1

Short-term modulation of Indian summer monsoon rainfall by 1

West Asian dust 2

3

4 V Vinoj1,2, Philip J Rasch1*, Hailong Wang1, Jin-Ho Yoon1, Po-Lun Ma1, Kiranmayi 5

Landu1 and Balwinder Singh1 6 7

1 Atmospheric Sciences and Global Change Division 8

Pacific Northwest National Laboratory 9

Richland, WA - 99352, USA 10

11 2 School of Earth, Ocean and Climate Sciences 12

Indian Institute of Technology Bhubaneswar 13

Odisha, INDIA 14

15

Email correspondence to philip.rasch@pnnl.gov* 16

17

18

2

Abstract 19

The Indian summer monsoon results from a complex interplay between radiative heating, 20

dynamics, and cloud-aerosol interactions. Despite increased scientific attention, the effect 21

of aerosols on monsoons still remains uncertain. Previous studies (using observational 22

and modeling approaches) have focused primarily on local climate effects on seasonal 23

timescales, but shorter-term and non-local links have not previously been explored or 24

identified. Here we present observational evidence and numerical modeling results to 25

demonstrate a link between aerosols over the Arabian Sea and West Asia and Indian 26

summer monsoon rainfall. Simulations using a state-of-the-art global climate model 27

support this remote link and indicate that variability in dust aerosol loadings influence 28

radiative heating rates that can induce larger scale circulation changes, modulating 29

moisture transport and convergence over Central India, and change monsoon rainfall on 30

relatively short time scales (about a week). Our simulations suggest that dust induced 31

solar heating over North Africa and West Asia increases low-level winds over the 32

Arabian Sea changing the moisture flowing into India. Our findings highlight the 33

importance of natural aerosols in influencing the strength of the Indian summer monsoon, 34

motivating additional research in how changes in background aerosols of natural origin 35

may be influencing monsoon precipitation. 36

37

38

39

40

Key Words: Dust, sea-salt, natural aerosol, Indian monsoon, climate change 41

3

1. Introduction 42

The inter-annual variability of the Indian summer monsoon rainfall is only about 10% 43

of the long term mean 1,2; but these relatively small variations have important impacts on 44

mankind and natural systems in India. A number of factors contribute to this variability, 45

including aerosols that act directly (via solar radiation) and indirectly (through cloud 46

microphysics) in influencing/modulating the monsoon strength. The role of aerosols in 47

modulating monsoon rainfall over India has gained increased scientific attention in the 48

recent past3-12. 49

Many recent studies have focused on anthropogenic and/or biomass burning aerosol 50

influences3,7,11-13. However, the high winds over Arabian Sea associated with the 51

monsoon not only act as a conduit for moisture from oceans, but also act as a source and 52

conduit of natural aerosols of marine (sea-salt) and continental (dust, mobilized by 53

Shamal winds over Arabia) origins that depend on both wind speed and direction 14-16. 54

These natural aerosols dominate the aerosol mass loading over the Arabian Sea and as we 55

show below, likely influence the monsoon precipitation on short time scales. 56

Observational studies exploring pre-monsoon (April/May) aerosol loading and its 57

effect on monsoon precipitation have found that maxima in aerosols lead precipitation6,17-58

19 by a month or more. There is observational evidence for an elevated heat pump 6 59

(EHP) associated with a warming by dust and black carbon aerosol layers, that 60

accumulates in the southern foothills of the Himalayas during the pre-monsoon period. 61

The hypothesis is that this elevated heat source can advance the monsoon circulation and 62

also strengthen rainfall during subsequent months, which has been a subject of scientific 63

debate 18. It has been challenging to obtain concomitant measurements of both aerosols 64

4

and precipitation at such large spatial scales (due to limitations in measurements and also 65

their interactive nature) to explore aerosol/rainfall interactions 20, and we know of no 66

existing observational studies that have used concomitant aerosol and precipitation 67

measurements on seasonal and shorter timescales. Aerosols can influence monsoons on 68

multiple timescales: aerosol absorption and scattering can influence heating rates and 69

stability in the atmospheric column and drive a relatively rapid monsoon response, but 70

the reduction in sunlight reaching the surface can also affect the monsoon more slowly. 71

