineEcosystems of Norway and the United States (MENU) and is partof the GLOBEC regional program, Ecosystem Studies of Sub-ArcticSeas (ESSAS).

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

UN

CO

RR

EC

T

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

Fig. 1. The four MENU areas where comparative studies were carried out.

2 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

At the initial MENU meeting, it was decided to compare thegeneral structure and function of the different marine ecosystemswith emphasis on the biological communities, which resulted inthree papers (Gaichas et al., this volume; Link et al., this volume,Megrey et al., this volume). A second decision was to compareand contrast the responses of both the physical oceanographyand the biology of the different ecosystems to recent changes inthe climate forcing. This paper examines the recent climate forcingand physical oceanographic responses while the biological re-

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

sponses are addressed a companion paper (Mueter et al., thisvolume).

In Section 2 we describe the geographic boundaries and topog-raphy of the different study areas and briefly summarize their cir-culation and hydrographic features. In Section 3 we discuss thedata used and give a brief description of the statistical analysis per-formed. A comparison of the time series of climate and oceano-graphic variables and indices between regions is provided inSection 4 with special attention paid to the recent changes. Section5 discusses the results of the comparisons.

2. Study areas

2.1. Geographical boundaries and areal extent

Within this study, the eastern Bering Sea (EBS) consists of theregion bounded by the Aleutian Island chain in the south, the coastof Alaska in the east, 61�N latitude in the north and the continentalslope in the west (Fig. 1). This continental shelf region is broad(500–800 km) and shallow with an average depth of about 70 m(Hunt and Megrey, 2005). Its area is approximately 0.5 millionkm2. Our Gulf of Alaska (GoA) is restricted to the narrow (typically100–200-km wide) shelf-slope region between approximately147�W and 175�W (Fig. 1). It consists of extremely irregular shelftopography dotted with numerous islands, banks and ridges, aswell as troughs or gullies cutting across the shelf. Depths typicallyrange from between 100 and 200 m on the shelf to over 3000–5000 m on the abyssal plain at the base of the slope. Its area isabout 0.37 million km2.

The Georges Bank/Gulf of Maine (GB/GoM) region, which in-cludes the Bay of Fundy, is situated off the northeastern UnitedStates between Cape Cod, Massachusetts, and Nova Scotia, Canada(Fig. 1). To the northeast of the Gulf is the Scotian Shelf and to thesouthwest is the Middle Atlantic Bight. The offshore border of theGB/GoM region is the continental slope. Georges Bank is shallow(<100 m) with a minimum depth of around 20 m. The Gulf ofMaine consists of three main basins, Georges, Jordan and Wilkin-son with depths exceeding 200 m and a 370 m maximum depthin Georges Basin. The area of the combined region is slightly overa 0.1 million km2.

The Norwegian and Barents Seas are situated off Norway in theNortheast Atlantic (Fig. 1). The Norwegian Sea (NS) is bounded bythe coast of Norway in the east, by the Barents Sea and Spitzbergenin the north, and by the Greenland and Iceland seas in the west. Itis separated from the rest of the North Atlantic in the south by theGreenland–Scotland Ridge. The Norwegian Sea consists of twodeep basins, the Norwegian Basin and the Lofoten Basin, and hasa surface area of about 1.1 million km2 with an average depth near1800 m (Skjoldal, 2004). The Barents Sea (BS) is bounded by the is-land of Novaya Zemlya in the east, the continental shelf break to-wards the deep Arctic Ocean in the north, the shelf breaktowards the Norwegian Sea in the west and by the coasts of Nor-way and Russia in the south (Fig. 1). It covers 1.4 million km2 withan average depth of 230 m. The maximum depth near the westernentrance is �500 m.

2.2. Circulation and hydrography

2.2.1. Eastern Bering SeaThe mean flow in the eastern Bering Sea is generally to the

northwest and consists of the Bering Slope Current along the con-tinental break (Schumacher and Reed, 1992), a generally weakerflow over most of the shelf, and the nearshore coastal current(Fig. 2). These flows originate from the Aleutian North Slope Cur-rent (Reed and Stabeno, 1999) and the water from the Gulf of Alas-ka passing between the Aleutian Islands (Schumacher et al., 2003;

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

UN

CO

RR

EC

T

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

Fig. 2. The mean surface circulation in the eastern Bering Sea and northern coastalGulf of Alaska (top), the Gulf of Maine and Georges Bank (middle) and theNorwegian and Barents Seas (bottom). The source of the material is fromthe University of Alaska (www.ims.uaf.edu), NMFS/NOAA in Woods Hole, Mass,and the Institute of Marine Research, Bergen, Norway.

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 3

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

Stabeno et al., 2002, 2005). Most of the flow over the shelf exitsthrough the Bering Strait into the Arctic Ocean (Schumacher andStabeno, 1998), while much of the Slope Current curves westwardand then southward to form part of the Kamchatka Current. Strongtidal currents on the shelf dominate the weak mean flow (Coach-man, 1986; Kowalik, 1999). These result in tidally induced circula-tion features in canyons along the slope (Schumacher and Reed,

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

1992; Kowalik and Stabeno, 1999) and around islands such asthe Pribilof Islands (Kowalik and Stabeno, 1999; Stabeno et al.,1999, 2008).

Based on summertime hydrography, the Eastern Bering shelf isdivided into the Coastal, Middle, and Outer Domains, correspond-ing roughly to the 50, 100, 170 m isobaths, respectively (Coach-man, 1986; Schumacher and Stabeno, 1998). The Coastal Domainexhibits weak stratification due to both tidal and wind mixing,the Middle Domain consists of a wind-mixed surface layer and atidally mixed bottom layer, and the Outer Domain has mixed upperand lower layers between which is a layer of slowly increasingdensity. These domains are separated horizontally by a system offronts (Iverson et al., 1979; Schumacher and Stabeno, 1998). Dur-ing winter, the domains are poorly defined since storms mix thewater column to >90 m. Vertical stratification near surface beginsto occur over the middle and outer shelf typically in late springwhen storms weaken. With the seasonal heating of the upper layer,the lower layer in the Middle Domain becomes insulated, resultingin a ‘‘cold pool” that contains temperatures typically <2 �C throughthe summer (Reed, 1995). Relatively low salinities in the region re-sult from melting of sea ice and local river runoff into the BeringSea from Alaska as well as low salinity waters transported intothe Bering Sea through the Aleutian chain (Stabeno et al., 2002,2005).

In the EBS, ice-free conditions generally persist from Junethrough October. As temperatures cool during autumn, ice formson the northern Bering Shelf and strong winds drive it southwarduntil it reaches water warm enough to melt the ice (Overlandand Pease, 1982). The maximum ice coverage, which extends intothe southeastern Bering Sea, typically occurs in March and retreatsnorthward through the spring and into the summer.

2.2.2. Gulf of AlaskaIn comparison to the sluggish mean currents on the Bering Sea

shelf, the coastal Gulf of Alaska is more dynamic (i.e. faster andstronger currents). The primary circulation feature is the AlaskaCoastal Current (ACC, Fig. 2). This persistent-, wind- and buoy-ancy-forced current follows the inner GoA shelf for 2500 km fromBritish Columbia to the Bering Sea with considerable mesoscalevariability resulting from eddies and meanders (Weingartneret al., 2002). Farther offshore, is the broader and less intense(0.05–0.1 m s�1) westward-moving Alaska Current that advectswarm, lower-latitude water into the northern Gulf of Alaska(Weingartner et al., 2002). To the west, it intensifies (0.3–1.0 m s�1) becoming the Alaskan Stream along the south side ofthe Alaska Peninsula and the Aleutian Arc. During winter, intensestorms produce strong alongshore winds that can increase thewestward flow and produce downwelling along the coast. In sum-mer, weak and variable winds cause much reduced downwelling,and occasionally upwelling, along the coast. High tidal elevationsin the GoA generate strong tidal currents, upwards of 0.3–0.7 m s�1 in some areas (Stabeno et al., 2004). These high tidalvelocities generate intense tidal mixing that in turn enhancescross-shelf exchange especially in regions of steep topographysuch as canyons (Ladd et al., 2005). The tidally generated residualcurrents in this region tend to flow counter to the observed large-scale mean circulation (Auad and Miller, 2008).

Freshwater input is important to the timing and strength ofstratification in the spring and summer (Weingartner et al.,2002). River runoff into the GoA reaches a maximum in the au-tumn due to the melting of snow and glacial ice and a minimumin winter, when much of the precipitation is stored as snow andthe glaciers freeze (Royer, 1982). An inner salinity front separatesthe Alaska Coastal Current from the mid-shelf region, which is sep-arated by another salinity front from the higher salinity waters far-ther offshore.

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

T

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

4 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR

EC

2.2.3. Gulf of Maine and Georges BankThe circulation in the Gulf of Maine area is part of a general

southward flow extending from the Canadian Arctic to Cape Hat-teras (Chapman and Beardsley, 1989). Surface waters flow intothe Gulf of Maine from off the Scotian Shelf and circulate anticlock-wise (Bigelow, 1927; Fig. 2). A minor component flows southwardonto the Middle Atlantic Bight east of Cape Cod but the majorityflows along the northern edge of Georges Bank where it splits, partflowing around the Bank and part rejoining the Gulf-wide circula-tion (Manning and Beardsley, 1993; Butman and Beardsley, 1987).Some of the flow on Georges Bank forms a leaky gyre over the topof the Bank, while the rest moves southward onto the MiddleAtlantic Bight (Butman and Beardsley, 1987).

Strong tidal currents persist throughout the Gulf of Maine dueto resonance near the diurnal tidal frequency resulting in the high-est measured tidal elevations in the world (>16 m) in the Bay ofFundy (Garrett, 1972). Consequently, maximum tidal velocities inthe GoM/GB region often exceed 1 m s�1 (Brown and Moody,1987) causing tidally generated residual currents, the most notablecontributing to the clockwise gyre over Georges Bank (Loder,1980).

The Gulf of Maine receives large quantities of cool, low salinitywater in its upper layers from the Scotian Shelf to the north (Smith,1983; Mountain, 2002) and higher salinity offshore waters at depththrough the Northeast Channel (Ramp et al., 1985). Runoff from lo-cal rivers further acts to lower the salinity in the surface layers.Large seasonal fluctuations in heat content in the upper 100 mare determined primarily by air-sea heat exchanges (Mountainet al., 1996). Hydrographic properties and stratification throughoutmuch of the Gulf are greatly influenced by the intense tidal mixing.This results in well-mixed waters throughout the year in the innerBay of Fundy, on the shallow coastal regions such as off southwestNova Scotia (Lurcher Shoals), off Cape Cod (Nantucket Shoals) andon Georges Bank (Garrett et al., 1978), as well as reduced stratifica-tion in most other regions of the Gulf of Maine. The tidal fronts sep-arating the tidally well-mixed waters from the stratified waters inthe deep regions are observed to vary seasonally being farther ontothe shallow banks and shoals in summer and farther offshore dur-ing the rest of the year (Mavor and Bisagni, 2001).

