sensitivity of the itcz location to ocean forcing via q...

Confidential manuscript submitted to replace this text with name of AGU journal

Sensitivity of the ITCZ Location to Ocean Forcing via q-flux Green’s Function 1

Experiments 2

Bryce E. Harrop1, Jian Lu

1, Fukai Liu

2, Oluwayemi Garuba

1, L. Ruby Leung

1 3

1Atmospheric Sciences and Global Change Division, Pacific Northwest National Laboratory, 4

Richland, WA, USA. 5

2Physical Oceanography Laboratory/CIMST, Ocean University of China and Qingdao National 6

Laboratory for Marine Science and Technology, Qingdao, China. 7

8

Corresponding author: Jian Lu ([email protected]) 9

Key Points: 10

The response of the ITCZ position to forcing is asymmetric depending on which 11

hemisphere the forcing is applied in. 12

The asymmetric ITCZ response comes from the nonlinear component of the response, 13

which has a similar magnitude to the linear response. 14

The nonlinear response pattern is insensitive to the location of the forcing, but its 15

magnitude is greatest for high latitude forcings. 16

17

mailto:[email protected])


Abstract 18

The sensitivity of the Intertropical Convergence Zone (ITCZ) position to forcing patterns is 19

important for understanding changes in tropical rainfall. A set of Green’s function experiments 20

reveals that this sensitivity is asymmetric depending on which hemisphere (northern or southern) 21

the forcing is applied. Northern hemisphere forcings produce a much larger response than 22

southern hemisphere forcings of similar magnitude. The response of the ITCZ position to 23

forcing can be broken into linear and nonlinear components, and it is shown that the asymmetry 24

arises from the nonlinear component. The linear and nonlinear response components have 25

similar magnitudes, but the nonlinear component is insensitive to the location of the forcing, 26

such that it amplifies the response to northern hemisphere forcings and dampens the response to 27

southern hemisphere forcings. This asymmetry hints at an intrinsic mode of the climate system 28

such as the ITCZ response to forcing depends on the current climate state. 29

30

Plain Language Summary 31

The tropical rain belt is a key component of the climate system and understanding its response to 32

heating and cooling patterns is a necessary step toward understanding tropical rainfall changes. 33

When the atmosphere is heated or cooled in one hemisphere, flow of energy across the equator in 34

the atmosphere shifts the tropical rain belt. Warming the northern hemisphere and cooling the 35

southern hemisphere produces a much stronger (northward) shift of the rain belt than warming 36

the southern hemisphere and cooling the northern hemisphere. The inclination towards a 37

northward shift of the rain belt, irrespective of the location of heating (or cooling), likely arises 38

from the intrinsic nonlinearity of the current climate. In other words, the asymmetry arises 39

because the rain belt response depends on the state of the current climate, which is asymmetric 40

about the equator because the northern hemisphere receives more energy than the southern 41

hemisphere on annual average. 42

43

1 Introduction 44

The distribution of tropical rainfall is one of the most interesting aspects of the climate 45

system, in part owing to its asymmetry about the equator. The northern hemisphere tropics sees 46

substantially more rainfall annually than the southern hemisphere tropics. The rainfall in the 47

tropics is concentrated in a band known as the intertropical convergence zone (ITCZ). Over the 48

past decade, a strong theoretical framework has been established for understanding the annual 49

mean position of the ITCZ through its relationship to the atmospheric transport of energy. The 50

canonical picture of the ITCZ is one in which precipitation is located at the ascending branch of 51

the Hadley circulation and energy is transported poleward in the upper branch of the circulation 52

(Held 2001), suggesting that the ITCZ ought to lie predominantly in the hemisphere with a net 53

export of energy in the annual mean. In the Earth’s atmosphere, there exists a cross-equatorial 54

flow of energy in the annual mean from the northern hemisphere to the southern hemisphere, 55

which counterbalances a northward flow of energy in the ocean (Kang et al. 2008; Frierson et al. 56

2013; Kang et al. 2015). 57

Changes in the cross-equatorial atmospheric energy transport, referred to as atmospheric 58

heat transport (AHTeq) for simplicity, have been shown to be a useful framework for 59

understanding changes in ITCZ location (Kang et al. 2008, 2009; Frierson and Hwang 2012; 60


Frierson et al. 2013; Donohoe et al. 2013, 2014; Seo et al. 2014; Bischoff and Schneider 2014, 61