The higher heat capacity of the surface leads to slower response in sea surface 72

temperatures over the northern Indian Ocean producing a reduction of north-south 73

temperature gradient with very important impacts on the monsoon11,21. Many climate 74

modeling studies have focused on aerosol–monsoon precipitation interactions3-8,10,11; 75

however, most of those concentrated on the slower forcing and response and the impact 76

of changes in anthropogenic aerosol emissions. Only a few studies have focused 77

specifically on natural aerosols 5 or variations that occur on the faster sub-seasonal time 78

scales. 79

In this paper we present evidence for the existence of a positive relationship between 80

natural (mostly dust) aerosol and monsoon precipitation using temporally collocated but 81

spatially separated satellite based aerosol and precipitation fields. The fields are 82

correlated at short timescales (within a week). This observational relationship is also 83

reproduced in simulations using an atmospheric global climate model (AGCM), which 84

further suggest that dust aerosols over the North Africa, West Asia and Arabian region 85

(operating through the fast forcing and response mechanism described above) modulate 86

the summer monsoon precipitation over India. 87

5

88

2. Aerosol-Monsoon Precipitation Relationships 89

We analyze anomalies (relative to monthly means) of weekly precipitation over 90

central India (averaged over 16.5-26.5oN, 74.5-86.6oE box shown in Fig. 1a selected to 91

be similar to an All India Rainfall index used in reference 22) and weekly aerosol optical 92

depth (AOD) at each grid cell over the domain covering the Indian sub-continent during 93

the monsoon season (June to August). We use the global precipitation climatology 94

project (GPCP) estimate for precipitation and AOD retrievals from the MODerate 95

resolution Imaging Spectroradiometer (MODIS) and Multi-angle Imaging 96

Spectroradiometer (MISR) (more details on the observational data, model and 97

methodology are provided in the supplementary materials). Figs. 1a and 1b show the 98

correlation coefficient between Central India monsoon precipitation and AOD at each 99

grid point over the Indian sub-continent based on satellite measurements (MODIS and 100

MISR AODs). Fig. 1c shows the simulated relationship using the Community 101

Atmosphere Model (CAM5) operating in “offline mode”, where the temperature, winds, 102

and air mass (pressure) fields (called ‘meteorology’ hereafter) are prescribed by a re-103

analysis product23 (ERA Interim reanalysis for 2000-2009), and the aerosols and water 104

variables evolve freely from the forcing provided by that meteorology (see 105

supplementary material for details). In all three panels, the central Indian rainfall is 106

strongly correlated with AOD over the Arabian Sea, West Asia and Saudi peninsular 107

regions, but they have low or even negative correlations over the Eastern India and Bay 108

of Bengal. A lead-lag analysis of weekly aerosols and precipitation variation indicates 109

6

that correlations are strongest (~0.6) for a zero-lag (Fig S3), indicating that aerosol and 110

precipitation vary in concert. 111

It is important to understand the source of these correlations. Although there are 112

clear relationships between pre-monsoon aerosol forcing, and monsoon precipitation 6,17, 113

the lifetime of aerosols is very short (about a week) near the Earth’s surface, and the 114

heavy precipitation during the monsoon effectively cleanses the atmosphere locally, so 115

the mechanisms that connect aerosol and precipitation on longer seasonal time scales are 116

probably not same as those evident on shorter timescales. In this study we focus on the 117

shorter term aerosol/precipitation relations identified earlier, attempting to tease out the 118

physical mechanism for the relationship, and the important aerosol types that contribute 119

to it. Hints about the type of aerosols responsible for the observed relationship (shown in 120

