Journal of Geophysical Research: Atmospheres

Influences of the Atlantic Ocean on the summer precipitation of the southeastern Tibetan Plateau

Authors

  • Gao Ya,

    Corresponding author
    1. Nansen-Zhu International Research Centre, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
    2. University of Chinese Academy of Sciences, Beijing, China
    3. Climate Change Research Center, Chinese Academy of Sciences, Beijing, China
    • Corresponding authors: G. Ya, Nansen-Zhu International Research Centre, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China. E-mail: (gaoy@mail.iap.ac.cn)

    Search for more papers by this author
  • Wang Huijun,

    Corresponding author
    1. Nansen-Zhu International Research Centre, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
    2. Climate Change Research Center, Chinese Academy of Sciences, Beijing, China
    • Corresponding authors: G. Ya, Nansen-Zhu International Research Centre, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China. E-mail: (gaoy@mail.iap.ac.cn)

    Search for more papers by this author
  • Li Shuanglin

    1. Nansen-Zhu International Research Centre, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing, China
    2. Climate Change Research Center, Chinese Academy of Sciences, Beijing, China
    Search for more papers by this author

Abstract

[1] The southeastern Tibetan Plateau is one of the predominant summer rainfall regions in the world and is also the crucial water vapor channel of the Asian summer monsoon. The rainfall variability in the region influences not only the local communities but also downstream communities in East Asia. However, previous studies have exhibited large rainfall biases in this region in state-of-the-art climate models. Understanding the observed rainfall variability provides an opportunity to identify the origin of model biases and to lay a foundation for improving model performance. In this study, the interannual variability of the summer precipitation (May–September) over the southeastern Tibetan Plateau was investigated based on National Centers for Environmental Prediction/National Center for Atmospheric Research reanalysis monthly mean data from 1979 to 2010. The associated atmospheric circulation anomalies of the southeastern Tibetan Plateau summer precipitation (SET_PR) display a North Atlantic Ocean-Europe-Asia teleconnection pattern, indicating a possible role of the Atlantic climate in the SET_PR. Further studies have revealed that the Atlantic sea surface temperature (SST) anomalies have the greatest influence on the SET_PR via the Rossby wave response, whereas the SST anomalies in the Indo-Pacific have less of an influence on the SET_PR because their main impacts are confined to the western North Pacific subtropical high and the monsoonal circulation there. This paper also documents the detailed spatial pattern of the atmospheric circulation anomalies associated with the SET_PR and Atlantic SST year-to-year variability.

1 Introduction

[2] The region around the southeastern Tibetan Plateau (20°N–30°N, 85°E–105°E) is one of the predominant summer rainfall regions in the world. The rainfall anomalies in this region often result in serious floods and droughts, which have caused numerous disasters. This region is also the most crucial water vapor channel of the Asian summer monsoon. The rainfall of the southeastern Tibetan Plateau is closely associated with the East Asian precipitation during boreal summer. Previous studies have illustrated large rainfall biases in this region in state-of-the-art models [Jiang et al., 2004; Xu et al., 2007]. These upstream model biases substantially influence downstream East Asia rainfall simulations and, thus, may play a crucial role in model biases in the simulated East Asian rainfall variability. Therefore, understanding the summer rainfall variability in the region is important for not only the local communities but also the remote downstream areas of East Asia. This understanding is also important for improving model performance in simulations of East Asian summer rainfall, which is still a challenge for the current generation of climate models.

[3] Geographically, the southeastern Tibetan Plateau primarily comprises Bangladesh, Burma, and southwestern China and is located between the South Asian and East Asian monsoon regions. There have been connections between the South and Southeast Asian, East Asian, and Indian summer monsoons [Kripalani and Kulkarni, 1997a, 1998, 2001]. The highly elevated Himalayas and Tibetan Plateau situated in the north and the water vapor transported by the summer monsoon winds from the ocean provide favorable conditions for the development of convection.