2016, Adam et al. 2016, 2017). Observations show that seasonal and interannual variations in 62

the ITCZ position are highly correlated with changes in cross-equatorial atmospheric energy 63

transport (Donohoe et al. 2013, 2014). The observed seasonal relationship between ITCZ 64

position and AHTeq yields a value of -2.7° PW-1

(compared to a value of -2.4° PW-1

in coupled 65

models (Donohoe et al. 2013). The displacement ratio is -2.9° PW-1

for interannual variations 66

(Donohoe et al. 2014). 67

The sensitivity of the ITCZ position to CO2 forcing is model dependent (Merlis et al. 68

2013), and is related to water vapor and cloud feedbacks (Kang et al. 2008; Shaw et al. 2015). 69

Additionally, it has been shown that forcing exerted in the extratropics invokes a larger shift of 70

the ITCZ than similar forcing applied within the tropics, owing at least in part to water vapor and 71

cloud feedbacks (Kang et al. 2009; Seo et al. 2014). Other factors have also been shown to be 72

important for understanding the sensitivity of ITCZ shifts to forcing, including (i) subtropical 73

wind-driven ocean circulation (Green and Marshall 2017) which tends to dampen the shifts in the 74

ITCZ, (ii) changes in solar constant (Smyth et al. 2017) which reduce seasonal migration of the 75

ITCZ because less cross-equatorial energy transport is needed to balance the interhemispheric 76

difference in energy input, and (iii) gross moist stability (Kang et al. 2009; Seo et al. 2017) 77

which relates fluxes of energy to fluxes of mass in the tropical overturning circulation and allows 78

for a decoupling of the relationship between ITCZ position and AHTeq. In short, the shift in the 79

ITCZ in response to changing AHTeq can be modulated by both oceanic responses as well as 80

gross moist stability changes in the atmosphere. 81

The same factors that impact AHTeq can also impact the ratio of heat transported by the 82

atmosphere to that transported by the ocean, known as Bjerknes compensation. Bjerknes 83

compensation can be greater than one where local feedbacks are positive (Liu et al. 2016; Yang 84

et al. 2016) or less than one where feedbacks are negative. Dai et al. (2017) used radiative 85

kernels to separate the role of individual climate feedbacks on Bjerknes compensation and 86

suggested that overcompensation is related to water vapor and cloud feedbacks, while 87

undercompensation is related to the longwave response to temperature increases in the 88

extratropics. 89

Here we build off of the work of (Seo et al. 2014) to examine the response of the ITCZ to 90

forcing perturbations at different latitudes taken from a series of Green’s function experiments. 91

The Green’s function approach allows us to systematically probe the atmospheric response to 92

forcing perturbations across the globe. The model experiments and methods used in this study 93

are outlined in section 2, results and discussion are presented in section 3, and a summary 94

statement is provided in section 4. 95

2 Model experiments and methods 96

The model used here is the slab ocean version of the Community Earth System Model 97

version 1.1 (CESM1.1-SOM). The model uses the Community Atmosphere Model version 5 98

(CAM5), the Community Land Model version 4 (CLM4), and the Community Ice Code (CICE) 99

components, as well as a motionless slab ocean mixed layer model. The use of a slab ocean 100

model means that the sea surface temperature (SST) in the model is prognostic, with its value 101

determined by the balance of ocean heat transport, radiant energy, and turbulent heat fluxes (both 102

sensible and latent) at the surface. Ocean heat transport and mixed layer depth are prescribed 103


based on a fully coupled CESM1.1 simulation. Ocean heat transport is allowed to vary in space 104

and has a repeating seasonal cycle, while mixed layer depth is only allowed to vary in space. 105

Horizontal grid spacing for CAM5 and CLM4 is 2.5°×1.9°, while for CICE and the slab ocean 106

model, horizontal grid spacing telescopes meridionally from 1° at high latitudes to ~0.3° at the 107

equator. 108

The Green’s function experiments are a series of branch runs from a 900-year control 109

simulation of CESM1.1-SOM run by the National Center for Atmospheric Research (NCAR), 110

forced with preindustrial greenhouse gases, aerosols, and solar insolation. For each experiment, a 111

q-flux perturbation is added or subtracted from the climatological q-flux profile over a limited 112

patch within the domain. In total, 97 patches are used for the q-flux perturbations. For each 113

patch, a positive and negative anomaly is added as a hump following (Barsugli and Sardeshmukh 114

2002). 115

2 2cos cos , for and2 2

othe0 r ise, w

k kw k w w k w

w w

Q

116

where Q is the magnitude of the q-flux perturbation at the patch center (λk, φk), set to be 12 Wm-