2.2.4. Norwegian and Barents SeasThe warm, salty Atlantic Current flows northward through the

Norwegian Sea as two separate branches, one that hugs the conti-nental shelf break off Norway and another more western offshorebranch (Blindheim and Østerhus, 2005; Fig. 2). As the Slope branchreaches northern Norway, it splits, part moving eastward into theBarents Sea, while the majority continues northward (Blindheimand Østerhus, 2005). Inshore on the shelf lays coastal water withlow salinities and seasonally varying temperatures. These watersare transported northward by the Norwegian Coastal Current(S�tre, 2007). A salinity front separates this Coastal water fromthe offshore Atlantic water, while the Arctic Front to the west sep-arates Atlantic Water from the colder waters in the Greenland andIceland seas.

Mean currents in the Barents Sea tend to be weaker than in theNorwegian Sea and are strongly influenced by bottom topography(Skjoldal and Rey, 1989). Transport of Atlantic water into theBarents Sea increases in winter (Ingvaldsen et al., 2004). Uponentering the Barents Sea, the Atlantic inflow splits, part transitingeastward through the Barents Sea and part recirculating to flowback into the Norwegian Sea (Ingvaldsen, 2005; Skagseth, 2008).The waters on route through the Barents Sea are modified due tomixing, atmospheric cooling, and the addition of discharge fromseveral rivers. To the south, inshore of the Atlantic inflow, are theeastward flowing low salinity Coastal waters, an extension of theNorwegian Coastal Current (Loeng, 1991). Cold Arctic Water, which

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

enters the Barents Sea from its northern and eastern entrances,dominates the northern Barents Sea. Most of this water flowssouthwestward eventually exiting the Barents Sea via the westernentrance (Blindheim, 1989; Loeng, 1991). The Polar Front, whichseparates Arctic and Atlantic waters, is strongly tied to the bottomtopography in the western region (Parsons et al., 1996; Harris et al.,1998) but in the east its position varies depending upon thestrength of the Atlantic inflow (Loeng, 1991).

Tidal currents in the Norwegian Sea are generally low but arestronger in the Barents Sea, typically exceeding the weak meanflows. Maximum tidal velocities occur over the banks in theBarents Sea, especially Bear Island, Spitzbergen and Hopen, wherethey produce tidally generated clockwise residual circulation pat-terns (Gjevik et al., 1994; Kowalik and Proshuntinsky, 1995).

Approximately 40% of the Barents Sea is ice-covered on an an-nual basis but with extensive seasonal variability (Loeng, 1979;Vinje and Kvambekk, 1991). Minimum ice coverage occurs in Au-gust/September when the ice edge can retreat as far north as theshelf break adjacent to the Arctic Ocean. By the end of October,ice begins to form due to decreasing temperatures and strongwinds. The ice cover expands gradually southward through a com-bination of advection of ice from the Arctic and local formation.Maximum coverage occurs in March or April when approximately60% of the Sea is covered, mainly limited to the northern and east-ern regions. In summer, a 5–30 m layer of sea-ice melt water over-lays the Arctic Water and at times extends over Atlantic waters(Loeng, 1991).

3. Data and methods

3.1. Data

To compare atmospheric forcing and oceanographic responsesbetween the various regions we sought similar datasets, wherepossible. For the atmospheric forcing, we chose the NCEP/NCARreanalysis data (Kalnay et al., 1996), which were obtained fromthe Climate Data Center, National Oceanic and AtmosphericAdministration (http://www.cdc.noaa.gov/cdc). At each grid pointwithin the MENU areas (Fig. 3), monthly mean air temperatures,wind speeds, vector components of the wind, and surface radiationterms, including the net surface heat fluxes, were abstracted forJanuary 1948 to June 2007. These data were then combined todetermine area-weighted monthly means for each ecosystem. Esti-mates for the Gulf of Maine and Georges Bank were individuallydetermined to test if there was any significant difference betweenthe two subregions. Because of the large differences in water massproperties in different regions of the Barents Sea (Loeng, 1991), itwas subdivided into the western (Atlantic Water), northern (ArcticWater) and eastern (mixed or Barents Sea Water) regions. Ob-served air temperature data from selected stations were also ob-tained to compare with the NCEP data. These stations includedSt. Paul for the eastern Bering Sea, Anchorage for the Gulf of Alaska,Boston for the Gulf of Maine, Bodø for the Norwegian Sea and BearIsland for the Barents Sea.

Sea surface temperature (SST) data from 1854 to 2006 were ta-ken from the NOAA National Climate Data Center (NCDC) ExtendedReconstruction Sea Surface Temperature (ERSST) dataset. The datawere abstracted from essentially similar regions to those for theNCEP data. Prior to 1900, the ERSST data had large estimated meansquare errors and low variance, therefore we restricted our analy-sis to 1900 onwards.

Subsurface temperature time series were available for compar-ison from four of the regions but at different depths or depthranges and of varying duration. In the eastern Bering Sea bottomtemperatures collected during the NOAA/NMFS Alaska FisheriesScience Center bottom trawl surveys between early June and early

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

TED

PR

OO

F342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

Fig. 3. The grid points from the NCEP database used to represent each of the regions. The upper left is the eastern Bering Sea (red dots) and Gulf of Alaska (green). The lowerleft is the Gulf of Maine (red) and Georges Bank (green). The upper right is the Barents Sea for the western or Atlantic region (green), the northern or Arctic region (red) andthe mixed or Barents Sea Water region (blue). The lower right is the Norwegian Sea (red). (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of this article.)

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 5

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR

ECAugust were available from 1982 to 2007. An area mean was esti-

mated from the average of the bottom measurements across allavailable stations (generally between 250 to slightly over 300 sta-tions approximately 20 nm apart within the area 54�300–62�N and179–158 �W). Anomalies or deviations from the mean value for the1991–2000 period normalized by the standard deviation wereavailable from the NOAA website http://www.beringclimate.noaa.gov/data/index.php. In the Gulf of Alaska a hydrographicmonitoring station (GAK1) is located at the mouth of ResurrectionBay near Seward, Alaska, in 263 m of water. Temperature profileshave been taken several times a year since December 1970 (seehttp://www.ims.uaf.edu/gak1/). For this study we examined thetemperatures from 200 m as well as those at the surface. A sea-sonal mean curve was fit to all of the available data (1970–2006)and then subtracted from the data to obtain anomalies. The anom-alies were averaged first by month for each year and then themonthly anomalies averaged to obtain annual means. In the Gulfof Maine, the NOAA Fisheries Service conducted between 3 and22 hydrographic surveys per year from 1977 to 2007. From eachsurvey area-averaged bottom and surface temperatures were esti-mates and temperature anomalies determined relative to the meanduring 1978–1987. These were the MARMAP (Marine ResourcesMonitoring and Prediction) program years and are often used asthe standard for the Gulf of Maine because of the extensive cover-age during those years (Mountain, 2003). The mean day of the sur-vey was used to assign the anomalies to a particular month. Annualanomalies for each year were then calculated using a simple meanof all the available cruises. We also averaged the data to obtainmonthly means when there was more than one cruise per monthand then averaged the monthly means to obtain the annual

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

means. There was no statistically difference in amplitude or var-iability and hence we used the mean based on the average of allof the available cruises. In the Barents Sea, Russia has occupiedthe Kola section north of the Kola Peninsula (along 33�300E) atleast monthly since 1900. Temperatures in the 0–200 m layer atstations 3–7 between 70�300N and 72�300N (Tereshchenko,1996) from 1900 to 2006 provided by PINRO Laboratory in Mur-mansk, Russia, were used in this study and are considered to re-flect temperature variability in the southern Barents Sea(Ingvaldsen et al., 2003).

Long-term salinity records are available from the Gulf of Alaska,Gulf of Maine and the Barents Sea. In the GoA, we use the salinitiesat 0 and 200 m available from the GAK1 monitoring site off Seward(http://www.ims.uaf.edu/gak1/). In the Gulf of Maine, salinity datawere derived from the NOAA Fisheries Service hydrographic sur-veys and treated in a similar manner to the temperatures describedabove. In the Barents Sea salinity measurements have been takenacross its western entrance (Fugløya–Bjørnoya Line) since 1975and were obtained from Randi Ingvaldsen (Institute of Marine Re-search, Bergen, personal communication).

Seasonal ice cover within the MENU regions only occurs in theeastern Bering Sea and the Barents Sea. The winter ice index for1981–2006 for the eastern Bering Sea was obtained from NOAA’s Pa-cific Marine Environmental Laboratory (http://www.beringclimate.noaa.gov/data). It is based on the average daily ice concentrationwithin the region 56–58�N, 163–165�W for January to May takenfrom the National Snow and Ice Data Center (NSIDC, http://nsidc.org/data/seaice/pm.html). The annual index is expressed asa normalized anomaly by subtracting the mean and dividing bythe standard deviation over the period 1981–2000. For the Barents

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

T

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

451451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

Fig. 4. Comparison of the latitudinal ranges (upper panel) and areas (lower panel)of the different study areas. EBS is the eastern Bering Sea, GoA is the Gulf of Alaska,GoM is the Gulf of Maine, GB is Georges Bank, NS is the Norwegian Sea and BS is theBarents Sea.

6 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR

EC

Sea, a winter sea-ice time series for 1967–2005 was obtained fromSorteberg and Kvingedal (2006). It is the normalized ice area southof 85�N between 20�E and 80�E for the months December, Januaryand February with concentrations greater than 15%. While thenorthern part of this area includes part of the Arctic Basin, thesea-ice variability is driven primarily by changes within theBarents Sea. Similar to the Bering Sea, normalized sea-ice anoma-lies were obtained for the Barents Sea by subtracting the long-termmean and dividing by the standard deviation (in this case over theperiod 1967–2005).

Large-scale climate indices were assembled in order to com-pare with the climate and oceanographic variables. Indices forthe Pacific and Arctic, including the Pacific Decadal Oscillation(PDO), the North Pacific Index (NPI), the Siberian–Aleutian Index(SAI) and Arctic Oscillation (AO), were obtained from the NOAAwebsite http://www.beringclimate.noaa.gov/data. The PDO (Man-tua et al., 1997) is the leading Principal Component of monthlySST anomalies in the North Pacific Ocean poleward of 20�N whilethe North Pacific Index is the area-weighted sea level pressureover the region 30–65�N, 160�E–140�W. We used the annualaverage based on the months January through December and alsothe summer average (June through August). The Siberian/AlaskanIndex (SAI) is the difference between the mean winter (DJFM)normalized 700-hPa anomalies in Siberia (55–70�N, 90–150�E)and over Alaska and the Yukon (60–70�N, 130–160�W). The ArcticOscillation (AO), also termed the Northern Annular Mode or NAM(Thompson and Wallace, 1998), is defined as the leading mode ofthe Empirical Orthogonal Function (EOF) analysis of monthlymean 1000 mb height during 1979–2000. For the Atlantic, theNorth Atlantic Oscillation (NAO) index was obtained fromhttp://www.cgd.ucar.edu/cas/jhurrell/indices.html. This index isthe mean atmospheric pressure difference at sea level for themonths December to March inclusive between Iceland and Lis-bon, Portugal (Hurrell, 1995).

3.2. Statistical methods

Correlation and regression analysis were carried out betweenvarious time series for the different regions. It was felt that thesewould identify any major relationships between climate andoceanographic responses as well as between ecosystems, althoughit is acknowledged that future work should undertake other quan-titative analyses.

