Impact of anthropogenic forcing on the Asian summer monsoon as simulated by eight GCMs



[1] The response of the Asian summer monsoon (ASM) to a transient increase in future anthropogenic radiative forcing is investigated by multi-model global warming experiments. Most models show that, in Asia, the summer monsoon rainfall increases significantly with global warming. On the other hand, the future change in the large-scale flow indicates a weakening of the ASM circulation. Enhanced moisture transport over the Asian summer monsoon region, associated with the increased moisture source from the warmer Indian Ocean, leads to a larger moisture flux convergence, which is responsible for the intensification of the mean rainfall. Pronounced warming over the tropics in the middle-to-upper troposphere causes a reduction in the meridional thermal gradient in the Asian region, which is consistent with the weakened monsoon circulation and eastward shift of the Walker circulation.

1. Introduction

[2] The Asian summer monsoon (ASM) constitutes the most spectacular manifestation of the global climate system resulting from land-sea thermal contrast and orographic features. Thus the ASM provides a major percentage of water sources over the most densely populated regions in the world. A number of assessments of ASM are currently being attempted using the projections of climate changes due to various scenarios of radiative forcing provided by transient runs with GCMs [Intergovernmental Panel on Climate Change (IPCC), 2001]. Several modeling studies with increased CO2 have shown that the final equilibrium of ASM precipitation is likely to increase [Douville et al., 2000], accompanied by intensified interannual variability [Meehl and Washington, 1993] and large regional differences [Lal and Harasawa, 2001].

[3] Kitoh et al. [1997] reported a paradox between increasing precipitation and decreasing circulation intensity in the South Asian summer monsoon region. Hu et al. [2000] indicated that an enhancement of the thermal gradient between the Asian continent and surrounding oceans is responsible for the intensified ASM. Several studies have revealed that the precipitation response cannot be inferred directly from the circulation changes alone. Meehl and Arblaster [2003] showed that the increase in the mean precipitation is due to increased moisture from the warmer Indian Ocean. May [2002] found that the increase in Indian rainfall is caused by the intensification of the moisture transport into the Indian region, which is partly counterbalanced by a decrease in the water vapor cycling rate [Douville et al., 2002].

[4] The above studies are based on single experiment; thus, the aim of this study is to investigate the impacts of anthropogenic climate changes on the ASM by use of multi-model global experiments, taking into consideration the moisture budget, the meridional thermal gradient (MTG) and the modulation of the Walker circulation.

2. Data and Method

[5] This study is a spin-off of the multi-model analyses conducted for the upcoming IPCC 4th assessment report [Meehl et al., 2005] under the auspices of the Coupled Model Intercomparison Project (CMIP) [Meehl et al., 2000]. The data sets used in this study are 20th century climate simulations (20th Century Climate in Coupled Models; 20C3M) and future climate simulations based on SRES-A1B scenarios [IPCC, 2001]. Most of the 20C3M experiments were integrated according to observed anthropogenic forcing from the late 19th century and ending in 2000.

[6] In this study, the climates during two time means, present (1981–2000, hereafter referred to as L20C3M) and future (2100–2200), are used for the intercomparison. Numerous details such as spatial resolution, physical processes and parameterizations vary with each model. Details are documented on the PCMDI website,, and references are given in Table 1. We also used the Climate Prediction Center (CPC) merged analysis of precipitation (CMAP) data of 2.5-degree grids derived from five kinds of satellite estimates (GPI, OPI, SSM/I scattering, SSM/I emission and MSU) for the period from 1981 to 2000 [Xie and Arkin, 1996].

Table 1. Climate Model Documentation and Their References
  • a

    D. Salas-Mélia et al. (Description and validation of the CNRM-CM3 global coupled model, submitted to Climate Dynamics, 2005).

CNRM-CM3FranceSalas-Mélia et al. [2005]a
ECHAM5/MPI-OMGermanyJungclaus et al. [2006]
FGOALS-g1.0ChinaYu et al. [2004]
GFDL-CM2.0USADelworth et al. [2006]
INM-CM3.0RussiaDiansky and Volodin [2002]
MIROC3.2 (medres)JapanHasumi and Emori [2004]
MRI-CGCM2.3.2JapanYukimoto et al. [2001]
UKMO-HadCM3UKGordon et al. [2000]

[7] The multi-model ensembles used in this study permit us to deduce the synthesis of the projected change. Generally averaging more simulations may reduce the effect of natural variability. To compare the strength of the climate change signals to natural variability noise, the “signal to noise ratio”, {ΔX}/σΔX, is calculated for moisture flux and MTG. A large ratio indicates that the signal stands out against to natural variability (“noise”).