117 2. The patch half-widths are 30° in the zonal direction (λw) and 12° in the meridional direction 118

(φw). Summing all 97 patches together gives a uniform 12 Wm-2

forcing over the global ocean, 119

except in the vicinity of the coasts or sea ice edge. 120

Each Green’s function experiment is integrated for 40 years, with the first 20 years 121

discarded as an adjustment period. Both the linear and non-linear parts of the response are 122

analyzed herein. The linear response is calculated as the response Rlinear = (R+

– R-)/2ΔQ, with 123

the superscript + (-) denoting a positive (negative) q-flux perturbation, and ΔQ being the total 124

forcing in PW from the q-flux area. Positive q-flux implies a heat source for the atmosphere-slab 125

coupled system. The nonlinear response is estimated as the response Rnonlinear = (R+ – 2R

0 + R

-126

)/2ΔQ, where superscript 0 denotes the control. Taking R to be the response anomaly from the 127

control implies R0 is zero, and hence the nonlinear response simplifies to Rnonlinear = (R

+ + R

-128

)/2ΔQ. The location of the ITCZ is measured as the latitude centroid of precipitation: 129

2

1

2

1

ITCZ

cos d

cos d

P

P

130

where φ1 =20°S and φ2 =20°N are the latitude bounds for the integration. AHTeq is estimated as 131

the average from 10°S-10°N. For calculation of AHT, the global mean value is removed from 132

the fluxes, such that AHT is zero at both poles. The global mean value of the net flux of energy 133

into the atmosphere is small and removing its value from the calculation of AHT does not impact 134

our conclusions. 135

3 Results 136


Figure 1 shows the AHT response to OHT at various latitudes, both for south-to-north 137

OHT and north-to-south OHT. It is immediately apparent that there is an asymmetry in the 138

atmospheric response depending on the direction of the flow of energy within the ocean. For 139

OHT that is oriented south-to-north, the atmosphere has a clear and strong return flow of energy 140

that is north-to-south, as expected based on previous studies (Seo et al. 2014). Like the results of 141

Seo et al. (2014), Figure 1 shows that high latitude forcing generates a larger AHT response than 142

low latitude forcing. Our results, however, differ from those of Seo et al. (2014) in two key 143

ways. First, the compensation by the atmosphere for south-to-north OHT is much larger (~400% 144

for mid-latitude forcing compared to ~140% in Seo et al. (2014). Second, we performed 145

experiments with north-to-south OHT, whereas Seo et al. (2014) only did south-to-north OHT 146

experiments. The north-to-south OHT experiments show a very weak and even negative 147

compensation (the AHT is in the same north-to-south direction as the OHT) for some cases at the 148

equator, revealing a key asymmetry simulated by the model. 149

What gives rise to the asymmetry shown in Figure 1? Figure 2 shows the breakdown of 150

the AHT into the feedback mechanisms responsible for the patterns shown in Figure 1. The 151

anomalous AHT from any given feedback is computed as 152

2 /2 22 2

/2 0 /2 0TOA SFC TOA SFC

AHT ( ) cos d d cos d dX

R R R Ra X a X

X X X X

153

154

where a is the radius of the earth, R is the radiative flux response, X is the feedback parameter of 155

interest, and 𝝏𝑹

𝝏𝑿 is the radiative kernel developed by Huang et al. (2017), which is calculated 156

using the rapid radiative transfer model of Mlawer et al. (1997) based on 6-hourly atmospheric 157

profiles from the ERA-Interim reanalysis (Dee et al. 2011). The global mean radiative flux 158

change is subtracted off such that the anomalous AHT goes to zero at both poles. All of the 159

individual feedbacks are important for understanding the total AHT response, except for the 160

albedo feedback. It is not to say that the albedo feedback is not important for modulating AHT, 161

rather that albedo only influences AHT indirectly through changes in the surface energy budget. 162

For the water vapor, temperature, cloud, and turbulent heat flux (THFLX; sensible plus latent 163

heat fluxes) components, the response is asymmetric between the different directions of OHT. In 164

all of the feedbacks, the magnitude of the response is stronger in the south-to-north OHT 165

experiments and weaker in the north-to-south OHT experiments. 166

The climatological ITCZ is in the northern hemisphere; as such, an intensification of that 167

pattern leads to stronger southward AHTeq. In the case of south-to-north OHT, this 168

intensification of southward AHTeq is expected, and the results confirm it. For north-to-south 169

