10.10.2014 Views

1 A Prototype Two-Decade Fully-Coupled Fine-Resolution CCSM ...

1 A Prototype Two-Decade Fully-Coupled Fine-Resolution CCSM ...

1 A Prototype Two-Decade Fully-Coupled Fine-Resolution CCSM ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

1<br />

2<br />

3<br />

4<br />

5<br />

6<br />

7<br />

8<br />

9<br />

10<br />

11<br />

12<br />

13<br />

14<br />

15<br />

16<br />

17<br />

18<br />

19<br />

20<br />

21<br />

22<br />

23<br />

24<br />

25<br />

26<br />

27<br />

28<br />

29<br />

30<br />

31<br />

32<br />

33<br />

34<br />

35<br />

36<br />

37<br />

38<br />

39<br />

40<br />

41<br />

42<br />

43<br />

44<br />

A <strong>Prototype</strong> <strong>Two</strong>-<strong>Decade</strong> <strong>Fully</strong>-<strong>Coupled</strong> <strong>Fine</strong>-<strong>Resolution</strong> <strong>CCSM</strong> Simulation<br />

Julie L. McClean 1,2 , David C. Bader 3,2 , Frank O. Bryan 4 , Mathew E. Maltrud 5 , John M. Dennis 4 ,<br />

Arthur A. Mirin 2 , Philip W. Jones 5 , Yoo Yin Kim 1 , Detelina P. Ivanova 2 , Mariana Vertenstein 4 ,<br />

James S. Boyle 2 , Robert L. Jacob 6 , Nancy Norton 4 , Anthony Craig 4 , and Patrick H. Worley 3<br />

1 Scripps Institution of Oceanography, La Jolla, CA 92093<br />

2 Lawrence Livermore National Laboratory, Livermore CA 94551<br />

3 Oak Ridge National Laboratory, Oak Ridge, TN 37831<br />

4 National Center for Atmospheric Research, Boulder, CO 80303<br />

5 Los Alamos National Laboratory, Los Alamos, NM 87545<br />

6 Argonne National Laboratory, Argonne, IL 60439<br />

PUBLISHED<br />

Ocean Modelling, 39, 10-30, 2011<br />

Corresponding author address:<br />

Dr Julie McClean<br />

Climate, Atmospheric Science and Physical Oceanography Division<br />

Scripps Institution of Oceanography<br />

University of California, San Diego<br />

9500 Gilman Drive, 0230<br />

La Jolla, CA 92093-0230, USA<br />

Ph: 858 534 3030/Fax: 858 534 9820<br />

Email: jmcclean@ucsd.edu<br />

1


44<br />

45<br />

46<br />

47<br />

48<br />

49<br />

50<br />

51<br />

52<br />

53<br />

54<br />

55<br />

56<br />

57<br />

58<br />

59<br />

60<br />

61<br />

62<br />

63<br />

64<br />

Abstract<br />

A fully coupled global simulation using the Community Climate System Model (<strong>CCSM</strong>) was<br />

configured using grid resolutions of 0.1° for the ocean and sea-ice, and 0.25° for the atmosphere<br />

and land, and was run under present-day greenhouse gas conditions for 20 years. It represents<br />

one of the first efforts to simulate the planetary system at such high horizontal resolution. The<br />

climatology of the circulation of the atmosphere and the upper ocean were compared with<br />

observational data and reanalysis products to identify persistent mean climate biases. Intensified<br />

and contracted polar vortices, and too cold sea surface temperatures (SSTs) in the subpolar and<br />

mid-latitude Northern Hemisphere were the dominant biases produced by the model. Intense<br />

category 4 cyclones formed spontaneously in the tropical North Pacific. A case study of the<br />

ocean response to one such event shows the realistic formation of a cold SST wake, mixed layer<br />

deepening, and warming below the mixed layer. Too many tropical cyclones formed in the North<br />

Pacific however, due to too high SSTs in the tropical eastern Pacific. In the North Atlantic<br />

anomalously low SSTs lead to a dearth of hurricanes. Agulhas eddy pathways are more realistic<br />

than in equivalent stand-alone ocean simulations forced with atmospheric reanalysis.<br />

Keywords:<br />

Numerical modelling, atmospheric circulation, ocean circulation, ocean eddies, tropical cyclones.<br />

2


65<br />

66<br />

67<br />

68<br />

69<br />

70<br />

71<br />

72<br />

73<br />

74<br />

75<br />

76<br />

77<br />

78<br />

79<br />

80<br />

81<br />

82<br />

83<br />

84<br />

85<br />

86<br />

87<br />

88<br />

89<br />

90<br />

91<br />

92<br />

93<br />

94<br />

95<br />

96<br />

97<br />

98<br />

99<br />

100<br />

101<br />

102<br />

103<br />

104<br />

105<br />

106<br />

107<br />

108<br />

109<br />

110<br />

1. Introduction<br />

Simulating the Earth System realistically under present day conditions, a necessary precursor to<br />

climate projection, represents a major scientific challenge. The veracity of such simulations is<br />

undermined by missing and inadequately represented processes in component models together<br />

with feedback errors within and among components (Large and Danabasoglu, 2006).<br />

Historically, a successful approach to reducing model biases related to poor fluid dynamical<br />

simulation is to increase the spatial resolution of the component models such that a greater<br />

fraction of the dynamical processes are explicitly resolved, and fewer are parameterized. Errors<br />

in the representation of physical processes may dominate systematic model biases and these<br />

process-related biases may not necessarily decrease in magnitude with increasing resolution.<br />

However, understanding how the explicit simulation of small-scale physical processes impacts<br />

the depiction of coupled processes as well as large-scale climate features in a fully coupled<br />

global system is an essential step towards reproducing a realistic present day climate system.<br />

Here we present results from a two-decade, present-day simulation of the Community Climate<br />

System Model version 4 (<strong>CCSM</strong>4) where a weather-scale resolution atmospheric model is<br />

coupled to an eddying ocean model using atmospheric and ocean/sea-ice components having<br />

approximately five and ten times the horizontal resolution of standard climate component<br />

models. Others have coupled combinations of high and low-resolution components in the <strong>CCSM</strong><br />

framework such as Gent et al. (2009) who use 0.5° atmospheric/land and nominal 1° oceanic/ice<br />

components. This is the first study, to our knowledge, where all components have fine enough<br />

horizontal resolution for both the ocean and the atmosphere to explicitly simulate turbulent<br />

instabilities of the large-scale circulation.<br />

Generally the use of finer resolution in atmospheric global climate models has led to systematic<br />

improvements in the fidelity of atmospheric features, particularly those associated with the largescale<br />

dynamical circulation (Hack et al. 2006). Williamson et al. (1995) showed that the scales of<br />

motion included at T63 (~1.9° transform grid) and higher resolutions were needed to capture the<br />

non-linear processes that appear to drive some larger scale flows. The coupled atmosphere-ocean<br />

dynamics of tropical cyclones are correctly simulated only by the use of very-high horizontal<br />

resolution in both the atmosphere and ocean. In the ocean, the amount of heat mixed downwards<br />

by typhoons may be an important component of global meridional heat transport (Sriver and<br />

Huber, 2007). Accordingly, their explicit simulation in global climate models may be key toward<br />

making assessments of the impact of anthropogenic forcing on mean climate, variability and<br />

extreme events (Emanuel, 2008; Korty et al. 2008; Bengtsson et al. 2007).<br />

Decreasing the horizontal mesh spacing to around 10 km in ocean-only simulations has<br />

improved the depiction of the mean state and variability of the ocean circulation relative to that<br />

in simulations using grid spacing of about 20-50 km (Smith et al., 2000; Hurlburt and Hogan,<br />

2000; McClean et al., 2002; Oschlies, 2002, Maltrud and McClean, 2005; Bryan et al., 2007).<br />

Bryan et al. (2007) discuss particulars of these studies and comment that the improvements<br />

appear to represent a regime shift in the flow dynamics as the grid spacing approaches roughly<br />

10 km. This is generally attributed to the first baroclinic mode Rossby radius being reasonably<br />

well resolved up to approximately 50 degrees latitude (Smith et al., 2000, Figure 1) by such a<br />

grid.<br />

3


111<br />

112<br />

113<br />

114<br />

115<br />

116<br />

117<br />

118<br />

119<br />

120<br />

121<br />

122<br />

123<br />

124<br />

125<br />

126<br />

127<br />

128<br />

129<br />

130<br />

131<br />

132<br />

133<br />

134<br />

135<br />

136<br />

137<br />

138<br />

139<br />

140<br />

141<br />

142<br />

143<br />

144<br />

145<br />

146<br />

147<br />

148<br />

149<br />

150<br />

151<br />

152<br />

153<br />

154<br />

155<br />

156<br />

Specifically improvements generally result when fast narrow (about 100-200 km) flows such as<br />

the Gulf Stream and most of the mesoscale variability (defined as having spatial scales of 50 to<br />

500 km on timescales between 20 and 150 days) are explicitly resolved, together with their eddymean<br />

flow interactions that hitherto were not simulated. Mesoscale variability accounts for about<br />

50% of the variability in the global ocean (Chelton et al., 2007). The Los Alamos Parallel Ocean<br />

Program (POP) ocean general circulation model has been shown to resolve much of the oceanic<br />

mesoscale, both in regional (Smith et al., 2000) and global domains (Maltrud and McClean,<br />

2005, McClean et al., 2006; 2008) when configured with a nominally 0.1º grid.<br />

Understanding the role and importance of mesoscale eddies in the meridional overturning<br />

circulation (MOC) is needed to improve the depiction of present-day conditions in climate<br />

simulations. Böning et al. (2008) argue that in the Southern Ocean eddy fluxes compensate<br />

increased Ekman transport during decadal-scale positive Southern Annular Mode (SAM) events.<br />

Screen et al. (2009) report eddy-driven warming to the south of the Polar Front in the southern<br />

ocean one to three years after a positive SAM anomaly from a 1/12° global ocean primitive<br />

equation model. They consider the enhanced poleward heat flux to be likely associated with the<br />

intensified mesoscale eddy activity. Another example of a mesoscale process of importance to<br />

climate is the Agulhas leakage at the southern tip of Africa, which is characterized by variability<br />

on intraseasonal to interannual timescales. Agulhas eddies, which are resolved in finer resolution<br />

simulations, but are absent in the oceans of standard climate models are the main source of the<br />

warm and salty waters carried towards the North Atlantic as the upper limb of the Atlantic MOC<br />

(Biastoch et al., 2008).<br />

The primary goal of this paper is to examine the general fidelity of air-ocean interaction<br />

processes in the fully coupled fine-resolution <strong>CCSM</strong>4 simulation. We start by describing the fine<br />

resolution coupled model set-up and then compare the state of the simulated atmosphere and the<br />

upper ocean with observational data and reanalysis products to identify persistent mean climate<br />

biases. Next we examine the explicit simulation of the ocean’s response to a spontaneously<br />

generated typhoon event in the North Pacific. Finally we compare the pathways of Agulhas<br />

eddies as depicted by <strong>CCSM</strong>4 and altimetry data to appreciate the role of an active atmosphere<br />

when modeling these climatically important mesoscale eddies.<br />

2. <strong>Coupled</strong> Model Description<br />

The global climate model used in this study is a prototype version of <strong>CCSM</strong>4; its components are<br />

the Community Atmospheric Model version 3.5 (CAM3.5), the Community Land Model version<br />

3 (CLM3; Dickinson et al., 2006), POP version 2.0 (POP2) and the Community Ice Code version<br />

4 (CICE4.0). CAM3.5 uses the Lin-Rood finite-volume dynamical core (Lin, 2004) and an<br />

updated CAM3 convection parameterization (Neale et al., 2008; Richter and Rasch; 2008). It<br />

was run on a 0.23° x 0.31° spherical grid with 26 levels. POP (Dukowicz and Smith, 1994) was<br />

configured on a tripole grid (Murray, 1996) that has poles in both Canada and Russia, resulting<br />

in a nearly isotropic distribution of resolution in the high northern latitudes. The horizontal grid<br />

spacing at the equator is 0.1º, with the latitudinal spacing decreasing with cosine (latitude).<br />

There are 42 vertical levels, varying smoothly in thickness from 10m at the surface to 250m at<br />

the maximum depth of 6000m. CICE 4.0 (LANL, 2008; Hunke and Dukowicz, 1997) is based on<br />

an Elastic-Viscous-Plastic (EVP) rheology, uses multiple ice thickness categories (Bitz and<br />

4


157<br />

158<br />

159<br />

160<br />

161<br />

162<br />

163<br />

164<br />

165<br />

166<br />

167<br />

168<br />

169<br />

170<br />

171<br />

172<br />

173<br />

174<br />

175<br />

176<br />

177<br />

178<br />

179<br />

180<br />

181<br />

182<br />

183<br />

184<br />

185<br />

186<br />

187<br />

188<br />

189<br />

190<br />

191<br />

192<br />

193<br />

194<br />

195<br />

196<br />

197<br />

198<br />

199<br />

200<br />

201<br />

202<br />

Lipscomb, 1999), and runs on the same horizontal grid as POP; CICE improvements are<br />

highlighted in Gent et al. (2009). Although this resolution may cause the EVP model to operate<br />

at its limits of validity, CICE has been extensively used at 9km resolution (Maslowski et al.,<br />