3. Precipitation-Wind Paradox

[8] Figure 1 shows the time series of simulated ASM precipitation from the late 19th century through 2200. To compare the performance of the models during the late 20th century, we also plotted the observed rainfall (dashed line). Though the simulated precipitation in all models is greater than the observed precipitation, most models show an increasing trend after the beginning of the 21st century, within a range of 0.4 ∼ 1.3 mm day−1. The mean precipitation change among models is about 0.8 mm day−1 toward the end of the 22nd century. An abrupt change occurs for MIROC3.2 in about 2060 due to a decrease in aerosols through changes in the anthropogenic radiative forcing of the atmosphere (T. Nozawa, NIES, personal communication, 2005).

Figure 1.

Time series of simulated JJA precipitation (9-year running mean) in millimeters per day, averaged for eq.–30°N, 60°–100°E. Dashed line denotes observed JJA rainfall from 1981 to 2000.

[9] The composite difference in the summertime precipitation in the 22nd century and L20C3M is depicted in Figure 2a. Although the regional details differ from one model to another (not shown), the composite difference clearly indicates an intensification of rainfall appearing over the North Indian Ocean through the western North Pacific, accompanied by a weakening of the low-level monsoon flows. The absolute values of signal-to-noise ratio for the rainfall (Figure 2b) and zonal wind (Figure 2c) are greater than 1.0 over those regions that are indicative of a robust change in monsoon circulation. The values of change in area-averaged JJA precipitation (equator to 30°N, 60° to 100°E) range from 0.2 to 1.1 mm/day and the multi-model ensemble mean is 0.8 mm/day (see Table 2).

Figure 2.

(a) Differences in JJA precipitation (mm/day) and 850 hPa winds (m/s) between the 22nd century and L20C3M. Shadings denote increase in rainfall greater than 0.5 mm/day. Absolute values of signal-to-noise ratio for (b) rainfall and (c) precipitation, in excess of 1.0, are depicted by shadings.

Table 2. Plausible Change of Area-Averaged JJA Precipitation, Zonal Winds, MTG, and Vertically Integrated Moisture Budgeta
ModelPr, mm/dayU, m/sMTG, mBudget, kg/ms
  • a

    JJA precipitation data were gathered over the equator to 30°N and 60° to 100°E; zonal winds over 0 to 20°N and 60°E–100°E; MTG for 500–200 hPa thickness over 35°–45°N minus 10°S–10°N along 60°–100°E sector; and vertically integrated moisture budget over 0–25°N and 60°–110°E.

MIROC3.2 (medres)1.1−2.4−45.0−23.7

4. Possible Causes for the Enhanced ASM Rainfall

4.1. Moisture Transport and Its Budget

[10] In order to account for the enhancement of ASM precipitation in the future climate despite the fact that the intensity of the low-level monsoon flows becomes weaker, the atmospheric moisture transport and its budget in the Asian monsoon region are considered. In general, the atmospheric wind pattern resembles the moisture flux. However, the future changes in the vertically integrated moisture flux during the summer monsoon season clearly indicate an enhancement of the moisture transport over the Arabian Sea into Southeast Asia (Figure 3a). The specific humidity increased remarkably over the Indian Ocean (Figure 3b), which is consistent with the intensified moisture transport. In Figure 3a, the signal-to-noise ratio for the water vapor flux is greater than 1.0 over the Arabian Sea through the Indochina Peninsula, including India and the Bay of Bengal, which is indicative of a robust change in moisture transport.

Figure 3.

Differences in multi-model ensembles of (a) vertically integrated water vapor flux (kgm−1s−1) and (b) vertically integrated specific humidity (kg m−2) between the 22nd century and L20C3M during June through August. Shading in the Figure 1a denotes a signal-to-noise ratio grater than 1.0.

[11] In order to elucidate quantitatively the moisture balance of the ASM in the future climate, the moisture budget is calculated at 0–25°N, 60°–110°E. The right column in Table 2 shows the difference in the moisture balance in the 22nd century and L20C3M. All models show a moisture flux convergence ranging from −2.9 ∼ −23.9 kgm−1s−1. The mean value of eight models is −12.2 kgm−1s−1, which is consistent with the enhanced moisture transports into the Asian monsoon region and the resultant increase in rainfall.

4.2. Meridional Thermal Gradient

[12] The establishment and maintenance of the ASM are driven by a large-scale thermal contrast between the Asian landmass and neighboring oceans [e.g., Li and Yanai, 1996; Ueda and Yasunari, 1998]. There have been a number of studies in which researchers have tried to capture the ASM intensity by the use of several monsoon indexes including those for India monsoon rainfall [Parthasarathy and Mooley, 1978], vertical wind shear [Webster and Yang, 1992] and the MTG [Kawamura, 1998]. In spite of the difference in spatial domain and vertical level, MTG should be similar to the wind shear index according to the geostrophic approximation theory.