[4] The Tibetan Plateau, with an average elevation of more than 4000 m above sea level and an area of 2.3 million square kilometers, is considered one of the most prominent features on the Earth. Because of its height and area, the plateau plays important roles in determining the formation and variation of regional weather and climate in eastern and southern Asia as well as the Northern Hemisphere atmospheric circulation in general [Manabe and Terpstra, 1974; Ye and Gao, 1979; Murakami, 1987; Tao and Chen, 1987; Manabe and Broccoli, 1990; Kutzbach et al., 1993]. In recent years, many studies have examined the interannual variability of precipitation over the Tibetan Plateau, which is associated with the upstream location of the signals and the India-Burma Trough [Bothe et al., 2010, 2011; Wang et al., 2011; Zhu et al., 2011; Yang et al., 2004]. The dynamical and thermal processes over the Tibetan Plateau influence the distinct local weather and climate phenomena over southwestern China on the eastern edge of the plateau, an area with complex topography and maximum cloud amount and minimum sunlight within China [Xu, 1991; Li and Gao, 2007; Li et al., 2010].

[5] The Tibetan Plateau, as Asia's “water tower,” plays an important role in the hydrology and agriculture in the downstream regions. So understanding the precipitation variability of the Tibetan Plateau and its surrounding area is crucial. There are three routes of moisture supply to the Tibetan Plateau [Simmonds et al., 1999]: the Arabian Sea and the Bay of Bengal, the South China Sea, and the midlatitude westerlies. These routes are closely linked to the Asian monsoon circulation and are further subject to the influence of long-distance teleconnections and stationary wave activity. A number of studies have shown a link from the North Atlantic and Europe to Asia and eastern Asia to North America through wave trains [Liu and Yin, 2001; Wakabayashi and Kawamura, 2004; Ding and Wang, 2005; Sato and Takahashi, 2006; Ding, 2007; Sun et al., 2008; Bothe et al., 2010, 2011, Sun and Wang, 2012].

[6] Although meteorological knowledge of the Tibetan Plateau has been acquired along with the continuous accumulation of various data collected over the plateau, our understanding of the climatic variation across the plateau is still far from complete. Previous studies suggest that the North Atlantic sea surface temperature (SST) exerts a significant impact on South Asian summer rainfall [Goswami et al., 2006; Lu et al., 2006; Li et al., 2008; Wang et al., 2009; Luo et al., 2011], which implies a possible impact from the North Atlantic SST. From the perspective of atmospheric circulation, this region is located beneath the subtropical upper tropospheric Asian jet originating from the subtropical Atlantic. The jet acts as a waveguide path for a wave train propagating from the North Atlantic toward East Asia [Chang et al., 2001; Yang et al., 2004; Ding and Wang, 2005; Bothe et al., 2010, 2011; Ding et al., 2011; Zhu et al., 2011], which also implies that the North Atlantic Ocean may be influential. However, there have been few studies on the association of the Atlantic Ocean with the Tibetan Plateau and surrounding areas. In other words, how the Atlantic Ocean affects the southeastern Tibetan Plateau is unclear. This consideration motivates the present study, which aims to study the connection between the Atlantic Ocean and the summer rainfall in the southeastern Tibetan Plateau, particularly how the Atlantic Ocean influences the spatiotemporal distribution of the summer precipitation of the southeastern Tibetan Plateau. These questions are significant for understanding the climate variability and predictability in Asia.

[7] The paper is organized as follows. Section 2 introduces the data sets and methods. Section 3 describes the spatiotemporal distribution of the southeastern Tibetan Plateau summer precipitation and the associated atmospheric circulation. Section 4 discusses the impacts of the Atlantic Ocean. Section 5 provides further analysis of the associated atmospheric circulation variability, and the last section summarizes the main conclusions.