2004). In addition, some issues that are manifest only at high resolution have previously been<br />

identified and corrected (Lipscomb et al. 2007).<br />

The POP configuration is essentially identical to that used by Maltrud et al. (2010), though we<br />

will highlight some key elements here. The ocean topography interpolated onto this grid was<br />

formed from a combination of ETOPO2 (http://www.ngdc.noaa.gov/mgg/fliers/01mgg04.html)<br />

and BEDMAP (http://www.antarctica.ac.uk//bas_research/data/access/bedmap/) bathymetric<br />

data sets, with the two smoothly blended between 68°S and 71°S. The interpolation used a<br />

weighting scheme that decreased quadratically with distance from the model grid point over a<br />

length scale equal to twice the local grid spacing. In order to improve topographic interactions<br />

with the flow, partial bottom cells (PBCs; Pacanowski and Gnanadesikan, 1998) were<br />

implemented, with partial cell thicknesses limited to be no thinner than the thickness of the<br />

uppermost cell (10 meters), or 10% of the full cell thickness, whichever was greater. Only a few<br />

key passages (the Strait of Gibraltar, Canadian Archipelago) were manually modified to<br />

accommodate sufficient transport, or to terminate large estuaries. The <strong>CCSM</strong>4 overflow<br />

parameterization (Danabasoglu et al., 2010) was not implemented in the version of POP used in<br />

this experiment.<br />

POP used a second order accurate centered difference advection scheme for both horizontal<br />

momentum and tracers. Although this method can result in numerical over- and under-shoots,<br />

tests using the flux-limiting advection scheme in <strong>CCSM</strong>4’s POP were judged to be too diffusive<br />

at high resolution. Subgrid scale horizontal mixing was parameterized using biharmonic<br />

operators for momentum and tracers. The viscosity and diffusivity values vary spatially with the<br />

cube of the average grid length for a given cell (see Maltrud et al., 1998) and have equatorial<br />

values, denoted by the subscript 0, of ν 0 = -2.7 x 10 10 m 4 /s and κ 0 = -0.3 x 10 10 m 4 /s, respectively.<br />

Vertical mixing coefficients for momentum and tracers were obtained from the K-Profile<br />

Parameterization (Large et al., 1994). Background values for vertical tracer diffusion range from<br />

10 −5 m 2 /s near the surface to 10 −4 m 2 /s at depth, with viscosity values an order of magnitude<br />

higher. Density is calculated from the equation of state for seawater derived by McDougall et al.<br />

(2003). Freshwater fluxes to the ocean are applied as a virtual salt flux. Unlike the simulation of<br />

Maltrud et al. (2010), the time step was 4.65 minutes to allow for higher amplitude and higher<br />

frequency forcing.<br />

POP was initialized with a two-step process. First, the model was started at rest using potential<br />

temperature and salinity from the World Hydrographic Program Special Analysis Center (WHP<br />

SAC) climatology (Gouretski and Koltermann, 2004). High amplitude transients required that a<br />

time step of approximately one minute be used for several simulated days to avoid numerical<br />

instability. This ocean state was then used to initialize a 2-year <strong>CCSM</strong> simulation configured<br />

with a 0.5° atmosphere, the final state of which became the initial condition for the simulation<br />

described here. The standard initialization option in CICE was used whereby the specified nonlinear<br />

sea-ice thickness distribution was the same at each grid point over the entire domain<br />

poleward of 70°N and 60°S. This initialization option was chosen rather than use the ice state<br />

from the existing 2-year simulation as the ice in that simulation was showing indications of poor<br />

5


203<br />

204<br />

205<br />

206<br />

207<br />

208<br />

209<br />

210<br />

211<br />

212<br />

213<br />

214<br />

215<br />

216<br />

217<br />

218<br />

219<br />

220<br />

221<br />

222<br />

223<br />

224<br />

225<br />

226<br />

227<br />

228<br />

229<br />

230<br />

231<br />

232<br />

233<br />

234<br />

235<br />

236<br />

237<br />

238<br />

239<br />

240<br />

241<br />

242<br />

243<br />

244<br />

245<br />

246<br />

247<br />

248<br />

performance. We experimented with starting from an initial state with no ice however that failed<br />

probably because significant shocks due to large fluxes in the Polar regions accompany such a<br />

start up. CAM was at rest initially, with prescribed profiles of temperature, vapor and trace gas<br />

concentrations and was allowed to come into equilibrium. Aerosol concentrations were<br />

prescribed in the same manner as in earlier <strong>CCSM</strong>3 simulations (Collins et al., 2006). Presentday<br />

conditions were represented by 1990’s green house gas (GHG) concentrations, particularly<br />

355 ppmv for CO 2 . The land surface in CLM was initialized with climatologically observed<br />

surface types and vegetation.<br />

The components of the coupled model exchanged information at different time intervals, with the<br />

atmosphere, sea ice, and land models coupling every time step (15 minutes), and the ocean every<br />

6 hours. The ocean coupling interval was chosen as a compromise between an improvement in<br />

stability and accuracy over daily coupling that is still used in lower resolution versions of <strong>CCSM</strong>,<br />

and efficiency considerations associated with the additional communication and component<br />

synchronization required by more frequent coupling. It should be noted that the fluxes supplied<br />

to the ocean are computed using much more frequent updates of the atmospheric variables, then<br />

averaged over the ocean’s coupling period. The coupled code executed at 0.43 simulated years<br />

per day on 512 nodes on the Lawrence Livermore National Laboratory’s Atlas supercomputer.<br />

This machine has 1152 total nodes, each containing eight, 2.4 GHz processors connected by a<br />

high speed network.<br />

3. Climatology of the <strong>Coupled</strong> System<br />

In this section we compare the climatology of the simulated circulation of the atmosphere and<br />

the upper ocean with observational data and reanalysis products to identify persistent mean<br />

climate biases. We use averages from model years 13-19 whenever possible. This period is<br />

chosen as a trade-off between using years that are no longer clearly dominated by the adjustment<br />

of the component models to their initial conditions and the need for a sufficiently long period for<br />

the statistics to be meaningful. In addition, comparisons will be made with two other simulations<br />

that use this POP model configuration to provide contrasting realizations of the ocean and<br />

ocean/ice without the complexity of an active atmosphere. One is a global coupled CICE 4.0 and<br />

POP2.0 simulation forced with the repeat annual cycle (normal-year) Coordinated Ocean-ice<br />

Reference Experiment (CORE) dataset (Large and Yeager, 2004; Griffies et al., 2009), with the<br />

6-hourly forcing averaged to monthly values. The other is the stand-alone 0.1° POP2.0<br />

simulation of Maltrud et al. (2010). We then offer explanations for the large-scale ocean and seaice<br />

biases in the subpolar oceans largely in terms of atmospheric circulation biases. Details of<br />

the CLM results will not be discussed here.<br />

The coupled model used the standard set of atmospheric parameter (tuning) constants employed<br />

in the Gent et al. (2009) simulations, thus the model’s climatology is very similar to those of<br />

earlier <strong>CCSM</strong> simulations (Collins et al. 2006; Bala et al., 2008). Significantly, the net energy<br />

imbalance at the top of the atmospheric model was a small net energy input of less than 1 Wm -2 .<br />

Similarly, the zonal mean distribution of net energy flux at the model top was very close to<br />

values from Clouds and Earth’s Radiant Energy System (CERES) observations, although the<br />

longwave and shortwave components had compensating mean zonal errors between 5 and 15<br />

Wm -2 (in a mean of ~250-300 Wm -2 ) in the mid-latitudes, which is in excess of the 1-4 Wm -2<br />

6


249<br />

250<br />

251<br />

252<br />

253<br />

254<br />

255<br />

256<br />

257<br />

258<br />

259<br />

260<br />

261<br />

262<br />

263<br />

264<br />

265<br />

266<br />

267<br />

268<br />

269<br />

270<br />

271<br />

272<br />

273<br />

274<br />

275<br />

276<br />

277<br />

278<br />

279<br />

280<br />

281<br />

282<br />

283<br />

284<br />

285<br />

286<br />

287<br />

288<br />

289<br />

290<br />

291<br />

292<br />

293<br />

294<br />

observational uncertainty in the CERES data (Loeb et al., 2008). The primary driver of the global<br />

atmospheric circulation is the net differential heating as a function of latitude (Sellers, 1969; and<br />

others). Since the net flux errors are small, we would not expect them to have a significant<br />

influence on the short seven-year climatology presented here. On the other hand, smaller but<br />

persistent longer term trends in the model biases may result from the component errors,<br />

especially when cloud feedbacks are considered in climate change scenarios. While the model<br />

climatology overall is very good considering that no tuning was conducted, below we highlight<br />

biases that would have implications for the mesoscale coupled atmosphere-ocean features.<br />

A comparison of precipitation climatologies from <strong>CCSM</strong>4 and Xie-Arkin (1997) observations<br />

reveal errors that exceed 5 mm day -1 across a long and narrow band in the tropical north Pacific<br />

(Fig. 1). It also shows that higher resolution does not improve the common problem of a double<br />

or split inter-tropical convergence zone (ITCZ) (Lin, 2007), manifested as the ~5 mm day -1<br />

precipitation deficit over the tropical western Pacific ocean and the Bay of Bengal. Neale and<br />

Slingo (2003) noted that increasing resolution alone does not solve the problem. More recent<br />

studies, including those of Wapler et al. (2010) and Boyle and Klein (2010), attempt to<br />

understand the complex interactions between microscale cloud physics and mesoscale<br />

atmospheric dynamics that are missing in GCM simulations. This remains an active topic of<br />

research, as the solution remains elusive. Less serious errors, such as the too extreme<br />

precipitation over the mountain ranges on the western coasts of the Americas, expose the<br />

requirement to adjust precipitation and cumulus parameterizations to properly reproduce the<br />

physics on the correct scales.<br />

The difference between climatological annual mean potential temperature from the upper-most<br />

ocean level (5m) for model years 13-19 and 0.25° daily Reynolds et al. (2007) satellite-derived<br />

SST data averaged over 1982-2008 (Fig 2a) is seen in Fig. 2b. Cold biases dominate in the<br />

Northern Hemisphere (NH) with extreme values of 10° and 6°C, respectively (Fig. 2b) occurring<br />

in the subpolar North Atlantic (NA) and Pacific (NP); in the subtropics they are up to 1°C. The<br />

highest warm biases are due to poleward displacements of the Gulf Stream (GS) and Kuroshio<br />

Current Extension (KE); more modest biases are found across much of the tropical NP (< 1°C)<br />

and in the Southern Ocean (SO) (up to 2°C). In upwelling zones along the west coasts of North<br />

and South America SST is generally up to 2°C too cold. Off the west coast of Africa surface<br />

waters between 0°and 20°S show a bias of the same magnitude but opposite sign; these<br />

contrasting biases are likely due to the errors in the representation of the dominant processes<br />

controlling the regional upwelling. Warm biases of greater than 7°C were reported in these<br />

upwelling zones in the standard <strong>CCSM</strong>3 (Collins et al. 2006). Gent et al. (2009) attribute their<br />

reduced SST biases in these regions to better-resolved orography that situates stronger winds<br />

closer to the coast, inducing increased upwelling.<br />

Equivalent seasonal climatological SST bias figures for years 13-19 (not shown) primarily show<br />

that the NH cold biases persist all year. In the NH subtropical gyres the extent and magnitude (up<br />

to 1.5°C) of the cold biases are greatest in boreal summer and fall, defined here as June-July-<br />

August (JJA) and September-October-November (SON), respectively. The positive biases<br />

associated with the KE and the GS are smaller in extent and magnitude relative to those in the<br />

December-January-February (DJF) and March-April-May (MAM) seasons. In the SO cold<br />

biases of generally up to 0.5°C are seen across the entire region in JJA and SON; the biases are<br />

7


295<br />

296<br />

297<br />

298<br />

299<br />

300<br />

301<br />

302<br />

303<br />

304<br />

305<br />

306<br />

307<br />

308<br />

309<br />

310<br />

311<br />

312<br />

313<br />

314<br />

315<br />

316<br />

317<br />

318<br />

319<br />

320<br />

321<br />

322<br />

323<br />

324<br />

325<br />

326<br />

327<br />

328<br />

329<br />

330<br />

331<br />

332<br />

333<br />

334<br />

335<br />

336<br />

337<br />

338<br />

339<br />

almost all positive in DJF (up to 2.5°C) with reduced warm biases occurring during MAM. In the<br />

eastern tropical Pacific the largest cold bias both in extent and magnitude (>2°C) is found in<br />

MAM. Maximum warm biases of >2.5° and >1°C, occur in JJA in the western tropical Indian<br />

Ocean and central tropical Pacific, respectively.<br />

Temperature and salinity biases also appear below the surface as SST and sea surface salinity<br />

(SSS) errors are transmitted into the interior by downwelling processes. Fig. 3 shows the<br />

magnitude of the subsurface drifts averaged over various latitude bands (78°S-50°S, 50°S-20°S,<br />