Fig. 1) are evident using the MODIS Angstrom Exponent (AE) product, an indicator of 121

aerosol size (with lower values indicating dominance of larger sized particles). A similar 122

pattern in AE of opposite sign (Fig. S2) suggests that coarse-mode particles dominated by 123

sea-salt and/or dust are responsible for the positive correlations shown in Fig. 1. A 124

similar analysis using a different MODIS product (i.e., fine-mode aerosol fraction) 125

suggests no significant positive correlation between central-India precipitation and 126

anthropogenic fine-mode aerosols over the Arabian Sea (as in Fig S2) and thus reinforces 127

the above finding. It is difficult to discriminate between dust and sea-salt aerosol induced 128

correlations using observations alone because both sources are sensitive to wind speed 129

and contribute to the coarse size mode. We therefore resort to climate model simulations 130

to separate the effects of different aerosols and expose the underlying physical 131

mechanisms. 132

7

133

3. Model Results and Discussion 134

While models are imperfect, and may have difficulty in simulating all aspects of 135

these correlations (especially on smaller spatial scales) they can provide significant 136

insight into the working of the complex climate system. Various configurations of the 137

CAM5 global climate model have been used (Table S1) to expose the physical 138

mechanism behind the observed and modeled relationships seen in Fig. 1. The model 139

includes treatments of the major aerosols species and their precursors, simulating dust, 140

sea-salt, black and organic carbon, and sulfate. Aerosols affect the liquid and ice 141

microphysics of stratiform clouds in the model (acting as nuclei for cloud liquid drop and 142

ice crystal formation) and the radiative transfer calculation24. As with most global climate 143

models, aerosols do not directly affect the CAM5 convective cloud parameterization’s 144

microphysics 25,26, but convection does respond to aerosol radiative forcing as it affects 145

stability and the surface buoyancy fluxes. CAM5 can be run in “offline mode” as 146

described earlier, or in its standard “free running” mode in which aerosols and 147

meteorological fields evolve and interact simultaneously unconstrained by observations. 148

The correlation between central Indian monsoon precipitation and AOD for the 149

free running CAM5 is shown in Fig. 2a. The spatial correlation pattern is similar to that 150

seen in Fig. 1, but the positive correlation extends over a much larger area, covering the 151

whole of Arabian Sea, Bay of Bengal and peninsular India. Modest changes to 152

meteorology can have a large impact on the aerosol spatial distribution, but the model 153

physical parameterizations (in both free running and offline mode) do reproduce the 154

general correlation pattern, positive over Arabian Sea and negative over Northeast India. 155

8

(Areas where the free running and offline simulations produce different correlation 156

patterns are regions where some caution is needed in inferring physical relationships from 157

the free running simulations). 158

Additional simulations have also been made removing sea-salt, dust and 159

anthropogenic aerosol sources individually, and in combination (see table S1 in 160

supplementary materials for more details of the model experiments). Exclusion of these 161

combinations of sources helps to determine the cause for the observed correlation 162

patterns (Fig. 2). Note that natural pre-industrial (PI) emissions of OC and EC fire 163

sources (wildfires), dimethyl sulfide (DMS) and background volcanic SO2 sources, and 164

SOA emissions from vegetation were always included as aerosols sources. 165

A statistically significant relationship exists (with slightly changed spatial pattern) 166

over the Arabian Sea with present-day (PD) anthropogenic aerosol sources when sea-salt 167

and dust emissions are individually excluded (Fig. 2b and 2c). However, when both dust 168

and sea-salt are removed (Fig. 2d), the amplitude of the correlation decreases and 169

statistical significance disappears, indicating that both sea-salt and dust aerosols together 170

contribute to the observed pattern. This also implies that anthropogenic aerosols alone are 171

not responsible for the correlations, which is consistent with our interpretation of the AE, 172

and fine/coarse mode partitioning of the aerosol retrievals (Fig. S2). Both dust and sea 173

salt influence clouds and radiative heating. Removal of one species may have an impact 174

on the other (through competition for water vapor, or vapor deposition of sulfuric acid 175

and organics versus new particle formation etc27,28 and/or through changes to 176

meteorology), affecting the precipitation downwind. Therefore, to identify the role of 177

anthropogenic aerosols (within the model), we also repeated the analysis with pre-178

