This article has been contributed to by US Government employees and their work is in the public domain in the USA.
It is well documented that the East Asian summer monsoon has been experiencing a steady weakening trend in recent decades. Because the Asian summer monsoon (including both East Asian monsoon and South Asian monsoon) is the largest and most pronounced monsoonal system in the world, its change in strength may exert a profound impact on global weather and climate systems, especially on the rainfall pattern in South and East Asia. On the other hand, as a vast elevated landmass, the Tibetan Plateau forms a huge heat source protruding into the free atmosphere. Setting against the backdrop of global climate change, whether or not does the change of this heating affect the change of Asian summer monsoon and thus rainfall distribution? Here we show that the apparent heat source over the Tibetan Plateau is closely correlated with the East Asian summer monsoonal circulation, and that the weakening of the East Asian summer monsoon is closely associated with the decreasing trend of the Tibetan Plateau apparent heat source. Further analysis indicates that the change of rainfall pattern in China in recent decades is consistent with the decreasing of the East Asian summer monsoon.
The Tibetan Plateau is often referred to as the roof of the world. It constitutes about one-sixth of Asia's total land area, and its average elevation is above 4000 m, reaching almost the mid-troposphere. In fact, the Tibetan Plateau plays a role of ‘the world water tower’ Xu et al., 2008. The related change in water resources and environmental in this region will affect socioeconomical development for almost 40% of world population.
The Asian summer monsoon is responsible for precipitation across most South and East Asia. In recent decades, it has been observed that the East Asian summer monsoon circulation exhibits an appreciable decreasing trend (Jiang and Wang, 2005; Ding et al., 2010; Duan et al., 2011; Liu et al., 2012). In response to this change, the precipitation pattern in East Asia also experiences significant changes (Wang et al., 2008; Zhou et al., 2008; Ashfaq et al., 2009; Duan et al., 2013).
Numerous studies have been conducted over the last decade to explain what causes the weakening of the Asian summer monsoon. Using the National Center for Atmospheric Research (NCAR) community atmospheric model version 3 (CAM3) and NOAA Geophysical Physical Fluid Dynamics Laboratory (GFDL) atmospheric model version 2.1 (AM2.1), Li et al. (2010) report that the recent warming in the Tropics, especially the warming associated with the tropical inter-decadal variability centered over the central and eastern Pacific, is a primary cause for the weakening of the Asian summer monsoon since the late 1970s. Lu and Dong (2008) attributed the weakening of the Asian summer monsoon to the weakening of Atlantic thermohaline circulation. Yu et al. (2004) found that the weakening of the Asian summer monsoon corresponded well to the cooling trend in the upper troposphere around 300 hPa in the East Asian region. However, all of these studies bypassed an important question, that is, what is the role of the Tibetan Plateau in the weakening of the Asian summer monsoon? As many studies have shown that the Tibetan Plateau serves as a main driving force for the Asian summer monsoon (Wu and Zhang, 1998; Xu et al., 2010), does the change of the monsoonal circulation reflect the change of the heat source induced by the Tibetan Plateau?
2. Data and data quality control
In this study, we used 2.5° × 2.5° latitude and longitude reanalysis data generated by US National Center for Environmental Prediction–National Center for Atmospheric Research (Kistler et al., 1996) for all atmospheric variable analyses. For precipitation and surface–air temperature analyses, we use a half-century surface station observational data (753 stations after quality control) archived at China Meteorological Administration. A gridded monthly precipitation data from CRU (the University of East Anglia Climatic Research Unit) with spatial resolution of 0.5° × 0.5° and covering the period 1901–2000 are also used (New et al., 2002).
