Revisiting Mountains and Stationary Waves: the Importance of the Zonal1

Mean Response2

R.H. White∗3

Barcelona Supercomputing Center, Barcelona, Spain.4

J.M. Wallace and D.S. Battisti5

Department of Atmospheric Sciences, University of Washington, Seattle, WA, USA6

∗Corresponding author address: R.H. White, Barcelona Supercomputing Center, Carrer de Jordi

Girona, 29, 31, 08034 Barcelona, Spain.

7

8

E-mail: rachel.white@cantab.net9

Generated using v4.3.2 of the AMS LATEX template 1

ABSTRACT

The impact of global orography on wintertime climate is revisited using a

state-of-the-art climate model, with sophisticated parameterization of the im-

pacts of sub-gridscale orography (the Whole Atmosphere Community Cli-

mate Model, WACCM). The main impact of orography is to amplify the

stationary waves produced by land-sea thermal contrasts, as evidenced by

the strong spatial resemblance between the stationary wave pattern in a no-

mountain (NM) simulation with that induced by orography. Orography am-

plifies extratropical NM stationary waves at all levels, with an amplification

factor of a∼ 1.6 in the mid-latitudes, increasing to a> 1.9 in the high-latitude

stratosphere. Orography also decelerates the zonal mean zonal wind pole-

ward of 40◦N throughout the troposphere and stratosphere. The effect of the

weakening of the zonal winds on stationary waves is twofold: it amplifies the

thermal forcing and it amplifies the planetary-wave response to a prescribed

thermal forcing.

10

11

12

13

14

15

16

17

18

19

20

21

22

23

2

1. Introduction24

Earth’s atmospheric stationary waves are seen in the planetary scale wave-like structure in cli-25

matological sea level pressure or geopotential height. In the Northern Hemisphere the dominant26

features of the wintertime stationary waves are so prominent that they have been given names: the27

Aleutian low, the Icelandic low, and the Siberian high. The forcing of stationary waves is provided28

by planetary features that are, unsurprisingly, largely stationary: orography (Charney and Eliassen29

1949; Bolin 1950), and zonal asymmetries in atmospheric heating, typically from land-sea con-30

trasts (e.g. Smagorinsky 1953; Manabe and Terpstra 1974; Hoskins and Karoly 1981; Valdes and31

Hoskins 1991).32

From the 1980s to 2010 numerous studies examined the relative contributions of ‘orographic’33

and ‘thermal’ forcing to the observed stationary waves. This research, collectively using linear and34

non-linear stationary wave models, as well as full general circulation models (GCMs), has shown35

that orography produces anywhere between 30% and 66% of the observed stationary wave ampli-36

tude, with recent estimates towards the lower end of this range (Held 1983; Chen and Trenberth37

1988a,b; Nigam et al. 1988; Valdes and Hoskins 1989; Ting et al. 2001; Held et al. 2002; Chang38

2009). The authors’ impression is that the value of 30% from the review of Held et al. (2002) is39

often taken as the best estimate of the role of mountains, even if it were not intended as such. The40

large range of results can be at least partially understood from differences in the background flow41

into which the orography was placed, including differences in low-level winds (e.g. Valdes and42

Hoskins 1989; Held and Ting 1990; Ringler and Cook 1995; Held et al. 2002), as well as model43

differences in atmospheric dissipation strength and imposed drag (Valdes and Hoskins 1989).44

Much less attention has been paid to the impact of orography on the axisymmetric flow. As Held45

et al. (2002) noted, the tropospheric zonal mean zonal wind structure is broadly similar between46

3

the Northern and Southern hemispheres (NH and SH), suggesting that the orography of the NH47

has relatively little impact in the troposphere, consistent with earlier modelling results (Manabe48

and Terpstra 1974). Conversely, a strong asymmetry exists between the wintertime zonal mean49

stratospheric jets in each hemisphere, caused by large differences in wave activity generated by50

land-sea contrasts and orography (Waugh and Polvani 2010; White et al. 2018). Early modelling51

results suggest that orography reduces the speed of the NH winter stratosphere jet by approxi-52

mately 15% (Manabe and Terpstra 1974), although large biases in that model existed. In a modern53

GCM, White et al. (2018) found that the presence of the ‘Mongolian mountains’ of Asia alone54

reduces zonal mean stratospheric wind speeds in winter by up to 40%.55

In addition to acting as a large-scale obstruction to the flow, orography impacts the climate56

through subgrid-scale orographic drag, via the forcing of orographic gravity waves, as well as the57

blocking of flow (e.g. McFarlane 1987; Shaw et al. 2009; Sigmond and Scinocca 2010; Pithan58

et al. 2016; Sandu et al. 2016, 2019). Including gravity wave parameterization of orographic59

effects substantially increases zonal mean sea level pressure in the high latitude NH (shown in the60

appendix of Broccoli and Manabe 1992), and will thus impact the low-level zonal mean zonal61

wind (U). Parameterizing the effects of orographic drag reduces tropospheric U (Sandu et al.62

2016) and alters stationary waves and their response to climate change (van Niekerk et al. 2017;63

