Projecting South Asian Summer Precipitation in CMIP3 Models: A Comparison of the Simulations with and without Black Carbon
  J. Meteor. Res.  2017, Vol. 31 Issue (1): 196-203   PDF    
http://dx.doi.org/10.1007/s13351-017-6101-y
The Chinese Meteorological Society
0

Article Information

LI Shuanglin, MAHMOOD Rashed . 2017.
Projecting South Asian Summer Precipitation in CMIP3 Models: A Comparison of the Simulations with and without Black Carbon. 2017.
J. Meteor. Res., 31(1): 196-203
http://dx.doi.org/10.1007/s13351-017-6101-y

Article History

Received June 4, 2016
in final form December 19, 2016
Projecting South Asian Summer Precipitation in CMIP3 Models: A Comparison of the Simulations with and without Black Carbon
Shuanglin LI1,2, Rashed MAHMOOD3,4     
1. Department of Atmospheric Science, China University of Geosciences, Wuhan 430074, China;
2. Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China;
3. School of Earth and Ocean Sciences, University of Victoria, Victoria BC, Canada;
4. Department of Meteorology, COMSATS Institute of Information Technology, Islamabad, Pakistan
ABSTRACT: Considering the importance of black carbon (BC), this study began by comparing the 20th century simulation of South Asian summer climate in IPCC CMIP3, based on the scenario of models with and without BC. Generally, the multi-model mean of the models that include BC reproduced the observed climate relatively better than those that did not. Then, the 21st century South Asian summer precipitation was projected based on the IPCC CMIP3 projection simulations. The projected precipitation in the present approach exhibited a considerable difference from the multi-model ensemble mean (MME) of IPCC AR4 projection simulations, and also from the MME of the models that ignore the effect of BC. In particular, the present projection exhibited a dry anomaly over the central Indian Peninsula, sandwiched between wet conditions on the southern and northern sides of Pakistan and India, rather than homogeneous wet conditions as seen in the MME of IPCC AR4. Thus, the spatial pattern of South Asian summer rainfall in the future may be more complicated than previously thought.
Key words: South Asian summer monsoon     black carbon     CMIP3 projection simulations    
1 Introduction

The role of black carbon (BC) in modulating the regional climate of South Asia has been identified in a number of previous modeling studies (e.g.,Ramanathan et al., 2005;Lau et al., 2006;Meehl et al., 2008;Wang et al., 2009;Mahmood and Li, 2013;Lau, 2016). A provisional consensus is that BC may be a primary contributor to the observed pre-monsoonal enhancement, and monsoonal decline, in rainfall (e.g.,Meehl et al., 2008, Figs. 9 and 10). The significant contribution of BC in modulating regional and/or remote climate arises from the fact that the induced strong heating in the lower troposphere can influence the local climate through changing atmospheric stability and/or general circulation systems (Jacobson, 2001;Menon et al., 2002;Jiang et al., 2013;Mahmood and Li, 2013,2014;Liao and Shang, 2015). While the physical mechanisms are still under debate, evidence suggests that aerosols may play a more important role in influencing South Asian climate than greenhouse gases (e.g.,Bollasina et al., 2011). Since some models in the World Climate Research Program's Coupled Model Intercomparison Project Phase 3 (CMIP3) historical simulation experiments include BC forcing, while others do not (Meehl et al., 2007), a comparison between the two groups' simulations may reveal the climatological impact of BC, instead of conducting additional experiments. One question we may naturally want to ask is: Do those CMIP3 models that include BC forcing produce relatively more realistic results? If so, what are the spatial and temporal distributions of projected precipitation from these models, given that the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report (AR4) projected a homogeneous rainfall enhancement in South Asia based on the multi-model ensemble mean (MME) of 21 of the CMIP3 models (Christensen et al., 2007, Fig. 11.9)? These considerations motivated the present study, which began by comparing the 20th century simulations of South Asian climate in two groups of models with and without the effect of BC. The results suggested that the models considering the effect of BC reproduce the observed climate more realistically. Therefore, the future South Asian climate was then projected by using these models, based on the IPCC AR4 projection simulations. The results projected a negative precipitation anomaly over central India in the MME of the models that include the effect of BC: a result that is in contrast to the homogeneous rainfall increase projected by the MME of the 21 IPCC AR4 models.