20°S-20N, 20°N-50°N, and 50°N-90°N). In the Polar regions, the temperature biases are<br />

strongest near the surface and show the same pattern as SST. However in the subtropics the<br />

trend is small at the surface and increases with depth as anomalously warm mode waters are<br />

ventilated into the gyres. In the tropics, the pattern reflects changes in the layered equatorial<br />

current system. The depth dependence of the salinity trend (Fig. 3b) is similar to that of<br />

temperature in the tropics and the Polar regions, while that of the subtropical NH is quite<br />

different to that of temperature. While the surface average shows a moderate freshening, the<br />

Atlantic and Pacific display opposite trends, with an overall increase in SSS in the Atlantic and<br />

decrease in the Pacific (not shown).<br />

Insight into some of the causes of the large-scale SST biases in the high and mid-latitudes are<br />

obtained by examining the atmospheric circulation. In Figs. 4a and c we show annual<br />

climatologies of 500 mb geopotential height and sea level pressure from European Center for<br />

Medium-Range Forecasts-European Reanalysis 40 (ECMWF-ERA40) along with their<br />

differences from <strong>CCSM</strong>4 (Figs 4b, d). In Fig. 5a surface wind stress vectors and their<br />

magnitudes from QuikSCAT (September 1999-August 2007) are shown together with the vector<br />

differences and their magnitudes between <strong>CCSM</strong>4 and QuikSCAT (Fig. 5b). The <strong>CCSM</strong>4 fields<br />

are averaged over model years 13-19. A major feature of this simulation is the intensification and<br />

contraction of both polar vortices through the depth of the troposphere. In Fig. 4b, the<br />

geopotential height anomalies are less than -100 m over Eurasia, while positive anomalies<br />

greater than 40 m can be seen over large areas between 40° and 60°S. The sea level pressure<br />

difference field (Fig. 4d) depicts the surface manifestation of this problem, with excessively deep<br />

low-pressure systems over the poles and intensification of the subtropical highs. The circulation<br />

around the intensified Azores High likely contributes to the anomalously cold SSTs in the<br />

tropical Atlantic. Similarly the wind stress vector differences between <strong>CCSM</strong>4 and QuikSCAT<br />

show that <strong>CCSM</strong>4 winds are too strong over the subpolar NP and NA oceans and over much of<br />

the SO (Fig. 5b). Particularly, <strong>CCSM</strong>4 wind stress magnitudes (plot not shown) exceed the<br />

QuikSCAT values (Fig. 5a) by up to 0.09 Nm -2 in the eastern subpolar NA where the winds are<br />

southwesterly. In this area root-mean-square (RMS) wind stress magnitude differences between<br />

QuikSCAT and the National Oceanography Center flux climatology are about 0.04 Nm -2 (Risien<br />

and Chelton, 2008); clearly here the model-data wind stress magnitude differences exceed the<br />

uncertainty in the QuikSCAT wind stresses. The contraction and intensification of the polar<br />

vortex, and resulting biases in the planetary wave structure, are described by Boville (1984). The<br />

problems most likely result from errors in the gravity-wave and mountain wave<br />

parameterizations, which lead to an anomalously strong polar-night jet trapping energy in the<br />

troposphere.<br />

8


340<br />

341<br />

342<br />

343<br />

344<br />

345<br />

346<br />

347<br />

348<br />

349<br />

350<br />

351<br />

352<br />

353<br />

354<br />

355<br />

356<br />

357<br />

358<br />

359<br />

360<br />

361<br />

362<br />

363<br />

364<br />

365<br />

366<br />

367<br />

368<br />

369<br />

370<br />

371<br />

372<br />

373<br />

374<br />

375<br />

376<br />

377<br />

378<br />

379<br />

380<br />

381<br />

382<br />

383<br />

384<br />

385<br />

The NH winter sea-ice pack extends eastward across the entire Labrador Sea and reaches across<br />

the East Greenland Sea into the Norwegian Sea (not shown). This is most likely caused by the<br />

overly intense polar atmospheric circulation rather than the zonally-symmetric initial ice<br />

configuration poleward of 70°N. In the Arctic the seasonal amplitude of total ice area (Fig. 6a) is<br />

seen to increase during the first 8 years of the simulation and then maintains the same too large<br />

magnitude relative to NASA Team Special Sensor Microwave/Imager (SSM/I) observations<br />

(Cavalieri et al., 1996; 2008) for the remaining years of the simulation. The SSM/I data is for the<br />

period 1980-1999. Similarly time series of area-averaged Arctic ice thickness and concentration<br />

(not shown) do not show any significant trends after the initial ice spin-up phase during the first<br />

decade of the simulation. In the Southern Hemisphere (SH) sea-ice is initially placed everywhere<br />

poleward of 60°S; however its area then shrinks rapidly over the first five years of the simulation<br />

and a seasonal cycle results whose amplitude is quite realistic (Fig. 6b). Here as well, time series<br />

of area-averaged ice concentration and ice thickness (not shown) in the second decade of the<br />

simulation show no significant trends and their respective seasonal amplitudes are generally<br />

repeating. Hence, memory of the sea-ice initial condition diminishes rapidly. In general the<br />

biases in the polar/subpolar winds, typically excessive northwesterlies (Fig. 5b), would produce<br />

excessive equatorward ice advection in the Nordic Seas whereas over the SO they would advect<br />

ice towards the South Pole.<br />

The too extensive sea-ice pack in <strong>CCSM</strong>4 and the displaced North Atlantic Current (NAC) cause<br />

anomalously cold surface waters to be located too far to the east in the NA subpolar gyre.<br />

Evaporative cooling of the surface water accompanies the overly strong subpolar winds that<br />

produce too strong Ekman currents that advect these cold surface waters toward the southeast,<br />

some of which enter the equatorward eastern boundary current (Fig.7). Flatau et al. (2003; Fig.<br />

12a) show the location of the 7° and 8° isotherms, a rough indicator of the NAC axis, during<br />

positive and negative phases of the North Atlantic Oscillation from SST observations during the<br />

1990s. In <strong>CCSM</strong>4 these isotherms (Fig. 7) are located even further eastwards than their observed<br />

counterparts during the positive NAO phase when anomalously strong wind stresses advect the<br />

NAC front towards the southeast. Subsequently in <strong>CCSM</strong>4 the anomalously cold water circulates<br />

around the subtropical gyre contributing to the cold bias seen in Fig. 2b. This same process is<br />

active in the subpolar NP.<br />

To appreciate the fidelity of the large-scale upper-ocean <strong>CCSM</strong>4 ocean circulation we compare<br />

mean sea surface height (SSH) for years 13-19 with the 0.5° mean dynamic ocean surface<br />

topography (MDOT) product of Maximenko et al. (2009) (Figs 8a and b). They combined<br />

satellite radar altimetry with the geoid model of the GRACE Recovery and Climate Experiment<br />

(GRACE) and synthesized upper-ocean velocities from 15-m drifting ocean buoys, hydrographic<br />

profiles, and ocean wind; inclusion of the latter data improved the depiction of the fine-scale<br />

dynamic topography. The product is global in coverage at latitudes equatorward of 65°. For<br />

contrast with <strong>CCSM</strong>4, we also consider the mean SSH field from stand-alone global 0.1° POP<br />

(Maltrud et al., 2010) for years 95-99 (Fig 8c). The patterns and magnitudes of mean SSH from<br />

<strong>CCSM</strong>4 agree best with MDOT (Fig. 8a). The difference plot between POP and MDOT (Fig.<br />

9b) highlight overly high POP mean SSH values in western boundary current frontal and<br />

recirculation regions: The Tasman Front in the southwestern Pacific near New Zealand, the<br />

Agulhas Current System, the KE and the GS-NAC. In the SO to the south of Australia and New<br />

Zealand and in the southeastern Atlantic sector POP values are more negative relative to MDOT.<br />

9


386<br />

387<br />

388<br />

389<br />

390<br />

391<br />

392<br />

393<br />

394<br />

395<br />

396<br />

397<br />

398<br />

399<br />

400<br />

401<br />

402<br />

403<br />

404<br />

405<br />

406<br />

407<br />

408<br />

409<br />

410<br />

411<br />

412<br />

413<br />

414<br />

415<br />

416<br />

417<br />

418<br />

419<br />

420<br />

421<br />

422<br />

423<br />

424<br />

425<br />

426<br />

427<br />

428<br />

429<br />

430<br />

<strong>CCSM</strong>4-MDOT (Fig. 9a) shows much more modest differences in these regions, except for<br />

narrow bands associated with the western boundary current and Antarctic Circumpolar Current<br />

(ACC) fronts. In the central subtropical regions of <strong>CCSM</strong>4 mean SSH values are order 10 cm<br />

lower than MDOT; these regions generally coincide with the cold SST biases identified in Fig<br />

2b. <strong>CCSM</strong>4 is slightly more negative (10-20 cm) relative to MDOT where the NAC fails to<br />

properly form the Northwest Corner and in the Labrador and Bering Seas.<br />

To further understand the realism of the ocean circulation in <strong>CCSM</strong>4 we integrated mean mass<br />

transports (years 13-19) through choke points and compared the results with available<br />

observational estimates (Table 1). We use directions here only to differentiate inwards and<br />

outwards flow across a section, channel or opening. The mean mass transport through Drake<br />

Passage is 169 Sv (1 Sv = 10 6 m 3 s -1 ); the observed value is 136.7 ± 7.8 Sv (Cunningham et al.,<br />

2003). The Indonesian Throughflow is -13.8 Sv; recent new measurements by Sprintall et al.<br />

(2009) estimate it to be -15 Sv (into the Indian Ocean). Transport through the Mozambique<br />

Channel is 18.5 Sv southward compared to 16.7 Sv southward observed by Ridderinkhof et al.<br />

(2010). The flow through the Florida- Bahamas Strait is 20.5 Sv while Larsen (1992)’s value<br />

from cable data is 32.3±3.2 Sv. The Windward Passage transport in <strong>CCSM</strong>4 is low (0.17 Sv into<br />

the Caribbean Sea) however observational estimates have shown significant variability; a more<br />

recent estimate is an inflow of 3.6 Sv (Smith et al., 2007). Across the Greenland-Scotland ridge<br />

(GSR) the net <strong>CCSM</strong>4 transports are all northward: 6.5 Sv (Denmark Strait), 1.0 Sv (Iceland-<br />

Faroe Ridge) and 5.5 Sv (Faroe-Scotland Ridge). Observational estimates vary with net estimates<br />

of 0.8 Sv northward (Østerhus et al., 2005) and 3 Sv southward (Olsen et al., 2008) through the<br />

Denmark Strait, 3.8 Sv northwards (Østerhus et al., 2005) and 2.1 Sv southward (Olsen et al.,<br />

2008) across the Iceland-Faroe Channel, and 3.8 northward (Østerhus et al., 2005) and 2.1<br />

southward (Olsen et al., 2008) through the Faroe-Scotland Channel. These biases, especially<br />

those across the GSR, are likely due in part to the overly strong polar atmospheric circulation.<br />

To provide a context for these mass transport results we examine the mean mass circulation in<br />

<strong>CCSM</strong>4 by considering the zonally-integrated flow: the MOC, and the closely related meridional<br />

heat transport (MHT). We use model years 15-19 as the covariances between velocity and<br />

temperature were not archived before year 15. Meridional overturning stream functions for the<br />

global and Atlantic oceans are shown in Fig. 10. The shallow low latitude cells with upwelling<br />

branches at the equator are the most intense features in Fig. 10a with maximum strengths of<br />

around 39 and 38 Sv centered at 60 m in the NH and Southern Hemisphere (SH), respectively,<br />

with the majority of the mass flux sinking within 30° of the equator. In the subpolar zones, the<br />

shallow surface cell is weak (5 Sv equatorwards) in the NH while the Deacon cell in the SH has<br />

a maximum of 38 Sv flowing towards the equator. The strong clockwise cell in the upper 2000-<br />

3000 m of the water column whose southward branch represents the transport of North Atlantic<br />

Deep Water (NADW) connects to the Deacon Cell at intermediate depths leading to strong<br />

upwelling at 60°S. Below it the counter-clockwise deep cell whose lower branch represents the<br />

transport of Antarctic Bottom Water (AABW) has a maximum value of 18 Sv in the SH centered<br />

at around 3500 m. The Atlantic overturning cell (Fig. 10b), driven by downwelling north of<br />

50°N, has a maximum strength of 18 Sv just south of 40°N centered at 1000m, 14Sv crosses the<br />

equator into the SH, and 15 Sv exits the South Atlantic. The strength of the bottom cell is 5 Sv.<br />

10


431<br />

432<br />

433<br />

434<br />

435<br />

436<br />

437<br />

438<br />

439<br />

440<br />

441<br />

442<br />

443<br />

444<br />

445<br />

446<br />

447<br />

448<br />

449<br />

450<br />

451<br />

452<br />

453<br />

454<br />

455<br />

456<br />

457<br />

458<br />

459<br />

460<br />

461<br />

462<br />

463<br />

464<br />

465<br />

466<br />

467<br />

468<br />

469<br />

470<br />

471<br />

472<br />

473<br />

474<br />

475<br />

476<br />

These values can be compared with those from the equivalent stand-alone POP simulation<br />