Sandu et al. 2019). The representation of such impacts within GCMs has improved in recent years,64

through both increased resolution, and improved parameterizations (Garcia et al. 2007; Richter65

et al. 2010; Garcia et al. 2016), although limitations and biases still exist (Sandu et al. 2019). The66

time therefore seems ripe to re-evaluate the role that orography plays in the wintertime climate67

using a state-of-the-art climate model.68

Using NCAR’s Whole Atmosphere Community Climate Model (WACCM), we quantify the to-69

tal role of orography on the observed stationary waves, including both local and remote changes70

4

in diabatic heating, as well as changes in SSTs due to air-sea heat fluxes. We present evidence71

for a substantial impact of orography on the zonal mean climate in both the troposphere and72

stratosphere, and highlight the remarkable similarity between the stationary wave patterns from73

‘land-sea forcing’ and from ‘orographic’ forcing. We offer a new perspective on how orography74

influences the stationary waves, with implications for our understanding of stationary waves in75

the present day and other climates: we propose that the most important impact of orography on76

stationary wave amplitude is to amplify the ‘land-sea’ stationary waves, in addition to forcing a77

stationary wave pattern of its own, as in the traditional view.78

2. Models and Experiments79

We use the WACCM4, an atmospheric dynamical model with 66 vertical levels, including a well-80

resolved stratosphere (Marsh et al. 2013). In addition to high resolution in the upper atmosphere,81

the WACCM has improved parameterizations of vertically propagating gravity waves and surface82

stress from unresolved topography (termed turbulent mountain stress, TMS), relative to the lower-83

top CAM model (Garcia et al. 2007; Richter et al. 2010; Marsh et al. 2013). A horizontal resolution84

of 1.9◦x 2.5◦is used.85

To investigate the role of orography in shaping the stationary waves, two WACCM simulations86

are performed, each 30 years in length (following a one year spin-up period). The ‘CTL’ sim-87

ulation has present-day topography, and the ‘NM’ (No-Mountain) simulation has all topography88

flattened to 50m above sea level. In addition to the topographic height, we reduce the variables89

associated with subgrid-scale variability in orography, ‘SGH’ and ‘SGH30’, to approximate values90

associated with regions of low orography in the CTL simulation (30 and 10 m respectively).91

Mountains are known to affect the ocean circulation (Kitoh 2002; Sinha et al. 2012) through92

impacts on wind stress and freshwater fluxes (Warren 1983; Emile-Geay 2003). Here we focus93

5

on the dynamical atmospheric response to global mountain ranges and do not consider dynamical94

ocean changes. Orographic forcing can also influence SSTs through air-sea heat fluxes (Okajima95

and Xie 2007; Chang 2009). This influence may affect the impact of orography on diabatic heating,96

although in the past this has been considered to be small (Held 1983). Since the WACCM is not97

currently coupled to a slab ocean, we allow atmospheric changes to affect SSTs by using simulated98

changes in monthly climatological SSTs from a slab-ocean model coupled to the CAM model99

with and without orography. These SST changes and their impacts on our results are discussed in100

Section 4.101

The NH extratropical SSTs for each experiment, as well as the CTL-NM difference, are shown102

in Fig. 8.103

The WACCM model reproduces tropospheric and stratospheric circulations well, albeit with a104

cold pole bias (Marsh et al. 2013). Throughout this paper we further evaluate the WACCM by105

comparing it with the ERA-interim re-analysis data (Dee et al. 2011).106

3. Results107

a. Stationary Wave Response108

The stationary wave pattern can be seen in the eddy geopotential height, Z′ = Z− Z, where109

the overbar denotes the zonal mean. Fig. 1 shows Z′ at 60◦N for ERA-I, the CTL and NM110

simulations, and the CTL-NM difference. The CTL simulation reproduces the observed pattern111

from ERA-interim re-analysis data well (compare Figs. 1a and b). The climate in the CTL and112

NM simulations, and thus the climate response to orography (CTL-NM), is broadly similar to that113

found in earlier studies (e.g. Nigam et al. 1986; Held 1983; Held et al. 2002).114

6

To provide a quantitative metric for comparing the relative contributions of orography and land-115

sea contrast, we calculate the ‘stationary wave strength’, defined as the longitudinal standard de-116

viation (σ ) in Z. The ‘orographic contribution’ is calculated as (σ(ZCT L)−σ(ZNM))/σ(ZCT L).117

There is substantial variation in the orographic contribution with latitude and height, shown in118

Fig. 2a, with a maximum value of∼ 85% in the high latitude stratosphere, and a latitude-weighted119

value in the NH mid-latitudes (30-70◦N) of approximately 45% throughout the troposphere and120

stratosphere.121

In idealized simulations, the stationary response to a single mountain has a zonal wavenumber122

of around 4-6 (Hoskins and Karoly 1981; Valdes and Hoskins 1991; Cook and Held 1992; Held123

et al. 2002; White et al. 2017). In contrast to these idealized studies, our CTL-NM response124

has a wavenumber, phase, and vertical structure remarkably similar to that in the NM simulation125