9

industrial (PI) emissions. The observed patterns are very similar to those seen with PD 179

anthropogenic emissions (Fig. 2a) when both dust and sea-salt natural emissions are 180

included (not shown). This again implies that anthropogenic aerosols alone cannot 181

produce these positive correlation patterns. However, as shown in the PI case (Figs. 3 a 182

&b), removal of any type of aerosol leads to a significant change in the observed 183

correlation patterns. The most striking change is produced by the removal of the dust 184

source (Fig. 3b), which leads to a complete disappearance and even change of sign in the 185

observed correlations. Correlations are still present, but weaker when sea-salt sources are 186

removed (Fig. 3a). The simulations suggest that the observed correlations can be 187

reproduced only when dust is present along with either sea-salt or anthropogenic aerosols 188

in the model. We note that in regions where aerosols are a minor player, and passively 189

responding to meteorological variability, removal of one species ought not to strongly 190

influence the correlation patterns. For example, removal of SO2 emissions affecting the 191

Arabian Sea does not strongly influence the correlation pattern there (not shown) and the 192

central Indian precipitation response on these very short time scales. 193

The modeled summer monsoon rainfall response to changes in dust and sea-salt 194

emissions is shown in Fig. 4. Sea-salt aerosol emissions lead to decreased precipitation 195

over the coastal southern peninsular and central Indian region (Fig. 4 a & c) and dust 196

aerosol emissions lead to increased precipitation (Fig. 4 b & d). Except for the intensity, 197

the responses are identical in both the stratiform and convective precipitation. Since the 198

convective precipitation microphysics is not linked to the presence of aerosols, the model 199

precipitation response must be a function of local or remote circulation and stability 200

changes (see next section), rather than a direct microphysical response (for example 201

10

through aerosol invigoration 29). South of 15oN, over oceanic regions remote from India, 202

the sea-salt correlates with increased rainfall, opposite behavior to that found over coastal 203

and central India. 204

Although the timescale cannot be inferred from the time averaged response shown in 205

Fig. 4,these seasonal changes in precipitation are consistent with the observed correlative 206

signatures found with shorter weekly scales shown in Fig. 1; increased precipitation over 207

central India is correlated with increased dust aerosols over Arabian Sea/West Asia and 208

the Saudi peninsula. To elucidate the short time scale response of aerosol-precipitation 209

interactions, another set of simulations called the Dust Pulse Experiment (DPE) was 210

performed. Two sets of 10-day model simulations were made starting from initial 211

conditions drawn during the monsoon period: one set of simulations included a dust 212

source, while the other simulation set had no dust source, so dust concentrations dropped 213

rapidly over the 10-day simulation period. All other aspects of the simulations were 214

identical (see Fig. S10 and corresponding text for details). Changes in dust burden (over 215

west Asia), and precipitation (over central India) were composited (by day of simulation) 216

to identify the impact on precipitation as the dust is gradually scavenged (Fig. 5). The 217

dust burden over West Asia/North Africa/Arabian region disappears quite rapidly, 218

decreasing to 10% of its climatological regional average within 5-7 days. Average 219

precipitation begins systematically decreasing over central India by day 7. The central 220

Indian precipitation drops as dust concentration decrease in the west Asian region. 221

222

223

224

11

4. The physical mechanism behind the remote dust aerosols and Indian 225

monsoon rainfall connection. 226

Dust is the largest contributor to the aerosol mass loading over northern Africa 227

and West Asia including the Arabian Peninsula. Observational studies have shown that 228

dust induced warming (clear sky) is the highest over Arabian Sea among the many dust 229

laden regions30. Our model estimates that dust aerosol increases top of the atmosphere 230

(TOA) absorption of solar insolation (+6 Wm-2), and decreases sunlight reaching the 231

surface (-30 Wm-2) producing a net atmospheric warming (+36Wm-2) and surface cooling 232

(dust aerosols can also change cloud microphysical properties, but this has a much 233

smaller effect on model radiative forcing and circulation changes). This combined 234

warming over regions with high dust concentrations leads to a reduction in surface 235

pressure thereby driving large-scale convergence over the North Africa/Arabian 236