OHT, one might expect the opposite, but our results show that AHTeq is still southward (albeit 170

much weaker than the south-to-north OHT cases). The southward AHTeq results from the water 171

vapor and cloud feedbacks (Figure 2). The structure of the cloud and water vapor feedback 172

responses, like the total AHT response, suggests the ITCZ is not shifting southward for the 173

north-to-south OHT experiments, but is instead shifting northward. 174


To further probe the asymmetric response of the atmosphere to heating and cooling 175

patterns, the shift in the ITCZ is separated into linear and nonlinear response components. For 176

this analysis, we focus on forcing at a single latitude instead of the forcing pairs examined above. 177

For example, the linear response is taken as the difference in positive and negative q-flux 178

perturbations at a single latitude. Figure 3a shows the linear response of the ITCZ for a forcing 179

at each latitude. For the linear response, the ITCZ shifts as expected toward the hemisphere 180

where the heating is applied (or away from in the case of cooling), with a strong correlation 181

(r2=0.98) between the ITCZ shift and AHTeq (here measured as the average energy flux between 182

10°S-10°N). Also as expected, the magnitude of the shift tends to be larger for higher-latitude 183

forcings than for lower-latitude forcings. Forcing right at the equator produces no shift of the 184

ITCZ. The slope of the response of the ITCZ shift to AHTeq is -3.8 ± 0.4° PW-1

. In observations 185

of the seasonal progression of the ITCZ, the slope of the ITCZ shift to AHTeq is only -2.7 ± 0.6° 186

PW-1

(Donohoe et al. 2013). As noted by Donohoe et al. (2013), the slope of the seasonal ITCZ 187

shift to AHTeq value varies substantially with mixed layer depth in models (from -2.0° PW-1

for a 188

2.4 m mixed layer depth to -6.1° PW-1

for a 50 m mixed layer depth). The -3.8° PW-1

value 189

retrieved from our experiments is well within that spread, and agrees with the findings of 190

Donohoe et al. (2013). For reference, the slab ocean mixed layer depth is variable in space for 191

our experiments, generally ranging from 6 m to 100 m (with a limited region of the North 192

Atlantic having a mixed layer depth of approximately 300 m). 193

While the linear response of the ITCZ position to forcing behaves as expected, the 194

nonlinear response is distinct. For the nonlinear response, shown in Figure 3b, the ITCZ shifts 195

northward regardless of which hemisphere the q-flux forcing is applied to. The net response 196

(linear plus nonlinear; Figure 3c) shows a northward shift of the ITCZ for all experiments except 197

the southern hemisphere high latitude forcing (55°S). The other four southern hemisphere 198

forcing bands (44°S-10°S) all have a net response of a northward ITCZ shift, despite the surface 199

heating being applied in the southern hemisphere. Furthermore, the sign of the q-flux forcing 200

does not matter for the nonlinear response by construction (either sign produces a northward shift 201

of the ITCZ). The nonlinear response magnitude is larger for high latitude forcings than low 202

latitude forcings. The forcing at 57°N is especially large, potentially as a result of the large 203

mixed layer depths encountered in the North Atlantic. Future work is needed to confirm or reject 204

this hypothesis. 205

The nonlinear response of the ITCZ suggests one or more modes of variability within the 206

control climate is in a local minimum, such that any forcing will create the same response as the 207

response of the system to natural fluctuations. The response is not sensitive to location either. 208

Even the equatorial q-flux forcing produces a northward ITCZ shift (see Figure 3b). The pattern 209

of the nonlinear component of surface temperature response is complex (unlike the linear 210

component which is far more zonally uniform, not shown) and looks like a combination of a 211

negative Interdecadal Pacific Oscillation (IPO) pattern, a cooling over the West Pacific warm 212

pool, a roughly zonally uniform southern ocean cooling, and a wavenumber one temperature 213

pattern in the arctic (Figure 4a). This complicated surface temperature structure gives rise to a 214

wavenumber 2 pattern in water vapor and precipitation across the tropics that is superimposed on 215

top of the zonal mean northward shift of those two moisture fields associated with the northward 216

shift of the ITCZ. The response of the longwave cloud radiative effect (LWCRE) measured at 217

the top of the atmosphere resembles the water vapor and precipitation response. The shortwave 218

cloud radiative effect (SWCRE) does as well, except over the stratocumulus regions, where 219


cooling surface temperatures lead to decreasing SWCRE. The THFLX (latent plus sensible heat 220