(Maltrud et al., 2010). In the Atlantic the maximum overturning strength is 25 Sv, 18 Sv crosses<br />

into the SH and then exits the South Atlantic, and the strength of the bottom cell is 3 Sv. The<br />

strength of the global bottom cell is 12 Sv and the Deacon Cell is 43 Sv. Relative to the forced<br />

stand-alone simulation the strength of the Atlantic mid-depth overturning cell is too weak and the<br />

bottom cell is too strong in <strong>CCSM</strong>4.<br />

The excessive NH ice cover will inhibit convective activity in the subpolar NA of <strong>CCSM</strong>4<br />

resulting in lower NADW transport. However in the SO convective activity will be enhanced.<br />

The too-strong SO zonal winds will cause increased evaporative cooling producing too cold<br />

surface waters at convection sites in the Ross and Weddell Seas. Fig. 2b and the seasonal SST<br />

bias figures for JJA and SON support this assertion since cold SST biases are found across the<br />

entire SO south of 60°S during these seasons. As well the formation of too-thick (although not<br />

too extensive) ice in the first decade of the simulation would also have contributed to producing<br />

overly dense surface waters by increasing the salinity of local surface waters. The resulting<br />

enhanced convective activity will produce excessive formation of Antarctic Bottom Water<br />

(AABW) hence the overly robust bottom cell circulation seen in the MOC plots.<br />

Further support for these assertions can be obtained by looking at the seasonal climatological<br />

mixed layer depths (MLDs) from <strong>CCSM</strong>4 (model years 13-19), stand-alone POP (model years<br />

95-99), and observations (Figs 11 and 12) using the 2° resolution global climatology by de Boyer<br />

Montégut et al. (2004). The criterion used to determine the MLD is a threshold value of<br />

temperature from a near-surface value at 10 m depth (ΔT=0.2°C). The temperature criterion<br />

rather than the density criterion is used, as the spatial coverage of the observed density based<br />

MLD estimates is very incomplete in some regions. Figs. 11 a-c show the winter (DJF) MLDs<br />

from observations, <strong>CCSM</strong>4, and POP; their difference fields are seen in Figs 11 d-e. Clearly the<br />

MLDs in the subpolar NA and Norwegian Sea are too shallow (Fig. 11d), while in POP (Fig 11e)<br />

they are too deep relative to observations. The <strong>CCSM</strong>4 MLDs are largely all less than POP in<br />

these regions (Fig. 11e). The shallow ML biases in <strong>CCSM</strong>4 are due to the too extensive ice cover<br />

inhibiting deep convection, and hence deep water formation, in the Labrador and Nordic Seas<br />

during this season. Conversely, too much convective activity may be taking place in POP. In the<br />

Arctic the <strong>CCSM</strong>4 MLDs are deeper than either the observed or POP values. In <strong>CCSM</strong>4 the<br />

Arctic temperature stratification is very low in the upper water column hence the overly deep<br />

MLDs are an artifact of the temperature-based criterion. It should be noted that in stand-alone<br />

POP the upper-level potential temperature and salinity was restored to climatology on a time<br />

scale of 30 days in regions of observed diagnosed/climatological ice.<br />

In the austral winter (JJA) MLDs in both POP and <strong>CCSM</strong>4 are too deep in the ACC relative to<br />

observations (Figs. 12a, b, c, d, and e). In the Pacific sector of the SO POP MLDs are generally<br />

deeper than those from <strong>CCSM</strong>4, while to the south of Australia and southeastwards of New<br />

Zealand <strong>CCSM</strong>4 has deeper MLDs (Fig. 12f). The reasons for these differences are complex and<br />

require further study. The wind stresses used to force POP did not include surface currents in the<br />

drag law calculation. In a recent study Hutchinson et al. (2010) using an eddy-resolving<br />

quasigeostrophic model of the SO showed that in simulations forced in this manner too much<br />

power is inputted to the SO eddy field such that it becomes too energetic. On the other hand,<br />

surface currents are included in the <strong>CCSM</strong>4 calculation however the polar atmospheric biases<br />

11


477<br />

478<br />

479<br />

480<br />

481<br />

482<br />

483<br />

484<br />

485<br />

486<br />

487<br />

488<br />

489<br />

490<br />

491<br />

492<br />

493<br />

494<br />

495<br />

496<br />

497<br />

498<br />

499<br />

500<br />

501<br />

502<br />

503<br />

504<br />

505<br />

506<br />

507<br />

508<br />

509<br />

510<br />

511<br />

512<br />

513<br />

514<br />

515<br />

516<br />

517<br />

518<br />

519<br />

520<br />

521<br />

522<br />

over-ride any benefit. Finally, the actual MLD definition may be misleading in an area of low<br />

stratification. In the Weddell and Ross Seas <strong>CCSM</strong>4 shows deeper MLDs than POP (Fig.12f),<br />

probably an outcome of the too cold, too saline surface waters conjectured above. AABW<br />

formation is too weak in POP while it is likely too strong in <strong>CCSM</strong>4.<br />

Together these results indicate a consistent understanding of the depiction of the global MOC<br />

(Fig 10a) from <strong>CCSM</strong>4. The wind biases in the two hemispheres result in too little convection in<br />

the NA and too much in the SO. In turn this leads to a weak mid-depth overturning cell with a<br />

too robust cell at depth. A comparison with the POP global MOC shows the <strong>CCSM</strong>4 deep cell to<br />

be intruding into waters as shallow as 2000 m at 38°S whereas in POP the mid-depth cell extends<br />

to depths of 2700 m at the same location.<br />

It is clear that the <strong>CCSM</strong>4 MOC biases primarily arise from the contracted and intensified polar<br />

atmospheric vortices. Another possible although more minor cause contributing to these biases<br />

may be the representation of the deep overflows in the NA. As is typical with z-level models,<br />

overflows of dense water tend to be diluted as they flow along the ocean floor. For example, the<br />

Denmark Strait Overflow Water (DSOW) is too warm by 1-2ºC in the Irminger Basin due to<br />

excess mixing with relatively warm overlying water. As noted in the model description section<br />

this version of POP does not use any overflow parameterization and tests using the diffusionbased<br />

Beckmann and Döscher (1997) scheme showed minimal improvement in overflow<br />

characteristics in the 0.1º global POP configuration of Maltrud and McClean (2005).<br />

So how realistic is the associated global total mean heat transport from <strong>CCSM</strong>4? We compare it<br />

with observational estimates based on atmospheric reanalyses by Trenberth and Carron (2000)<br />

and from stand-alone POP (Fig. 13a). The <strong>CCSM</strong>4 MHT is larger (smaller) than POP to the<br />

south of 30°S (from 30°S to the equator) and lies almost entirely outside of the observed<br />

uncertainty in the SH. It increases to the north of the equator but underestimates the observed<br />

and POP MHT maxima at 15°N. Further north it decreases to a level between those of<br />

observations and POP, again mostly not within the observed uncertainty. Further we decompose<br />

the MHT from the two simulations into their respective total, mean and eddy heat transport<br />

components (Fig 13b). Equatorward of 45°S both the mean and eddy contributions are generally<br />

lower in <strong>CCSM</strong>4 than in POP especially in the ACC and the NH mid-latitude western boundary<br />

currents, consistent with the weaker MOC in <strong>CCSM</strong>4. Between 45° and 60° S the <strong>CCSM</strong>4<br />

southward eddy fluxes are consistently larger than those in POP. In the NH subtropics (20° and<br />

40° N) POP eddy heat fluxes range from 0.1 to 0.2 PW southward whereas in <strong>CCSM</strong>4 they are<br />

up to 0.15 PW northwards. These two latitude ranges coincide with those previously identified as<br />

being dominated by cold SST biases, and hence are the likely explanation for the behavior of the<br />

<strong>CCSM</strong>4 eddy fluxes in those regions.<br />

4. Tropical Cyclone Event<br />

In this section we move away from the mean zonally-averaged construct used earlier to<br />

understand ocean heat flux. Here we consider a specific tropical cyclone event and calculate the<br />

ocean heat convergence associated with its passage. Our context however continues to be about<br />

understanding the important mechanisms controlling the movement of mass and heat in this<br />

simulation.<br />

12


523<br />

524<br />

525<br />

526<br />

527<br />

528<br />

529<br />

530<br />

531<br />

532<br />

533<br />

534<br />

535<br />

536<br />

537<br />

538<br />

539<br />

540<br />

541<br />

542<br />

543<br />

544<br />

545<br />

546<br />

547<br />

548<br />

549<br />

550<br />

551<br />

552<br />

553<br />

554<br />

555<br />

556<br />

557<br />

558<br />

559<br />

560<br />

561<br />

€ 562<br />

563<br />

564<br />

565<br />

566<br />

567<br />

568<br />

A very positive feature of the simulation is its ability to generate and propagate intense tropical<br />

cyclones when environmental conditions favor their development. To identify events in model<br />

years 13 through 19, we created animations of the 850 mb relative vorticity field (ζ 850 ), which<br />

was calculated from the 850 mb winds interpolated to a spectral resolution T63 grid (Bengtsson,<br />

et al. 2006). Possible cyclone events were associated with persistent vorticity centers exceeding 5<br />

x 10 -5 s -1 . More detailed analyses of individual storm events was performed using the full high<br />

resolution output.<br />

Compared to climatology, few storms were generated in the tropical NA, due to the anomalously<br />

low SSTs (Fig. 2b) and accentuated Azores High (Fig. 4b). The animation of ζ 850 shows easterly<br />

waves forming off Africa, however no hurricanes developed in the tropical Atlantic. A few<br />

waves that survived upon reaching the Gulf of Mexico did develop into storms as a result of the<br />

more favorable local conditions. Conversely, the too-high SSTs in the central tropical North<br />

Pacific (Fig. 2b; 160°E - 120°W) permitted too many tropical storms to form. Nevertheless, the<br />

high resolution of the simulation permits the spontaneous generation and recurvature of Saffir-<br />

Simpson Category 4 typhoons, as seen in an event with sustained 850 mb wind speeds of 66.5<br />

ms -1 from July-August of year 19 over the western tropical Pacific.<br />

The SST response under and to the right of the track is seen in Fig. 14a. Observational studies of<br />

the upper ocean response to the passage of extreme tropical storms such as those by Jacob et al.<br />

(2000) and Jaimes and Shay (2009) guide our interpretations. The cold SST "wake” is produced<br />

by the entrainment of cold subsurface water either through Ekman pumping or by instability of<br />

the vertical shear of forced near-inertial oscillations (Jaimes and Shay, 2009). A surface<br />

temperature drop of more than 3.6°C is seen at station 9 (133°E, 26°N) on 08/04/19, a day after<br />

the passage of the storm over this location, relative to the pre-storm conditions on 07/29/19 (Fig.<br />

14b). The ocean mixed layer deepened from about 40 m to 80 m, while the temperature below<br />

the bottom of the mixed layer and above 150 m increased by up to 2°C. Surface salinities<br />

increased from about 34.6 to 34.73 (not shown). By 08/18/19 the near-inertial oscillations had<br />

largely abated (not shown), and the mixed layer shallowed to almost pre-storm conditions and<br />

warmed by 0.8°C.<br />

The amount of net surface heating, Q, needed to restore the upper water column to pre-storm<br />

temperatures can be approximated by:<br />

Q = ∫∫∫ ρ 0<br />

C P<br />

ΔTdhdWdL = 5.8 ×10 20 J<br />

where ΔT is the magnitude of the negative part of the temperature anomaly, h is the depth to<br />

which the negative part of the temperature anomaly extends, and dW and dL are the cross-track<br />

and along-track extents, respectively (Emanuel, 2001). The cross-track extent was identified by<br />

differencing SST before the storm and 24 hours afterwards; h was calculated using a 0.2° C<br />

threshold value of temperature from the 10 m value 24 hours after the storm’s passage. Emanuel<br />

(2001) obtained a similar value of Q using approximate values for one NA hurricane event. The<br />

amount of heat mixed downwards into the oceans by typhoons may be an important component<br />

of the global meridional heat transport (Sriver and Huber, 2007). A calculation of Q for all<br />

typhoon events and a subsequent integration in time (such as a year) to estimate the average<br />

13


569<br />

570<br />

571<br />

572<br />

573<br />

574<br />

575<br />

576<br />

577<br />

578<br />

579<br />

580<br />

581<br />

582<br />

583<br />

584<br />

585<br />

586<br />

587<br />

588<br />

589<br />

590<br />

591<br />

592<br />

593<br />

594<br />

595<br />

596<br />

597<br />

598<br />

599<br />

600<br />

601<br />

602<br />

603<br />

604<br />

605<br />

606<br />

607<br />

608<br />

609<br />

610<br />

611<br />

612<br />

613<br />

614<br />

cyclone-induced ocean heat convergence awaits a simulation where the number of hurricane<br />

occurrences more correctly reflect current climatic conditions.<br />

5. Ocean Mesoscale and Agulhas Eddies<br />

The majority of ocean variability is generally accounted for by mesoscale fluctuations; Chelton<br />

et al. (2007) report that more than 50% of the variability over much of the World Ocean is<br />

accounted for by eddies with amplitudes of 5-25 cm and diameters of 100-200 km. Fig. 15 shows<br />

the mesoscale RMS sea surface height anomaly (SSHA) from <strong>CCSM</strong>4 for years 15-19 and the<br />

AVISO-blended altimetry product from TOPEX/POSEIDON and ERS 1 and 2 (TPERS) for<br />