2 Data and methods

The observational rainfall datasets used in this study include: the land-only precipitation from the Climate Research Unit (CRU) of the University of East Anglia (CRU TS3.1;Mitchell and Jones, 2005); the University of Delaware terrestrial precipitation, version 2.01 (Legates and Willmott, 1990); and global precipitation from the Global Precipitation Climatology Project (GPCP), version 2.2 (Adler et al., 2003). The mean sea level pressure (SLP) was from the ECMWF's ERA-40 reanalysis (Uppala et al., 2005).

Two sets of CMIP3 model simulations were used: one from the 20th century simulations (20C3M) and the other from the 21st century simulations with SRESA1B scenarios. The simulations of a total of 22 models were analyzed according to two groups: one involving 9 models that consider the effect of BC (in short, CMIP3_BC), and the other involving 13 models that do not consider the effect of BC (in short, CMIP3_NOBC) (see Table 1). The methodology for grouping the models was similar to that of Purich and Son (2012) and Allen et al. (2012), but just based on BC.Purich and Son (2012) grouped the models based on considerations of ozone when studying the influence of ozone depletion and recovery on Southern Hemisphere climate.Allen et al. (2012) grouped the CMIP3 models based on BC and tropospheric ozone when studying the two forcings' contributions to Northern Hemispheric tropical expansion.

Table 1 Grouping of the 22 CMIP3 models used in this study. The names of the corresponding institutions and countries are given in brackets. CMIP3_NOBC refers to the CMIP3 models that do not consider black carbon (BC), while CMIP3_BC refers to those that do
CMIP3_NOBC CMIP3_ BC
1 CSIRO Mk3.0 (CSIRO Atmospheric Research Group, Australia) 1 CCSM3 (National Center for Atmospheric Research, USA)
2 CSIRO Mk3.5 (CSIRO Atmospheric Research Group, Australia) 2 CNRM-CM3 (Meteo France/Centre National de Recherches Météorologiques, France)
3 ECHAM5_MPI (Max Plank Institute for Meteorology, Germany) 3 GFDL CM2.0 (U.S. Dept. of Commerce/NOAA/Geophysical Fluid Dynamics Laboratory, USA)
4 FGOALS-g1.0 (Institute of Atmospheric Physics, China) 4 GFDL CM2.1 (U.S. Dept. of Commerce/NOAA/Geophysical Fluid Dynamics Laboratory, USA)
5 IPSL-CM4 (L'Institut Pierre Simon Laplace, France) 5 GISS-E2-H (NASA/Goddard Institute for Space Studies, USA)
6 INM-CM3.0 (Institute for Numerical Mathematics, Russia) 6 GISS-E2-R (NASA/Goddard Institute for Space Studies, USA)
7 GISS-AOM (NASA/Goddard Institute for Space Studies, USA) 7 MIROC3.2 (hires) [Center for Climate System Research (University of Tokyo), National Institute for Environmental Studies and Frontier Research Center for Global Change, Japan]
8 NCAR_PCM1 (National Center for Atmospheric Research, USA) 8 MIROC3.2 (medres) [Center for Climate System Research (University of Tokyo), National Institute for Environmental Studies and Frontier Research Center for Global Change, Japan]
9 HadCM3 (Hadley Centre for Climate Prediction and Research/Met Office, UK) 9 HadGEM1 (Hadley Centre for Climate Prediction and Research/Met Office, UK)
10 CGCM3.1 [T47] (Canadian Centre for Climate Modeling & Analysis, Canada)
11 CGCM3.1 [T63] (Canadian Centre for Climate Modeling & Analysis, Canada)
12 INGV_ECHAM4 (Instituto Nazionale di Geofisica e Vulcanologia, Italy)
13 MRI-CGCM2.3.2 (Meteorological Research Institute, Japan)

To assess the potential summer (June–September, JJAS) precipitation responses in the future, the results of the SRESA1B experiment were subtracted from the 1950–99 climatology of the 20C3M experiment.Seager et al. (2007) calculated similar climatological differences to assess the possible future aridity in southwestern North America. Due to the different spatial resolutions of the various models, bilinear interpolation was performed to interpolate all the modeled precipitation onto a uniform 5°×5° grid. Only the first run (i.e., RUN1) of the model output was analyzed for both types of experiments (i.e., 20C3M and SRESA1B), except for two models (FGOALS-g1.0 and GISS-E2-H) in which RUN2 of the SRESA1B experiment was used due to the unavailability of RUN1.