The assumption in estimating significance levels for correla-tions is that all of the points used in the analysis are statisticallyindependent. This is not the case for most geophysical time seriesas they can exhibit strong persistence (high autocorrelation). Toprovide a better estimate of the number of statistically indepen-dent number of points in the time series, we used the method ofGarrett and Toulany (1981), i.e.

Ni�1 ¼ N�1 þ 2 � N�2 �XðN � jÞAj ðj ¼ 1;2;3 . . .Þ

where N is the original number of points in the time series, Ni is thenumber of statistically independent points and Aj is the autocorre-lation at lag j. The summation was over j = 1–3 or the first zerocrossing (Aj = 0), whichever came first. Taking higher j values gener-ally made little difference in the value of Ni. For cross correlationbetween series X and Y, Aj is taken as the autocorrelation of the ser-ies X * Y. The ratio of Ni/N depended upon the variable, and morespecifically the persistence the variable exhibited. Winds were theleast persistent of the variables and the ratio varied between 0.33and 1. For heat fluxes, the ratio varied from 0.33 to 0.8. Air temper-atures showed higher autocorrelation resulting in Ni/N ratios of0.3–0.6 and SSTs exhibited the highest autocorrelations reducingthe ratio to 0.1–0.2.

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

4. Regional comparisons

4.1. Latitude, areal extent and depth

A comparison of the four regions in terms of their latitudinalboundaries is presented in Fig. 4. The Barents and Norwegian Seasare the most northerly and the Gulf of Maine the most southerly. Itis worth noting that the southern boundary of the Barents and Nor-wegian Seas is north of the northernmost limits of both the BeringSea and Gulf of Alaska. Similarly, the two Pacific regions lie to thenorth of the Georges Bank/Gulf of Maine region. The areas of boththe Norwegian and the Barents Seas are 2–3 times that of the EBS,4–5 times that of GoA and at least 10 times that of the GB/GoM(Fig. 4). In terms of depth, the Norwegian Sea and the GoA regionare the deepest regions, with mean depths over 1000 m, whilethe remainder are shallow continental shelves, with the order ofdeepest to shallowest being the Barents Sea, the Gulf of Maine,the eastern Bering Sea and Georges Bank.

4.2. Air temperatures

The long-term (1948–2007) annual mean NCEP air tempera-tures show the warmest conditions in the Gulf of Maine and cold-est in the Barents Sea, roughly following the latitudinal gradient(Fig. 5). The outlier is the Norwegian Sea, which is warmer thanthe EBS although it lies north of the Bering Sea. This is largelydue to warm air masses carried by the low pressure systems acrossthe North Atlantic, which move in a northeastward direction fromthe eastern seaboard of North America into the Nordic Seas.

The time series of the NCEP annual mean air temperaturesmatch well with the observed temperatures from the monitoringstations for each region (r = 0.8–0.9; p < 0.01; plots not shown).Data from the Pacific ecosystems show no significant linear trendover the past 30 years, but do exhibit a low in the 1970s followed

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

T495

496

497

498

499

500

501

502

503

504

505

506

507

508

509

510

511

512

513

514

515

516

517

518

519

520

521

522

523

524

525

526

527

528

529

530

531

532

533

534

535

536

537

538

539

540

541

542

543

544

545

546

547

548

549

550

551

552

553

554

555

556

557

558

559

560

561

562

563

564

565

566

567

568

569

570

571

572

573

574

575

576

577

578

579

580

581

582

583

584

585

586

587

588

589

590

591

592

593

594

595

596

597

598

599

Fig. 5. The long-term (1948–2006) annual mean air temperature from the differentMENU regions and subregions plotted as a function of latitude. The letters indicatethe region or subregion: GoM (Gulf of Maine), GB (Georges Bank), GoA (Gulf ofAlaska), EBS (eastern Bering Sea), NS (Norwegian Sea), BSW (Barents Sea west), BSE(Barents Sea east) and BSN (Barents Sea north).

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 7

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR

EC

by a rapid rise in the late 1970s (Fig. 6), the timing of which is con-sistent with the well known regime shift that occurred at that time(Mantua et al., 1997). Another dramatic rise occurred in the late1990s and into the 2000s, which peaked in 2005 and since thenhas declined. The maximum values in the Gulf of Maine andGeorges Bank region were in the 1950s, declined to a minimumin the 1960s and then have shown a gradual increase with highvariability and a peak in 2006 (Fig. 6). The Norwegian and BarentsSeas also show a minimum in the 1960s and since then have variedon approximately decadal scales or longer superimposed upon agradual rising trend (Fig. 6). Air temperatures since 2000 have dis-played high variability but have been at or near record values in re-cent years.

High and strongly significant (r = 0.70–0.97, p < 0.01) correla-tions were found between annual mean air temperatures (1948–2006) for adjacent ecosystems (EBS with GoA, and withinsubregions of both the Barents Sea and Gulf of Maine regions),weak but significant (r = 0.35–0.47, p = 0.02–0.05) correlationsacross ocean basins (GB/GoM with NS and BS) and low and non-significant correlations between oceans (EBS, GoA with GoM, NS,and BS) (Table 1). Correlations of the air temperatures over theNorwegian Sea with those over the Barents Sea subregions werehighly significant (r = 0.46–0.71, p < 0.01), the highest values beingwith the western Barents Sea region.

4.3. Winds

The long-term mean annual wind speeds (Ws) for the MENU re-gions based on NCEP data generally show a decrease with increas-ing latitude (Fig. 7 and Table 2). The primary exception to thistrend is the Gulf of Alaska, which has the lowest wind speeds. Allof the regions exhibit seasonal variability with the highest speedsin winter and lowest during the summer (Table 2). The GoM andGB, however, have a secondary maximum in July resulting in twominima, one in April and a lower one in September. The windspeeds in all of the MENU regions exhibit high interannual variabil-ity with no significant linear trends (p levels >0.05) over the periodfrom 1948 to 2006 (not plotted). There is, however, suggestion of aslight increase in wind speed in the EBS, GoA and the BSN from1948 to 2006 (p = 0.05–0.10).

The velocity components (Uw, east–west component, positiveeastward; Vw, north–south component, positive northward) ineach of the regions were also investigated. For wind directions

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

we adopted the oceanographic convention, that is a northwardwind is one in which the winds are blowing towards the north.The prevailing winds in the EBS are southwestward and westwardin the GoA, but in both regions they switch to the northeastward insummer. The interannual variability of the anomalies in Uw overthe EBS and GoA shows a step jump in the mid to late 1970s fromweaker to stronger westward winds (Fig. 8). The Vw anomalies ex-hibit high variability with a general increase in northward winds inthe GoA after the late 1970s but no dominant trend (Fig. 8). In theGoM/GB region the prevailing winds are southeastward except insummer when they switch to northeastward. The anomalies ofboth Uw and Vw show lower variability than most of the other re-gions but again with no significant trends. In the NS the dominantwinds are northeastward but switch to the southeastward inspring while in the Barents Sea they are southwestward through-out most of the year and veer slightly eastward in summer in thenorthern area. In both regions there is high interannual variabilitybut no strong trends in the anomaly of the wind components.

Correlations of Uw, Vw and Ws (1948–2006) are statisticallysignificant (p < 0.05) between adjacent ecosystems or within sub-regions but non-significant between ecosystems of the Pacificand those in the Atlantic (Table 3). Across the Atlantic the only sig-nificant relationships are for the Vw component between the NSwith both the GoM and GB and much weaker between the BSNwith GB.

The time series of Uw and Vw were also correlated with the lo-cal air temperatures for each of the regions. Correlations with Vwwere significant in all regions except the Gulf of Maine with stron-ger northward winds lead to higher temperatures (Table 4). Thehighest correlation coefficients were in the Norwegian and BarentsSeas. In the GoA, air temperatures are dependent almost equallyupon the Uw and Vw components, with northwestward windsresulting in high temperatures.

4.4. Air-sea surface heat fluxes

The net surface heat flux (Qt) and its components (Qsw, shortwave radiation; Qlw, long wave radiation; Ql, latent heat; andQs, sensible heat) were obtained from the NCEP data base. In all re-gions there is a net heat loss from the oceans to the atmosphere(+Qt) in winter while in the summer the oceans absorb heat fromthe atmosphere (�Qt) (Fig. 9). This seasonal change is dominatedby Qsw and while Ql and Qs also contribute to seasonal variabilityin Qt, their contribution is much less than Qsw. Qt and the timingof the switch in spring and autumn between a net heat gain and anet loss for the oceans are largely determined by the latitudinal po-sition of the regions (Fig. 9). An exception to the latitudinal depen-dence of Qt is the northern Barents Sea, which is farthest north butwhose annual net heat loss is less than from the Norwegian Sea(Fig. 9). Similarly, net heat loss is larger from the Gulf of Alaskathan from the eastern Bering Sea. In both cases this is believed tobe due to sea-ice cover in the Bering and Barents Seas that insu-lates the underlying water and prevents additional heat losses dur-ing the winter. Only in the Gulf of Maine and Georges Bank regionis there a net heat gain from the atmosphere to the ocean throughthe year on average, while the rest of the regions experience a netheat loss (Fig. 9).

Correlations of annual Qt (1948–2006) show significant rela-tionships within and between adjacent ecosystems (Table 5). How-ever, between ecosystems of the Pacific and the Atlantic orbetween ecosystems across the Atlantic, the correlations are notsignificant.

In the Pacific regions, the time series of Qt anomalies reveals ashift from greater heat loss from the oceans (+anomalies) prior tothe late 1970s to less heat loss (�anomalies) after (Fig. 10). Sincethe mid-1990s in the Pacific, there has been no overall trend in

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

CO

RR

EC

TED

PR

OO

F

600

601

602

603

Fig. 6. Annual (light line) and 5-year running mean (solid line) of NCEP air temperatures for the various MENU regions and subregions. The letters indicate the region orsubregion: EBS (eastern Bering Sea), (Gulf of Alaska), GoM (Gulf of Maine), GB (Georges Bank), NS (Norwegian Sea), BSW (Barents Sea west), BSN (Barents Sea north) and BSE(Barents Sea east).

8 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

NQt, but slightly more heat loss than in the 1980s and early 1990s. Inthe Gulf of Maine, there is no overall trend in Qt but in Georges

605

606

607

608

609

610

611

612

613

614

615

Table 1Statistically significant correlations (p < 0.05; bold values p < 0.01) of mean annual airtemperatures (1948–2006) between regions and subregions. Blank cells indicate non-significant correlations.

GoA GoM GB NS BSW BSN BSE

EBS 0.757GoAGoM 0.921 0.351GB 0.424 0.463 0.474 0.353NS 0.713 0.594 0.464BSW 0.939 0.806BSN 0.806

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

Bank Qt has been increasing, indicative of higher heat loss by theocean. Since 2000, positive Qt anomalies have dominated onGeorges Bank with anomalies near or at record highs. In contrast,the Norwegian and Barents Seas have shown a slight tendencyfor negative annual anomalies over the last couple of decades, indi-cating reduced heat losses to the atmosphere. This trend towardsnegative heat flux anomalies is particularly strong in summer,which means at that time of the year there is a net heat gain inthe waters of Barents Sea, something that has not been observedin the other MENU regions. The interannual variability in Qt, unlikethe seasonal variability, largely reflects changes in both Qs and Ql.Qs and Ql are strongly correlated with one another, but the ampli-tude of Qs is typically 2–9 times that of Ql, therefore Qs accountsfor most of the interannual variability in Qt.