Care must be taken when NCEP-NCAR reanalysis data is used in this study. Two questions are pertinent here: (1) how good is the reanalysis data for the Tibetan Plateau region (Figure 1), as there might be less data available to be assimilated in the model that generated these reanalyses? and (2) was there an artificial trend existed in the reanalysis data due to the ingestion of satellite data since 1979? (Kistler et al., 1996) To ensure that our analyses are not subject to these problems, we compared reanalysis data with rawinsond data recorded at 12 sounding launching stations over the Tibetan Plateau region since 1957. We divided the data into two periods (as pre- and post-satellite eras): 1957–1979 and 1979–2004. Correlations were calculated for spring (March–April–May) temperature between the observed and corresponding temperature calculated from reanalysis data for these two periods respectively. The results show reasonably high correlations between the two datasets, with the correlation coefficients R2 are 0.5369 and 0.6676, respectively, for the two periods (figures not shown). The improved correlation for the latter period is understandable for the fact that with ingest of satellite data, the reanalysis data becomes more accurate. However, this improvement of reanalysis data quality is one-time jump in data quality and should have minimal impact on our trend analysis. We also conducted correlation analysis between apparent heat source (will be defined in the next section) from reanalysis data and surface-sensible heat calculated from surface station data over the Tibetan Plateau (archived in China Meteorological Administration from 1957 to 2004) for two periods: 1957–1978 and 1979–2004. The correlation coefficients R2 are 0.116 and 0.154, respectively. This latter result further supports our hypothesis that the calculated long-term trend of apparent heat source from reanalysis data is trustworthy.
Another issue related to this long-term trend analysis is whether some interannual variability signals, such as El Nino-Southern Oscillation (ENSO), will affect the analyses. Chen et al. (2008) developed an ENSO-removal technique for the trend analysis using NCEP-NCAR reanalysis data, for the purpose of discriminating global warming signal. Li et al. (2007) suggested that ENSO served as one of the factors for driving late-spring or early-summer heating anomalies over the Tibetan Plateau. Other studies [such as Barnett et al. (1989); Li and Yanai (1996); Zhao and Chen (2001); Shaman and Tziperman (2005); and Li et al. (2007)] also pointed to delicate interactions between ENSO and heating over the Tibetan Plateau from various aspects. Therefore, we believe that these ENSO signals are important part of heating anomalies over the Tibetan Plateau, but only as interannual variabilities that superimposed on the long-term trend.
3. Anomalies and long-term trends in spring heating over the Tibetan Plateau
We first calculate the apparent heat source over the Tibetan Plateau. Following Yanai and Johnson (1993) and Yanai and Tomita (1998), the apparent heat source (Q1) and apparent moisture sink (Q2) are defined as, respectively,
where T is the air temperature, q the specific humidity, V = (u, v) the horizontal wind vector, p the pressure, ω the vertical velocity, θ the potential temperature, p0 = 1000 hPa, L and Cp are the latent heat of water vaporization and the specific heat capacity of air at constant pressure, respectively, and κ = R/Cp with R being the gas constant. With simple mathematical manipulations (e.g. see Yanai and Tomita, 1998), one can show that Q1 includes Q2 plus radiative heating and heat sink due to vertical transport of dry-static energy and latent heat. (The whole-column Q1 and Q2 are in units of w m−2).
The color background in Figure 1 indicates land elevation in the eastern hemisphere. The Tibetan Plateau region stands out in dark red, where the average land elevation exceeds 4000 m. The rectangle encompasses the area where Q1 and Q2 are calculated (80–105°E, 25–37.5 N). Although latent heating is very important in controlling the East Asian summer monsoon, as Q1 includes Q2 and radiative heating, we concentrate only on the collective effect of apparent heating (Q1) over the Tibetan Plateau in this paper. In fact, we calculated 50 years' trend for both Q1 and Q2. They essentially showed the similar long-term trends (figure not show). On the basis of this analysis, we decide that only Q1 is calculated to gauge the change of heat source over the Tibetan Plateau in this study.
We now examine how this heat source has changed over the last 60 years. In order to do this, we first calculated time series of Q1 from 1950 to 2010. We then compute an anomaly index of Q1 by conducting a standardization treatment at every grid point in the box in Figure 1, i.e.
where is the standardized Q1 (or standardized index), and s are values of average and standard deviation of Q1. We define the Q1 after standardization as an anomaly index of apparent heat source.