3 Results and discussion 3.1 Reasonableness of the 20C3M simulations

We began our analysis by plotting 5-yr running means of observed and simulated summer precipitation averaged over (25°–35°N, 65°–95°E) for the period 1950–99 (Fig. 1a), along with the corresponding 50-yr statistics (Fig. 1b). A key feature in Fig. 1 is the CMIP3_BC's closer resemblance than CMIP3_NOBC to the observation. The fact that the modeled precipitation in CMIP3_BC has amplitudes similar to the observed signifies the potential effect of BC on South Asian climate, as found in many previous studies (e.g.,Menon et al., 2002;Mahmood and Li, 2013,2014). However, both model groups failed to reproduce the observed precipitation trend for the selected domain.

Fig. 1 (a) Five-year running mean of area-averaged summer (June–September, JJAS) precipitation (1950–99) over South Asia (25°–35°N, 65°–95°E). The dotted lines represent the 50-yr least squares linear trend. The blue, green, and cyan lines represent observations from the Climatic Research Unit (CRU), the University of Delaware (UDEL), and the average of CRU and UDEL (AVEOB), respectively. The dark brown and red lines represent CMIP3_BC and CMIP3_NOBC, respectively. (b) Summary of JJAS mean precipitation from 1950 to 1999 averaged over the same region as in (a). The upper and lower edges of each box represent maximum and minimum values, respectively. Black dots represent the 50-yr mean values, with upward- and downward-pointing red triangles representing the corresponding +1 standard deviation. Units: mm day–1.

The results shown in Fig. 1 provide interesting information related to the dominance of CMIP3_BC over CMIP3_NOBC in reproducing the observed summer precipitation. However, this is likely to be domain dependent, and therefore, the spatial distributions of the observed and simulated JJAS climatologies for the period 1980–99 were compared (Fig. 2). In observations, the strongest and most spatially extended precipitation center is over Myanmar, Bangladesh, and eastern India (Figs. 2a–c). Another prominent precipitation center is near the western side of the Indian Peninsula (Figs. 2a–c). The fact that the two model groups simulated these strong precipitation centers comparatively better may provide a basic assessment for the reasonableness of the simulations in CMIP3_BC and CMIP3_NOBC.Annamalai et al. (2007) used a similar approach to test the performance of several individual models. A general comparison between the observed (Figs. 2a–c) and simulated (Figs. 2d and 2e) precipitation suggested that there was a marked discrepancy between the two groups in reproducing the above observed precipitation centers. Despite that, it appeared that the underestimation was much higher in the case of CMIP3_NOBC (see their difference depicted in Fig. 2f). From Fig. 2f, almost all of the land regions, including the two strong precipitation centers in observations, were comparatively stronger in the case of CMIP3_BC, and thus featured less severe underestimation. Interestingly, the strongest difference between CMIP3_BC and CMIP3_NOBC (about 4 mm day–1) was located approximately in the strongest precipitation region in the observation. In the northwestern Indian Ocean (off the Somali Coast), there was a relatively stronger precipitation bias in CMIP3_BC. Such a model bias was investigated recently by Bollasina and Ming (2013), who attributed it to the extra sensitivity of models to the meridional sea surface temperature gradients. Also, for southern Indo-China, both groups appeared to overestimate the observed precipitation. Nevertheless, it may be argued that, at least for the South Asian region, the MME of the CMIP3_BC models may represent observed precipitation comparatively better than that of the CMIP3_NOBC models.

Fig. 2 Comparison between observed and modeled 20-yr (1980–99) JJAS mean precipitation: (a) CRU (Climatic Research Unit, University of East Anglia); (b) UDEL (University of Delaware); (c) GPCP (Global Precipitation Climatology Project); (d) CMIP3_BC; (e) CMIP3_NOBC; (f) difference between (d) and (e). The model values were interpolated onto a uniform 5°×5° grid to take multi-model means. Units: mm day–1.