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

EC

T

616

617

618

619

620

621

622

623

624

625

626

627

628

629

630

631

632

633

634

635

636

637

638

639

640

641

642

643

644

645

646

647

648

649

650

651

652

653

654

655

656

657

658

659

660

661

662

663

664

665

666

667

668

669

670

671

672

673

674

675

676

677

678

679

680

681

682

683

684

685

686

687

688

689

690

691

692

693

694

695

696

697

698

699

700

701

702

Fig. 7. The long-term (1948–2006) annual mean wind speeds from the differentMENU regions and subregions plotted as a function of latitude. The letters indicatethe region or subregion: GoM (Gulf of Maine), GB (Georges Bank), GoA (Gulf ofAlaska), EBS (eastern Bering Sea), NS (Norwegian Sea), BSW (Barents Sea west), BSE(Barents Sea east) and BSN (Barents Sea north).

Table 2The long-term monthly mean wind speeds in m s�1 by region as well as the annualmeans and standard deviations of the annual means based on the monthly means.

EBS GOA GoM GB NS BSW BSN BSE

J 3.8 3.1 4.1 4.9 3.2 2.8 3.3 2.9F 3.9 2.9 3.9 4.7 3.6 3.3 2.8 3.4M 3.0 2.4 3.0 3.7 3.0 2.6 2.7 2.6A 2.5 1.9 2.2 2.9 2.4 2.3 2.3 2.0M 2.1 1.9 1.8 2.1 2.1 1.9 1.8 1.6J 1.7 1.5 2.6 2.9 2.2 1.8 1.6 1.6J 1.8 1.3 2.9 3.2 1.9 1.9 1.5 1.6A 2.6 1.7 2.3 2.2 2.2 2.0 1.5 1.8S 2.5 2.0 1.6 1.5 2.2 2.1 2.1 2.0O 3.1 1.8 2.0 2.5 2.9 2.4 3.0 2.3N 2.9 2.5 2.7 3.4 3.2 2.9 3.2 3.0D 3.3 2.6 3.7 4.5 3.2 3.0 3.3 3.0

Annual 2.76 2.14 2.73 3.20 2.68 2.42 2.40 2.32Std. dev. 0.71 0.56 0.81 1.09 0.56 0.49 0.71 0.63

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 9

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR4.5. Sea-surface ocean temperatures (ERSSTs)

The long-term annual mean temperatures generally decreasewith latitude with the warmest conditions (9 �C) in the Gulf ofMaine/Georges Bank region and the coldest (0.3 �C) in the northernBarents Sea (Fig. 11). Monthly mean temperatures peak in Augustand are at a minimum in March except in the northern Barents Seawhere the minimum is in April. The largest range between mini-mum and maximum monthly means is in the GoM/GB (13 �C)and the smallest is in the Barents and Norwegian Seas (around5 �C).

The only statistically significant correlation between the annualmean temperatures (1900–2006) in the different MENU regionswas between the eastern Bering Sea and the Gulf of Alaska. Thereis also significant correlation among the subregions of the Norwe-gian/Barents Seas (Table 6). However, time series plots show thatthe annual mean temperatures in all regions during the past dec-ade or so have been at or near record highs (Figs. 12 and 13). Inthe eastern Bering Sea and the Gulf of Alaska, the warmest yearon record was in 2003, but since then temperatures have declinedalthough still remain above the long-term mean. In the Gulf ofMaine, recent years have seen the warmest sea surface tempera-tures since the late 1940s and early 1950s when peak tempera-tures were recorded. The peak temperatures in 1999, 2002 and2006 represent the 4th to the 6th warmest years, respectively, inthe record. In the Norwegian and Barents Sea, the largest annual

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

mean temperature anomalies have been observed during the pastdecade with 2006 being particularly warm. Fig. 14 shows the max-imum temperature anomaly in each region or subregion during thepast 10 years of available data (1997–2006) and the normalizedanomaly (anomaly divided by the standard deviation over the per-iod 1900–2006). The highest absolute anomaly (>1.5 �C) is in theeastern Bering Sea and the lowest in the Barents Sea/NorwegianSeas (�0.5 �C). Relative to the interannual variability as measuredby their standard deviations, the maximum anomalies in the mostrecent decade are between 1.5 and 2.5 standard deviations abovetheir long-term means. In five of the regions or subregions (EBS,GoA, NS, BSW, BSN) the maximum temperature anomaly duringthis recent period set a new all time record, while in the GoM/GBregion it was the 4th warmest on record and in the BSE, the thirdwarmest.

Correlations of the interannual variability in the monthly SSTanomalies with the overall annual SST anomalies (not shown) re-veal that summer and spring conditions contributed most to thevariability of the annual averages. This is consistent with observa-tions that these seasons have seen the largest warming during re-cent years and that the temperature difference between summerand winter has been increasing.

For the eastern Bering Sea and the Gulf of Alaska the annual netheat fluxes account for approximately 25% of the variance of theannual mean sea-surface temperatures (p = 0.01–0.02), but forthe Atlantic regions correlations were low and non-significant(p > 0.05) between annual means of Qt and SST.

4.6. Subsurface ocean temperatures

Time series of subsurface temperature data were readily avail-able for the EBS, GoA, GoM and the BSW. The subsurface meantemperature variability was similar to that at the surface (ERSSTs)in the EBS and the BSW (r = 0.7–0.8, p < 0.01), but only weakly re-lated in the GoM (r = 0.4, p = 0.05). However, for the GoM, the an-nual surface and bottom temperatures anomalies, calculated fromthe same NOAA surveys, were strongly related (r = 0.86, p < 0.01)but neither were statistically related to the ERSST annual mean.This may be due to the lower temporal resolution of the surveyscompared to the sea surface temperatures. In the GoA, the surfaceand subsurface time series at the GEK1 site were statisticallydifferent.

The time series of subsurface temperatures from the four re-gions are shown in Fig. 15. In the eastern Bering Sea, similar tothe surface values, the highest summer bottom temperatures onrecord were measured in 2003. Temperatures remained high until2005, but more recently have declined to below normal values. Inthe Gulf of Alaska, temperatures at the GAK1 site at 200 m rose onthe order of 1 �C or more from 1970 to the present. In the Gulf ofMaine the bottom temperatures generally have been high since1999 with a peak in 2002, although 2004 and 2005 were belowthe long-term mean. In the Barents Sea, the Kola temperaturesshow a general trend of increasing temperatures, although withhigh variability, reaching a maximum value in 2006 and onlyslightly lower temperatures in 2007, the last year of the availablerecord.

4.7. Salinities

The time series of salinity anomalies for several of the MENU re-gions are shown in Fig. 16. In the Gulf of Alaska salinities at GAK1show a decreasing trend in the top 30 m from the early 1970s topresent with no significant trend deeper in the water column(50–250 m). The lowest salinities were recorded in 2005 but haveincreased in 2006 and again in 2007. In the Gulf of Maine, salinitieshave shown a general decreasing trend from 1977 to the present.

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

CO

RR

EC

TED

PR

OO

F

703

704

705

706

707

708

709

710

711

712

713

714

715

716

717

718

719

720

721

722

723

724

725

726

Fig. 8. The time series of annual mean Uw (left) and Vw (right) wind velocity components for the different MENU regions and subregions. The letters indicate the region orsubregion: EBS (eastern Bering Sea), (Gulf of Alaska), GoM (Gulf of Maine), GB (Georges Bank), NS (Norwegian Sea), BSW (Barents Sea west), BSN (Barents Sea north) and BSE(Barents Sea east).

10 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

Strong negative salinity anomalies have dominated the recent re-cords especially during the years 1995–2000 and again in 2004–2006. In late 2006 and 2007 salinities have increased again. In con-trast to these two regions, the Barents Sea salinities have beenincreasing since the mid-1970s with the highest salinity anomalieson record relative to the 1970–2005 mean being recorded in 2006(>0.1 psu). Generally above normal salinity anomalies have beenobserved since 1999.

4.8. Sea ice

Sea-ice coverage, which only occurs in the eastern Bering Seaand the Barents Sea in the MENU regions, exhibits high interannual

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

variability (Fig. 17). In the EBS there is no significant trend over theentire dataset. The minimum ice coverage, which occurred be-tween 2000 and 2005 coincided with extensive warming (Figs.12 and 13). With the return to cooler conditions in the EBS during2006 and 2007 (not plotted) there was a rapid increase in ice cov-erage and indeed the maximum extent over the entire record oc-curred in 2007. In contrast to the Bering, in the Barents Sea,there has been a gradual decline in sea ice over the available recordas well as high interannual variability (Fig. 17). Minimum coveragewas observed in 2005, the last year for which the time series isavailable. However, ice cover remained low during 2006 and2007 (although the particular index has not been calculated forthese years, the ice charts have revealed low ice extent during this

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

CTED

PR

OO

F727

728

729

730

731

732

733

734

735

736

737

738

739

740

741

742

743

744

745

746

747

748

749

750

751

752

753

754

755

756

757

Table 3Statistically significant correlations (p < 0.05; bold values p < 0.01) of mean annualUw, Vw and Ws (1948–2006) between regions and subregions. Blank cells indicatenon-significant correlations.

GoA GoM GB NS BSW BSN BSE

EBS Uw 0.762Vw 0.364Ws 0.391

GoA UwVwWs

GoM Uw 0.946Vw 0.863 0.437Ws 0.912

GB UwVw 0.441 0.278Ws

NS Uw 0.503Vw 0.490 0.573Ws

BSW Uw 0.691 0.790Vw 0.933 0.827Ws 0.398 0.529

BSN Uw 0.894Vw 0.823Ws 0.587

Table 4Statistically significant correlations (p < 0.05; bold values p < 0.01) between the meanannual air temperatures with Uw and Vw (1948–2006) for the MENU regions andsubregions. Blank cells indicate non-significant correlations.

EBS GoA GoM GB NS BSW BSN BSE

Uw �0.474Vw 0.385 0.385 0.271 0.627 0.689 0.725 0.449

Fig. 9. The monthly mean net heat fluxes (+ represents a heat gain to theatmosphere and loss to the ocean) for the MENU regions and subregions (top panel).The mean annual net heat flux as a function of latitude (middle panel). The meanday of the year that the heat flux changes from + to � in the spring (grey diamonds)and from � to + in the autumn (black squares) as a function of latitude (bottompanel). The letters indicate the region or subregion: EBS (eastern Bering Sea), (Gulfof Alaska), GoM (Gulf of Maine), GB (Georges Bank), NS (Norwegian Sea), BSW(Barents Sea west), BSN (Barents Sea north) and BSE (Barents Sea east).

Table 5Statistical significant correlations (p < 0.05; bold values p < 0.01) of mean annual netsurface heat fluxes (1948–2006) between regions and subregions. Blank cells indicatenon-significant correlations.

GoA GoM GB NS BSW BSN BSE

EBS 0.614GoAGoM 0.714GBNS 0.734 0.589BSW 0.423 0.626BSN 0.478

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 11

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR

Etime). The correlation between the winter sea-ice indices for theBering and Barents Seas (1979–2005) are weak (r = �0.06) andnot statistically significant.