(compare Figs. 1c and d), with a pattern of approximately wavenumber 2 in the troposphere, and126

wavenumber 1 in the stratosphere. That is, the first order effect of the mountains is to amplify127

the mid-latitude stationary wave pattern that is already present as a consequence of the land-sea128

contrast.129

This amplification of the land-sea stationary wave pattern (NM) in the CTL simulation can also130

be seen in maps of Z′ (Fig. 3). A comparison of Z′ between ERA-I and our CTL simulation shows131

that the WACCM simulates the observed spatial pattern of stationary waves well at both 300 and132

1000 mb (c.f. Fig. 3a with 3b and Fig. 3d and 3e; colours show the eddy component, whilst black133

contours show the full fields). The NM eddy fields (Figs. 3c and f) exhibit a spatial structure134

similar to that of the CTL fields, but with muted amplitude; the first order effect of orography is to135

greatly amplify this pattern.136

We calculate the Pearson correlation coefficient between the CTL-NM and NM fields, with137

masking for points that are below the surface in the CTL simulation (Fig. 2b. At 1000, 300, and138

7

30 mb respectively, the latitude-weighted correlation coefficients for the mid-latitudes are 0.33,139

0.60, and 0.81, confirming that a large component of the CTL-NM response is to amplify the140

background stationary wave pattern present in the NM simulation, particularly above the lower141

troposphere. To provide a quantitative metric for the strength of this amplification by orography142

we define an ‘amplification factor’, a, as the slope of the linear regression between the CTL and143

NM fields; we can thus write Z′CT L = aZ′NM +Z′M, where Z′M is the orography-induced change in144

eddy geopotential height that does not project onto the NM pattern. This amplification factor is145

show in Fig. 2c, with a latitude-weighted mean value of a ∼ 1.6 throughout the troposphere and146

stratosphere. Vertical cross-sections of both the correlation coefficient and amplification factor are147

shown in Figs. 2b and c.148

b. Zonal mean response to orography149

In addition to the stationary wave response, the WACCM simulations show that orography sub-150

stantially shapes the wintertime zonal mean zonal wind climatology. Fig. 4 shows a pole-to-pole151

meridional cross-section of U (colours), and Z (black contours) for winter in each hemisphere,152

i.e. DJF in the NH and JJA in the SH. Fields are masked between 10◦S and 10◦N to emphasise153

the discontinuity at the equator. The WACCM reproduces the general structure of the wintertime154

geopotential and zonal jets in both hemispheres well, including the difference in strength of the155

NH and SH stratospheric jets (c.f. Figs 4a and b). The NM fields (Fig. 4c) show much greater156

symmetry between the NH and SH, as expected. At 10 mb the NH wintertime stratospheric jet157

in the NM simulation is only ∼ 10% weaker than its SH counterpart: orography thus plays a158

dominant role in the hemispheric differences in the stratospheric winter jet (see also White et al.159

2018).160

8

Fig. 4d shows the CTL-NM differences in the zonal mean fields. The orographic impact on161

the NH stratospheric jet is substantial: at 10 mb the jet reaches 82 m/s in the NM simulation,162

in contrast to 39 m/s in the CTL (c.f. Figs. 4b and c). Consistent with previous understanding163

(e.g. Held et al. 2002), changes in the sub-tropical jet are small. There is, however, a significant164

orographically-induced weakening of tropospheric U north of 40◦N, extending all the way down165

to the surface. Whilst the absolute magnitude of the near-surface changes is small relative to166

the stratospheric changes (∼ 3m/s at 925 mb between 50-80◦N), it has a profound effect on the167

character of the low-level flow.168

To explore the mechanisms behind the weakening of U in the CTL simulation relative to NM,169

Fig. 5 shows the Eliassen-Palm (EP) wave activity flux and its divergence (Eliassen and Palm170

1961; Andrews and McIntyre 1976; Edmon et al. 1980). Arrows are scaled (see caption), and thus171

should not be used to estimate divergence. The EP fluxes are calculated based on climatological172

wintertime data, that is, they show the propagation of wintertime stationary wave activity, and the173

impacts of the wintertime stationary waves on the zonal mean flow. Positive divergence of the EP174

fluxes signifies a strengthening of the zonal mean flow by the waves (Andrews and McIntyre 1976;175

Edmon et al. 1980).176

The CTL simulation reproduces the ERA-I spatial pattern of stationary wave EP fluxes, and their177

divergence, relatively well (c.f. Figs. 5a and b), although the magnitude of EP flux divergence in178

the mid-latitude troposphere is underestimated; we thus avoid making quantitative conclusions179

based on this analysis. The EP flux arrows show a reduction in the propagation of wave activity180

strength away from the surface in the NM simulation relative to the CTL (c.f. Figs 5b and c).181

Stationary wave activity is still present in the troposphere in the NM simulation: the land-sea182

forcing is sufficient to produce substantial tropospheric wave activity at lower latitudes; however183

the strong EP flux divergence north of 40◦N is absent in the NM simulation. There is also a184