Figure 2 shows the annual time series of anomaly index of Q1 (blue curve) from 1950 to 2010. The fitted average of this heat source index is marked by the solid black curve, depicting the multi-decadal trend of this field. It is clear that since early 1970s, the Tibetan Plateau apparent heat source has presented a decreasing trend. In contrast, before early 1970s, this field displays an increasing trend. It is also noted that there seems a reversal of the decreasing trend for the recent decade (after 2002), so the trend can effectively be divided into three periods (1950–1972, 1973–2002, and 2003–2010). Superimposed on these trends is interannual variability, a combination of ENSO and other climate variability signals. In subpanel Figure 2(A)–(F), the interannual anomalies of summer precipitation (June–July–August—JJA) are composited. The blue and purple-colored areas are the strongest rainbands. It exhibited some correlations with the heating trends over the Tibetan Plateau. As Q1 is in the strengthening phase (upward trending, such as in A, B, and F panels), the rainbands march and extend to further north of China. While in the weakening phase of Q1 (downward trending, such as in C, D, and E panels), the rainbands retreat back to the south.
In this study, we define the years for which the anomaly indices Q1 > 1 and Q1 < −1 as the strong and weak, respectively, anomaly years of apparent heating over the Tibetan Plateau. From Figure 2, we can collect that the years for Q1 > 1 are 1964, 1967–1969, 1974, 1978, 1981, and 1989, and the years for Q1 < −1 are 1950, 1971, 1994, and 2002. We then calculated averaged Q1 distribution for these anomaly years. Figure 3(a) and (b) are vertical-meridional (from 20–80°N) cross-section of Q1, averaged over 95–100°E. It can be seen that the apparent heating over the Tibetan Plateau during Q1 > 1 year (Figure 3(a)) is apparently stronger than that during Q1 < −1 year (Figure 3(b)). So this is true for the apparent heating in the vertical-latitudinal (from 0 to 180°E) distribution (Figure 3(c) vs (d).
4. Anomalous patterns in China summer precipitation
Are there any distinct summer precipitation patterns in correspondence to abnormal spring heating over the Tibetan Plateau? In Figure 4, we computed summer rainfall from 753 rain-gauge stations across China averaged over the years for Q1 > 1 and Q1 < −1, respectively. It can be seen that the ensuing summer precipitation in China presents significant differences when the spring apparent heating over the Tibetan Plateau is abnormal. During the strong heating years, summer precipitation in China can be characterized as ‘North wet–South dry’, while during the weak heating years, the pattern tends to be reversed, that is, ‘North dry–South wet’.
5. Change of monsoonal moisture transport in response to anomalous heating over the Tibetan Plateau
From above analyses, we see a clear lagged correlation between the abnormal spring heating and distinct summer precipitation patterns in China. What is the physical process that connects and dictates these correlations? As we discussed in the introduction, heating over the Tibetan Plateau has long been regarded as the driving force for Asian summer monsoon. In particular, the heat source over the Plateau serves as an ‘air pump’ that attracts warm-moist air from the low-latitude oceans towards the lands in the north. This moisture air is then deflected to the east, which becomes the source of summer rainfall in China and entire East Asia (Xu and Lu, 2010).
To see how moisture transport in response to long-term trends of spring heating over the Tibetan Plateau, we calculate a whole-column moisture transfer vector, whose components are defined as
where g is gravitational acceleration, (u, v) are respectively the zonal and meridional wind components, q the specific humidity (in unit: g kg−1), and Ps the surface pressure. (qu and qv are in units of kg m− 1 s− 1.)
With these defined physical variables, we can calculate the correlation between the apparent heat source over the Tibetan Plateau and the water vapor fluxes. Figure 5 is the calculated horizontal distribution of correlation vector between the apparent heat source over the Tibetan Plateau and the whole column water vapor flux averaged over March to August of 1950–2010. The yellow (or green) color highlights the positive (or negative) areas that passed the 90% significance tests for the correlations. The northward correlation vectors in the yellow region in the south of the Tibetan Plateau, especially in the Bay of Bangle, represent water vapor fluxes that are ‘pumped up’ by the heat source of the Tibetan Plateau. These vectors then turn slightly eastward, becoming the classical southwestly moisture fluxes of the East Asian summer monsoon. From the distribution of these correlation vectors, one can clearly see that the heat source over the Tibetan Plateau (Q1) correlated well with the Asian summer monsoonal flow.