4.9. Relationship with climate indices

Each of the time series of climate and oceanographic variableswere correlated with the climate indices, PDO, NPI, SAI, AO andNAO and the statistically significant (p < 0.05) relationships arelisted in Table 7. For the Pacific regions of the EBS and the GoA,many of the annual means of the climate and ocean variables weresignificantly correlated with the annual PDO index including airtemperatures, winds, heat fluxes, SSTs and salinity indices. In the+PDO phase, sea surface temperature (SST) anomalies are positive(warmer than normal) in the Gulf of Alaska and eastern Bering Searegions and colder than normal during the negative phase of thePDO. There is also less heat loss from the ocean during the +PDOphase, most notably in the GoA and wind speeds tend to increase,with stronger westward wind velocities in both regions and stron-ger northward in the GoA. Subsurface salinities in the GoA tend todecrease during the +PDO phase as well as the �NPI phase. Positivevalues of SAI are associated with stronger-than-normal southwest-ward winds, which in turn result in lower temperatures and hea-vier than normal ice cover in the EBS.

In the GoM and GB no variables were significantly correlatedwith the climate indices (Table 7). In the Norwegian Sea only thewinds were correlated with the climate indices (Table 7). A higherNAO/AO leads to strong wind speeds and increased eastward andnorthward winds. In the Barents Sea, the AO again is the dominant

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

climate index affecting the SSTs in all three subregions, the air tem-peratures in the eastern Barents and the subsurface temperaturesand salinities in the western Barents. Interestingly, the SAI alsoseems to be important in the northern Barents (affecting air tem-

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

UN

CO

RR

EC

TED

PR

OO

F

758

759

760

Fig. 11. The long-term mean SSTs by MENU region as a function of latitude. Theletters indicate the region or subregion: GoM/GB (Gulf of Maine/Georges Bank);GoA (Gulf of Alaska), EBS (eastern Bering Sea), NS (Norwegian Sea), BSW (BarentsSea west), BSE (Barents Sea east) and BSN (Barents Sea north).

Table 6Statistically significant correlations (p < 0.05; bold values p < 0.01) of annual meanSSTs (1900–2005) between regions and subregions. Blank cells indicate non-significant correlations.

GoA GoM/GB NS BSW BSN BSE

EBS 0.820GoAGoM/GBNS 0.583BSW 0.749 0.770BSN 0.835

Fig. 10. The time series of the net heat flux (Qt) anomalies for the regions or subregions: EBS (eastern Bering Sea), (Gulf of Alaska), GoM (Gulf of Maine), GB (Georges Bank),NS (Norwegian Sea), BSW (Barents Sea west), BSN (Barents Sea north) and BSE (Barents Sea east).

12 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

peratures and sea ice) and the eastern Barents (affecting winds).Correlations of the SAI with the winds tend to be positive whilefor the sea ice they are negative.

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

NC

OR

REC

TED

PR

OO

F

761

762

763

764

765

766

767

768

769

770

771

772

773

774

775

776

777

778

779

780

781

Fig. 12. Mean annual sea surface temperature anomalies for the different MENU regions and subregions. The smooth lines represent 11-year running mean. The lettersindicate the region or subregion: EBS (eastern Bering Sea), (Gulf of Alaska), GoM/GB (Gulf of Maine/Georges Bank), NS (Norwegian Sea), BSW (Barents Sea west), BSN (BarentsSea north) and BSE (Barents Sea east).

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 13

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

U5. Discussion

In this paper we have compared the mean and variability of sev-eral climate and oceanographic variables across four high latitudeNorth Hemisphere regions (eastern Bering Sea, Gulf of Alaska, Gulfof Maine/Georges Bank and Norwegian/Barents Seas). The four re-gions differ greatly in their areal extent with the largest region(Norwegian and Barents Sea each) approximately 10 times thatof the smallest (Georges Bank/Gulf of Maine). Depths also differ,from deep basins (1800–2000 m; Norwegian Sea and Gulf of Alas-ka) to shallow banks (shallowest is Georges Bank at <100 m). In

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

spite of these large differences some trends and similarities in cli-mate forcing and physical oceanographic responses are observed.For example, differences in the long-term mean air and sea tem-peratures, as well as net heat fluxes, are roughly determined by lat-itudinal differences with decreasing temperatures and greater heatlosses in the more northern regions. Wind speeds also show a lat-itudinal dependence with decreasing amplitudes northward owingto the peak in the winds caused by the mid-latitude westerlies.One of the effects of latitudinal differences between the regions,which has not been discussed here but is especially importantfor the biology, is that of light levels. For example the Barents

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

UN

CO

RR

EC

TED

PR

OO

F

782

783

784

785

786

787

788

789

790

791

792

793

794

795

796

797

798

799

800

801

802

803

804

805

806

Fig. 13. Monthly sea surface temperature anomalies (long-term monthly means removed) for the different regions and subregions.

Fig. 14. The maximum sea surface temperature anomaly over the decade 1997–2006 by MENU region or subregion and the same anomalies normalized by thestandard deviation of the entire time series of anomalies (1900–2006).

14 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

and Norwegian Seas experience 24 h of sunlight in the summer and24 h of darkness in winter. This results in a more compact biolog-ically productive season in these northern regions compared tothose farther south.

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

All of the MENU regions are highly advective systems, with thestrongest throughflows in the Gulf of Alaska and the weakest in theBarents Sea. For the Gulf of Maine/Georges Bank, the advectivepathway is primary southward bringing cold, low salinity waterinto the region and is in contrast to the other three regions wherethe flows are northward, carrying warmer waters into those re-gions. The effect of northward flows is to produce warmer condi-tions relative to typical latitudinal means and colder conditionswhere the flow is southward. The most extreme example is inthe Norwegian and Barents Seas. Another important physicaloceanographic feature that appears in all of the regions with theexception of the Norwegian Sea, is that of strong tidal currents.These intense tidal currents produce areas of well-mixed watersin shallow areas and associated tidal fronts that separate the tidallymixed waters from the deeper stratified waters. In addition, tidalcurrents produce tidally rectified flows around topographic fea-tures such as banks and islands. Tidal flows therefore modify boththe hydrography and the circulation of these regions. Tides play anespecially important role on the eastern Bering Shelf, on the shelfin the Gulf of Alaska, throughout most of the Gulf of Maine andBay of Fundy, and on several of the shallow banks in the Barents

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

TED

PR

OO

F807

808

809

810

811

812

813

814

815

816

817

818

819

820

821

822

823

824

825

826

827

828

829

830

831

832

833

834

835

836

837

838

839

840

841

842

843

844

845

846

847

848

849

850

851

852

853

854

855

856

857

858

859

860

861

862

863

864

865

866

867

868

869

870

871

872

873

874

Fig. 15. Time series of subsurface temperature anomalies for four the MENU regions. The EBS (eastern Bering Sea) and GoM (Gulf of Maine) are bottom temperatures derivedfrom NOAA surveys, the GoA (Gulf of Alaska) is from 200 m at the GEK1 monitoring station and the BSW (Barents Sea west) are average temperatures from 0 to 200 m on theKola section. See text for further details on the data.

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 15

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR

EC

Sea. Biologically, intense tidal mixing helps to replenish near sur-face nutrient levels throughout the summer thereby increasing to-tal primary production compared to regions without such mixing(Horne et al., 1989).

Comparison of annual mean time series of the winds and netsurface heat fluxes from 1948 to 2006 using the NCEP data basefrom the different MENU regions revealed that there was high sim-ilarity within regions (GoM with GB, and the three subregions ofthe BS) and between adjacent regions (e.g. EBS with GoA and BSwith NS, see Table 2), but no significant correlation between basins(Pacific with Atlantic) nor across the Atlantic Ocean (GoM/GB withBS/NS). This correlation pattern was also true for air temperatures,except for a weak but still significant correlation across the Atlan-tic, i.e. between Georges Bank and the Gulf of Maine with the Nor-wegian and Barents Seas. The correlation within and betweenadjacent ecosystems, but generally not farther, is consistent withthe typical spatial scale of high and low pressure systems (storms)of a few 1000 km or more. The connection in air temperatures be-tween Georges Bank and the Norwegian and Barents Seas likelyarises because low pressure storms, carrying with them their asso-ciated air masses, tend to track northeast from the United StatesAtlantic seaboard through to the Nordic Seas.

The only significant correlations of mean annual SSTs betweenregions were restricted to the adjacent ecosystems of the GoAand EBS and within both the GoM/GB and the Barents Sea regions.This holds whether using the entire dataset (1900–2006; Table 6)or the more restricted time series corresponding to the climatedataset (1948–2006; not shown). There was also a lack of similar-ity in the annual time series of the subsurface temperature dataand salinities between different MENU areas. Sea-ice time seriesfrom the eastern Bering Sea and the Barents Sea were alsounrelated.

Air temperature and SST time series for the same regions showhigh similarity (correlations) suggesting the two are related. How-

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ever, correlations of SSTs with the NCEP net heat fluxes (Qt) suggestthat about 25% of the variance in SSTs in the EBS and the GoA can beaccounted for by air-sea fluxes. In the Atlantic, correlations be-tween SSTs and Qt are very low and not significant suggesting thatincreases in SSTs in the GoM/GB and NS/BS are generally not due toatmospheric heat exchanges but are most likely a result of advec-tive processes. This is consistent with observations that interannualvariability in sea temperatures in the southern Barents Sea isstrongly influenced by changes in the properties and volume ofthe Atlantic inflow (Loeng, 1991; Ingvaldsen et al., 2003) and recentestimates of the heat budget which suggests the heat content in theBarents Sea is primarily determined by advection rather than atmo-spheric heat exchanges (A.B. Sandø, Nansen Environmental and Re-mote Sensing Center, Bergen Norway, personal communication). Inthe Gulf of Maine, Mountain et al. (1996) estimated the surface heatflux for the period 1979–1987 using air temperature, water temper-ature and wind speed data measured at a NOAA buoy in the centralGulf and solar radiation data from coastal stations and climatolog-ical parameters. They found that in the western Gulf the interan-nual variability in water temperature was significantly correlatedwith the heat flux variations but in the eastern Gulf, no relationshipwas found and the temperature variability was believed dominatedby advective changes. Our results, which cover most of the Gulf ofMaine and Georges Bank region, are consistent with findings ofMountain et al. (1996) for the eastern Gulf but not for the westernGulf. This could mean that the temperature variability averagedover the larger area we investigated is dominated by changes inthe eastern area. Other possibilities are differences in the heat fluxestimates between the NCEP and those based on local measure-ments, or to temporal variability in the processes controlling tem-perature fluctuations. Further studies will be required to sort outthe major cause.

Additional support for the strong role of advection in the Atlan-tic comes from available salinity data. These data show general

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

RR

EC

TD

PR

OO

F

875

876

877

878

879

880

881

882

883

884

885

886

887

888

889

890

891

892

893

894

895

896

897

898

899

900

901

902

903

904

905

906

907

908

909

910

911

912

913

914

915

916

917

918

919

920

921

922

923

924

925

926

927

928

929

930

931

Fig. 16. The salinity anomalies for (A) the Gulf of Alaska based on the GEK1 Station:surface (solid line) and 200 m (dashed line), (B) the Gulf of Maine based on the area-averaged bottom values from the NOAA groundfish surveys and (C) the Barents Seawest based on the 50–200 m layer at the western entrance to the Barents Sea. Thediagonal lines represent linear trend lines. The trend for the 200 m Gulf of Alaskadata was not significantly different from zero and was not plotted.