flux) is positively correlated with SWCRE over oceans, suggesting the increase (decrease) of 221

solar forcing at the surface is partly offset by increasing (decreasing) THFLX into the 222

atmosphere. 223

While the magnitude of the nonlinear AHT response depends on the latitude of the 224

forcing (high latitude forcings produce a larger response than low latitude forcings, as expected), 225

it is remarkable that the zonal mean spatial pattern of the nonlinear AHT response is about the 226

same for all of the forcing locations (see Figure 5). Like the total nonlinear AHT response, the 227

individual feedbacks also only vary by magnitude between forcing locations (the spatial patterns 228

are the same). The nonlinear AHT response acts to converge energy into the SH 229

tropics/subtropics (anomalous northward transport south of ~20°S and anomalous southward 230

transport north of ~20°S). Different feedbacks dominate these anomalies at different latitudes. 231

The southward AHTeq in the nonlinear response is the culmination of forcings at both 232

high and low latitudes. At global scales, the temperature feedback induces a northward flow of 233

energy in the atmosphere, while the turbulent heat flux (THFLX; sensible plus latent heat fluxes) 234

induces a southward flow of energy in the atmosphere with a relative minimum near the equator. 235

Both of these broad features can be understood in terms of the asymmetric temperature response 236

seen in Figure 4a. The southern hemisphere cools relative to the northern hemisphere which 237

weakens both radiative cooling of the atmosphere in the SH relative to the NH (ignoring 238

emissivity changes) and the THFLX into the SH relative to the NH. On top of these responses, 239

large changes to the energy fluxes in the tropics occur owing to the response of water vapor, 240

clouds, and THFLX. The tropical changes are all consistent with a northward shift of the ITCZ, 241

though it is impossible to determine causality in this simple diagnostic framework. 242

The water vapor and cloud feedback response of AHT is consistent with energy 243

convergence between 15°S-0°, and energy divergence between 0°-20°N. The response of AHT 244

to THFLX is the exact opposite (energy divergence 15°S-0°, energy convergence 0°-20°N). The 245

response of AHT to clouds comes from changes in the atmospheric cloud radiative effect. In the 246

tropics, the atmospheric cloud radiative effect is dominated by LWCRE (Slingo and Slingo 247

1988). Thus, the cloud feedback portion of the AHT response results from changing high cloud 248

patterns at low latitudes. Figure 4f shows the northward shift of tropical high clouds, consistent 249

with the ITCZ shift, and the resulting heating/cooling dipole about the equator. The AHT 250

response to water vapor is similar to that of clouds at low latitudes. Changes in water vapor can 251

impact radiative cooling in multiple ways. The increase in column water vapor generally 252

increases atmospheric cooling as a result of increased downwelling longwave radiation 253

(Pendergrass and Hartmann 2013). The vertical redistribution of water vapor by convection, 254

however, decreases atmospheric cooling as a result of reduced outgoing longwave radiation 255

(Harrop and Hartmann 2015). Given the similarity of the water vapor and cloud responses in the 256

tropics in Figure 5, it appears that the convective redistribution of water vapor is the bigger 257

factor in the nonlinear response. 258

Figure 4c shows that THFLX is negatively correlated with precipitation over a large 259

portion of the tropics, meaning THFLX increases where precipitation decreases. The correlation 260

between precipitation and THFLX suggests that the THFLX changes are directly linked to the 261

precipitation changes, though it is unclear whether that is because of changes in boundary layer 262


humidity, circulation, or other factors. Zonal mean easterlies increase near the surface at, and 263

just south, of the equator, with a decrease in the easterlies just north of the equator (not shown). 264

These changes in zonal wind imply increasing THFLX south of the equator and decreasing 265

THFLX north of the equator, which Figure 4 and Figure 5 confirm. 266

It is also important to consider that the relative importance of each feedback process 267

tends to be model dependent. RCP8.5 simulations show consistent responses of AHTeq and 268

ITCZ shift for similar feedbacks, but the magnitudes vary dramatically (McFarlane and Frierson 269