Another metric used for evaluating model performance was the 1980–99 JJAS mean SLP, which in observations was represented by ECMWF (ERA-40) reanalysis data (Fig. 3). For the observed climatology, one important feature is the existence of a low pressure center over Pakistan and western India (Fig. 3a). This low pressure is further extended up to the Bay of Bengal in the east and to the eastern Arabian Peninsula in the west. Another important feature is the relatively higher pressure over the Tibetan Plateau (TP). So, we further evaluated the performance of the two model groups based on how well they performed in reproducing this dipole-like SLP structure. First, all the models in both groups appeared to extend the low pressures too far north over the TP (Figs. 3b and 3c), and thus there was less of a blocking effect of the TP, which may have contributed to the overall underestimation in simulated precipitation in South Asia and an overestimation in southern Indo-China. Again, the difference between the two model groups (Fig. 3d) suggested that CMIP3_BC had a dipole-like pattern, with positive SLP over the TP and negative SLP over the subcontinent. From these analyses, the CMIP3_BC simulations arguably correlated better than those of CMIP3_NOBC with the observed/reanalysis results in the South Asia region.

Fig. 3 (a–c) Summer (June–September) mean climatology (1980–99) of sea level pressure, and (d) the difference between CMIP3_BC and CMIP3_NOBC.

However, it is important to note that both groups struggled with reproducing the spatial distributions of observed precipitation and SLP. In addition, although the focus of this study was on the potential influence of BC on climatological precipitation in South Asia, it is highly likely that other natural and anthropogenic forcings also play a role in modulating observed precipitation. Similarly, the missing indirect effects of aerosols could contribute too. Furthermore, the inherent differences in individual model structures and assumptions for parameterizations may also influence the model simulations.

3.2 Future precipitation projections

We evaluated the projection of future precipitation based on SRESA1B simulations. Although there are several uncertainties associated with model simulations of future climate (e.g.,Knutti, 2008) and questionable credibility due to model discrepancies and the assumptions about future physical conditions (Schwartz et al., 2007;Tebaldi and Knutti, 2007;Knutti, 2008), they nevertheless may provide a basic assessment (or a calculated guess) on the possible future conditions (e.g.,Christensen et al., 2007;Seager et al., 2007;Mariotti et al., 2008;Purich and Son, 2012). Thus, our next objective was to evaluate future precipitation in the simulations with and without BC, which may be useful when interpreting the possible future climate conditions in the subcontinent.

Figure 4 shows the JJAS mean precipitation difference (i.e., 2050–99 minus 1950–99) for land-only regions. The difference derived from CMIP3_BC (Fig. 4a) suggested that the mean summer precipitation may increase in Bangladesh, eastern and southern India, and northern Pakistan, along with a relative increase in aridity in central India; while that from CMIP3_NOBC suggested increasingly wet conditions over the whole of South Asia (Fig. 4b). These results highlighted a major difference over central India, with a relatively dryer future according to CMIP3_BC (Fig. 4a) versus a wetter future according to CMIP3_NOBC (Fig. 4a). This future summer dryness in central India according to CMIP3_BC may reach up to a maximum of 8% relative to the mean conditions during 1950–99 (Fig. 5). Interestingly, such dry conditions appear to be in line with the observed drying trends over this region during the last 50 years of the 20th century (cf.Figs. 5 and 6). Another notable point is the distribution of the future precipitation anomaly over Pakistan, where both model groups suggested wetness, although the spatial distribution appeared more similar to the observed trends, especially for northern Pakistan, in the past decades in CMIP3_BC. CMIP3_BC suggested up to a 20% increase in precipitation in northern Pakistan, which is interesting in terms of the past moistening trends in some of the observations (e.g., in CRU; cf.Figs. 5 and 6). Similarly, CMIP3_NOBC also suggested a positive sign in this region, but its center shifted towards the southern coastal areas. However, it should be noted that these trends and anomalous patterns are not directly comparable, since they span over different timescales and are therefore subject to change. Additionally, the current study is limited in terms of identifying the sole source of such distributions, which could be the result of model artifacts and/or missing forcings in the simulations.

Fig. 4 Difference in climatological mean (2050–99) summer (June–September) precipitation relative to mean 1950–99 conditions. (a) CMIP3_BC; (b) CMIP3_NOBC. Units: mm day–1.
Fig. 5 Percentage change in climatological mean (2050–99) summer (June–September) precipitation relative to mean 1950–99 conditions. (a) CMIP3_BC; (b) CMIP3_NOBC.
Fig. 6 Observed 50-yr (1950–99) least squares linear trends of summer (June–September) mean precipitation in (a) CRU TS3.1 (Climatic Research Unit, University of East Anglia) and (b) UDEL v2.01 (University of Delaware). Units: mm day–1 (50 yr)–1.