Peninsula region (Fig. S7&S8a). The climatological winds are mostly southwesterly over 237

Indian region during the summer monsoon season. The convergent flow created by the 238

dust heating over land regions to the north and west of Arabian Sea strengthens the 239

northward component of the monsoon westerly (as a consequence of strengthening of the 240

pressure gradient) over Arabian Sea (see Figs. 6, S7a, S8a & S9a) leading to increased 241

moisture convergence (Fig. S9a) and precipitation (Fig. 4b) over the Indian region. This 242

mechanism is consistent for both the observed seasonal relationships and the DPE and is 243

summarized in Fig.S11 (see supplementary material for more information). The 244

proposed mechanism operates as a result of remote dust aerosol forcing far from central 245

India on a time scales of about a week, showing a different signature from the one the 246

EHP hypothesis5 addresses. 247

12

248

5. Synthesis, Integration, Implications. 249

We have used independent satellite retrievals to document a strong correlation 250

between Arabian Sea AOD and central Indian monsoon precipitation occurring on short 251

timescales. Satellite estimates of Angstrom Exponent and the Fine Mode Fraction (both 252

characterizing particle size) suggest that the aerosol involved is either dust or sea salt. 253

Our model shows a similar correlation, and our analysis indicates that desert dust aerosols 254

(originating over North Africa / West Asia/Arabian region and appearing as high aerosol 255

loading over the Arabian Sea) are responsible for the co-variability. This study suggests 256

that the interaction operates as a “direct radiative effect” where the radiative heating near 257

dust sources modulates the climatological monsoon westerly over the Arabian Sea with a 258

resultant change in precipitation (~3 to 6%) over central India. This aerosol effect on 259

precipitation through thermodynamics and large-scale circulation operates on timescales 260

of less than a week, and it does not involve a change to sea surface temperatures (SSTs). 261

While aerosols have been shown to be important in modulating sea surface temperatures, 262

significant ocean temperature change is not expected on these very short timescales. The 263

observed changes must primarily be a consequence of the atmospheric response to the 264

aerosol induced radiative effect. Previous studies have shown that anthropogenic 265

aerosols lead to decreases in Indian monsoon rainfall 3,11,12. These studies indicate that 266

anthropogenic aerosols over India (or adjacent ocean regions) also act to affect local 267

heating rates, and/or SSTs. We believe that the monsoon response to SST changes is very 268

important (perhaps more important than the interaction we are documenting), but they 269

operate on longer time scales. The present study shows that remote changes to natural 270

13

dust emissions as a consequence of anthropogenic and/or natural causes over regions to 271

the west of the Indian land mass, may also be modulating the rainfall over the Indian 272

region, highlighting the need to understand the effect of both natural and anthropogenic 273

aerosols and their interactions on precipitation in order to provide better predictions of 274

changes in monsoon precipitation. 275

276

14

6. Methods Summary 277

A weekly rainfall (using Global Precipitation Climatology Project daily dataset) 278

time series (the region averaged over 16.5 to 26.5oN and 74.5 to 86.6oE) was created for 279

the period 2000 to 2009 (June to August). This box was chosen as past studies have 280

shown that the inter-annual variability of rainfall over this region is similar to that of all 281

India rainfall (AIR) 22. The spatial and temporal homogeneity of rainfall22, consistency 282

with previous studies3 and ultimately its relationship with all India rainfall were the 283

important factors that influenced the choice of this box over India. The climatological 284

monthly mean rainfall over this region was subtracted from each time point 285

(corresponding to the month) to create a new time series of anomalies, forming the 286

primary index of rainfall over central India for this study. Similar anomaly time series 287

was created for the satellite-based AOD (MODIS and MISR) at each grid point over the 288