2017). Also, OHT responses counterbalance some of the AHTeq response, further reducing 270

model agreement on the response of the climate to forcing perturbations. Our results show that 271

while the magnitude of the nonlinear response may be sensitive to forcing location, the pattern is 272

not. We caution that the spatial pattern of the total nonlinear AHT response may vary across 273

models. Despite the potential model-dependency of the nonlinear spatial pattern, its robustness 274

to various forcings within the same model means it can serve as a fingerprint for understanding 275

the atmospheric response to external or oceanic forcings. 276

4 Concluding Remarks 277

The results of this study reveal a key asymmetry of the tropical precipitation response to 278

forcing within the simulated climate system. The linear response of AHT to a forcing 279

perturbation shifts the ITCZ toward the hemisphere in which the heating is applied. The 280

nonlinear response, however, acts to shift the ITCZ northward, regardless of the sign of the 281

forcing or in which hemisphere it is applied. The sensitivity of the ITCZ to high latitude 282

forcings, for both the linear and nonlinear responses, is greater than the corresponding responses 283

to low latitude forcings, partly attributable to the amplifying effects from the cloud and water 284

vapor feedbacks (consistent with previous work). Intriguingly, the spatial pattern of the 285

nonlinear response is the same regardless of the location of the forcing, hinting at an intrinsic 286

mode of the climate system such that the response of the ITCZ to any forcing will be similar to 287

the response of the system in its current mean state to natural fluctuations. 288

Future effort is needed to understand the physical processes responsible for the nonlinear 289

response and their robustness across different climate models. Additional work is needed to 290

determine the role of the ocean in modulating the nonlinear response in a coupled model with a 291

dynamic ocean. The land-sea geometry may also prove to be an important factor in shaping the 292

pattern of the nonlinear response. Nevertheless, the similarity in magnitude of the linear and 293

nonlinear responses of the ITCZ to forcing underlines the importance of the nonlinear response 294

and the inherent nonlinearity of the climate system, which has not been a focus of climate 295

sensitivity research. Last, efforts are also needed to determine whether the patterns uncovered in 296

our modeling analysis may be found in observations. 297

Acknowledgments, Samples, and Data 298

This study is supported by the U.S. Department of Energy Office of Science Biological and 299

Environmental Research (BER) as part of the Regional and Global Climate Modeling Program. 300

We acknowledge the use of computational resources of the National Energy Research Scientific 301

Computing Center, a DOE Office of Science User Facility supported by the Office of Science of 302

the U.S. Department of Energy under Contract No. DE-AC02-05CH11231. All the model data 303


output has been archived on NERSC High Performance Storage System (HPSS) and can be 304

accessed via the following URL (TBD). The Pacific Northwest National Laboratory is operated 305

for the Department of Energy by Battelle Memorial Institute under contract DE-AC05-306

76RL01830. 307

References 308

Adam, O., T. Bischoff, and T. Schneider, 2016: Seasonal and Interannual Variations of the 309

Energy Flux Equator and ITCZ. Part I: Zonally Averaged ITCZ Position. J. Clim., 29, 310

3219–3230, doi:10.1175/JCLI-D-15-0512.1. http://journals.ametsoc.org/doi/10.1175/JCLI-311

D-15-0512.1. 312

——, T. Schneider, and F. Brient, 2017: Regional and seasonal variations of the double-ITCZ 313

bias in CMIP5 models. Clim. Dyn., 0, 0, doi:10.1007/s00382-017-3909-1. 314

http://dx.doi.org/10.1007/s00382-017-3909-1. 315

Barsugli, J. J., and P. D. Sardeshmukh, 2002: Global atmospheric sensitivity to tropical SST 316

anomalies throughout the Indo-Pacific basin. J. Clim., 15, 3427–3442, doi:10.1175/1520-317

0442(2002)015<3427:GASTTS>2.0.CO;2. 318

Bischoff, T., and T. Schneider, 2014: Energetic constraints on the position of the intertropical 319

convergence zone. J. Clim., 27, 4937–4951, doi:10.1175/JCLI_D_13_00650.1. 320

——, and ——, 2016: The Equatorial Energy Balance, ITCZ Position, and Double-ITCZ 321

Bifurcations. J. Clim., 29, 2997–3013, doi:10.1175/JCLI-D-15-0328.1. 322

http://journals.ametsoc.org/doi/10.1175/JCLI-D-15-0328.1. 323

Dai, H., H. Yang, and J. Yin, 2017: Roles of energy conservation and climate feedback in 324

Bjerknes compensation: a coupled modeling study. Clim. Dyn., 49, 1513–1529, 325

doi:10.1007/s00382-016-3386-y. 326

Dee, D. P., and Coauthors, 2011: The ERA-Interim reanalysis: configuration and performance of 327

the data assimilation system. Q. J. R. Meteorol. Soc., 137, 553–597, doi:10.1002/qj.828. 328

http://doi.wiley.com/10.1002/qj.828 (Accessed April 28, 2014). 329

Donohoe, A., J. Marshall, D. Ferreira, and D. Mcgee, 2013: The relationship between ITCZ 330

location and cross-equatorial atmospheric heat transport: From the seasonal cycle to the last 331

glacial maximum. J. Clim., 26, 3597–3618, doi:10.1175/JCLI-D-12-00467.1. 332

——, ——, ——, K. Armour, and D. McGee, 2014: The Interannual Variability of Tropical 333