Comparing the present results with those shown in IPCC AR4 (Christensen et al., 2007, Fig. 11.9), CMIP3_BC presents a contrasting pattern over central India: a potential future of increasingly dry rather than wet conditions over the whole of the South Asia region. Moreover, the heterogeneity in the spatial detail of the projected precipitation over South Asia in our analysis suggests that, perhaps, it may not be wise to simply calculate the area mean for a region as large as the whole South Asian domain, as shown by Giorgi and Bi (2005,Fig. 1), to predict future regional precipitation.Giorgi and Bi (2005) concluded an increase in future precipitation over South Asia in the wet season, which can obscure the more complicated spatial patterns as shown in the current analysis.

4 Concluding remarks

We investigated the CMIP3 simulations of South Asian past and future summer precipitation by dividing the models into two groups based on inclusion/exclusion of BC in these simulations. First, the results from the 20C3M experiment were compared with observations to assess which model group resembles the observations more closely. The comparisons suggested that the MME of the group with BC (CMIP3_BC) represents the observed situation relatively better than that of the group without BC (CMIP3_NOBC). Next, we investigated what these model groups suggest in terms of future precipitation under the SRESA1B experiment. The most interesting result was the contrasting anomalous future precipitation projection in central India, where CMIP3_BC suggested increasingly dry conditions while CMIP3_NOBC implied increasingly wet conditions. The projected pattern of the former was also in contrast with that established by the MME of 21 models in IPCC AR4 and in previous studies (e.g.,Giorgi and Bi, 2005). Based on this analysis of CMIP3 models, we can expect more anomalously complex spatial precipitation patterns in South Asia, rather than a uniform increase in future precipitation.

Readers will undoubtedly be wondering why the simulations of CMIP5, instead of CMIP3, were not used in the present analysis. The answer is that all historical simulations from CMIP5 models include BC forcing. Thus, they were unsuitable for the present comparison. Nevertheless, the present results provide indications about the complexity of future projections of South Asian summer rainfall in these models.

Possible caveats to the present analysis include the different model configurations, resolutions, and basic atmospheric physics in individual models, all of which can contribute differently to the simulated results (e.g.,Mahmood et al., 2016). Additionally, from the current study alone, it is not possible to explicitly assign BC as the sole contributor to the simulated results (as discussed above):a matter that requires more objectively based multi-model simulations.

Acknowledgements . We acknowledge the modeling groups, the PCMDI, and the WCRP's Working Group on Climate Modeling, for making the WCRP CMIP3 multi-model dataset available. Support of this dataset is provided by the Office of Science, U.S. Department of Energy.