1997-2001. The ocean model fields were subsampled every 7 days and binned to match the<br />

spatial and temporal sampling of the data. The mesoscale energy is obtained by band-pass<br />

filtering the SSHA between 20-150 days; we are aware that it also contains energy at scales<br />

longer than a few hundred kilometers. The overall comparison generally shows good agreement<br />

in terms of the magnitude and spatial distribution of the variability, however in the ACC and the<br />

KE the mesoscale variability is unrealistically large in places; this is likely due in part to a<br />

barotropic intraseasonal response to the overly strong winds in that region. As well, at latitudes<br />

higher than 50° sampling limitations and the objective analysis procedure produce an<br />

underestimated TPRES signal (Brachet et al., 2004).<br />

A significant improvement in <strong>CCSM</strong>4 is the depiction of the variability associated with Agulhas<br />

eddies in the South Atlantic. The unrealistic band of strong SSH variability extending<br />

northwestward across the basin that was seen in two earlier global stand-alone 0.1° POP<br />

simulations, without (Maltrud and McClean, 2005; Fig. 4b) and with PBCs (Maltrud et al., 2010)<br />

on dipole and tripole grids, respectively, and forced with different synoptic fluxes, is not seen in<br />

<strong>CCSM</strong>4. This bias was attributed to Agulhas eddies predominately following the same toonorthwestward<br />

path across the South Atlantic and reaching waters offshore of South America at<br />

about 12-15°S.<br />

We hypothesize that the two-way coupling of 0.1° POP to the high-resolution atmosphere in<br />

<strong>CCSM</strong>4 has produced more realistic Agulhas eddy pathways and surface dissipation locations.<br />

Examination of the <strong>CCSM</strong>4 time-evolving latent and sensible heat fluxes, wind stress curl and<br />

Ekman pumping velocities reveals the signature of Agulhas eddies in the Cape Basin region<br />

(west of South Africa) in all these fields; to the west their signatures are barely discernable.<br />

These exchanges and feedbacks likely reduce the range of the eddy sea surface height<br />

displacements causing them to start to spin-down. This assertion is supported by the significant<br />

reduction in mesoscale variability westward of the Cape Basin region (Fig. 15a).<br />

The POP/CICE and the two stand-alone POP simulations were forced with atmospheric fluxes<br />

that were in part derived from National Center for Environmental Prediction/National Center for<br />

Atmospheric Research (NCEP/NCAR) reanalysis products (Kalnay et al., 1996). Uppermostlevel<br />

model potential temperature was included in the calculation of the net surface heat flux<br />

along with the NCEP turbulent fluxes. The 40-year NCEP reanalysis products have a spectral<br />

resolution of triangular wavenumber 62 (T62) or about 210-km grid spacing. Rouault et al.<br />

(2003) found that the NCEP model tends to underestimate latent and sensible transfers from the<br />

Agulhas Current and the Retroflection. They attribute this to the model’s inability to represent<br />

14


615<br />

616<br />

617<br />

618<br />

619<br />

620<br />

621<br />

622<br />

623<br />

624<br />

625<br />

626<br />

627<br />

628<br />

629<br />

630<br />

631<br />

632<br />

633<br />

634<br />

635<br />

636<br />

637<br />

638<br />

639<br />

640<br />

641<br />

642<br />

643<br />

644<br />

645<br />

646<br />

647<br />

648<br />

649<br />

650<br />

651<br />

652<br />

653<br />

654<br />

655<br />

656<br />

657<br />

658<br />

659<br />

660<br />

the fluxes over the warmest waters in the core of the current. Since the core is only 80-100 km<br />

wide they suggest that the one-degree SST product used to force the atmospheric model is<br />

inadequate to resolve the Agulhas Current and its mesoscale variability. Underestimated<br />

turbulent fluxes and the lack of feedbacks would result in less eddy spin-down in the forced<br />

simulations, consistent with the strong band of mesoscale variability found across much of the<br />

South Atlantic.<br />

To further quantify the causes of the improvement in <strong>CCSM</strong>4 we compared Agulhas eddy tracks<br />

from satellite altimeter-derived sea surface height anomaly from AVISO for 2004-2008 and from<br />

<strong>CCSM</strong>4 for model years 15-19 (Figs. 16a and b). We identified and tracked eddies using the<br />

Okubo-Weiss parameter, W, which measures the relative importance of deformation and rotation<br />

and is given by the sum of squares of the normal and shear components of strain minus the<br />

square of the relative vorticity (Isern-Fontanet et al., 2003; 2006). Eddies were identified by<br />

closed contours of W = -0.8 x 10 -12 s -2 . AVISO shows three repeating (41, 42%, and 17%)<br />

northwesterly-directed pathways located progressively equatorward of each other by 2.5°of<br />

latitude; they terminate progressively eastward of each other at 32°W and 24°S, 25°W and 26°S,<br />

and 18°W and 26°S. <strong>CCSM</strong>4 shows two repeating tracks (78% and 13%) directed toward the<br />

northwest about 4° apart; they terminate at 23°W, 25°S and 25°W, 21°S. Some latitudinal spread<br />

is seen in the cluster of pathways constituting the primary pathway. Their signature disappeared<br />

slightly to the east of the observed dissipation locations. For contrast we show tracks from the<br />

coupled POP/CICE simulation forced with normal year CORE forcing. It has two eddy tracks<br />

(90% and 10%) that extend all the way across the South Atlantic (Fig. 16c). The dominant<br />

northwestward pathway has too much of a northward component; the cluster of tracks<br />

comprising it have almost no latitudinal spread and they all terminate at about 14°S.<br />

To support our assertion that the POP eddies are both generated with higher amplitudes and<br />

maintain higher amplitudes during their passage across the Cape Basin than their observed or<br />

<strong>CCSM</strong>4 counterparts we have calculated the evolution of the along-track amplitude of SSHA for<br />

long-lived anticyclonic eddies from AVISO, <strong>CCSM</strong>4, and POP/CICE together with their<br />

respective means (Fig. 17). The mean SSHA amplitudes at the formation locations from AVISO<br />

and <strong>CCSM</strong>4 are about 40-50 cm while it is 100 cm from POP/CICE. These values decrease to<br />

about 30 cm after 300 days both for AVISO and <strong>CCSM</strong>4 but only drop to 50 cm in the case of<br />

POP/CICE. Clearly POP/CICE eddies are shed from the Agulhas Retroflection with higher<br />

SSHA amplitudes than those derived from observations or in <strong>CCSM</strong>4. After 300 days their mean<br />

amplitudes decrease to the values of the other cases in their formation regions.<br />

Finally we examined the subsurface signature of a particular <strong>CCSM</strong>4 Agulhas eddy that<br />

followed the black track in Fig. 16b by identifying it in time-evolving fields of 10°C isotherm<br />

depth (not shown). The eddy was shed from the Agulhas retroflection at the beginning of year 15<br />

and was last identified at the surface using the tracking procedure above in March of year 17 at<br />

about 23°W, 25°S. The subsurface signature of the eddy can be identified as a coherent feature<br />

for at least 8-9 months after it was no longer identifiable at the surface and is last seen at depth at<br />

about 35°W, 23°S.<br />

6. Discussion and Conclusions<br />

15


661<br />

662<br />

663<br />

664<br />

665<br />

666<br />

667<br />

668<br />

669<br />

670<br />

671<br />

672<br />

673<br />

674<br />

675<br />

676<br />

677<br />

678<br />

679<br />

680<br />

681<br />

682<br />

683<br />

684<br />

685<br />

686<br />

687<br />

688<br />

689<br />

690<br />

691<br />

692<br />

693<br />

694<br />

695<br />

696<br />

697<br />

698<br />

699<br />

700<br />

701<br />

702<br />

703<br />

704<br />

705<br />

706<br />

We examined the general fidelity of air-ocean interaction processes in a prototype two-decade<br />

fully coupled fine resolution <strong>CCSM</strong>4 simulation. The simulation’s climate is very reasonable<br />

considering that no tuning was done to correct for finer model resolution. As noted by Pope and<br />

Stratton (2002), any changes in resolution require significant testing to properly adjust the<br />

multiple constants in resolution dependent parameterizations. In other versions of <strong>CCSM</strong>, the<br />

scale adjustments for parameterization constants (model tuning) required many simulations, each<br />

lasting several decades, which was beyond the scope and resources of this investigational study.<br />

This simulation was extremely computationally expensive and only one simulation could be<br />

conducted, therefore we needed to minimize risk by choosing component setups with minimal<br />

changes from those that had already been successfully conducted. Our choices for the coupled<br />

model setup therefore were guided by our experience with a forced stand-alone 0.1° POP<br />

simulation on this same grid (Maltrud et al. 2010), low resolution fully-coupled <strong>CCSM</strong><br />

simulations (Collins et al., 2006; Bala et al. 2008), and the Gent et al. (2009) <strong>CCSM</strong> simulation<br />

that was configured with higher than standard-resolution atmosphere/land components (0.5°).<br />

Given the resource constraints tests of increases in vertical resolution and other subgrid scale<br />

parameterizations in both the atmosphere and ocean await investigation in future simulations.<br />

The major high latitude biases in the simulation were dominated by problems in the atmospheric<br />

model’s circulation, which were likely a major cause in the drift and worsening SST errors. The<br />

intensified and contracted polar vortices point to problems with the scale sensitivity of the<br />

parameterization constants associated with gravity and mountain wave sub-grid scale schemes,<br />

exacerbated by the non-linearities associated with the polar filter required to keep the solutions<br />

stable on the fine spacing on the spherical grid. Early test simulations of atmosphere-only runs<br />

forced by climatological surface conditions showed some signature of the high-latitude<br />

problems, but they were not as severe and the belief was that a fully coupled simulation would<br />

bring the atmosphere into a more realistic state. In reality, the biases worsened as a result of the<br />

coupling and introduced problems into the ocean.<br />

Higher resolution does little to correct the large-scale biases in the radiation and precipitation<br />

fields that most likely result from subgrid-scale parameterization weaknesses (Pope and Stratton,<br />

2002). The anomalous high latitude atmospheric circulation was reflected in surface wind<br />

stresses that were too strong in the subpolar NA, NP, and SO. As a result, cold SST biases and<br />

excessive ice were created in the subpolar NH. These anomalously low SSTs caused a dearth of<br />

hurricanes in the North Atlantic, while too high SSTs in the tropical eastern Pacific resulted in an<br />

excessive number of tropical systems forming in the NP.<br />

Encouragingly however the coupled simulation provided an important example of an improved<br />

depiction of mesoscale processes important to climate relative to the forced ocean simulations.<br />

The amplitude of SSHA in eddies being shed from the Agulhas Retroflection are unrealistically<br />

high in the forced simulations; their amplitudes remain unrealistically high as they move across<br />

the South Atlantic basin following a too northwesterly path. These higher amplitudes are<br />

attributed to the coarseness of the NCEP atmospheric fluxes used to force the ocean model and<br />

the SST product used as a bottom boundary condition for the NCEP atmospheric prediction<br />

model. Exchanges and feedbacks between the high-resolution atmosphere and ocean are<br />

implicated as the cause for the more realistic <strong>CCSM</strong>4 Agulhas eddy SSH displacements at the<br />

time of their formation and their spin-down in the Cape Basin. Therefore it is likely that the<br />

16


707<br />

708<br />

709<br />

710<br />

711<br />

712<br />

713<br />

714<br />

715<br />

716<br />

717<br />

718<br />

719<br />

720<br />

721<br />

722<br />

723<br />

724<br />

725<br />

726<br />

727<br />

728<br />

729<br />

730<br />

731<br />

732<br />

733<br />

734<br />

735<br />

736<br />

737<br />

738<br />

739<br />

740<br />

741<br />

742<br />

743<br />

744<br />

745<br />

746<br />

747<br />

748<br />

749<br />

750<br />

751<br />

752<br />

resulting respective energetics of the Agulhas eddies in the coupled versus the stand-alone<br />

models determine their differing pathways and surface dissipation locations. Unrealistically high<br />

mean SSH was seen in the western boundary currents and recirculation regions, including the<br />

Agulhas System, of a forced stand-alone high-resolution ocean simulation. On the other hand,<br />

<strong>CCSM</strong>4 in these regions was in much better agreement with the observed dynamic topography<br />

estimate from the combined altimetry, GRACE geoid, and surface drifter data. Again the low<br />

resolution atmospheric fluxes forcing the stand-alone model are likely to be a contributor to these<br />

biases. Hence demonstrable improvements in the depiction of both the mean and mesoscale<br />

variability of the upper ocean circulation were found in the fully coupled fine resolution<br />

simulation relative to the stand-alone ocean simulations.<br />

In all, this prototype simulation demonstrates the computational feasibility and model readiness<br />

to execute coupled high resolution simulations and highlights the future potential of very high<br />

resolution, long coupled climate runs to examine such features as decadal variability or the<br />

impact of ocean mixing by tropical storms on ocean heat transport. More importantly, the<br />

simulation shows the potential for the spontaneous simulation of important coupled mesoscale<br />

features that were previously impractical in a global coupled model, particularly the generation<br />

of intense tropical cyclones under the correct environmental conditions.<br />

Toward the goal of improving the performance of the ocean model in the coupled system, we are<br />

conducting stand-alone ocean or coupled ocean/ice simulations configured as consistently as<br />

possible with the coupled model set-up. In this way, we can shed light on ocean biases without<br />

the complexity of an active atmosphere. These results can then help us understand the ocean’s<br />

behavior in the fully coupled system. Hence we are running stand-alone 0.1° POP with the model<br />

ocean surface velocities incorporated into the drag law calculation as is done in <strong>CCSM</strong>4. We<br />

wish to know if the eddy field will be weakened as found by Hutchinson et al. (2010). The<br />

current wind stress forcing formulation will also likely provide too much momentum to the<br />

western boundary currents hence we hypothesize that the biases seen in the mean SSH of POP<br />