2 Data and Methods

[8] This study focuses on the boreal summer from May to September during the period from 1979 to 2010. The monthly National Centers for Environmental Prediction (NCEP)-National Center for Atmospheric Research (NCAR) data set [Kalnay et al., 1996], with a horizontal resolution of 2.5° latitude by 2.5° longitude and 17 vertical pressure levels, is used. The precipitation data are derived from the Global Precipitation Climatology Project Version 2.2 monthly precipitation analysis [Adler et al., 2003], which has a horizontal resolution of 2.5° latitude by 2.5° longitude. The SST data set provided by NOAA (NOAA Extended Reconstructed Sea Surface Temperature V3b) [Smith et al., 2008] has a horizontal resolution of 2° latitude by 2° longitude. The outgoing longwave radiation (OLR) data set, which has a horizontal resolution of 2.5° latitude by 2.5° longitude, is also provided by NOAA (NOAA Interpolated Outgoing Longwave Radiation) [Liebmann, 1996]. All of the above monthly data sets are first calculated to the seasonal mean and then are detrended before analysis.

3 The Spatiotemporal Distribution of the Southeastern Tibetan Plateau Summer Precipitation and Associated Atmospheric Circulation

[9] As the interannual variability of the southeastern Tibetan Plateau summer precipitation from May to September during the period of 1979–2010 is focused in this study (Figure 1a), all the following analyses are based on detrended data. Figure 1c shows the detrended and normalized time series of the Precipitation Index (PRI), defined as the regional mean of the southeastern Tibetan Plateau precipitation (SET_PR) over the domain (20°N–30°N, 85°E–105°E). The correlation coefficient between the PRI and the first empirical orthogonal function (EOF) principal component of the SET_PR is 0.97 for the 1979–2010 period. Thus, the PRI can illustrate the spatial and temporal distribution of the SET_PR well. There are two strong climatological rainfall centers located in northern Burma and Bangladesh near 25°N–26°N, respectively (Figure 1b), and main rainfall occurs in regions with lower topography than 1500 m. One additional feature is the homogeneous variability of the SET_PR in the entire region.

Figure 1.

(a) Temporal evolution of the monthly mean precipitation averaged over the southeast Tibetan Plateau region from 1979 to 2010 (range: 20°N–30°N, 85°E–105°E). (b) The climatology of the southeast Tibetan Plateau summer precipitation from 1979 to 2010; the black lines represent the geopotential height at the surface (unit: m). (c) Detrended and normalized time series of the southeast Tibetan Plateau summer precipitation (as the Precipitation Index-PRI) during the 1979–2010 period (range: 20°N–30°N, 85°E–105°E).

[10] In summer, the Asian summer monsoon circulation brings a large amount of moisture from the surrounding oceans to the southeastern Tibetan Plateau. Moisture from the Arabian Sea and the Bay of Bengal can be transported onto the Tibetan Plateau via the Indian summer monsoon. Meanwhile, the southeastern Asian monsoon and the western North Pacific subtropical high transport warm vapor from the South China Sea and the western Pacific toward the Tibetan Plateau [Sun et al., 2011; Zhu et al., 2011; Wang and Chen, 2012]. Figure 2 shows the regression patterns of the sea level pressure (SLP) and 850 hPa wind (the wind field is significant at a 95% confidence level) against the PRI. Negative (positive) pressure anomalies are displayed over Europe, West Asia, central Africa, the tropical eastern Atlantic Ocean, and the southern Atlantic Ocean, corresponding to the positive (negative) phase of the PRI. An anomalous cyclonic circulation exists over the west of the Tibetan Plateau at the lower level. The upper level presents an opposite pattern, in which an anomalous anticyclonic circulation exists over the west of the Tibetan Plateau, and an anomalous cyclonic circulation is located over the east of the plateau (Figure 3); this feature corresponds to a southerly anomaly at the lower level and a northerly anomaly at the higher level near 90°E. The figure shows a vertical circulation with an updraft over the southeastern Tibetan Plateau and increasing rainfall over the region, while a downdraft and less precipitation occurs over the southern Bay of Bengal. The entire tropospheric wind shows an anomalous easterly from the Arctic Ocean along the northern edge of Eurasia south to 40°N. Thus, corresponding to more rainfall over the southeastern Tibetan Plateau, the atmospheric circulation forms a profound anomalous easterly that expands from the Arctic Ocean to the North Atlantic Ocean along the edge of northern Eurasia through the entire troposphere. The easterly anomalies along the edge of northern Eurasia turn to an anomalous westerly over the Atlantic Ocean and are bifurcated through the northern Mediterranean region. One anomaly, going through the Black Sea, forms an anomalous anticyclone in the Mediterranean area over the lower level, corresponding to the 200 hPa wind and pressure field over the higher level. The other anomaly, via Arabia to the Red Sea and the Gulf of Aden, converges with the moisture from the Arabian Sea and induces more rainfall over the southeastern Tibetan Plateau along its southern edge. Compared with the 850 hPa wind field climatology (figure not shown), Figure 2 demonstrates that the lower level wind path is more northward, so the moisture transport is more northward than normal, thus reducing the moisture transport into the Indo-China Peninsula and the Maritime Continent over Southeast Asia.

