The large-scale Asian summer monsoon circulation has experienced a weakening tendency in recent decades. Using observed data and output from model experiments with the atmospheric component of the NCEP Climate Forecast System, the authors show that a relatively smaller warming in Asia compared to the surrounding regions may be a plausible reason for this change in the monsoon. Although the surface temperature over Asia has increased, the landmass has become a relative “heat sink” because of the larger warming in other regions of the world. Indeed, over Asia, the vertically integrated tropospheric temperature in the most recent decades is colder than that in the earlier decades, a feature different from the characteristics outside Asia.
The Asian summer monsoon, which exerts a large impact on agricultural production and hence the economy within and outside Asian countries, is in concord with the seasonal reversal of larger-scale atmospheric heating (Webster et al. 1998; Ding and Chan 2005). The fundamental mechanism of the Asian monsoon is land–sea thermal contrast (Webster 1987). By the early 1900s, the major advances in monsoon study include the discovery of the strong links of monsoon with the Himalayan and Eurasian snow extent and with El Niño–Southern Oscillation (ENSO). The former indicates the thermal condition over the Asian landmass (Blanford 1884; Walker 1924). The Asian summer monsoon has also been considered as a low-frequency baroclinic–Rossby wave response to the heating over Asia. Based on the close relationship between atmospheric heating and regional vertical wind shear (Webster 1972; Gill 1980), Webster and Yang (1992) defined an index (the W–Y index) by the vertical shear of zonal winds at 850- and 200-hPa levels averaged over 0°–20°N, 40°–110°E. It depicts the thermally driven nature of monsoon and is widely used in both research and operational prediction of the Asian summer monsoon. ENSO is closely related to the Asian monsoon on interannual time scales through influencing the east–west displacement of large-scale heat sources in the tropics (Walker 1924; Rasmusson and Carpenter 1983).
During the last decades, global ocean and land surface temperatures have increased and influenced the atmospheric circulation on various scales (Hoerling et al. 2008; Wang et al. 2008). The weakening Asian summer monsoon and the monsoon–ENSO relationship may be one of the consequences (Krishnamurthy and Goswami 2000; Kripalani et al. 2001; Kinter et al. 2002; Wu 2005; Kucharski et al. 2006; Wang and Ding 2006; Fan et al. 2010; Wu et al. 2010). Kripalani and Kulkarni (1997) showed that the decadal variability of the Indian summer monsoon rainfall might modulate the ENSO–monsoon relationship. For example, the impact of El Niño (La Niña) is stronger during the below (above)-normal rainfall epochs. Kumar et al. (1999) viewed that global warming led to a warmer Eurasian continent and a more southeastward shift of the Pacific Walker circulation, playing an important role that favors normal Asian monsoon conditions after the late 1970s. Chang et al. (2001) questioned the impact of interdecadal change of the Walker cell and emphasized the importance of Eurasian surface temperature for the Asian monsoon after the late 1970s. Zhao et al. (2010) documented that the decrease in temperature over the Tibetan Plateau and adjacent areas under the global warming background weakened the land–sea thermal contrast, accompanied by weakened Asian summer monsoon circulation. Other studies, however, questioned the weakening of the Asian summer monsoon and ENSO–monsoon relationship and the impact of global warming on the Asian monsoon (e.g., Kripalani et al. 2003; Goswami and Xavier 2005; Zhou et al. 2009).
Furthermore, the impact of tropospheric temperature (TT) anomalies over the Eurasian continent on the Asian summer monsoon has received substantial attention during the past decades (Verma 1980; Parthasarathy et al. 1990; Li and Yanai 1996; Singh and Chattopadhyay 1998; Liu and Yanai 2001; Yu et al. 2004; Zhou and Zhang 2009). It has been documented that the Asian monsoon is positively and significantly correlated with the preceding May and June–August (JJA) TTs over Asia and that enhanced (weakened) land–sea thermal contrast in the upper troposphere occurs during the summer months of strong (weak) monsoon (Ding et al. 2009). Motivated by the previous studies on weakening monsoon–ENSO relationship and the importance of thermal condition over the Eurasian continent for the Asian monsoon, we conduct this study to examine the influence of the thermal condition over Asia under the global warming background on the broad-scale Asian summer monsoon circulation measured by the W–Y index.
2. Data and methods
The datasets used in this study include the monthly air temperature and winds from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) reanalysis (Kalnay et al. 1996) and the National Oceanic and Atmospheric Administration (NOAA) Climate Prediction Center gauge-based monthly global land precipitation data (Chen et al. 2002). They also include the Climate Precipitation Center monthly land surface air temperature analysis (Fan and Van den Dool 2008). Simulations with eight ensemble members by the atmospheric general circulation model of the NCEP Climate Forecast System version 2 (CFSv2) with monthly varying CO2 are used to determine the impact of global warming on the Asian monsoon as well. All datasets cover the same time period of 1950–2010.