9

stark absence of wave activity (and thus divergence) reaching into the stratosphere in the NM185

simulation. Calculation of EP fluxes on daily data confirms that stationary waves dominate the186

stratospheric CTL-NM response, consistent with a relatively small propagation of transient waves187

into the stratosphere (Li et al. 2011).188

4. Discussion189

The impact of global orography on the climatological extratropical circulation is examined with190

the WACCM, a state-of-the-art climate model, revisiting the seminal work of the 1980s and 1990s.191

The influence of orography on the NH mid-latitude stationary waves is generally comparable to192

land-sea thermal contrasts, with orography dominating in some parts of the troposphere and strato-193

sphere. This is consistent with previous work, but suggests a larger role for orography than the194

most recent estimates (Ting et al. 2001; Held et al. 2002; Chang 2009). We highlight the remark-195

able similarity between the NM stationary wave pattern and the CTL-NM response, concluding196

that the first-order role of orography is to amplify the stationary wave pattern that would be present197

by virtue of the land-sea contrast alone. The amplification factor, a measure of how much orog-198

raphy amplifies the NM stationary wave pattern, is approximately 1.6 for the troposphere, and 1.8199

for the stratosphere.200

The impact of orography on large-scale extratropical circulation is not limited to the stationary201

waves - orography has a strong impact on the zonal mean flow poleward of 40◦N. Without orog-202

raphy, the NH wintertime polar night jet is much stronger than in observations and the WACCM203

CTL simulation, and is comparable to the strength of the SH wintertime stratospheric jet. The204

strong sensitivity of the NH stratospheric winter jet to the inclusion of mountains is consistent205

with the concept of a positive feedback in stratospheric zonal wind: faster stratospheric flow limits206

the waves that can propagate into the stratosphere to smaller wavenumbers (Charney and Drazin207

10

1961); this reduces the Eliassen-Palm flux convergence in the stratosphere, further enhancing the208

zonal wind. Our results suggest that, in the absence of orography, very little stationary wave ac-209

tivity is able to propagate into the stratosphere. The impact of orography on zonal winds extends210

all the way down to the surface poleward of ∼ 50◦N; gravity wave drag and frictional drag from211

the presence of mountains impact the near surface flow.212

This profound impact of orography on the zonal mean flow can be connected to the orographic213

amplification of the land-sea stationary wave that we have highlighted. For low-level extratrop-214

ical heating, the amplitude of the stationary wave response to a given diabatic heating profile is215

inversely proportional to the low-level mean zonal wind (Held and Ting 1990). The linear model216

of Held and Ting (1990) shows a stationary wave amplitude response to low-level wind changes217

of the same order of magnitude that we see in our simulations: an amplification of ∼1.6 for a218

reduction in U from, for example, ∼ 6 m/s (NM) to 2 m/s (CTL).219

In addition to the reduction in stationary wave amplitude for a fixed diabatic heating, decel-220

eration of the zonal mean flow allows stronger atmospheric zonal temperature gradients (from221

land-sea contrast) to be maintained; in the limit of infinite zonal flow, no zonal temperature gra-222

dients could be maintained. Thus a reduced zonal flow can enhance the diabatic heating pattern223

from land-sea contrast, leading to a further amplification of the ‘thermal’ stationary wave pattern.224

Fig. 6 shows that the diabatic heating patterns are indeed enhanced in the CTL simulation relative225

to NM.226

The positive feedback of stratospheric zonal wind speed, discussed above, may provide an addi-227

tional mechanism by which orography amplifies the existing stationary wave in the stratosphere:228

in the presence of orography, enhanced stationary wave EP flux convergence in the stratosphere229

weakens the stratospheric jet; this in turn enhances the probability of waves generated by land-sea230

contrast propagating into the stratosphere. This mechanism differs somewhat from that discussed231

11

by White et al. (2018), who removed individual mountain regions; this difference suggests non-232

linear interactions between different orographic regions. The amplification of the NM stationary233

wave pattern by the orography may also be affected by the location of the main orographic fea-234

tures of the NH. Both the North American Rockies and the dominant orography in Asia are located235

approximately 30◦ longitude upstream of the eastern edge of a continent; the mechanical forcing236

of the orography results in northerly flow and cold advection downstream of the mountain range,237

which is thus, rather coincidentally, located such that it enhances the existing land-sea temperature238

contrast (e.g. Brayshaw et al. 2009).239

The impact of orography is reduced when SSTs are held fixed to the observed climatology in240

both the CTL and NM simulations (c.f. Figs. 2a) and 7a). The degree of correlation between241

the CTL-NM and NM patterns (7b), and the orographic amplification factor 7c) are also reduced242

(although still present). The imposed changes in SSTs, as simulated by the CAM coupled to a slab243

ocean model with q-fluxes modified to reproduce observed SSTs in the CTL simulation, are shown244

in Fig. 8. The strong cooling in the west of each ocean basin is consistent with cold air advec-245

tion induced by the orographically forced stationary wave. The strong warming in the northeast246

Atlantic is associated with sea-ice changes. The influence of this lower boundary condition may247

be part of the reason that this amplification effect has remained largely unnoticed up until now. In248

retrospect, amplification can be seen to some degree in earlier studies, even with SSTs were fixed249

to observed values in both CTL and NM simulations. Held (1983), Held et al. (2002) and Chang250