Fig. 17. The normalized annual winter ice area for (A) the eastern Bering Sea and(B) the Barents Sea.

16 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

COtrends of increasing salinity in the Barents Sea but decreasing in

the Gulf of Maine region. The source of higher salinities into theBarents is advection of higher salinity Atlantic Waters from thesouth (Skagseth et al., 2008; Holliday et al., 2008). These highersalinities in the Atlantic Waters, in turn, are thought to be linkedto increased salinities in the subtropics caused by higher evapora-tion in recent years which are subsequently advected northward(Bethke et al., 2006). On the other side of the Atlantic, lower salin-ities in the Gulf of Maine are tied to fresher flow originating fromthe north (Drinkwater et al., 2003b). These are thought to be ulti-mately related to Arctic outflow through the Canadian Archipelagothat carry relatively low salinity water of Pacific origin southwardalong the continental shelves off eastern Canada and onto the east-ern seaboard of the United States (Jones et al., 2003; Greene et al.,2008). In the GoA in the Pacific, salinities also have been decreasingrecently, but here the cause is believed to be due mostly to localprocesses primarily through higher river runoff (Royer, 2005).

In spite of the lack of correlation of SSTs between the differentregions over the entire dataset (1900–2006), recent sea surfacetemperatures show high similarity between regions. In general,from the late 1990s to present, ocean temperatures have been at

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

Eor near record temperatures in the majority of the MENU regions.This temperature increase is not limited to surface waters but alsois observed in subsurface temperature data collected either atmonitoring sites or on broad-scale surveys of the different regions.Record temperatures have been observed in the Barents and Nor-wegian Seas while in the Gulf of Maine and Georges Bank temper-atures have increased but remain below the peak temperatures ofthe 1950s. Record temperatures in the eastern Bering Sea and theGulf of Alaska were reached during the 2003–2004. Along withthe higher sea temperatures in the more northern regions (Beringand Barents Seas) there were significant reductions in sea ice. Thisis consistent with earlier studies that show reductions in sea iceunder warm conditions (Beverton and Lee, 1965; Vinje, 2001; Gre-bmeier et al., 2006; Sorteberg and Kvingedal, 2006). These warmconditions have also had important ecological impacts (Overlandand Stabeno, 2004; Grebmeier et al., 2006; Mueter et al., thisvolume).

The cause of increased warming during the last decade differsbetween MENU regions. In the Pacific, the Aleutian Low was dis-placed northward and also strengthened which accounted for in-creased northward and westward winds that carried with themwarmer air masses (Overland and Wang, 2005). This led to in-creased warming of the waters, and in the Eastern Bering Seareductions in sea ice. In the Norwegian and Barents Seas, recordhigh temperatures (and salinities) appear to be related to advec-tion. The inflowing Atlantic waters have been warmer and moresaline than normal and are thought to be due to increased heatingand evaporation in the subtropics (Bethke et al., 2006). In the Gulfof Maine, temperatures have been relative high but salinities havebeen low. The latter are attributed to advection from the north asdiscussed above. Net surface heat flux anomalies suggest increasedheat loss from the ocean to the atmosphere in recent years, thushigher temperatures during this time must be due to advection.This could be caused by reduced cold water flux from off the Sco-tian Shelf, either due to increased temperatures in these waters orfrom reduced transport. Another possibility is increased influx of

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

T

932

933

934

935

936

937

938

939

940

941

942

943

944

945

946

947

948

949

950

951

952

953

954

955

956

957

958

959

960

961

962

963

964

965

966

967

968

969

970

971

972

973

974

975

976

977

978

979

980

981

982

983

984

985

986

987

988

989

990

991

992

993

994

995

996

997

998

999

1000

1001

1002

1003

1004

1005

1006

1007

1008

1009

1010

1011

1012

1013

1014

1015

1016

1017

1018

1019

1020

1021

1022

1023

1024

1025

1026

1027

1028

1029

1030

1031

Table 7Significant correlations (p < 0.05) between climate indices with climate and ocean-ographic variables by region and subregion. Maximum correlations are shown.Significance levels of p < 0.01 are marked in bold. Empty cells indicate non-significantrelationships and shaded regions indicate there were no data.

EBS GoA GoM GB NS BSW BSN BSE

Air temp. SAI PDO SAI AO�0.517 0.580 0.365 0.449PDO0.480

Ws SAI0.306

Uw PDO PDO AO�0.603 �0.664 0.491

Vw PDO NAO SAI0.401 0.359 0.329

Qt AO PDO�0.370 �0.681

ERSSTs PDO PDO AO AO AO0.566 0.545 0.461 0.435 0.474

Subsurfacetemp.

SAI AO�0.435 0.511

Salinity NPI0.520PDO�0.428

Sea ice SAI SAI0.524 �0.446

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 17

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

UN

CO

RR

EC

warm Slope Water from offshore that is then mixed into the sur-face layers through strong tidal mixing. However, this is unlikelysince the warm Slope Water tends to have high salinity and couldnot explain the observed lower-than-normal salinities throughoutthe Gulf of Maine over the past couple of decades.

It has been suggested that the warming observed in the highlatitude northern regions we are studying is a manifestation ofanthropogenic-induced global warming (IPCC, 2007). While thismay be true, part of the observed signal is also likely to be dueto local natural variability. Indeed, although ocean and air temper-atures continue to remain high in most of the MENU regions, in theeastern Bering Sea there has been significant cooling and increasedsea-ice coverage since 2004. It is not clear whether this is a shortterm phenomenon or part of a longer-term trend and only timewill tell.

All of the MENU regions are strongly influenced by large-scaleatmospheric pressure patterns, and much of the observed variabil-ity in the physical features as well as the underlying biology ofthese seas can be accounted for by changes in these patterns (Hur-rell, 1995; Francis et al., 1998; Hare and Mantua, 2000; Hollowedet al., 2001; Drinkwater et al., 2003a). Not surprisingly then, thecomparison of climate and oceanographic variables with large-scale climate indices revealed several linkages. In the EBS andthe GoA, the predominant link of both climate and oceanographicvariables is with the PDO, a result consistent with earlier studies(Mantua et al., 1997; Hare and Mantua, 2000; Mantua and Hare,2002). In 1977, the PDO shifted from a cold to a warm phase withfar reaching effects, including major shifts in the Bering Sea andGulf of Alaska ecosystems (Hare and Mantua, 2000; Benson andTrites, 2002). Almost all of the climate indices for the EBS andGoA exhibit a significant change in 1977 consistent with thechange in the PDO. This change has been identified as a regimeshift by Francis et al. (1998) and Hare and Mantua (2000), amongothers. Such a shift was observed during our analysis in air temper-atures, wind velocity components, net heat fluxes, sea surface tem-peratures and sea ice.

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

One of the unexpected results in the exploration of climate indi-ces was the statistically significant importance of the Siberian–Aleutian Index (SAI) on the winds and sea ice in the Barents Sea.This connection, albeit weak, is consistent with the results of Sorte-berg and Kvingedal (2006). They were able to account for 77% ofthe interannual variability in winter sea-ice coverage in theBarents Sea from cyclone activity over eastern Siberia, 1–2 yearsearlier. They suggested the relationship was through the cycloneseffect on transport of ice from the Arctic into the Barents Sea.The correlations found in the present study were positive for Vwwinds and negative for sea ice. Thus a stronger northward windwould lead to reduced ice and a stronger southward wind, moreice. However, a direct influence of the SAI index over the BarentsSea is unlikely. It may be reflect the intensity or number of cy-clones but confirmation of such a relationship will require furtherinvestigation.

While this study has indicated the importance of advection incontrolling increases in temperature in the Barents and that air-sea heat exchanges are more important in the Bering, it must becautioned that these conclusions are largely based on NCEP data.Renfrew et al. (2002) in a study of heat fluxes over the LabradorSea found that the NCEP model overestimates the bulk fluxes by51% for Qs and 27% for Ql. The roughness length formula used inthe NCEP reanalysis project was found to be inappropriate formoderate to high wind speeds. Its failings are acute for situationsof large air–sea temperature difference and high wind speed, thatis, for areas of high sensible heat fluxes. However, it was not clearwhether this would greatly influence the interannual variability inthese variables or of Qt or just their absolute values. Future workshould examine the heat fluxes using other atmospheric datasets,such as the new CORE (Coordinated Ocean Reference Experiments)data (Large and Yeager, 2008).

In summary, comparison of a select number of climate andoceanographic variables from the eastern Bering Sea, the Gulf ofAlaska, the Gulf of Maine/Georges Bank and the Norwegian/BarentsSeas over the period 1948–2006 (1900–2006 for SSTs) has revealedsimilar patterns of variability in adjacent regions or subregions butlittle similarity across the Atlantic or between high latitude regionsin the Atlantic and Pacific. An exception to the lack of similarity isthe fact that all of the regions have experienced generally warmconditions during the last decade. However, the processes respon-sible for the warming across the regions differ. While air-sea heatfluxes are believed to play an important role in the warming in theeastern Bering Sea and the Gulf of Alaska, advection is consideredto be the dominant process in the Atlantic regions. This conclusionis supported by the observed salinity changes. While much of thehydrographic and circulation variability can be linked to changesin the large-scale atmospheric patterns, further research is re-quired to determine the exact processes responsible for these link-ages. The physical oceanographic and climate changes described inthis paper are already having strong impacts on the marine ecologyof the MENU regions. These changes are influencing many trophiclevels, from phytoplankton up to the fish, marine mammals andseabirds, through affects on their distribution, phenology, abun-dance and community structure (Grebmeier et al., 2006; Mueteret al., this volume). Improved understanding of the linkages be-tween climate, physical oceanography and ecological responsesshould be a high research priority as it is essential before adequateprojections of the changes to marine ecosystems under increasedanthropogenic climate change will be realized.

Acknowledgements

The authors wish to thank the Research Council of Norway(RCN) for funding the initial meetings of MENU. K.D. also wishesto acknowledge the Bjerknes Centre for Climate Research and the

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

1032

1033

1034

1035103610371038103910401041104210431044104510461047104810491050105110521053105410551056105710581059106010611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113

111411151116

18 K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

RNC-funded NESSAS project for their support for the research andwriting of this paper.

T

11171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178117911801181118211831184118511861187118811891190119111921193119411951196119711981199

UN

CO

RR

EC

References

Auad, G., Miller, A.J., 2008. The role of tidal forcing in the Gulf of Alaska’s circulation.Geophysical Research Letters 35, L04603. doi:10.1029/2007GL032727.

Benson, A.J., Trites, A.W., 2002. Ecological effects of regime shifts in the Bering Seaand eastern North Pacific Ocean. Fish and Fisheries 3, 95–113.

Bethke, I., Furevik, T., Drange, H., 2006. Towards a more saline North Atlantic and afresher Arctic under global warming. Geophysical Research Letters 33, L21712.