Figure 2.

Regression patterns of the detrended SLP and 850 hPa wind fields with the PRI. (The light shading denotes regions with correlations that are positively or negatively significant at the 95% confidence level. The heavy shading denotes regions with correlations that are positively or negatively significant at the 99% confidence level. The entire wind field denotes regions significant at the 95% confidence level.)

Figure 3.

Regression patterns of detrended (a) 200 hPa potential height and anomalies of Plumb's stationary wave flux at 200 hPa ([m2/s2], arrows) and (b) 500 hPa potential height fields with the PRI. (The light (heavy) shading denotes regions with correlations that are positively (negatively) significant at the 95% confidence level.)

[11] Figure 3b shows the regression pattern of the 500 hPa height field with the PRI, which emerges from a North Atlantic Ocean-Europe-Africa meridional teleconnection from north to south. There is an anomalous anticyclone from the Arctic Ocean to Iceland, an anomalous cyclone over Europe, anomalous anticyclones over the Mediterranean and the plateau of Iran, and an anomalous cyclone over central Africa. Moreover, a zonal circumglobal teleconnection exists in the middle latitudes, but it is not significant everywhere at the 95% confidence level. The corresponding 200 hPa height field displays the North Atlantic Ocean-Europe-Central-West Asia teleconnection pattern even more prominently (Figure 3a). Moreover, a circumglobal zonal wave pattern also exists in the middle latitudes at the 200 hPa height, and the pattern is coherent with the pattern Ding and Wang [2005] found. By analyzing 56 year NCEP-NCAR reanalysis data, they revealed a recurrent circumglobal teleconnection (CGT) in the midlatitudes of the Northern Hemisphere during summertime that supports the remote influence on the Tibetan Plateau. Additionally, Ding and Wang proposed that the circumglobal teleconnections are best defined in August, when the strongest connection between the Atlantic and central Asia exists. This pattern in August is similar to that shown in Figure 3a, but the teleconnection pattern in Figure 3a is not significant at the 95% confidence level over eastern central Asia in the middle latitudes. Thus, in contrast to the CGT pattern, the southeastern Tibetan Plateau summer precipitation is linked to the Atlantic Ocean-Europe-Central-West Asia teleconnection pattern in the upper level. Furthermore, the EOF analysis is applied to the detrended 200 hPa height field in the Northern Hemisphere during summertime. We find that the spatial distribution of the third EOF mode resembles Figure 3b, where the correlation coefficient of the third principle component with the PRI is 0.53 and is significant at the 99% confidence level. Thus, the Atlantic Ocean-Europe-Central-West Asia meridional pattern is coherent with the southeastern Tibetan Plateau summer precipitation. The dynamic of the southeastern Tibetan Plateau precipitation is related to stationary wave activity bridging the North Atlantic and the Eurasian continent (Figure 3a). In the Bothe et al. [2010] study, the dry case is related to the stationary and transient eddy activity bridging the North Atlantic and the Eurasian continent. Compared with Bothe et al., the stationary wave from the North Atlantic Ocean to the Tibetan Plateau is discontinuous at the Iranian plateau, and the route is more eastward in this study. There are two parts of energy reaching the southeastern Tibetan Plateau. One portion is via the North Atlantic Ocean-Europe-Central-West Asia wave train, which transports the stationary wave energy from the North Atlantic Ocean to the south of the Mediterranean Sea and the Iranian plateau. The other portion is from the Arabia and the Iranian plateau to the Tibetan Plateau.