The above result simply re-demonstrates the classic description of the Tibetan Plateau acting as a heat source to drive the monsoonal circulation. We now examine how the change of this heat source impacts the monsoonal moisture transport. In Figure 1, we have shown that an upward trend of heating over the Tibetan Plateau during 1950–1972, followed by a downward trend of heating during 1973–2008. By calculating the anomalies of water vapor fluxes and summer precipitations corresponding to these two time periods, we can see two different patterns (Figure 6(a) vs (b)) of monsoonal moisture transport and summer precipitation. During the period of strengthening of the Tibetan apparent heat source (1950–1972), the water vapor fluxes display a normal monsoonal flow pattern, similar to that shown in Figure 6. However, corresponding to the decreasing period (1973–2000; we only have CRU data up to 2000) of the apparent heat source over the Tibetan Plateau, northeastly water vapor fluxes in the eastern China and the Bay of Bangle are found, which present an anti-monsoonal flow pattern. During this period, southeast and west China are abnormally wet, whereas Northeast China extending all the way to the Southeast Asian Peninsula presents an abnormally dry pattern. These results imply that the decreasing trend of apparent heat source over the Tibetan Plateau is a possible cause for the weakening of Asian summer monsoon. In response to this weakening of monsoon, the entire Asian continent may experience abnormal rainfall patterns.
6. Long-term trend of precipitation distribution over China
Finally, we carefully computed long-term trend of summer precipitation change rate during 1957–2010, shown in Figure 7. The red (blue) dots indicate a positive (negative) change rate, while the size of the dots depicts the amount of rainfall change over a decade (change rate). The precipitation trend in China displays three distinctive regions, which is not entirely in agreement with the anomalous precipitation distribution shown in Figures 4 and 6. However, summer rainfall distribution in Figure 4 is calculated based on anomaly indices Q1 > 1 and Q1 < −1, corresponding to two opposite heating anomaly periods, while Figure 7 representing the distribution of long-term precipitation trend over 1957–2010. Region A presents a ‘strong wet-trend’ area; region B presents a ‘strong dry-trend’ area; and region C presents a ‘weak wet-trend’ area. The trend of rain-suppression in the B belt is consistent with the decreasing of southwesterly moisture transport, corresponding to the weakening of East Asian monsoonal circulation.
7. Conclusion and discussions
In this study, we analyzed correlation between the heat source over the Tibetan Plateau and moisture transport due to East Asian summer monsoon and precipitation patterns in China. We first found that anomalies of spring heating over the Tibetan Plateau result in distinctive summer precipitation patterns in China. These anomalous precipitation patterns can be traced back to the response of moisture transport to the anomalous heating over the Tibetan Plateau, which seems consistent with the classic understanding of Plateau-monsoon dynamics. Further analysis of the long-term trend of the apparent heat source over the Tibetan Plateau indicates that the change of the heat source results in two different anomalous monsoonal flows. In particular, the decreasing trend of the apparent heat source over the Tibetan Plateau in recent decades may be responsible for the weakening of the East Asian summer monsoon. Another side discovery of this long-term trend analysis, which has not been discussed in this study, is that there has been a reversal of the decreasing trend of apparent heating over the Tibetan Plateau since 2003 to present. Of course, longer record of observational data is needed to confirm the reversed trend.
In response to the weakening of East Asian summer monsoon, China's precipitation presents a pattern with three-trend regions. The rain-diminishing region is consistent with the decreased moisture transport because of the weakening of Asian summer monsoon, which forms a dry-trend belt in the north and central China. This result may have a profound implication for China's environment and sustainability developments.
This research was Jointly supported by projects of Nature Science Fund of China (No. 41130960, 41275050), the China-Japan intergovernmental cooperational project (JICA), and the Key Project of Basic Scientific Research and Operation fund of the Chinese Academy of Meteorological Sciences (2011Z001).