(2009) mention a constructive interference between the thermal forcing and the orography forcing251

between 120E and 60W, but did not explore this further.252

The amplification of the land-sea stationary wave as a large part of the orographic contribution253

suggests a greater importance for the land-sea thermal contrast that previously understood. This254

has potential implications for stationary waves under global warming: during the coming century255

12

the land will warm more than the ocean surface, reducing the land-sea thermal contrast during win-256

ter. A greater importance of this land-sea contrast for forcing stationary waves may help explain257

the projected reduction in low wavenumber stationary waves seen in Simpson et al. (2016).258

This work offers a new perspective of how orography shapes Earth’s stationary waves: orogra-259

phy provides a substantial amplification of the background, land-sea generated, stationary wave260

pattern, through the impact of orography on the zonal mean flow, as well as forcing an additional261

stationary wave pattern.262

Acknowledgments. RHW received funding from the European Unions Horizon 2020 research263

and innovation programme under the Marie Skodowska-Curie Grant Agreement 797961, from264

the National Science Foundation grant AGS-1665247, and from the Tamaki Foundation. DSB265

was supported by the Tamaki Foundation. Simulations were performed at the Department of266

Atmospheric Sciences at the University of Washington, and on the NCAR’s Cheyenne computer267

under grant UWAS0064. The code to produce the figures shown in this paper can be found on268

RHW’s github account (rhwhite): https://git.io/fjumK.269

References270

Andrews, D. G., and M. E. McIntyre, 1976: Planetary waves in horizontal and vertical shear:271

The generalized Eliassen-Palm relation and the mean zonal acceleration. J. Atmos. Sci., 33 (11),272

2031–2048.273

Bolin, B., 1950: On the influence of the earth’s orography on the general character of the wester-274

lies. Tellus, 2 (3), 184–195.275

Brayshaw, D. J., B. Hoskins, and M. Blackburn, 2009: The basic ingredients of the north atlantic276

storm track. part i: Land–Sea contrast and orography. J. Atmos. Sci., 66 (9), 2539–2558.277

13

Broccoli, A. J., and S. Manabe, 1992: The effects of orography on midlatitude northern hemi-278

sphere dry climates. J. Clim., 5 (11), 1181–1201.279

Chang, E. K. M., 2009: Diabatic and orographic forcing of northern winter stationary waves and280

storm tracks. J. Clim., 22 (3), 670–688.281

Charney, J. G., and P. G. Drazin, 1961: Propagation of planetary-scale disturbances from the lower282

into the upper atmosphere. J. Geophys. Res., 66 (1), 83–109.283

Charney, J. G., and A. Eliassen, 1949: A numerical method for predicting the perturbations of the284

middle latitude westerlies. Tellus, 1 (2), 38–54.285

Chen, S.-C., and K. E. Trenberth, 1988a: Forced planetary waves in the northern hemisphere286

winter: Wave-Coupled orographic and thermal forcings. J. Atmos. Sci., 45 (4), 682–704.287

Chen, S.-C., and K. E. Trenberth, 1988b: Orographically forced planetary waves in the northern288

hemisphere winter: Steady state model with Wave-Coupled lower boundary formulation. J.289

Atmos. Sci., 45 (4), 657–681.290

Cook, K. H., and I. M. Held, 1992: The stationary response to Large-Scale orography in a general291

circulation model and a linear model. J. Atmos. Sci., 49 (6), 525–539.292

Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: configuration and performance of293

the data assimilation system. Q.J.R. Meteorol. Soc., 137 (656), 553–597.294

Edmon, H. J., B. J. Hoskins, and M. E. McIntyre, 1980: Eliassen-Palm cross sections for the295

troposphere. J. Atmos. Sci., 37 (12), 2600–2616.296

Eliassen, A., and E. Palm, 1961: Eliassen and palm (1961). Geofys. Publ.297

14

Emile-Geay, J., 2003: Warren revisited: Atmospheric freshwater fluxes and “Why is no deep water298

formed in the North Pacific”. J. Geophys. Res., 108 (C6), 243.299

Garcia, R. R., D. R. Marsh, D. E. Kinnison, B. A. Boville, and F. Sassi, 2007: Simulation of300

secular trends in the middle atmosphere, 1950–2003. J. Geophys. Res., 112 (D9), 2113.301

Garcia, R. R., A. K. Smith, D. E. Kinnison, A. d. l. Camara, and D. J. Murphy, 2016: Modifica-302

tion of the gravity wave parameterization in the whole atmosphere community climate model:303

Motivation and results. J. Atmos. Sci., 74 (1), 275–291.304

Held, I. M., 1983: Stationary and quasi-stationary eddies in the extratropical troposphere: theory.305

Large-Scale Dynamical Processes in the Atmosphere, R. P. P. B J Hoskins, Ed., Academic Press,306

127–167.307

Held, I. M., and M. Ting, 1990: Orographic versus Thermal Forcing of Stationary Waves: The308