(Figs. 8 and 9) may be reduced once ocean surface currents are included in the stress calculation.<br />

Global coupled 0.1° POP/CICE is also being run and evaluated to improve the fidelity of the<br />

coupled ice/ocean system, especially the ice cover in the Southern Hemisphere in austral<br />

summer. Finally, the effects of using the baroclinic restratification parameterization developed<br />

by Fox-Kemper et al. (2008) and Fox-Kemper and Ferrari (2008) will be tested in future<br />

simulations. It is based on the concept that ageostrophic baroclinic instabilities form in the ocean<br />

surface mixed layer at horizontal fronts and act to re-stratify the ocean. This submesoscale<br />

parameterization has already been evaluated in stand-alone 0.1º POP (Fox-Kemper et al., 2011),<br />

and in the coupled context should help to re-stratify the upper ocean following the passage of<br />

typhoons.<br />

The promising results of the present investigation have lead to a multi-institutional effort to<br />

develop a CMIP (<strong>Coupled</strong> Model Intercomparison Project)-capable high resolution version of<br />

<strong>CCSM</strong>, involving multiple dynamical cores and systematic evaluation of the scale dependence of<br />

sub-grid scale schemes. We hope to perform accurate hindcasts of twentieth century climate, as<br />

well as projection of twenty-first century climate, which will explicitly simulate climate change<br />

at small scales, including climatologies of extreme events. These simulations will provide the<br />

means to study mesoscale air-sea coupling over long time scales and understand its importance<br />

17


753<br />

754<br />

755<br />

756<br />

757<br />

758<br />

759<br />

760<br />

761<br />

762<br />

763<br />

764<br />

765<br />

766<br />

767<br />

768<br />

769<br />

770<br />

771<br />

772<br />

773<br />

774<br />

775<br />

776<br />

777<br />

778<br />

779<br />

780<br />

781<br />

782<br />

783<br />

784<br />

785<br />

786<br />

787<br />

788<br />

789<br />

790<br />

791<br />

792<br />

793<br />

794<br />

795<br />

796<br />

to climate change. A recent study by Bryan et al. (2010) is a forerunner of such analyses. They<br />

found by comparing this <strong>CCSM</strong> simulation with others having lower resolution components that<br />

it is only when the ocean model largely resolves eddies that the characteristics of observed<br />

frontal scale ocean –atmosphere interaction are captured.<br />

Acknowledgements<br />

This work was supported by the Office of Science (BER), U.S. Department of Energy at the<br />

Lawrence Livermore, Oak Ridge, Argonne, and Los Alamos National Laboratories under<br />

contracts DE-AC52-07NA27344 (AAM, DCB, DPI, JSB, JLM), DE-AC05-00OR22725 (PHW,<br />

DCB), DE-AC02-06CH11357 (RLJ), DE-AC52-06NA25396 (PWJ, MEM, and as an SIO<br />

subcontract: JLM and YYK), respectively and by grants DE-FG02-05ER64119 (JLM, YYK),<br />

and DE-PS02-07ER0706 (JMD). NASA contract 1273575 (via U. Maine) also supported JLM<br />

and YYK. Participation of FOB, JMD, MV, NN, and AN were supported by the National<br />

Science Foundation through its sponsorship of NCAR. Computer time was provided by the<br />

Department of Energy as part of the Multiprogramatic and Institutional Initiative at Lawrence<br />

Livermore National Laboratory. Altimeter products are from Ssalto/Duacs, Aviso and Cnes,<br />

ERA-40 reanalysis came from ECMWF and NCAR, and QuikSCAT winds are from the<br />

Scatterometer Climatology of Ocean Wind (SCOW) data set found at<br />

http://cioss.coas.oregonstate.edu/scow/. We thank both reviewers and the guest editor of this<br />

special edition for their helpful comments and suggestions.<br />

References<br />

Bala G., Rood, R.B., Mirin, A., McClean, J., Achuta Rao, K. Bader, D., Gleckler, P., Neale, R.,<br />

Rasch, P., 2008. Evaluation of a <strong>CCSM</strong>3 Simulation with a Finite Volume Dynamical Core<br />

for the Atmosphere at 1°x 1.25° Longitude <strong>Resolution</strong>. J. Clim. 21, 1467-1486.<br />

Beckmann, A., and Döscher, R.,1997: A method for improved representation of dense water<br />

spreading over topography in geopotential-coordinate models. J. Phys. Oceanogr. 27, 581–<br />

591.<br />

Bengtsson, L., Hodges, K.I., Roeckner E., 2006. Storm tracks and climate change. J. Clim. 19,<br />

3518-3543.<br />

Bengtsson, L., Hodges, K.I., Esch, M., Keenlyside, N., Kornblueh, L. Luo, J.-J., Yamagata T.,<br />

2007. How may tropical cyclones change in a warmer climate. Tellus A 59, 539–561.<br />

Biastoch, A., Lutjeharms, J.R.E., Böning, C.W., Scheinert, M., 2008. Mesoscale perturbations<br />

control inter-ocean exchange south of Africa. Geophys. Res. Lett. 35, L20602,<br />

doi:10.1029/2008GL035132.<br />

Bitz, C. M., Lipscomb, W.H., 1999. An energy-conserving thermo-dynamic sea ice model for<br />

climate study, J. Geophys. Res., 104, 15,669 – 15,677.<br />

Bryan, F.O., Hecht, M.W., Smith, R.D., 2007. <strong>Resolution</strong> convergence and sensitivity studies<br />

with North Atlantic circulation models. Part I: The western boundary current system. Ocean<br />

Modell. 16, 141-159.<br />

Bryan, F.O., Tomas, R., Dennis, J.M., Chelton, D.B., Loeb, N.G., McClean, J.L., 2010. Frontal<br />

scale air-sea interaction in high-resolution coupled climate models. J. Clim. 23, 6277-6291.<br />

18


797<br />

798<br />

799<br />

800<br />

801<br />

802<br />

803<br />

804<br />

805<br />

806<br />

807<br />

808<br />

809<br />

810<br />

811<br />

812<br />

813<br />

814<br />

815<br />

816<br />

817<br />

818<br />

819<br />

820<br />

821<br />

822<br />

823<br />

824<br />

825<br />

826<br />

827<br />

828<br />

829<br />

830<br />

831<br />

832<br />

833<br />

834<br />

835<br />

836<br />

837<br />

838<br />

839<br />

840<br />

841<br />

842<br />

Böning, C. W., Dispert, A.,Visbeck, M., Rintoul, S.R., Schwarzkopf, F.U., 2008. The response<br />

of the Antarctic Circumpolar Current to the recent climate change. Nature Geosci. 1, 864–<br />

869.<br />

Boville, Byron A., 1984. The Influence of the Polar Night Jet on the Tropospheric Circulation in<br />

a GCM. J. Atmos. Sci. 41, 1132-1142.<br />

Boyle, J., Klein, S.A., 2010. Impact of horizontal resolution on climate model forecasts of<br />

tropical precipitation and diabatic heating for the TWP-ICE period, J. Geophys. Res. 115,<br />

D23113, doi:10.1029/2010JD014262.<br />

Brachet, S., Le Traon, P.Y., Le Provost, C., 2004. Mesoscale variability from a high-resolution<br />

model and from altimeter data in the North Atlantic Ocean. J. Geophys. Res. 109,<br />

doi:10.1029/2004JC002360.<br />

Cavalieri, D., Parkinson, C., Gloersen, P., Zwally, H. J., 1996, updated 2008. Sea ice<br />

concentrations from Nimbus-7 SMMR and DMSP SSM/I passive microwave data, [1979<br />

2002]. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.<br />

Chelton, D. B., Schlax, M.G., Samelson, R.M., de Szoeke, R. A., 2007. Global obervations of<br />

large oceanic eddies. Geophys. Res. Lett. 34, L15606, doi:10.1029/2007GL030812.<br />

Collins, W. D., Bitz, C.M., Blackmon, M.L., Bonan, G.B., Bretherton, C.S., Carton, J.A., Chang,<br />

P., Doney, S.C., Hack, J.J., Henderson, T.B., Kiehl, J.T., Large, W.G., McKenna, D.S.,<br />

Santer, R. D. Smith, R.D., 2006. The Community Climate System Model Version 3<br />

(<strong>CCSM</strong>3). J. Clim. 19, 2122-2143.<br />

Cunningham, S.A., Alderson, S.G., King, B.A., Brandon, M.A., 2003. Transport and variability<br />

of the Antarctic Circumpolar Current in Drake Passage. J. Geophys. Res. 108(C5), 8084,<br />

doi:10.1029/2001JC001147.<br />

Danabasoglu, G., Large, W.G., and Briegleb, B.P., 2010: Climate impacts of parameterized<br />

Nordic Sea overflows. J. Geophys. Res., 115, C11005, doi:10.1029/2010JC006243.<br />

de Boyer Montégut, C., Madec, G., Fischer, A.S., Lazar, A., Iudicone, D., 2004. Mixed layer<br />

depth over the global ocean: An examination of profile data and a profile-based climatology.<br />

J. Geophys. Res.109, C12003, doi:10.1029/2004JC002378.<br />

Dickinson, R.E., Oleson, K.W., Bonan, G., Hoffman, F., Thornton, P., Vertenstein, M., Yang,<br />

Z.-L., Zeng, X., 2006. The Community Land Model and its climate statistics as a component of<br />

the Community Climate System Model. J. Clim. 19, 2302-2324.<br />

Dukowicz, J.K., Smith, R.D., 1994. Implicit free-surface method for the Bryan-Cox-Semtner<br />

ocean model. J. Geophys. Res. 99, 7991-8014.<br />

Emanuel, K., 2001. Contribution of tropical cyclones to meridional heat transport by the oceans.<br />

J. Geophys. Res.-Atmos. 106, 14771-14781.<br />

Emanuel, K., 2008. The hurricane—climate connection. Bull. Amer. Meteorol. Soc. 89, ES10–<br />

ES20.<br />

Flatau, M. K., Talley, L., Niiler, P.P., 2003. The North Atlantic Oscillation, surface current<br />

velocities, and SST changes in the Subpolar North Atlantic. J. Clim. 16, 2355-2369.<br />

Fissel, D.B., Birch J.R., Melling, H., Lake, R.A., 1998. Non-tidal flows in the Northwest<br />

Passage. Canadian Tech. Rep. of Hydrography and Ocean Sciences, 98, 143pp.<br />

Fox-Kemper, B., Ferrari, R., Hallberg, R.W., 2008. Parameterization of mixed layer eddies. Part<br />

I: Theory and diagnosis. J. Phys. Oceanogr. 38, 1145-1165.<br />

Fox-Kemper, B., Danabasoglu, G., Ferrari, R., Griffies, S.M., Hallberg, R.W., Holland, M.M.,<br />

Maltrud, M.E., Peacock, S. Samuels, B. L., 2011. Parameterization of mixed layer eddies. III:<br />

Implementation and impact in global ocean climate simulations. Ocean Modell., in press.<br />

19


843<br />

844<br />

845<br />

846<br />

847<br />

848<br />

849<br />

850<br />

851<br />

852<br />

853<br />

854<br />

855<br />

856<br />

857<br />

858<br />

859<br />

860<br />

861<br />

862<br />

863<br />

864<br />

865<br />

866<br />

867<br />

868<br />

869<br />

870<br />

871<br />

872<br />

873<br />

874<br />

875<br />

876<br />

877<br />

878<br />

879<br />

880<br />

881<br />

882<br />

883<br />

884<br />

885<br />

886<br />

887<br />

888<br />

Fox-Kemper, B., Ferrari, R., 2008. Parameterization of mixed layer eddies. Part II: Prognosis<br />

and impact. J. Phys.Oceanogr. 38, 1166-1179.<br />

Gent, P. R., Yeager, S.G., Neale, R.B., S. Levis, S., Bailey, D.A., 2009. Improvements in a halfdegree<br />

atmosphere/land version of the <strong>CCSM</strong>. Clim. Dyn. doi:10.1007/s00382-009-0614-8:<br />

Published online.<br />

Gouretski, V.V., Koltermann, K.P., 2004. WOCE global hydrographic climatology. Technical<br />

35, Bundesamtes für Seeschiffahrt und Hydrographie, Hamburg, Germany.<br />

Griffes, S.M., Biastoch, A., Böning, C., Bryan, F., Danabasoglu, G., Chassignet, E.P., England,<br />

M.H., Gerdes, R., Haak, H., Hallberg, R.W., Hazeleger, W., Jungclaus, J., Large, W.G.,<br />

Madec, G., Pirani, A., Samuels, B.L., Scheinert, M., Gupta, A.S., Severijns, C.A., Simmons,<br />