Indian sub-continent region shown in Fig.1. The Correlation coefficient between the 289

rainfall index over central India and the time series of aerosol optical depth at each grid 290

point was calculated and shown in Fig.1 (a&b). The statistical significance was 291

calculated using the standard student-t test. A dust pulse experiment was used to show 292

the short time scale nature of the aerosol-monsoon link. 293

294

15

Figure Captions 295

Fig. 1: Correlation coefficient between precipitation (over central India as indicated 296

by the box) and aerosol optical depth over the Indian region. (a) MODIS Terra (550 297

nm) and (b) MISR (558 nm) and (c) CAM5 in “offline mode” (550 nm). The black dots 298

represent significance at 95% confidence level. 299

Fig. 2: Correlation between precipitation over central India and aerosol optical 300

depth in free running simulations with present day emissions. Simulations include: 301

(a) All species, (b) No Sea-salt, (c) No Dust and (d) No Sea-salt and dust for present day 302

anthropogenic aerosol emissions. The black dots represent significance at 95% 303

confidence level. 304

Fig. 3: Correlation between precipitation over central India and aerosol optical 305

depth in free running simulations with pre-industrial emissions. Simulations include: 306

(a) No Sea-salt and (b) No Dust for pre-industrial aerosol emissions. The black dots 307

represent significance at 95% confidence level. 308

Fig. 4: The precipitation response associated with natural aerosol emissions (mm 309

day-1). The figure shows the control minus the simulation with suppressed dust or sea salt 310

emissions) Left panels (a & c) indicate the role of sea-salt aerosols. Right panels (b & d) 311

indicate the role of dust emissions. The top panels (a & b) show total precipitation 312

(convective + stratiform precipitation) and bottom panels (c & d) only the stratiform 313

precipitation. The black dots indicate a statistically significant change at the 95% 314

confidence level. 315

Fig. 5: The short-time response of precipitation to dust forcing as estimated from the 316

composite of two 19-member sets of 10-day simulations (the Dust Pulse 317

16

Experiment). The blue line corresponds to changes in precipitation over central India 318

and red line shows dust burdens over West Asia when the dust emissions are disabled. 319

The grey envelope represents the range of precipitation change between the 25th and 75th 320

percentiles of the 19 member ensembles. The dotted red line shows the mean dust 321

loading during the monsoon period from the control simulation. 322

Fig. 6: Schematic representation of circulation response (the aerosol induced 323

anomalies) suggested in the paper from dust aerosols over West Asia/Arabian Sea. 324

Land colors indicate altitude above mean sea level and vectors indicate the 325

climatological direction of winds between June and August (from the NCEP Reanalysis). 326

The darker elliptic region indicates the areas of increased warming from high dust 327

concentrations that lead to a reduction in surface pressure that induces large-scale 328

convergence over the North Africa/Arabian Peninsula region. The convergent flow 329

created by the dust heating to the north and west of Arabian Sea strengthens the 330

northward component of the monsoon westerly (brown arrows) as a consequence of 331

strengthening of the pressure gradient over Arabian Sea. This leads to an increased 332

moisture convergence and precipitation over the Indian region. 333

334

17

335

336 Fig.1 337

338

18

339

340

Fig. 2 341 342

19

343

Fig. 3 344 345

20

346

Fig. 4 347 348

21

349 Fig. 5 350

22

351

Fig. 6 352 353 354

355

23

References 356

357 1 Webster, P. J. et al. Monsoons: Processes, predictability, and the prospects for 358