Importance of the Mean Low-Level Wind. J. Atmos. Sci., 47 (4), 495–500.309

Held, I. M., M. Ting, and H. Wang, 2002: Northern winter stationary waves: Theory and modeling.310

J. Clim., 15 (16), 2125–2144.311

Hoskins, B. J., and D. J. Karoly, 1981: The steady linear response of a spherical atmosphere to312

thermal and orographic forcing. J. Atmos. Sci., 38 (6), 1179–1196.313

Kitoh, A., 2002: Effects of Large-Scale mountains on surface climate —A coupled Ocean-314

Atmosphere general circulation model study. Journal of the Meteorological Society of Japan.315

Ser. II, 80 (5), 1165–1181.316

Li, Q., H.-F. Graf, and X. Cui, 2011: The role of stationary and transient planetary waves in the317

maintenance of stratospheric polar vortex regimes in northern hemisphere winter. Adv. Atmos.318

Sci., 28 (1), 187–194.319

15

Manabe, S., and T. B. Terpstra, 1974: The effects of mountains on the general circulation of the320

atmosphere as identified by numerical experiments. J. Atmos. Sci., 31 (1), 3–42.321

Marsh, D. R., M. J. Mills, D. E. Kinnison, J.-F. Lamarque, N. Calvo, and L. M. Polvani, 2013:322

Climate change from 1850 to 2005 simulated in CESM1(WACCM). J. Clim., 26 (19), 7372–323

7391.324

McFarlane, N. A., 1987: The effect of orographically excited gravity wave drag on the general325

circulation of the lower stratosphere and troposphere. J. Atmos. Sci., 44 (14), 1775–1800.326

Nigam, S., I. M. Held, and S. W. Lyons, 1986: Linear simulation of the stationary eddies in a327

general circulation model. Part I: The No-Mountain model. J. Atmos. Sci., 43 (23), 2944–2961.328

Nigam, S., I. M. Held, and S. W. Lyons, 1988: Linear simulation of the stationary eddies in a329

GCM. part II: The “mountain” model. J. Atmos. Sci., 45 (9), 1433–1452.330

Okajima, H., and S.-P. Xie, 2007: Orographic effects on the northwestern pacific monsoon: Role331

of air-sea interaction. Geophys. Res. Lett., 34 (21), 1989.332

Pithan, F., T. G. Shepherd, G. Zappa, and I. Sandu, 2016: Climate model biases in jet streams,333

blocking and storm tracks resulting from missing orographic drag. Geophys. Res. Lett., 43 (13),334

2016GL069 551.335

Richter, J. H., F. Sassi, and R. R. Garcia, 2010: Toward a physically based gravity wave source336

parameterization in a general circulation model. J. Atmos. Sci., 67 (1), 136–156.337

Ringler, T. D., and K. H. Cook, 1995: Orographically induced stationary waves: Dependence on338

latitude. J. Atmos. Sci., 52 (14), 2548–2560.339

16

Sandu, I., P. Bechtold, A. Beljaars, A. Bozzo, F. Pithan, T. G. Shepherd, and A. Zadra, 2016:340

Impacts of parameterized orographic drag on the northern hemisphere winter circulation. J Adv341

Model Earth Syst, 8 (1), 196–211.342

Sandu, I., and Coauthors, 2019: Impacts of orography on large-scale atmospheric circulation. npj343

Climate and Atmospheric Science, 2 (1), 10.344

Shaw, T. A., M. Sigmond, T. G. Shepherd, and J. F. Scinocca, 2009: Sensitivity of simulated345

climate to conservation of momentum in gravity wave drag parameterization. J. Clim., 22 (10),346

2726–2742.347

Sigmond, M., and J. F. Scinocca, 2010: The influence of the basic state on the northern hemisphere348

circulation response to climate change. J. Clim., 23 (6), 1434–1446.349

Simpson, I. R., R. Seager, M. Ting, and T. A. Shaw, 2016: Causes of change in northern hemi-350

sphere winter meridional winds and regional hydroclimate. Nat. Clim. Chang., 6 (1), 65–70.351

Sinha, B., A. T. Blaker, J. J.-M. Hirschi, S. Bonham, M. Brand, S. Josey, R. S. Smith, and352

J. Marotzke, 2012: Mountain ranges favour vigorous Atlantic meridional overturning. Geophys.353

Res. Lett., 39 (2), L02 705.354

Smagorinsky, J., 1953: The dynamical influence of large-scale heat sources and sinks on the quasi-355

stationary mean motions of the atmosphere. Q.J Royal Met. Soc., 79 (341), 342–366.356

Taguchi, M., and D. L. Hartmann, 2006: Increased Occurrence of Stratospheric Sudden Warmings357

during El Nino as Simulated by WACCM. J. Clim., 19 (3), 324–332.358

Ting, M., H. Wang, and L. Yu, 2001: Nonlinear stationary wave maintenance and seasonal cycle359

in the GFDL R30 GCM. J. Atmos. Sci., 58 (16), 2331–2354.360

17

Valdes, P. J., and B. J. Hoskins, 1989: Linear stationary wave simulations of the Time-Mean361

climatological flow. J. Atmos. Sci., 46 (16), 2509–2527.362

Valdes, P. J., and B. J. Hoskins, 1991: Nonlinear orographically forced planetary waves. J. Atmos.363