References
R. F. Adler, G. J. Huffman, A. Chang, et al ,2003: The version-2 Global Precipitation Climatology Project (GPCP) monthly precipitation analysis (1979–present). J. Hydrometeor. , 4 , 1147–1167.
R. J. Allen, S. C. Sherwood, J. R. Norris, et al ,2012: Recent Northern Hemisphere tropical expansion primarily driven by black carbon and tropospheric ozone. Nature , 485 , 350–354.
H. Annamalai, K. Hamilton, K. R. Sperber ,2007: The South Asian summer monsoon and its relationship with ENSO in the IPCC AR4 simulations. J. Climate , 20 , 1071–1091.
M. A. Bollasina, Y. Ming ,2013: The general circulation model precipitation bias over the southwestern equatorial Indian Ocean and its implications for simulating the South Asian monsoon. Climate Dyn. , 40 , 823–838. DOI:10.1007/s00382-012-1347-7
M. A. Bollasina, Y. Ming, V. Ramaswamy ,2011: Anthropogenic aerosols and the weakening of the south Asian summer monsoon. Science , 334 , 502–505.
Christensen, J. H., B. Hewitson, A. Busuioc, et al., 2007: Regional climate projections.Climate Change 2007: The Physical Science Basis. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, S. Solomon et al., Eds., Cambridge University Press, 996pp.
F. Giorgi, X. Bi ,2005: Updated regional precipitation and temperature changes for the 21st century from ensembles of recent AOGCM simulations. Geophys. Res. Lett. , 32 , L21715. DOI:10.1029/2005GL024288
M. Z. Jacobson ,2001: Strong radiative heating due to the mixing state of black carbon in atmospheric aerosols. Nature , 409 , 695–697.
Y. Q. Jiang, H. Liu X., Q. Yang X., et al ,2013: A numerical study of the effect of different aerosol types on East Asian summer clouds and precipitation. Atmos. Environ. , 70 , 51–63.
R. Knutti ,2008: Should we believe model predictions of future climate change?. Phil. Trans. Roy. Soc. A Math. Phys. Eng. Sci. , 366 , 4647–4664.
K. M. Lau, M. K. Kim, K. M. Kim ,2006: Asian summer monsoon anomalies induced by aerosol direct forcing: The role of the Tibetan Plateau. Climate Dyn. , 26 , 855–864.
W. K. M. Lau ,2016: The aerosol-monsoon climate system of Asia: A new paradigm. J. Meteor. Res. , 30 , 1–11. DOI:10.1007/s13351-015-5999-1
D. R. Legates, C. J. Willmott ,1990: Mean seasonal and spatial variability in gauge-corrected, global precipitation. Int. J. Climatol. , 10 , 111–127.
H. Liao, J. J. Shang ,2015: Regional warming by black carbon and tropospheric ozone: A review of progresses and research challenges in China. J. Meteor. Res. , 29 , 525–545. DOI:10.1007/s13351-015-4120-0
R. Mahmood, S. L. Li ,2013: Delay in the onset of South Asian summer monsoon induced by local black carbon in an AGCM. Theor. Appl. Climatol. , 111 , 529–536. DOI:10.1007/ s00704-012-0681-3
R. Mahmood, S. L. Li ,2014: Remote influence of South Asian black carbon aerosol on East Asian summer climate. Int. J. Climatol. , 34 , 36–48. DOI:10.1002/joc.3664
von Salzen Mahmood, Flanner R., et al ,2016: Seasonality of global and Arctic black carbon processes in the Arctic monitoring and assessment programme models. J. Geophys. Res. Atmos. , 121 , 7100–7116. DOI:10.1002/2016JD024849
A. Mariotti, N. Zeng, J.-H. Yoon, et al ,2008: Mediterranean water cycle changes: Transition to drier 21st century conditions in observations and CMIP3 simulations. Environ. Res. Lett. , 3 , 044001. DOI:10.1088/1748-9326/3/4/044001
G. A. Meehl, C. Covey, K. E. Taylor, et al ,2007: The WCRP CMIP3 multimodel dataset: A new era in climate change research. Bull. Amer. Meteor. Soc. , 88 , 1383–1394.
G. A. Meehl, J. M. Arblaster, W. D. Collins ,2008: Effects of black carbon aerosols on the Indian monsoon. J. Climate , 21 , 2869–2882. DOI:10.1175/2007JCLI1777.1
S. Menon, J. Hansen, L. Nazarenko, et al ,2002: Climate effects of black carbon aerosols in China and India. Science , 297 , 2250–2253.
T. D. Mitchell, P. D. Jones ,2005: An improved method of constructing a database of monthly climate observations and associated high-resolution grids. Int. J. Climatol. , 25 , 693– 712 .
A. Purich, S.-W. Son ,2012: Impact of Antarctic ozone depletion and recovery on Southern Hemisphere precipitation, evaporation, and extreme changes. J. Climate , 25 , 3145– 3154 .
V. Ramanathan, C. Chung, D. Kim, et al ,2005: Atmospheric brown clouds: Impacts on South Asian climate and hydrological cycle. Proc. Natl. Acad. Sci. U. S. A. , 102 , 5326–5333. DOI:10.1073/pnas.0500656102
S. E. Schwartz, J. R. Charlson, H. Rodhe ,2007: Quantifying climate change—too rosy a picture?. Nat. Rep. Climate Change , 2 , 23–24. DOI:10.1038/climate.2007.22
R. Seager, M. F. Ting, I. Held, et al ,2007: Model projections of an imminent transition to a more arid climate in southwestern North America. Science , 316 , 1181–1184.
C. Tebaldi, R. Knutti ,2007: The use of the multi-model ensemble in probabilistic climate projections. Phil. Trans. Roy. Soc. A Math. Phys. Eng. Sci. , 365 , 2053–2075.
Kållberg Uppala, Angeles S., et al ,2005: The ERA-40 Re-Analysis. Quart. J. Roy. Meteor. Soc. , 131 , 2961–3012. DOI:10.1256/qj.04.176
C. E. Wang, D. Kim, A. M. L. Ekman, et al ,2009: Impact of anthropogenic aerosols on Indian summer monsoon. Geophys. Res. Lett. , 36 , L21704. DOI:10.1029/2009GL040114