Precipitation and Interhemispheric Energy Transport. J. Clim., 27, 3377–3392, 334

doi:10.1175/JCLI-D-13-00499.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-335

00499.1 (Accessed April 24, 2014). 336

Frierson, D. M. W., and Y.-T. Hwang, 2012: Extratropical Influence on ITCZ Shifts in Slab 337

Ocean Simulations of Global Warming. J. Clim., 25, 720–733, doi:10.1175/JCLI-D-11-338

00116.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-11-00116.1. 339


Frierson, D. M. W., and Coauthors, 2013: Contribution of ocean overturning circulation to 340

tropical rainfall peak in the Northern Hemisphere. Nat. Geosci., 6, 940–944, 341

doi:10.1038/ngeo1987. http://www.nature.com/doifinder/10.1038/ngeo1987. 342

Green, B., and J. Marshall, 2017: Coupling of trade winds with ocean circulation damps itcz 343

shifts. J. Clim., 30, 4395–4411, doi:10.1175/JCLI-D-16-0818.1. 344

Harrop, B. E., and D. L. Hartmann, 2015: The Relationship between Atmospheric Convective 345

Radiative Effect and Net Energy Transport in the Tropical Warm Pool. J. Clim., 28, 8620–346

8633, doi:10.1175/JCLI-D-15-0151.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-347

15-0151.1. 348

Held, I., 2001: The partitioning of the poleward energy transport between the tropical ocean and 349

atmosphere. J. Atmos. Sci., 58, 943–948. http://journals.ametsoc.org/doi/abs/10.1175/1520-350

0469(2001)058%3C0943:TPOTPE%3E2.0.CO;2 (Accessed August 16, 2013). 351

Huang, Y., Y. Xia, and X. Tan, 2017: On the pattern of CO2 radiative forcing and poleward 352

energy transport. J. Geophys. Res. Atmos., 122, 10,578-10,593, doi:10.1002/2017JD027221. 353

Kang, S. M., I. M. Held, D. M. W. Frierson, and M. Zhao, 2008: The response of the ITCZ to 354

extratropical thermal forcing: Idealized slab-ocean experiments with a GCM. J. Clim., 21, 355

3521–3532, doi:10.1175/2007JCLI2146.1. 356

——, D. M. W. Frierson, and I. M. Held, 2009: The Tropical Response to Extratropical Thermal 357

Forcing in an Idealized GCM: The Importance of Radiative Feedbacks and Convective 358

Parameterization. J. Atmos. Sci., 66, 2812–2827, doi:10.1175/2009JAS2924.1. 359

http://journals.ametsoc.org/doi/abs/10.1175/2009JAS2924.1. 360

——, R. Seager, D. M. W. Frierson, and X. Liu, 2015: Croll revisited: Why is the northern 361

hemisphere warmer than the southern hemisphere? Clim. Dyn., 44, 1457–1472, 362

doi:10.1007/s00382-014-2147-z. 363

Liu, Z., H. Yang, C. He, and Y. Zhao, 2016: A Theory for Bjerknes Compensation: The Role of 364

Climate Feedback. J. Clim., 29, 191–208, doi:10.1175/JCLI-D-15-0227.1. 365

http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-15-0227.1. 366

McFarlane, A. A., and D. M. W. Frierson, 2017: The role of ocean fluxes and radiative forcings 367

in determining tropical rainfall shifts in RCP8.5 simulations. Geophys. Res. Lett., 44, 8656–368

8664, doi:10.1002/2017GL074473. 369

Merlis, T. M., M. Zhao, and I. M. Held, 2013: The sensitivity of hurricane frequency to ITCZ 370

changes and radiatively forced warming in aquaplanet simulations. Geophys. Res. Lett., 40, 371