Sci., 48 (18), 2089–2106.364

van Niekerk, A., J. F. Scinocca, and T. G. Shepherd, 2017: The modulation of stationary waves,365

and their response to climate change, by parameterized orographic drag. J. Atmos. Sci., 74 (8),366

2557–2574.367

Warren, B. A., 1983: Why is no deep water formed in the North Pacific? J. Mar. Res., 41 (2),368

327–347.369

Waugh, D. W., and L. M. Polvani, 2010: Stratospheric polar vortices. The Stratosphere: Dynamics,370

Transport, and Chemistry, L. M. Polvani, A. H. Sobel, and D. W. Waugh, Eds., American371

Geophysical Union, 43–57.372

White, R. H., D. S. Battisti, and G. H. Roe, 2017: Mongolian Mountains Matter Most: Impacts of373

the Latitude and Height of Asian Orography on Pacific Wintertime Atmospheric Circulation. J.374

Clim., 30 (11), 4065–4082.375

White, R. H., D. S. Battisti, and A. Sheshadri, 2018: Orography and the Boreal Winter Strato-376

sphere: The Importance of the Mongolian Mountains. Geophys. Res. Lett., 45 (4), 2088–2096.377

18

LIST OF FIGURES378

Fig. 1. Eddy geopotential height, Z′, at 60◦N: a. ERA-I, b. CTL, c. NM, and d. CTL - NM . . . . 20379

Fig. 2. Cross-section of DJF metrics of orographic impact on stationary waves. a. orographic con-380

tribution to stationary wave strength ((σ(ZCT L)−σ(ZNM))/σ(ZCT L)); b. Pearson correla-381

tion coefficient between CTL-NM and NM fields; c. orographic amplification factor (the382

slope of the linear regression between the CTL and NM fields). . . . . . . . . . 21383

Fig. 3. DJF Z, full fields (black contours; zero contour bold) and zonal anomalies (Z′; colours), at384

1000 mb (left), and 300 mb (right). Top row: ERA-I; centre: CTL, and bottom row: NM.385

Prominent highs and lows in the full fields are denoted with H and L respectively. At 300386

mb the contour interval for the full fields for CTL and NM is 200 m, whilst at 1000 mb the387

contour interval is 20 m. . . . . . . . . . . . . . . . . . . . 22388

Fig. 4. Wintertime (DJF for NH, JJA for SH) U (colours), and Z (black contours, contour interval389

200 m for a-c and 100 m for d; negative values dashed and 0 contour bold), for a. ERA-I, b.390

CTL, c. NM, and d. CTL - NM. Wintertime fields are shown for both hemispheres; values391

are masked between 10◦S to 10◦N. . . . . . . . . . . . . . . . . . 23392

Fig. 5. Wintertime (DJF for NH, JJA for SH) stationary wave Eliassen-Palm flux (arrows, m/s2) and393

flux divergence (colours, m/s/day) for a. ERAI, b. CTL and c. NM. The EP flux follows the394

scaling of Edmon et al. (1980) for log-pressure coordinates, and divides by the square root395

of 1000/pressure to aid visualization (Taguchi and Hartmann 2006). Wintertime fields are396

shown for both hemispheres; values are masked between 10◦S to 10◦N. . . . . . . . 24397

Fig. 6. DJF diabatic heating fields for ERA-I (top row), CTL (centre row) and NM (bottom row).398

Left: mass-weighted average between 1000-850 mb; right: mass-weighted average between399

850-500 mb. . . . . . . . . . . . . . . . . . . . . . . . 25400

Fig. 7. As Fig. 2, but for simulations where SSTs are fixed to observed values in both CTL and NM401

experiments. . . . . . . . . . . . . . . . . . . . . . . . 26402

Fig. 8. DJF sea surface temperature for a. CTL and b. CTL-NM differences. CTL values are from a403

CAM simulation coupled to a slab ocean with q-fluxes adjusted to simulate observed SSTs.404

NM values taken from an identical simulation but with flattened orography. . . . . . . 27405

19

0 60E 120E 180 120W 60W1000

500

200

100

50

20

10

Pressure (mb)

a. ERA-I

0 60E 120E 180 120W 60W1000

500

200

100

50

20

10

Pressure (mb)

b. CTL

0 60E 120E 180 120W 60W1000

500

200

100

50

20

10

Pressure (mb)

c. NM

0 60E 120E 180 120W 60W1000

500

200

100

50

20

10

Pressure (mb)

d. CTL-NM

−450 −300 −150 0 150 300 450m

FIG. 1. Eddy geopotential height, Z′, at 60◦N: a. ERA-I, b. CTL, c. NM, and d. CTL - NM.

20

30N 40N 50N 60N 70N 80N1000

500

200

100

50

20

10

pressure (mb)

a.

−1.0−0.8−0.6−0.4−0.20.00.20.40.60.8

30N 40N 50N 60N 70N 80N1000

500

200

100

50

20

10pressure (mb)

b.