3. Weakening Asian summer monsoon during 1950–2010
Figure 1a presents the normalized time series of the W–Y index and the corresponding linear trend observed for 1950–2010. An apparent feature is the observed negative trend, which seems to be related mainly to the major shift that occurred in 1995. Positive values mainly appeared in 1950–94, and negative values frequently occurred in 1995–2010. Indeed, the averaged value of the W–Y index was 25.9 before 1995 and 24.2 afterward.
In this study, we analyze the dynamical monsoon index, instead of the monsoon rainfall index, because the dynamical index measures the large-scale features of monsoon—especially the monsoon circulation pattern—more appropriately. We choose the W–Y monsoon index because it depicts well the variability of summer monsoon circulation over the broad tropical Asian regions (Wang and Fan 1999; Lau et al. 2000). In contrast, the popular all-India monsoon rainfall (AIMR) index is appropriate for measuring the variability of monsoon over India but does not include the rainfall over the Bay of Bengal, where the maximum monsoon heating is located. Additionally, it cannot represent the variability of the East Asian summer monsoon. Nevertheless, since precipitation is one of the important meteorological parameters describing the variability of monsoon climate, we also discuss the difference in JJA precipitation between 1995 and 2010 and 1950 and 1994. Figure 1b clearly shows negative anomalies of rainfall observed around the Bay of Bengal, which indicates a reduction in monsoon rainfall in 1995–2010 and is generally in agreement with the weakened monsoon circulation. The observed rainfall anomalies over much of India are also negative. In addition, rainfall anomalies show a positive–negative pattern across the Yangtze–Huaihe River basin (YHRB) in eastern China. The enhanced rainfall over the south of YHRB and the reduced rainfall over the north of YHRB generally correspond to the weakened East Asian summer monsoon circulation (e.g., Zhou et al. 2009).
4. Relationship between Asian thermal condition and weakening Asian monsoon
Warmer Asian landmass usually leads to a larger land–sea thermal gradient in summer and thereby a stronger Asian summer monsoon, especially during the recent period of weakening of the monsoon–ENSO relationship (Kumar et al. 1999). In this context, a weakened Asian summer monsoon should be associated with a colder condition over the Asian continent. However, it is seen from the normalized time series of surface air temperature over Asian land (15°–45°N, 60°–120°E; referred to as STA) that the STA has a positive, instead of negative, trend during the recent 61 yr (Fig. 2; data from Fan and Van den Dool 2008). Therefore, there seems to be a paradox about the relationship between the warming Asian landmass and the weakening Asian summer monsoon.
The purple line in Fig. 2 shows the normalized time series of global mean surface air temperature over land (averaged over 60°S–80°N, 0°–360°E). The global mean temperature has been increasing during the recent 61 yr, consistent with a trend toward warmer temperatures. Global mean temperature is also highly correlated with STA (R = 0.91), indicating that the STA varies with the global warming. However, it is also observed that the positive trend of global mean land surface air temperature is larger than that of STA. That is, although the surface temperature over Asia has increased, the continent has become “relatively cooler” when the changes in temperatures in other regions are factored in. The blue line in Fig. 2 shows the normalized time series of residual surface air temperature over land in which the global mean has been removed. Surprisingly, a negative trend appears. Thus, under the background of global warming, the Asian landmass tends to become a relative heat sink, accompanied by a reduced land–sea thermal contrast and weakened Asian summer monsoon.
Since the Asian summer monsoon is a vertically deep system and is related strongly to the diabatic heating of the atmosphere and the tropospheric temperature, we also examine the vertically integrated tropospheric temperature from 850 to 200 mb in JJA. The vertically integrated temperature is supreme to surface temperature for measuring the thermal condition of the entire troposphere. Figure 3 demonstrates that while the tropospheric temperature increases across most of the globe, it decreases clearly over the Asian continent (smaller decreases in TT are also observed over northern Europe and a small portion of the North Pacific). That is, the meridional and zonal thermal contrasts between the Asian continent and its adjacent regions are becoming smaller. The TT averaged over Asia (20°–55°N, 60°–120°E) is positively and significantly correlated with the residual STA with the global means removed (exceeding the 95% confidence level). The correlation between the TT of the negative anomaly center (40°N, 110°E) and the residual STA even significantly exceeds the 99.9% confidence level for 1950–2010. While a tropospheric warming over Asia and a cooling over tropical oceans and the North Pacific should lead to a stronger Asian summer monsoon, the opposite features shown in Fig. 3 give rise to weakened Asian summer monsoon.