H.L., Treguier, A.M., Winton, M., Yeager, S., Yin, J., 2009. Coordinated ocean-ice reference<br />

experiments (COREs). Ocean Modell., 26, 1-46.<br />

Hack, J. J., Caron, J.M., Danabasoglu, G., Oleson, K.W., Bitz, C., Truesdale, J.E., 2006. <strong>CCSM</strong>-<br />

CAM3 climate simulation sensitivity to changes in horizontal resolution. J. Clim. 19, 2267–<br />

2289.<br />

Hurlburt, H.E., Hogan, P.J., 2000. Impact of 1/8° to 1/64° resolution on Gulf Stream model-data<br />

comparisons in basin-scale subtropical Atlantic models. Dyn. Atmos. Ocean. 32, 283–329.<br />

Hutchinson, D.K., Hogg, A., Blundell, J.R., 2010. Southern Ocean response to relative velocity<br />

wind stress forcing. J. Phys. Oceanogr., 40,326-339.<br />

Hunke, E. C., Dukowicz, J.K., 1997. An elastic-viscous-plastic model of sea ice dynamics, J.<br />

Phys. Oceanogr., 27, 1849 – 1867.<br />

Ingvaldsen, R. B., Asplin, L., Loeng, H., 2004. The seasonal cycle in the Atlantic transport to<br />

the Barents Sea during the years 1997-2001. Cont. Shelf. Res. 24, 1015-1032.<br />

Isern-Fontanet, J., Garcia-Ladona, E., Font, J., 2003. Identification of marine eddies from<br />

altimetric maps. J. Atmos. Ocean. Technol. 20, 772–778.<br />

Isern-Fontanet, J., Garcia-Ladona, E., Font, J., 2006. Vortices of the Mediterranean Sea: An<br />

altimetric perspective. J. Phys. Oceanogr. 36, 87–103.<br />

Jacob, S. D., Shay, L.K., Mariano, A.J., Black, P.G., 2000: The 3D oceanic mixed layer response<br />

to hurricane Gilbert. J. Phys. Oceanogr., 30, 1407-1429<br />

Jaimes, B., Shay, L.K., 2009. Mixed layer cooling in mesoscale oceanic eddies during hurricanes<br />

Katrina and Rita. Mon. Wea. Rev. 137, 4188-4207.<br />

Johnson, G., Sloyan, B., Kessler, W., McTaggart, K., 2002. Direct measurements of upper ocean<br />

currents and water properties across the tropical Pacific during the 1990s. Prog. Oceanogr. 52,<br />

Pergamon, 31-61.<br />

Kalnay, E., Kanamitsu, M., Kistler, R., Collins, W., Deaven, D., Gandin, L., Iredell, M., Saha,<br />

S., White, G., Woollen, J., Zhu, Y., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J.,<br />

Mo, K.C., Ropelewski, C., Leetmaa, A., Reynolds, R., Jenne, R., 1996. The NCEP/NCAR<br />

40-year reanalysis project. Bull. Amer. Meteorol. Soc. 77, 437-471.<br />

Kessler, W.S., Johnson, G.C., D.W. Moore, 2003. Sverdrup and nonlinear dynamics of the<br />

Pacific equatorial currents. J. Phys. Oceanogr. 33, 994-1008.<br />

Korty, R.L., Emanuel, K.A., Scott, J.R., 2008. Tropical cyclone–induced upper-ocean mixing<br />

and climate: Application to equable climates. J. Clim. 21, 638–654.<br />

Large, W. G., McWilliams, J.C., Doney, S.C., 1994. Oceanic vertical mixing: a review and a<br />

model with nonlocal boundary layer parameterization. Rev. Geophys. 32, 363-403.<br />

Large, W.G., Yeager S.G., 2004. Diurnal to decadal global forcing for ocean and sea-ice models:<br />

the datasets and flux climatologies. NCAR Technical Note TN-460+STR, National Center for<br />

20


889<br />

890<br />

891<br />

892<br />

893<br />

894<br />

895<br />

896<br />

897<br />

898<br />

899<br />

900<br />

901<br />

902<br />

903<br />

904<br />

905<br />

906<br />

907<br />

908<br />

909<br />

910<br />

911<br />

912<br />

913<br />

914<br />

915<br />

916<br />

917<br />

918<br />

919<br />

920<br />

921<br />

922<br />

923<br />

924<br />

925<br />

926<br />

927<br />

928<br />

929<br />

930<br />

931<br />

932<br />

933<br />

934<br />

Atmospheric Research.<br />

Large, W.G., Danabasoglu, G., 2006. Attribution and impacts of upper-ocean biases in <strong>CCSM</strong>3.<br />

J. Clim. 19, 2325-2346.<br />

Larsen, J. C., 1992. Transport and heat flux of the Florida Current at 27°N derived from crossstream<br />

voltages and profiling data: theory and observations. Phil. Trans. Roy. Soc. London<br />

338A, 169-236.<br />

Lin, S. J., 2004. A “Vertically Lagrangian” finite-volume dynamical core for global models.<br />

Mon. Wea. Rev. 132, 2293–2307.<br />

Lin, J.-L., 2007: The double-ITCZ problem in IPCC AR4 coupled GCMs: Ocean-atmosphere<br />

feedback analysis. J. Clim. 20, 4497-4525.<br />

Lipscomb, W. H., Hunke, E.C., Maslowski, W., Jakacki, J., 2007. Ridging, strength, and<br />

stability in high-resolution sea ice models, J. Geophys. Res., 112, C03S91,<br />

doi:10.1029/2005JC003355.<br />

Loeb, N. G., Wielicki, B.A., Doelling, D.R., Smith, G.L., Keyes, D.F., Kato, S., Manalo-Smith,<br />

N., Wong, T., 2009. Toward optimal closure of the Earth's top-of-atmosphere radiation budget,<br />

J. Clim., 22, 748–766. doi: 10.1175/2008JCLI2637.1<br />

Maltrud, M.E., Smith, R.D., Semtner, A.J., Malone, R.C., 1998. Global eddy-resolving ocean<br />

simulations driven by 1985–1995 atmospheric fields. J. Geophys. Res. 103, 30,825–30,853.<br />

Maltrud, M. E., McClean, J.L., 2005. An eddy resolving global 1/10 degree ocean simulation.<br />

Ocean Modell. 8, 31-54.<br />

Maltrud, M., Bryan, F., Peacock, S., 2010. Boundary impulse response functions in a centurylong<br />

eddying global ocean simulation. Environ. Fluid Mech. 10(1-2), 275-295.<br />

doi:10.1007/s10652-009-9154-3.<br />

Maslowski, W., Marble, D., Walczowski, W., Schauer, U., Clement, J.L, Semtner, A.J., 2004.<br />

On climatological mass, heat, and salt transports through the Barents Sea and Fram Strait from<br />

a pan-Arctic coupled ice-ocean model simulation, J. Geophys. Res., 109, C03032,<br />

doi:10.1029/2001JC001039.<br />

Maximenko, N., Niiler, P., Rio, M.-H., Melnichenko, O., Centuroni, L., Chambers, D., Zlotnicki,<br />

V., Galperin, B., 2009. Mean dynamic topography of the ocean derived from satellite and<br />

drifting buoy data using three different techniques. J. Atmos. Ocean. Technol. 26,1910-1919.<br />

McClean, J.L., Poulain, P.-M., Pelton, J.W., Maltrud, M.E., 2002. Eulerian and Lagrangian<br />

statistics from surface drifters and a high-resolution POP simulation in the North Atlantic. J.<br />

Phys. Oceanogr. 32, 2472-2491.<br />

McClean, J. L., Maltrud, M.E., Bryan, F.O., 2006. Measures of the fidelity of eddying ocean<br />

models. Oceanography, 19, 104-117.<br />

McClean, J. L., Jayne, S., Maltrud, M.E., Ivanova, D.P., 2008. The fidelity of ocean models with<br />

explicit eddies. In “Ocean Modeling in an Eddying Regime”. M. Hecht and H. Hasumi,<br />

Editors, AGU Geophys. Monogr. Ser. 170, pp.149-163.<br />

McDougall, T.J., Jackett, D.R., Wright, D.G., Feistel, R., 2003: Accurate and computationally<br />

efficient algorithms for potential temperature and density of seawater. J. Atmos. Ocean.<br />

Technol. 20, 730-741.<br />

Melling, H., 2000. Exchanges of freshwater through the shallow straits of the North American<br />

Arctic. The Freshwater Budget of the Arctic Ocean, 497-502, Kluwer Academic Publishers,<br />

the Netherlands.<br />

Murray R. J., 1996. Explicit generation of orthogonal grids for ocean models. J. Comput. Phys.<br />

126, 251-273.<br />

21


935<br />

936<br />

937<br />

938<br />

939<br />

940<br />

941<br />

942<br />

943<br />

944<br />

945<br />

946<br />

947<br />

948<br />

949<br />

950<br />

951<br />

952<br />

953<br />

954<br />

955<br />

956<br />

957<br />

958<br />

959<br />

960<br />

961<br />

962<br />

963<br />

964<br />

965<br />

966<br />

967<br />

968<br />

969<br />

970<br />

971<br />

972<br />

973<br />

974<br />

975<br />

976<br />

977<br />

978<br />

979<br />

980<br />

Neale, Richard, Slingo, Julia, 2003. The Maritime Continent and Its Role in the Global Climate:<br />

A GCM Study. J. Climate, 16, 834–848.<br />

doi:10.1175/15200442(2003)0162.0.CO;2<br />

Neale, R. B., Richter, J.H., Jochum, M., 2008. The impact of convection on ENSO: from a<br />

delayed oscillator to a series of events. J. Clim. 21, 5904-5924.<br />

Olsen, S. M., Hansen, B., Quadfasel, D. Osterhus, S., 2008. Observed and modelled stability of<br />

overflow across the Greenland-Scotland ridge. Nature 455, 519-523.<br />

Oschlies, A., 2002. Improved representation of upper-ocean dynamics and mixed-layer depths in<br />

a model of the North Atlantic on switching from eddy-permitting to eddy resolving grid<br />

resolution. J. Phys. Ocean. 32, 2277–2298.<br />

Østerhus, S., Turrell, W.R., Jónsson, S., Hansen B., 2005. Measured volume, heat, and salt<br />

fluxes from the Atlantic to the Arctic Mediterranean. Geophys. Res. Lett. 32, L07603,<br />

doi:10.1029/2004GL022188.<br />

Otto, L., Zimmerman, J.T.F., Furnes, G.K., Mork, M., Saetre, R., Becker, G., 1990. Review of<br />

the physical oceanography of the North Sea. Neth. J. Sea Res., 26, 161-238.<br />

Pacanowski, R.C., Gnanadesikan, A., 1998. Transient response in a z-level ocean model that<br />

resolves topography with partial cells. Mon. Wea. Rev. 126, 3248-3270.<br />

Pope, V., Stratton, R., 2002. The processes governing horizontal resolution sensitivity in a<br />

climate model. Clim. Dyn. 19, 211-236.<br />

Qu, T., 2000. Upper-layer circulation in the South China Sea. J. Phys. Oceanogr., 30, 1450-1460.<br />

Reynolds, R. W., Smith, T. M., Chunying, L., Chelton, D.B., Casey, K.S., Schlax, M.G., 2007.<br />

Daily high-resolution-blended analyses for sea surface temperature. J. Clim. 20, 5473-5496.<br />

Richter J. H., Rasch, P.J., 2008. Effects of convective momentum transport on the atmospheric<br />

circulation in the Community Atmosphere Model, version 3. J. Clim. 21, 1487– 1499.<br />

doi:10.1175/2007JCLI1789.1<br />

Ridderinkhof, H., van der Werf, P.M., Ullgren, J.E., van Aken, H. M., van Leeuwen, P.J., W. P.<br />

de Ruijter, M., 2010. Seasonal and interannual variability in the Mozambique Channel from<br />

moored current observations. J. Geophys. Res. 115, C06010, doi:10.1029/2009JC005619.<br />

Risien, C. M., Chelton, D.B., 2008: A global climatology of surface wind and wind stress fields<br />

from eight years of QuikSCAT Scatterometer data. J. Phys. Oceanogr. 38, 2379-2413.<br />

Rouault, M., Reason, J.C., Lutjeharms, J.R., Beljaars, A.C.M., 2003. Underestimation of latent<br />

and sensible heat fluxes above the Agulhas Current in NCEP and ECMWF analyses. J. Clim.<br />

16, 776-782.<br />

Sadler, H.E, 1976. Water, heat, and salt transports through Nares Strait, Ellesmere Island. J. Fish.<br />

Res. Board of Canada 33, 2286-2295.<br />

Schauer, U., Fahrbach, E., Osterhus, S., Rohardt, G., 2004. Arctic warming through the Fram<br />

Strait: Oceanic heat transport from 3 years of measurements. J. Geophys. Res. 109,<br />

doi:10.1029/2003JC001 823.<br />

Screen, J. A., Gillett, N.P., Stevens, D.P., Marshall, G.J., Roscoe, H.K., 2009. The role of eddies<br />

in the Southern Ocean temperature response to the Southern Annular Mode. J. Clim. 22, 806-<br />

818.<br />

Sellers, W.D. 1969. A Global Climatic Model Based on the Energy Balance of the Earth-<br />