−1.0−0.8−0.6−0.4−0.20.00.20.40.60.8

30N 40N 50N 60N 70N 80N1000

500

200

100

50

20

10

pressure (mb)

c.

0.00.20.40.60.81.01.21.41.61.8

FIG. 2. Cross-section of DJF metrics of orographic impact on stationary waves. a. orographic contribution to

stationary wave strength ((σ(ZCT L)−σ(ZNM))/σ(ZCT L)); b. Pearson correlation coefficient between CTL-NM

and NM fields; c. orographic amplification factor (the slope of the linear regression between the CTL and NM

fields).

406

407

408

409

21

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

L

L

H

a.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

L

L

H

b.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

LH

H

H

c.

−150 −100 −50 0 50 100 150

m

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

L

d.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

L

e.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

L

f.

−300 −200 −100 0 100 200 300

m

ERA-I

CTL

NM

1000 mb 300 mb

FIG. 3. DJF Z, full fields (black contours; zero contour bold) and zonal anomalies (Z′; colours), at 1000 mb

(left), and 300 mb (right). Top row: ERA-I; centre: CTL, and bottom row: NM. Prominent highs and lows in

the full fields are denoted with H and L respectively. At 300 mb the contour interval for the full fields for CTL

and NM is 200 m, whilst at 1000 mb the contour interval is 20 m.

410

411

412

413

22

80N60N40N20N020S40S60S80S1000

500

200

100

50

20

10pre

ssure

(m

b)

a. ERA-I

80N60N40N20N020S40S60S80S1000

500

200

100

50

20

10

pre

ssure

(m

b)

b. CTL

80N60N40N20N020S40S60S80S1000

500

200

100

50

20

10

pre

ssure

(m

b)

c. NM

−72

−48

−24

0

24

48

72

m/s

80N60N40N20N020S40S60S80S1000

500

200

100

50

20

10

pre

ssure (mb)

d. CTL - NM

−32

−24

−16

−8

0

8

16

24

32m/s

FIG. 4. Wintertime (DJF for NH, JJA for SH) U (colours), and Z (black contours, contour interval 200 m for

a-c and 100 m for d; negative values dashed and 0 contour bold), for a. ERA-I, b. CTL, c. NM, and d. CTL -

NM. Wintertime fields are shown for both hemispheres; values are masked between 10◦S to 10◦N.

414

415

416

23

5m2/s25m2/s2

80N60N40N20N020S40S60S80S1000

500

200

100

50

20

10Press

ure (mb)

a. ERAI

5m2/s25m2/s2

80N60N40N20N020S40S60S80S1000

500

200

100

50

20

10

Pre

ure (mb)

b. CTL

5m2/s25m2/s2

80N60N40N20N020S40S60S80S1000

500

200

100

50

20

10

Pre

ure (mb)

c. NM

−5.00 −3.75 −2.50 −1.25 0.00 1.25 2.50 3.75 5.00m/ /day

FIG. 5. Wintertime (DJF for NH, JJA for SH) stationary wave Eliassen-Palm flux (arrows, m/s2) and flux

divergence (colours, m/s/day) for a. ERAI, b. CTL and c. NM. The EP flux follows the scaling of Edmon

et al. (1980) for log-pressure coordinates, and divides by the square root of 1000/pressure to aid visualization

(Taguchi and Hartmann 2006). Wintertime fields are shown for both hemispheres; values are masked between

10◦S to 10◦N.

417

418

419

420

421

24

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

a.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

d.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

b.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

e.

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

c.

−6 −3 0 3 6

K/day

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

f.

−3.0 −1.5 0.0 1.5 3.0

K/day

Q1000− 850mb Q500− 850mbERA-I

CTL

NM

FIG. 6. DJF diabatic heating fields for ERA-I (top row), CTL (centre row) and NM (bottom row). Left:

mass-weighted average between 1000-850 mb; right: mass-weighted average between 850-500 mb.

422

423

25

30N 40N 50N 60N 70N 80N1000

500

200

100

50

20

10

pressure (mb)

a.

−1.0−0.8−0.6−0.4−0.20.00.20.40.60.8

30N 40N 50N 60N 70N 80N1000

500

200

100

50

20

10pressure (mb)

b.

−1.0−0.8−0.6−0.4−0.20.00.20.40.60.8

30N 40N 50N 60N 70N 80N1000

500

200

100

50

20

10

pressure (mb)

c.

0.00.20.40.60.81.01.21.41.61.8

FIG. 7. As Fig. 2, but for simulations where SSTs are fixed to observed values in both CTL and NM experiments.

26

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

a.

0

4

8

12

16

20

24

28

030W

60W

90W

120W

150W180

30E

60E

90E

120E

150E

b.

−4.5

−3.0

−1.5

0.0

1.5

3.0

4.5

FIG. 8. DJF sea surface temperature for a. CTL and b. CTL-NM differences. CTL values are from a CAM

simulation coupled to a slab ocean with q-fluxes adjusted to simulate observed SSTs. NM values taken from an

identical simulation but with flattened orography.

424

425

426

27