4109–4114, doi:10.1002/grl.50680. 372

Mlawer, E. J., S. J. Taubman, P. D. Brown, M. J. Iacono, and S. A. Clough, 1997: Radiative 373

transfer for inhomogeneous atmospheres: RRTM, a validated correlated-k model for the 374

longwave. J. Geophys. Res., 102, 16663–16682, doi:10.1029/97JD00237. 375

http://www.agu.org/pubs/crossref/1997/97JD00237.shtml. 376


Pendergrass, A. G., and D. L. Hartmann, 2013: The atmospheric energy constraint on global-377

mean precipitation change. J. Clim., 2, 130916120136005, doi:10.1175/JCLI-D-13-00163.1. 378

http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-00163.1 (Accessed January 20, 379

2014). 380

Seo, J., S. M. Kang, and D. M. W. Frierson, 2014: Sensitivity of Intertropical Convergence Zone 381

Movement to the Latitudinal Position of Thermal Forcing. J. Clim., 27, 3035–3042, 382

doi:10.1175/JCLI-D-13-00691.1. http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-383

00691.1. 384

——, ——, and T. M. Merlis, 2017: A model intercomparison of the tropical precipitation 385

response to a CO 2 doubling in aquaplanet simulations. Geophys. Res. Lett., 44, 993–1000, 386

doi:10.1002/2016GL072347. http://doi.wiley.com/10.1002/2016GL072347. 387

Shaw, T. A., A. Voigt, S. M. Kang, and J. Seo, 2015: Response of the intertropical convergence 388

zone to zonally asymmetric subtropical surface forcings. Geophys. Res. Lett., 42, 9961–389

9969, doi:10.1002/2015GL066027. http://doi.wiley.com/10.1002/2015GL066027. 390

Slingo, A., and J. M. Slingo, 1988: The response of a general circulation model to cloud 391

longwave radiative forcing. I: Introduction and initial experiments. Q. J. R. Meteorol. Soc., 392

114, 1027–1062, doi:10.1002/qj.49711448209. 393

http://doi.wiley.com/10.1002/qj.49711448209. 394

Smyth, J. E., R. D. Russotto, and T. Storelvmo, 2017: Thermodynamic and dynamic responses of 395

the hydrological cycle to solar dimming. Atmos. Chem. Phys., 17, 6439–6453, 396

doi:10.5194/acp-17-6439-2017. https://www.atmos-chem-phys.net/17/6439/2017/. 397

Yang, H., Y. Zhao, and Z. Liu, 2016: Understanding Bjerknes Compensation in Atmosphere and 398

Ocean Heat Transports Using a Coupled Box Model. J. Clim., 29, 2145–2160, 399

doi:10.1175/JCLI-D-15-0281.1. http://journals.ametsoc.org/doi/10.1175/JCLI-D-15-0281.1. 400

401

402


403 Figure 1. (a) The implied OHT needed to balance the Green’s function heating/cooling 404

perturbations; (b) the AHT response; and (c) the compensation ratio (AHT/OHT). Colors denote 405

the latitude at which the forcing is applied ( 50N 50S 50S 50N;50N R R S0 R R5 ). 406


407 Figure 2. The breakdown of the AHT (left) and compensation (right) for the various components 408

including the water vapor feedback, temperature feedback (both Planck and lapse rate), albedo 409

feedback, cloud feedback, and the change in total heat flux (surface sensible plus latent). Colors 410

denote the latitude at which the forcing is applied ( 50N 50S 50S 50N;50N R R S0 R R5 ). 411

412

413

414

Figure 3. (a) The shift of the ITCZ for the linear response (415

50N 50S 50S 50N;50N R R S0 R R5 ). (b) The shift of the ITCZ for the nonlinear response (416

50N 50S 50S 50N;50N R R S0 R R5 ). (c) The shift of the ITCZ for the sum of the linear and 417

nonlinear responses. Colors denote the latitude at which the forcing is applied. 418


419 Figure 4: Maps of nonlinear response for (a) surface temperature, (b) precipitable water, (c) total 420

(latent + sensible) heat flux (d) total precipitation rate, (e) shortwave cloud radiative effect, and 421

(f) longwave cloud radiative effect, and (f) combined sensible and latent heat flux (positive 422

downward). For SWCRE and LWCRE, both are computed at top-of-atmosphere. 423

424


425

426

427 Figure 5. The AHT anomalies from the nonlinear response for (a) total, (b) water vapor 428

feedback, (c) temperature feedback (lapse rate + Planck), (d) albedo feedback, (e) cloud 429

feedback, (f) latent + sensible heat feedback. Colors denote the latitude at which the forcing is 430

applied ( 50N 50S 50S 50N;50N R R S0 R R5 ). 431

sensitivity of the itcz location to ocean forcing via q...

Documents