prediction. J. Geophys. Res. 103, 14451-14510, doi:10.1029/97jc02719 (1998). 359

2 Gadgil, S. The Indian monsoon and its variability. Annual Review of Earth and 360

Planetary Sciences 31, 429-467, doi:10.1146/annurev.earth.31.100901.141251 361

(2003). 362

3 Bollasina, M., Ming, Y. & Ramaswamy, V. Anthropogenic Aerosols and the 363

Weakening of the South Asian Summer Monsoon. Science 334, 502-505, 364

doi:10.1126/science.1204994 (2011). 365

4 Chung, C. E. & Ramanathan, V. Weakening of North Indian SST gradients and 366

the monsoon rainfall in India and the Sahel. J Climate 19, 2036-2045 (2006). 367

5 Lau, K. M., Kim, M. K. & Kim, K. M. Asian summer monsoon anomalies 368

induced by aerosol direct forcing: the role of the Tibetan Plateau. Climate 369

Dynamics 26, 855-864, doi:10.1007/s00382-006-0114-z (2006). 370

6 Lau, K. M. & Kim, K. M. Observational relationships between aerosol and Asian 371

monsoon rainfall, and circulation. Geophys Res Lett 33, 372

doi:10.1029/2006gl027546 (2006). 373

7 Meehl, G. A., Arblaster, J. M. & Collins, W. D. Effects of Black Carbon Aerosols 374

on the Indian Monsoon. Journal of Climate 21, 2869-2882, 375

doi:10.1175/2007jcli1777.1 (2008). 376

8 Menon, S., Hansen, J., Nazarenko, L. & Luo, Y. Climate effects of black carbon 377

aerosols in China and India. Science 297, 2250-2253, 378

doi:10.1126/science.1075159 (2002). 379

24

9 Patra, P. K. et al. The Indian summer monsoon rainfall: interplay of coupled 380

dynamics, radiation and cloud microphysics. Atmospheric Chemistry and Physics 381

5, 2181-2188 (2005). 382

10 Wang, C., Kim, D., Ekman, A. M. L., Barth, M. C. & Rasch, P. J. Impact of 383

anthropogenic aerosols on Indian summer monsoon. Geophys Res Lett 36, 384

doi:10.1029/2009gl040114 (2009). 385

11 Ramanathan, V., C. Chung, D. Kim, T. Bettge, L. Buja, J. T. Kiehl, W. M. 386

Washington, Q. Fu, D. R. Sikka, M. Wild. Inaugural Article: Atmospheric brown 387

clouds: Impacts on South Asian climate and hydrological cycle. Proceedings of 388

the National Academy of Sciences 102, 5326-5333, doi:10.1073/pnas.0500656102 389

(2005). 390

12 Ganguly, D., Rasch, P. J., Wang, H. L. & Yoon, J. H. Climate response of the 391

South Asian monsoon system to anthropogenic aerosols. J Geophys Res-Atmos 392

117, doi:Artn D13209 393

Doi 10.1029/2012jd017508 (2012). 394

13 Menon, S., Hansen, J., Nazarenko, L. & Luo, Y. F. Climate effects of black 395

carbon aerosols in China and India. Science 297, 2250-2253 (2002). 396

14 Vinoj, V. & Satheesh, S. K. Measurements of aerosol optical depth over Arabian 397

Sea during summer monsoon season. Geophys Res Lett 30, 398

doi:10.1029/2002GL016664 (2003). 399

15 Satheesh, S. K. & Srinivasan, J. Enhanced aerosol loading over Arabian Sea 400

during the pre-monsoon season: Natural or anthropogenic? Geophys Res Lett 29, 401

doi:10.1029/2002gl015687 (2002). 402

25

16 Vinoj, V., Satheesh, S. K. & Moorthy, K. K. Optical, radiative, and source 403

characteristics of aerosols at Minicoy, a remote island in the southern Arabian 404

Sea. Journal of Geophysical Research 115, doi:10.1029/2009jd011810 (2010). 405

17 Bollasina, M., Nigam, S. & Lau, K. M. Absorbing Aerosols and Summer 406

Monsoon Evolution over South Asia: An Observational Portrayal. Journal of 407

Climate 21, 3221-3239, doi:10.1175/2007jcli2094.1 (2008). 408

18 Nigam, S. & Bollasina, M. “Elevated heat pump” hypothesis for the aerosol-409

monsoon hydroclimate link: “Grounded” in observations? Journal of Geophysical 410

Research 115, doi:10.1029/2009jd013800 (2010). 411

19 Manoj, M. G., Devara, P. C. S., Safai, P. D. & Goswami, B. N. Absorbing 412

aerosols facilitate transition of Indian monsoon breaks to active spells. Climate 413