5. Weakening monsoon simulated by NCEP model
Previous studies have shown that the NCEP CFS, which is an atmosphere–ocean–land fully coupled climate prediction system, has noticeable skills in simulating and predicting the Asian monsoon (e.g., Yang et al. 2008, 2011). The simulated variations of the W–Y monsoon index in the mean of eight ensemble members of Atmospheric Model Intercomparison Project (AMIP)-type experiments using the atmospheric model of NCEP CFSv2 with time-varying CO2 are showed in Fig. 4a. The agreement between observation and model output is evident, especially in the low-frequency component, including the positive values before 1995 and the negative values afterward. The simulated liner trend is also close to the observed. This model–observation agreement suggests that the long-term change in the observed W–Y index is related to the varying CO2 forcing. The CFSv2 atmospheric model with varying CO2 also simulates the reduced rainfall over most of India and the Bay of Bengal. It simulates the positive–negative precipitation anomalies across the YHRB in China as well (Fig. 4b), although the pattern is less apparent than the observed. That is, with varying CO2, the CFSv2 atmospheric model captures the weakening tendency of the Asian summer monsoon in the recent 61 yr, representing a link of the weakening monsoon to the global warming. As concluded in the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment Report, most of the observed increase in global averaged temperatures since the midtwentieth century is likely due to the observed increase in anthropogenic greenhouse gas concentrations (Alley et al. 2007). Since CO2 is the main factor of greenhouse gases, the above feature obtained by varying CO2 validates that the weakening Asian summer monsoon is related to the global warming. We also note, however, that since the observed sea surface temperature is specified in the AMIP simulations and sea surface temperature variations themselves may be a consequence of the increase in CO2 and in turn may affect atmospheric variability, our analysis does not separate the direct and the indirect roles of CO2.
The NCEP–NCAR reanalysis and output from AMIP experiments with the atmospheric model of NCEP CFSv2 with varying CO2, among others, are used to study the change in Asian summer monsoon and its relationship with the thermal condition over Asia under the global warming background. It is demonstrated that the large-scale Asian summer monsoon circulation measured by the W–Y index has exhibited a weakening tendency during the recent decades, and that this monsoon tendency is related to the relative “cooling trend” of the thermal condition over Asia. Temperatures have increased nearly globally in the past decades. However, under this global warming background, the increase in the surface temperature over the Asian landmass is relatively small compared to its adjacent regions. Over this “relative cold” landmass, heat fluxes from the land surface to the troposphere are relatively small. Indeed, the tropospheric temperature over Asia has become colder in the recent decades. As a consequence, the meridional and zonal land–sea thermal contrasts are reduced and the Asian summer monsoon becomes weaker. These features are captured successfully by the atmospheric model of the NCEP CFSv2 with varying CO2, indicating that the weakening of the Asian summer monsoon is related to the spatially uneven effects of greenhouse gases. The concept of differential warming is similar to the argument proposed by Vecchi et al. (2008), in which the Atlantic hurricane activity is determined by the relative tropical Atlantic SST anomalies that are different from SST anomalies averaged over global tropical oceans and that it does not depend solely on the local SST anomalies in the tropical Atlantic.
It should also be pointed out that the weakening of monsoon is not significantly related to the long-term change in the tropical central-eastern Pacific SST that is often used to measure ENSO. Moreover, the weakening of the broad-scale Asian summer monsoon (measured by the W–Y index) is more strongly related to the weakening of monsoon over some regions, such as Southeast Asia, than that over other regions. Details of these features will be reported in a separate paper.
Nevertheless, a question that has not been answered is why the warming over Asian landmass is smaller than the warming over other places under the global warming background. Although previous studies have documented an impact of the North Atlantic SST on global warming and the thermal condition over the Eurasian continent (e.g., Chang et al. 2001; Zhang et al. 2007; Kravtsov and Spannagle 2008; Rajeevan and Sridhar 2008), other influences should also be factored in. Further investigations to understand the other possible factors, such as aerosols (Lau et al. 2006; Lau and Kim 2006; Massimo et al. 2011) related to the relatively small warming over Asia and thus the weakening Asian summer monsoon, are undoubtedly necessary.
The authors appreciate the several constructive discussions with Dr. Dong Xiao of the China Meteorological Administration (CMA). Comments from three anonymous reviewers have improved the overall quality of this paper. This study was jointly supported by the National Natural Science Foundation of China under Grant 40921003, the International S&T Cooperation Project of the Ministry of Science and Technology of China under Grant 2009DFA21430, the Basic Research Fund of the Chinese Academy of Meteorological Sciences under Grant 2010Z001, and the NOAA-CMA bilateral program.