Beverton, R.J.H., Lee, A.J., 1965. Hydrographic fluctuations in the North AtlanticOcean and some biological consequences. In: Johnson, C.G., Smith, L.P. (Eds.),The Biological Significance of Climate Changes in Britain. Symposia of theInstitute of Biology, vol. 14. Academic Press, London, pp. 79–109.

Bigelow, H.B., 1927. Physical oceanography of the Gulf of Maine. Bulletin of theUnited States Bureau of Fisheries 40, 511–1027.

Blindheim, J., 1989. Ecological features of the Norwegian Sea. In: Rey, L., Alexander,V. (Eds.), Proceedings of the Sixth Conference of the Comite ArctiqueInternational, 13–15 May, 1985. E.J. Brill, New York, pp. 366–401.

Blindheim, J., Østerhus, S., 2005. The Nordic Seas, main oceanographic features. In:Drange, H., Dokken, T., Furevik, T., Gerdes, R., Berger, W. (Eds.), The Nordic Seas:An Integrated Perspective. Geophysical Monograph, vol. 158. AmericanGeophysical Union, Washington, USA, pp. 11–38.

Brown, W., Moody, J., 1987. Tides. In: Backus, R.H. (Ed.), Georges Bank. MIT Press,Cambridge, USA, pp. 100–107.

Butman, B., Beardsley, R.C., 1987. An introduction to the physical oceanography ofGeorges Bank. In: Backus, R.H. (Ed.), Georges Bank. MIT Press, Cambridge, USA,pp. 88–98.

Chapman, D.C., Beardsley, R.C., 1989. On the origin of shelf water in the MiddleAtlantic Bight. Journal of Physical Oceanography 19, 389–391.

Clarke, A., Murphy, E.J., Meredith, M.P., King, J.C., Peck, L.S., Barnes, D.K.A., Smith,R.C., 2007. Climate change and the marine ecosystem of the western AntarcticPeninsula. Philosophical Transactions of the Royal Society of London Series B:Biological Sciences 362, 149–166.

Coachman, L.K., 1986. Circulation, water masses and fluxes on the southeasternBering Sea shelf. Continental Shelf Research 5, 23–108.

Copeland, L., Mueller, D.R., Weir, L., 2007. Rapid loss of the Ayles Ice Shelf, EllesmereIsland, Canada. Geophysical Research Letters 34, L21501.

Drinkwater, K.F., Belgrano, A., Borja, A., Conversi, A., Edwards, M., Greene,C.H., Ottersen, G., Pershing, A.J., Walker, H., 2003a. The response ofmarine ecosystems to climate variability associated with the NorthAtlantic Oscillation. In: Hurrell, J., Kushnir, Y., Ottersen, G., Visbeck, M.(Eds.), The North Atlantic Oscillation, Climatic Significance andEnvironmental Impact. Geophysical Monograph Series, vol. 134. AGUPress, pp. 211–234.

Drinkwater, K.F., Petrie, B., Smith, P.C., 2003b. Climate variability on the ScotianShelf during the 1990s. ICES Marine Science Symposium 219, 40–49.

Francis, R.C., Hare, S.R., Hollowed, A.B., Wooster, W.S., 1998. Effects ofinterdecadal climate variability on the Northeast Pacific. FisheriesOceanography 7, 1–21.

Gaichas, S., Skaret, G., Falk-Petersen, J., Link, J.S., Overholtz, W., Megrey, B.A.,Gjøs�ter, H., Stockhausen, W.T., Dommasnes, A., Friedland, K.D., Aydin, K.Y.,this volume. A comparison of community and trophic structure in five marineecosystems based on energy budgets and system metrics. Progress inOceanography.

Garrett, C., 1972. Tidal resonance in the Bay of Fundy and Gulf of Maine. Nature 238,441–443.

Garrett, C., Toulany, B., 1981. Variability of the flow through the Strait of Belle Isle.Journal of Marine Research 39, 163–189.

Garrett, C.J.R., Keeley, J.R., Greenberg, D.A., 1978. Tidal mixing versus thermalstratification in the Bay of Fundy and Gulf of Maine. Atmosphere – Ocean 16,403–423.

Gjevik, B., Nøst, E., Straume, T., 1994. Model simulations of the tides in the BarentsSea. Journal of Geophysical Research 99 (C2), 3337–3350.

Grebmeier, J.M., Overland, J.E., Moore, S.E., Farley, E.V., Carmack, E.C., Cooper, L.W.,Prey, K.E., Helle, J.H., McLaughlin, F.A., McNutt, S.L., 2006. A major ecosystemshift in the northern Bering Sea. Science 311, 1461–1464.

Greene, C.H., Pershing, A.J., Cronin, T.M., Ceci, N., 2008. Arctic climate change and itsimpacts on the ecology of the North Atlantic. Ecology 89, S24–S38. doi:10.1890/07-0550.1.

Haas, C., Pfaffling, A., Hendricks, S., Rabenstein, L., Etienne, J.-L., Rigor, I., 2008.Reduced ice thickness in Arctic Transpolar Drift favors rapid ice retreat.Geophysical Research Letters 35, L17501. doi:10.1029/2008GL034457.

Hare, S.R., Mantua, N.J., 2000. Empirical evidence for North Pacific regime shifts in1977 and 1989. Progress in Oceanography 47, 103–145.

Harris, C.L., Plueddemann, A.J., Gawarkiewicz, G.C., 1998. Water mass distributionand polar front structure in the western Barents Sea. Journal of GeophysicalResearch 103 (C2), 2905–2917.

Holliday, N.P., Hughes, S.L., Bacon, S., Beszczynska-Möller, A., Hansen, B., Lavin, A.,Loeng, H., Mork, K.A., Østerhus, S., Sherwin, T., Walczowski, W., 2008. Reversalof the 1960s to 1990s freshening trend in the northeast North Atlantic andNordic Seas. Geophysical Research Letters 35, L03614. doi:10.1029/2007GL032675.

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

Hollowed, A.B., Hare, S.R., Wooster, W.S., 2001. Pacific Basin climate variability andpatterns of Northeast Pacific marine fish production. Progress in Oceanography49, 257–282.

Horne, E.P.W., Loder, J.W., Harrison, W.G., Mohn, R., Lewis, M.R., Irwin, B., Platt, T.,1989. Nitrate supply and demand at the Georges Bank tidal front. ScientiaMarina 53, 145–158.

Hunt Jr., G.L., Megrey, B.A., 2005. Comparison of the biophysical and trophiccharacteristics of the Bering and Barents Seas. ICES Journal of Marine Science62, 1245–1255.

Hunt Jr., G.L., Stabeno, P., Walters, G., Sinclair, E., Brodeur, R.D., Napp, J.M., Bond,N.A., 2002. Climate change and control of the southeastern Bering Sea pelagicecosystem. Deep-Sea Research Part II 49, 5821–5853.

Hunt Jr., G.L., Stabeno, P.J., Strom, S., Napp, J.M., 2008. Patterns of spatial andtemporal variation in the marine ecosystem of the southeastern Bering Sea,with special reference to the Pribilof Domain. Deep-Sea Research Part II 55,1919–1944.

Hurrell, J.W., 1995. Decadal trends in the North Atlantic Oscillation: regionaltemperatures and precipitation. Science 269, 676–679.

Ingvaldsen, R., 2005. Width of the North Cape Current and location of the PolarFront in the western Barents Sea. Geophysical Research Letters 32, L16603.doi:10.1029/2005GL023440.

Ingvaldsen, R., Loeng, H., Ottersen, G., Ådlandsvik, B., 2003. Climate variability in theBarents Sea during the 20th century with focus on the 1990s. ICES MarineScience Symposia 219, 160–168.

Ingvaldsen, R.B., Asplin, L., Loeng, H., 2004. Velocity field of the western entrance tothe Barents Sea. Journal of Geophysical Research 109, C03021. doi:10.1029/2003JC001811.

IPCC, 2007. Climate Change 2007: the physical science basis. Solomon, S., Qin, D.,Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.),Contribution of Working Group I to the Fourth Assessment Report of theIntergovernmental Panel on Climate Change. Cambridge University Press,Cambridge, UK, p. 996.

Iverson, R.L., Coachman, L.K., Cooney, R.T., English, T.S., Goering, J.J., Hunt Jr., G.L.,Macauley, M.C., McRoy, C.P., Reeburgh, W.S., Whitledge, T.E., 1979. Ecologicalsignificance of fronts in the southeastern Bering Sea. In: Livingston, R.J. (Ed.),Ecological Processes in Coastal and Marine Systems. Plenum Press, New York,NY, pp. 437–466.

Jones, E.P., Swift, J.H., Anderson, L.G., Lipizer, M., Civitarese, G., Falkner, K.K., Kattner,G., McLaughlin, F., 2003. Tracing Pacific water in the North Atlantic Ocean.Journal of Geophysical Research 108 (C4), 3116. doi:10.1029/2001JC001141.

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredall, M.,Saha, S., Withe, G., Woollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgens, W.,Janowiak, J., Mo, K.C., Ropelewski, C., Wang, J., Leetma, A., Reynolds, R., Jenne, R.,Joseph, D., 1996. The NCEP/NCAR 40-year reanalysis project. Bulletin of theAmerican Meteorological Society 77, 437–471.

Kowalik, Z., 1999. Bering Sea tides. In: Loughlin, T.R., Ohtani, K. (Eds.), The BeringSea: Physical, Chemical and Biological Dynamics. Alaska Sea Grant Press,Fairbanks, AK, pp. 93–127.

Kowalik, Z., Proshuntinsky, A.Y., 1995. Topographic enhancement of tidal motion inthe western Barents Sea. Journal of Geophysical Research 100 (C2), 2613–2637.

Kowalik, Z., Stabeno, P.J., 1999. Trapped motion around the Pribilof Islands in theBering Sea. Journal of Geophysical Research 104, 25667–25684.

Ladd, C., Stabeno, P., Cokelet, E.D., 2005. A note on the cross-shelf exchange in thenorthern Gulf of Alaska. Deep-Sea Research Part II 52, 667–679.

Large, W.G., Yeager, S.G, 2008. The global climatology of an interannually varyingair-sea flux data set. Climate Dynamics. doi: 10.1007/s00382-008-0441-3.

Link, J.S., Stockhausen, W.T., Skaret, G., Overholtz, W., Megrey, B.A., Gjøs�ter, H.,Gaichas, S., Dommasnes, A., Falk-Petersen, J., Kane, J., Mueter, F.J., Friedland,K.D., Hare, J.A., this volume. A comparison of biological trends from four marineecosystems: synchronies, differences, and commonalities. Progress inOceanography.

Loder, J.W., 1980. Topographic rectification of tidal currents on the sides of GeorgesBank. Journal of Physical Oceanography 10, 1399–1416.

Loeng, H., 1979. A review of the sea ice conditions of the Barents Sea and the areawest of Spitsbergen. Fisken og Havet 1979 (2), 29–75 (In Norwegian withEnglish abstract).

Loeng, H., 1991. Features of the physical oceanographic conditions of the BarentsSea. Polar Research 10, 5–18.

Manning, J.P., Beardsley, R.C., 1993. Assessment of Georges Bank recirculation fromEulerian current observations in the Great South Channel. Deep-Sea ResearchPart II 43, 1575–1600.