Dynamics 37, 2181-2198, doi:10.1007/s00382-010-0971-3 (2010). 414

20 Bollasina, M. & Nigam, S. Absorbing aerosols and pre-summer monsoon 415

hydroclimate variability over the Indian subcontinent: The challenge in 416

investigating links. Atmos Res 94, 338-344, doi:10.1016/j.atmosres.2009.06.008 417

(2009). 418

21 Ganguly, D., Rasch, P. J., Wang, H. L. & Yoon, J. H. Fast and slow responses of 419

the South Asian monsoon system to anthropogenic aerosols. Geophys. Res. Lett. 420

39, doi:10.1029/2012gl053043 (2012). 421

22 Goswami, B. N., Venugopal, V., Sengupta, D., Madhusoodanan, M. S. & Xavier, 422

P. K. Increasing trend of extreme rain events over India in a warming 423

environment. Science 314, 1442-1445, doi:10.1126/science.1132027 (2006). 424

26

23 Ma, P.-L. et al. The role of circulation features on black carbon transport into the 425

Arctic in the Community Atmosphere Model version 5 (CAM5). Journal of 426

Geophysical Research: Atmospheres 118, 4657-4669, doi:10.1002/jgrd.50411 427

(2013). 428

24 Liu, X. et al. Toward a minimal representation of aerosol direct and indirect 429

effects: model description and evaluation. Geosci. Model Dev. Discuss. 4, 3485-430

3598, doi:10.5194/gmdd-4-3485-2011 (2011). 431

25 Zhang, G. J. & McFarlane, N. A. Sensitivity of climate simulations to the 432

parameterization of cumulus convection in the Canadian Climate Center general-433

circulation model. Atmosphere Ocean 33, 407-446 (1995). 434

26 Park, S., Bretherton, C. S. & Rasch, P. The global cloud simulations in the 435

community atmosphere model CAM5. In Preparation (2012). 436

27 Ghan, S. J., Guzman, G. & Abdul-Razzak, H. Competition between sea salt and 437

sulfate particles as cloud condensation nuclei. Journal of the Atmospheric 438

Sciences 55, 3340-3347 (1998). 439

28 O'Dowd, C. D., Lowe, J. A. & Smith, M. H. Coupling sea-salt and sulphate 440

interactions and its impact on cloud droplet concentration predictions. Geophys 441

Res Lett 26, 1311-1314 (1999). 442

29 Rosenfeld, D. et al. Flood or drought: how do aerosols affect precipitation? 443

Science 321, 1309-1313, doi:10.1126/science.1160606 (2008). 444

30 Zhu, A., Ramanathan, V., Li, F. & Kim, D. Dust plumes over the Pacific, Indian, 445

and Atlantic oceans: Climatology and radiative impact. J Geophys Res-Atmos 446

112, doi:Artn D16208 447

27

Doi 10.1029/2007jd008427 (2007). 448

449

450

28

Supplementary Information is linked to the online version of the paper at 451

www.nature.com/ngeo 452

453

Acknowlegements: This work was supported by the Pacific Northwest National 454

Laboratory (PNNL) Program for Laboratory Directed Research and the U.S. Department 455

of Energy (DOE) Office of Science (BER) Earth System Modeling Program. PNNL is 456

operated for the DOE by Battelle Memorial Institute under contract DE-AC05-76RLO 457

1830. The MODIS, MISR and GPCP mission scientists and associated NASA personnel 458

are acknowledged for the production of the data used in this research effort. 459

460

Author Contributions 461

VV and PJR primarily conceived the idea. VV carried out the data analysis and free 462

running model simulations. PM and JHY carried out the offline and DPE model 463

simulations respectively. All authors contributed through discussions and interpretation 464

of the results. VV wrote a first draft of the paper with subsequent input from all the 465

coauthors. 466

467

Author Information Reprints and permissions information is available at 468

www.nature.com/reprints. Correspondence should be addressed to P.J.R. 469

(philip.rasch@pnnl.gov). 470

471

Competing Interests The authors declare that they have no competing financial interests. 472

473