Atmosphere System. J. Appl. Meteor. 8, 392-400, doi:10.1175/1520-<br />

0450(1969)0082.0.CO;2<br />

Smith, R. D., Maltrud, M. E., Bryan, F. O., Hecht, M.W., 2000. Numerical simulation of the<br />

22


981<br />

982<br />

983<br />

984<br />

985<br />

986<br />

987<br />

988<br />

989<br />

990<br />

991<br />

992<br />

993<br />

994<br />

995<br />

996<br />

997<br />

998<br />

999<br />

1000<br />

1001<br />

1002<br />

1003<br />

1004<br />

1005<br />

1006<br />

1007<br />

1008<br />

1009<br />

1010<br />

1011<br />

1012<br />

1013<br />

1014<br />

1015<br />

1016<br />

1017<br />

1018<br />

1019<br />

1020<br />

1021<br />

1022<br />

1023<br />

1024<br />

1025<br />

1026<br />

North Atlantic Ocean at 1/10°. J. Phys. Oceanogr. 30, 1532-1561.<br />

Smith, R. H., Johns, W.E., Johns, E.M., (2007), Volume transport and variability at Windward<br />

Passage, EOS Trans. AGU, 88(23), Jt. Assem. Suppl., Abstract OS52A-08.<br />

Sprintall J., Wijffels, S., Molcard, R., Jaya, I., 2009. Direct estimates of the Indonesia<br />

Throughflow entering the Indian Ocean. J. Geophys. Res.114, C07001, doi:<br />

10.1029/2008JC005257.<br />

Sriver, R. L., Huber, M., 2007. Observational evidence for an ocean heat pump induced by<br />

tropical cyclones. Nature, 447, 577-580.<br />

Tian, J., Yang, Q., Liang, X., Xie, L., Hu, D., Wang, F., Qu, T., 2006. Observation of Luzon<br />

Strait transport. Geophys. Res. Lett. 33, L19607, doi:10.1029/2006GL026272.<br />

Trenberth, K. E., Caron, J. M., 2000. Estimates of meridional atmosphere and ocean heat<br />

transports. J. Clim., 14, 3433-3443.<br />

Wang, Y.-H., Chiao, L.-Y., Lwiza, K.M.M., Wang, D.-P., 2004. Analysis of flow at the gate of<br />

Taiwan Strait. J. Geophys. Res., 109, C02025,doi:10.1029/2003JC001937.<br />

Wapler, K., Lane, T.P., May, P.T., Jakob, C., Manton, M.J., Siems, S.T., 2010. Cloud-System-<br />

Resolving Model Simulations of Tropical Cloud Systems Observed during the Tropical Warm<br />

Pool-International Cloud Experiment. Mon. Wea. Rev., 138 55-73,<br />

doi:10.1175/2009MWR2993.1<br />

Whitworth, T., Peterson, R.G., 1985. Volume transport of the Antarctic Circumpolar Current<br />

from bottom pressure measurements. J. Phys. Oceanogr. 15, 810-816.<br />

Williamson D. L, Kiehl, J.T., Hack, J.J., 1995. Climate sensitivity of the NCAR community<br />

climate model (CCM2) to horizontal resolution. Clim. Dyn. 11: 377–397.<br />

Woodgate, R. A., Aagaard, K., Weingartner, T. J., 2005. Monthly temperature, salinity, and<br />

transport variability of the Bering Strait through flow. Geophys. Res. Lett. 32, L04601,<br />

doi:10.1029/2004GL021880.<br />

Xie, P.P., and Arkin, P.A, 1997: Global Precipitation. A 17-year monthly analysis based on<br />

gauge observations, satellite estimates, and numerical model outputs. Bull of Amer. Meteorol.<br />

Soc. 78, 2539-2558.<br />

Figure Captions<br />

Fig. 1. Annual average precipitation rate (mm/day) from Xie-Arkin observations (1997) (top)<br />

and <strong>CCSM</strong>4 (bottom) for model years 13-19.<br />

Fig. 2. (a) Climatology of daily 0.25° Reynolds et al. (2007) satellite-derived SST (°C) data<br />

averaged for 1982-2008. (b) Difference field (°C) between the mean potential temperature for<br />

model years 13-19 from the upper-most model level (5m) and the observed SST climatology.<br />

Contour line is 0°C.<br />

Fig. 3. <strong>CCSM</strong>4 ocean potential temperature (left) and salinity (right) change from model year 19<br />

– model year 1. Red line depicts high latitude regions (50º-90º), green lines are mid-latitudes<br />

(20º-50º), blue line is the tropics (20°S-20°N), and black lines are global. Thick and thin lines<br />

represent the Northern Hemisphere and Southern Hemispheres, respectively.<br />

Fig. 4. (a) Annual average of geopotential height at 500 mb from ERA40 and the (b) difference<br />

(m) of the model annual average geopotential height at 500 mb (m) and the ERA40 field. (c)<br />

23


1027<br />

1028<br />

1029<br />

1030<br />

1031<br />

1032<br />

1033<br />

1034<br />

1035<br />

1036<br />

1037<br />

1038<br />

1039<br />

1040<br />

1041<br />

1042<br />

1043<br />

1044<br />

1045<br />

1046<br />

1047<br />

1048<br />

1049<br />

1050<br />

1051<br />

1052<br />

1053<br />

1054<br />

1055<br />

1056<br />

1057<br />

1058<br />

1059<br />

1060<br />

1061<br />

1062<br />

1063<br />

1064<br />

1065<br />

1066<br />

1067<br />

1068<br />

1069<br />

1070<br />

1071<br />

1072<br />

Annual average of sea level pressure (mb) from ERA40 and the (d) difference (mb) of the model<br />

annual average sea level pressure and the ERA40 field.<br />

Fig. 5. (a) Annual average wind stress vectors (Nm -2 ) from QuikSCAT superimposed on their<br />

vector magnitudes (color contours). (b) Wind stress difference vectors (Nm -2 ) between annuallyaveraged<br />

<strong>CCSM</strong>4 and QuikSCAT superimposed on the magnitudes of their vector differences<br />

(color contours). Model wind stresses are averaged for model years 13-19.<br />

Fig. 6. Time series of (a) Arctic and (b) Antarctic total ice area (km 2 ) from <strong>CCSM</strong>4 for years 0-<br />

19 (red) and from SSM/I (blue) for 1980-1999. The SSM/I data are for the period 1980-1999.<br />

Fig. 7. North Atlantic mean (a) velocity (ms -1 ) and temperature (°C) from the uppermost level<br />

(5m) of the ocean component of <strong>CCSM</strong>4 and (b) wind stress (Nm -2 ) and 5m temperature (°C) for<br />

years 13-19. Thin black contour line is 7°C.<br />

Fig. 8. (a) Mean dynamic ocean surface topography (cm) from Maximenko et al. (2009). Mean<br />

sea surface height (cm) from (b) <strong>CCSM</strong>4 for model years 13-19 and (c) a stand-alone tripole 0.1°<br />

POP simulation for model years 95-99. The same grid is used in the two simulations.<br />

Fig. 9. Difference fields (cm) between (a) <strong>CCSM</strong>4 mean sea surface height (model years 13-19)<br />

and mean dynamic ocean surface topography (MDOT), and between (b) stand-alone tripole 0.1°<br />

POP sea surface height (model years 95-99) and MDOT.<br />

Fig. 10: Mean meridional overturning streamfunction (Sv) for model years 15-19 for the (a)<br />

global ocean and (b) the Atlantic Ocean. Negative contours are dashed and denote a counterclockwise<br />

sense of flow.<br />

Fig. 11. Mixed layer depth (m) climatologies for December-January-February from de Boyer<br />

Montégut et al. (2004) observations, (b) <strong>CCSM</strong>4 (model years 13-19), and (c) POP (model years<br />

95-99). Differences (m): (a) <strong>CCSM</strong>4 – data, (b) POP – data, and (c) <strong>CCSM</strong>4 – POP. MLD<br />

criterion is based on a ΔT=0.2°C from the near-surface temperature value at 10 m.<br />

Fig. 12. Mixed layer depth (m) climatologies for June-July-August from de Boyer Montégut et<br />

al. (2004) observations, (b) <strong>CCSM</strong>4 (model years 13-19), and (c) POP (model years 95-99).<br />

Differences (m): (a) <strong>CCSM</strong>4 – data, (b) POP – data, and (c) <strong>CCSM</strong>4 – POP. MLD criterion is<br />

based on a ΔT=0.2°C from the near-surface temperature value at 10 m.<br />

Fig. 13: (a) Zonally-integrated global ocean total meridional heat transport (PW) from <strong>CCSM</strong>4<br />

(model years 15-19), POP (model years 95-99) and Trenberth and Caron (2000) atmospheric<br />

reanalysis estimates. (b) Zonally-integrated global ocean total, mean, and eddy meridional heat<br />

transports from <strong>CCSM</strong>4 and POP.<br />

Fig. 14. (a) Track of a Category 4 tropical cyclone event in July-August of model year 19<br />

superimposed on the SST (°C) cold water "wake” under and to the right of the track of the center<br />

of the storm in the tropical northwest Pacific. (b) Vertical profiles of temperature (°C) at station<br />

9 (133°E, 26°N) that correspond to pre-storm conditions (07/29/19), a day after the passage of<br />

24


1073<br />

1074<br />

1075<br />

1076<br />

1077<br />

1078<br />

1079<br />

1080<br />

1081<br />

1082<br />

1083<br />

1084<br />

the storm’s center (08/04/19), and the water column once the near-inertial oscillations have<br />

largely abated (08/18/19).<br />

Fig. 15. Mesoscale RMS SSHA (cm) from (a) the ocean component of <strong>CCSM</strong>4 for years 15-19<br />

and (b) the AVISO-blended (TOPEX/POSEIDON and ERS 1 and 2) altimetry for 1997-2001.<br />

Fig. 16. Agulhas eddy pathways from (a) AVISO altimetry, (b) <strong>CCSM</strong>4, and the (c) coupled<br />

POP/CICE simulation. Numbers are percentage of times that eddies follow the same track.<br />

Fig. 17. Evolution of along-track amplitude of SSHA (cm) as a function of time since eddy<br />

formation for AVISO (upper), <strong>CCSM</strong>4 (middle), and POP/CICE (lower).<br />

25


1084<br />

Channel/Strait/Opening<br />

<strong>CCSM</strong>4 Observed<br />

(Sv)<br />

(Sv)<br />

1 Sv = 10 6 m 3 s -1<br />

References for<br />

observational<br />

estimates<br />

Barents Opening 3.1 northward 1.5±0.1<br />

northward<br />

Ingvaldsen et al.<br />

(2004)<br />

Bering Strait 1.4 northward 0.8±0.16<br />

northward<br />

Woodgate and<br />

Aagaard (2005)<br />

Canadian Archipelago 1.2 southward 0.7-2.0<br />

southward<br />

Sadler (1976),<br />

Fissel et al. (1998)<br />

Melling (2000)<br />

Denmark Strait<br />

6.5 net<br />

southward<br />

0.8 net<br />

northward<br />

3 net southward<br />

Østerhus et al.<br />

(2005)<br />

Olsen el al. (2008)<br />

Drake Passage 168.9 eastward 136.7±7.8<br />

eastward<br />

Cunningham et al.<br />

(2003)<br />

English Channel 0.17 northward 0.1-0.2 Otto et al. (1990)<br />

northward<br />

Pacific Equatorial<br />

30.7 eastward 26 eastwards Kessler et al. (2003)<br />

Undercurrent (140°W)<br />

Faroe-Scotland Channel 5.5 net<br />

northward<br />

3.8 net<br />

northwards<br />

2.1 net<br />

southward<br />

Østerhus et al.<br />

(2005)<br />

Olsen el al. (2008)<br />

Florida-Bahamas Strait 20.5 northward 32.3±3.2 Larsen (1992)<br />

Fram Strait<br />

3.1 net<br />

southward<br />

4±2 & 2±2 net<br />

southward<br />

Schauer et al.<br />

(2004)<br />

Iceland-Faroe Channel 1.0 net<br />

northward<br />

3.8 net<br />

northward<br />

1 net southward<br />

Østerhus et al.<br />

(2005)<br />

Olsen el al. (2008)<br />

Indonesian Throughflow 13.8 westward 15.0 westward Sprintall et al.<br />

(2009)<br />

Mozambique Channel 18.5 southward 16.7 southward Ridderinkhof et al.<br />

(2010)<br />

Taiwan-<br />

Luzon Straits:<br />

2.2 northwards<br />

3.4 westwards<br />

1.6 southward<br />

6±3 westward<br />

3.0 westward<br />

Wang et al. (2004)<br />

Tian et al. (2006)<br />

Qu (2000)<br />

Windward Passage 0.17 southward 3.6 southward Smith et al. (2007)<br />

1085<br />

1086<br />

Table 1: Depth-integrated mass transports (Sv) through selected channels, straits, and openings<br />

from <strong>CCSM</strong>4 and observations.<br />

26


Fig. 9: Difference fields (cm) between (a) <strong>CCSM</strong>4 mean sea surface height (model years 13-<br />

19) and mean dynamic ocean surface topography (MDOT), and between (b) stand-alone<br />

tripole 0.1° POP sea surface height (model years 95-99) and MDOT.<br />

9

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!