[13] Kitoh et al. [1997] indicated that the weakening of vertical wind shear at the times of increased CO2 is caused by a northward shift of the monsoon circulation. To further facilitate the driving force of ASM in the future climate, we present the changes of 200–500 hPa thickness in Figure 4. Most models show an increase in air temperature in the middle-to-upper troposphere over the Asian continent, chiefly confined to the Tibetan Plateau except for the CNRM-CM3 and GFDL-CM2.0 models. The warming is also salient in the tropical region especially over the equatorial Indian Ocean. It is quite interesting to note that the temperature increase over the tropics is relatively larger than those on the Eurasian continent.

Figure 4.

As in Figure 2a, except for 200–500 hPa thickness.

[14] According to Kawamura [1998], MTG is defined by an area-averaged 200–500 hPa thickness between the Tibetan Plateau (20°–40°N, 50°–100°E) and the north Indian Ocean (eq–20°N, 60°–100°E). Since the ASM is also associated with an inter-hemispheric thermal contrast, we redefined the MTG index for the future climate between the Tibetan Plateau (35°–45°N, 60°–100°N) and the tropical Indian Ocean (10°S–10°N, 60°–100°E), along with precipitation and wind changes, which is summarized in Table 2. All models clearly show that the final equilibrium of ASM precipitation increases, while the low-level monsoon flows become weak. The decreasing trend of the MTG is conceivable in all models, which is collocated with the reduction of monsoon circulation.

4.3. Walker Circulation

[15] To further explore the change in the Walker circulation in the future climate, velocity potential (χ) at the 200-hPa level is examined. Figure 5 is a plot of the difference of ∣200-hPa χ∣ between 22nd century and L20C3M, during the boreal summer (JJA) along a latitudinal band of 10°S–25°N. In Figure 5, it can be seen that divergence in the monsoon region (∼60°–150°E) is reduced remarkably, which is consistent with the weakened monsoon circulation seen in Figures 2 and 4. In addition, the descending branch of the Walker circulation over the eastern Pacific Ocean becomes weak. Thus, the tropical east-west circulation in the future climate exhibits a reduced Walker circulation type [Tanaka et al., 2004], including the weakened monsoon circulation and the intensified upward motion over the eastern Pacific Ocean.

Figure 5.

Differences in 200-hPa velocity potential, averaged for a latitudinal band of 10°S–25°N during June through August, between the 22nd century and L20C3M. Thick bold line denotes multi-model ensemble mean. Positive or negative values indicate divergence or convergence respectively.

5. Summary and Discussion

[16] The ASM response to global warming is examined by the use of state-of-the-art atmosphere and eight ocean-coupled GCMs. It is demonstrated that the increase in GHG forcing causes enhanced ASM precipitation by about 13% in the 22nd century, while the large-scale ASM flow becomes weak.

[17] The enhanced moisture transport into the Asian monsoon region is a key mechanism that is responsible for the intensified rainfall. On the other hand, future warming of the air temperature in the middle-to-upper troposphere over the tropical Indian Ocean exceeds that of the Asian continent, which can explain the decrease in MTG and the resultant weaker ASM circulation.

[18] The multi-model analysis of the velocity potential in the upper troposphere reveals that the ASM intensity and Walker circulation over the Pacific Ocean become weak in the 22nd century, which is conjunction with the previous studies [e.g., May, 2004].

[19] Another point to be considered is that an increase in atmospheric water content from the warmer oceans could be more significant for the understanding of the ASM precipitation change [Kitoh et al., 1997]. Several papers suggest that in situ precipitation efficiency associated with evapotranspiration is a more important factor in the rainfall change than the horizontal moisture transport [Douville et al., 2002; Ashrit et al., 2003]. These issues require further clarification and will be examined in future papers.


[20] We acknowledge the international modeling groups for providing their data for analysis, the Program for Climate Model Diagnosis and Intercomparison (PCMDI) for collecting and archiving the model data, the JSC/CLIVAR Working Group on Coupled Modeling (WGCM) and their Coupled Model Intercomparison Project (CMIP) and Climate Simulation Panel for organizing the model data analysis activity, and the IPCC WG1 TSU for technical support. The IPCC Data Archive at Lawrence Livermore National Laboratory is supported by the Office of Science, U.S. Department of Energy. This research was supported by Core Research for Evolutional Science and Technology (JST/CREST) and Grants-in-Aids for Young Scientists (B) (15740288) of Japanese Ministry of Education, Science, Sports and Culture. The reviewers' comments were very helpful in the revision.