Mantua, N.J., Hare, S.R., 2002. The Pacific Decadal Oscillation. Journal ofOceanography 58, 35–44.

Mantua, N.J., Hare, S.R., Zhang, Y., Wallace, J.M., Francis, R.C., 1997. A Pacificinterdecadal climate oscillation with impacts on salmon production. Bulletin ofthe American Meteorological Society 78, 1069–1079.

Massom, R.A., Stammerjohn, S.E., Lefebvre, W., Harangozo, S.A., Adams, N., Scambos,T., Pook, M.J., Fowler, C., 2008. West Antarctic Peninsula sea ice in 2005:extreme ice compaction and ice edge retreat due to strong anomaly withrespect to climate. Journal of Geophysical Research 113, C02S20. doi:10.1029/2007JC004239.

Mavor, T.P., Bisagni, J.J., 2001. Seasonal variability of sea-surface temperature frontson Georges Bank. Deep-Sea Research Part II 48, 215–243.

Megrey, B.A., Hare, J.A., Stockhausen, W.T., Dommasnes, A., Gjøs�ter, H., Overholtz,W., Gaichas, S., Skaret, G., Falk-Petersen, J., Link, J.S., Friedland, K.D., this volume.A cross-ecosystem comparison of spatial and temporal patterns of covariation

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003

T

12001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237123812391240124112421243124412451246124712481249125012511252125312541255125612571258

12591260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296129712981299130013011302130313041305130613071308130913101311131213131314131513161317

K.F. Drinkwater et al. / Progress in Oceanography xxx (2009) xxx–xxx 19

PROOCE 871 No. of Pages 19, Model 5G

16 April 2009 Disk UsedARTICLE IN PRESS

RR

EC

in the recruitment of functionally analogous fish stocks. Progress inOceanography.

Mountain, D.G., 2002. Variability of the water properties in NAFO subareas 5 and 6during the 1990s. Journal of the Northwest Atlantic Fishery Science 34, 103–112.

Mountain, D.G., 2003. Variability in the properties of Shelf Water in the MiddleAtlantic Bight, 1977–99. Journal of Geophysical Research 108 (C1), 3014.doi:10.1029/2001JC001044.

Mountain, D.G., Strout, G.A., Beardsley, R.C., 1996. Surface heat flux in the Gulf ofMaine. Deep-Sea Research Part II 43, 1533–1546.

Mueter, F.J., Broms, C., Drinkwater, K.F., Friedland, K.D., Hare, J.A., Hunt Jr., G., Melle,W., Taylor, M., this volume. Comparison of 4 Northern Hemisphere regions:ecosystem responses to recent oceanographic variability. Progress inOceanography.

Overland, J.E., Pease, C.H., 1982. Cyclone climatology of the Bering Sea and itsrelation to sea ice extent. Monthly Weather Review 110, 5–13.

Overland, J.E., Stabeno, P.J., 2004. Is the climate of the Bering Sea warming andaffecting the ecosystem? EOS Transactions of the American Geophysical Union85, 309–316.

Overland, J.E., Wang, M., 2005. The third Arctic climate pattern: 1930s and early200s. Geophysical Research Letters 32, L23808. doi:10.1029/2005GL024254.

Parsons, A.R., Bourke, R.H., Muench, R.D., Chiu, C.-S., Lynch, J.F., Miller, J.H.,Plueddemann, A.J., Pawlowicz, R., 1996. The Barents Sea Polar Front insummer. Journal of Geophysical Research 101 (C6), 14201–14221.

Polyakov, I.V., Beszczynska, A., Carmack, E.C., Dmitrenko, I.A., Fahrback, E., Frolov,I.E., Gerdes, R., Hansen, E., Holfort, J., Ivanov, V.V., Johnson, M.A., Karcher, M.,Dauker, F., Morison, J., Orvik, K.A., Schauer, U., Simmons, H.L., Skagseth, Ø.,Sokolov, V.T., Steele, M., Timokhov, L.A., Walsh, D., Walsh, J.E., 2005. One morestep toward a warmer Arctic. Geophysical Research Letters 32, L17605.doi:10.1029/2005GL023740.

Ramp, S.R., Schlitz, R.J., Wright, W.R., 1985. The deep flow through NortheastChannel, Gulf of Maine. Journal of Physical Oceanography 15, 1790–1808.

Reed, R.K., 1995. On the variable subsurface environment of fish stocks in the BeringSea. Fisheries Oceanography 4, 317–323.

Reed, R.K., Stabeno, P.J., 1999. The Aleutian North Slope Current. In: Loughlin, T.R.,Ohtani, K. (Eds.), Dynamics of the Bering Sea. University of Alaska Sea Grant, AK-SG-99-03, pp. 177–192.

Renfrew, I.A., Moore, G.W.K., Guest, P.S., Bumke, K., 2002. A comparison of surfacelayer and surface turbulent flux observations over the Labrador Sea withECMWF analyses and NCEP reanalyses. Journal of Physical Oceanography 32,383–400.

Royer, T.C., 1982. Coastal fresh water discharge in the northeast Pacific. Journal ofGeophysical Research 87, 2017–2021.

Royer, T.C., 2005. Hydrographic responses at a coastal site in the northern Gulf ofAlaska to seasonal and interannual forcing. Deep-Sea Research Part II 52, 267–288.

S�tre, R. (Ed.), 2007. The Norwegian Coastal Current. Tapir Academic Press,Trondheim, Norway. 159p.

Scambos, T., Hulbe, C., Fahnestock, M., 2003. Climate-induced ice shelfdisintegration in the Antarctic Peninsula. In: Domack, E., Leventer, A., Burnett,A., Bindschadler, R., Conley, P., Kirby, M. (Eds.), Antarctic Peninsula ClimateVariability: Historical and Paleoenvironmental Perspective. AmericanGeophysical Union, Washington, DC, pp. 335–347.

Schumacher, J.D., Reed, R.K., 1992. Characteristics of currents over thecontinental slope of the eastern Bering Sea. Journal of GeophysicalResearch 97, 9423–9433.

Schumacher, J.D., Stabeno, P.J., 1998. The continental shelf of the Bering Sea. In:Robinson, A.R., Brink, K.H. (Eds.), The Sea – The Global Coastal Ocean: RegionalStudies and Synthesis, vol. 11. Wiley, New York, NY, pp. 789–822.

UN

CO 1318

Please cite this article in press as: Drinkwater, K.F., et al. Recent climate forcinreview and comparison of four marine ecosystems. Prog. Oceanogr. (2009), do

ED

PR

OO

F

Schumacher, J.D., Bond, N.A., Brodeur, R.D., Livingston, P.A., Napp, J.M., Stabeno, P.J.,2003. Climate change in the southeastern Bering Sea and some consequencesfor biota. In: Hempel, G., Sherman, K. (Eds.), Large Marine Ecosystems of theWorld-Trends in Exploitation, Protection and Research. Elsevier, Amsterdam,pp. 17–40.

Skagseth, Ø., 2008. Recirculation of Atlantic Water in the western Barents Sea.Geophysical Research Letters 35, L11606. doi:10.1029/2008GL033785.

Skagseth, Ø., Furevik, T., Ingvaldsen, I., Loeng, H., Mork, K.A., Orvik, K.A., Ozhigin, V.,2008. Volume and heat transports to the Arctic Ocean via the Norwegian andBarents Seas. In: Dickson, R.R., Meincke, J., Rhines, P. (Eds.), Arctic–SubarcticOcean Fluxes. Springer, Dordrecht, The Netherlands, pp. 45–64.

Skjoldal, H.R., 2004. An introduction to the Norwegian Sea ecosystem. In: Skjoldal,H.R. (Ed.), The Norwegian Sea Ecosystem. Tapir Academic Press, Trondheim, pp.15–32.

Skjoldal, H.R., Rey, F., 1989. Pelagic production and variability of the Barents Seaecosystem. In: Sherman, K., Alexander, L.M. (Eds.), Biomass Yields andGeography of Large Marine Ecosystems. AAAS Selected Symposium, vol. 111.Westview Press, Colorado, USA, pp. 241–296.

Smith, P.C., 1983. The mean and seasonal circulation off southwest Nova Scotia.Journal of Physical Oceanography 13, 1034–1054.

Sorteberg, A., Kvingedal, B., 2006. Atmospheric forcing on the Barents Sea winter iceextent. Journal of Climate 19, 4772–4784.

Stabeno, P.J., Schumacher, J.D., Salo, S.A., Hunt Jr., G.L., Flint, M., 1999. Physicalenvironment around the Pribilof Islands. In: Loughlin, T.R., Ohtani, K. (Eds.),Dynamics of the Bering Sea: Physical, Chemical and Biological Characteristicsand a Synopsis of Research on the Bering Sea. University of Alaska Sea Grant,AK-SG-99-03, Fairbanks, pp. 193–215.

Stabeno, P.J., Reed, R.K., Napp, J.M., 2002. Transport through Unimak Pass, Alaska.Deep-Sea Research Part II 49, 5919–5930.

Stabeno, P.J., Bond, N.A., Hermann, A.J., Overland, J.E., 2004. Meteorology andoceanography of the Northern Gulf of Alaska. Continental Shelf Research 24,697–859.

Stabeno, P.J., Kachel, D.G., Kachel, N.B., Sullivan, M.E., 2005. Observations frommoorings in the Aleutian Passes: temperature, salinity and transport. FisheriesOceanography 14, 39–54.

Stabeno, P.J., Kachel, N.B., Mordy, C.W., Righi, D., Salo, S.A., 2008. An examination ofthe physical variability around the Pribilof Islands in 2004. Deep-Sea ResearchPart II 55, 1701–1716.

Stroeve, J., Serreze, M., Drobot, S., Gearheard, S., Holland, M., Maslanik, J., Meier, W.,Scambos, T., 2008. Arctic sea ice extent plummets in 2007. EOS Transactions ofthe American Geophysical Union 89, 13–14.

Tereshchenko, V.V., 1996. Seasonal and year-to-year variations of temperature andsalinity along the Kola meridian transect. ICES CM 1996/C:11, 24p.

Thompson, D.W.J., Wallace, J.M., 1998. The Arctic Oscillation signature in thewintertime geopotential height and temperature fields. Geophysical ResearchLetters 25, 1297–1300.

Vinje, T., 2001. Anomalies and trends of sea-ice extent and atmospheric circulationin the Nordic Seas during the period 1864–1998. Journal of Climate 14, 255–267.

Vinje, T., Kvambekk, Å., 1991. Barents Sea drift ice characteristics. Polar Research 10,59–68.

Weingartner, T.J., Coyle, K., Finney, B., Hopcroft, R., Whitledge, T., Brodeur, R.,Dagg, M., Farley, E., Haidvogel, D., Haldorson, L., Hermann, A., Hinckley,S., Napp, J., Stabeno, P., Kline, T., Lee, C., Lessard, E., Royer, T., Strom, S.,2002. The Northeast Pacific GLOBEC Program: Coastal Gulf of Alaska.Oceanography 15, 48–63.

Wouters, B., Chambers, D., Schrama, E.J.O., 2008. GRACE observes small-scale massloss in Greenland. Geophysical Research Letters L20501. doi:10.1029/2008GL034816.

g and physical oceanographic changes in Northern Hemisphere regions: Ai:10.1016/j.pocean.2009.04.003