[12] To better illustrate the associated large-scale atmospheric circulation of the southeastern Tibetan Plateau summer precipitation, according to Figure 3a, we choose three key points to perform the regression with the 200 hPa height field (Figure 4). The three indices are defined as the area mean geopotential heights of Iceland (25°W–15°W, 63°N–67°N), the Mediterranean (20°E, 32.5°N), and the plateau of Iran (70°E, 32.5°N). Figure 4a shows the regression pattern of the Iceland Index against the 200 hPa height field with an evident Atlantic Ocean-Europe-Central Asia teleconnection pattern similar to that in Figure 3b. However, the Northern Hemisphere circumglobal teleconnection signal is very strong to the east of 90°E. Figures 4b and 4c show the regression patterns of the Mediterranean Index and the Plateau of Iran Index against the 200 hPa height field, respectively. Compared with Figure 4a, Figures 4b and 4c both show that the North Atlantic high is weak and the circumglobal teleconnections in the Northern Hemisphere are strong. Additionally, the correlation coefficients of the Iceland Index, the Mediterranean Index, and the Plateau of Iran Index with the PRI are 0.59, 0.47, and 0.47, respectively, all of them exceeding the 99% significance level. Hence, the primary effect of the atmospheric circulation on the southeastern Tibetan Plateau summer precipitation comes from the North Atlantic. Besides, the central and western Asia also influence the SET_PR, but they excessively emphasize the South Asia-North Pacific Ocean-North America patterns and exaggerate the zonal distribution. Thus, the main impact of the atmospheric circulation on the southeastern Tibetan Plateau summer precipitation comes from the North Atlantic Ocean. In the upper level of the troposphere, it appears that the South Asia high is more westward than normal and exhibits the Iranian high pattern.

Figure 4.

One-point regression maps showing the regression pattern of the 200 hPa height with base points at (a) Iceland (25°W–15°W, 63°N–67°N), (b) the Mediterranean (20°E, 32.5°N), and (c) the plateau of Iran (70°E, 32.5°N). (The light shading denotes regions with correlations that are positively or negatively significant at the 95% confidence level. The heavy shading denotes regions with correlations that are positively or negatively significant at the 99% confidence level.)

4 The Impact of the Atlantic Ocean on the Southeastern Tibetan Plateau Summer Precipitation

[13] We now analyze the SST characteristics associated with the SET_PR. The tropical sea surface temperature anomalies, especially the Indian Ocean Dipole (IOD) Mode [ÝSAJI et al., 1999] and the El Niño–Southern Oscillation (ENSO) [Webster and Yang, 1992], can influence the South Asian precipitation and the moisture supply for the Tibetan Plateau. Hong et al. [2008] illustrated that the positive phase of the IOD or ENSO can reduce the moisture inflows from the origins of the southerly water vapor to the Tibetan Plateau, while the inflows are strengthened during negative episodes. The regression pattern of the detrended SST field with the PRI is shown in Figure 5. The figure displays five key positive SST anomaly sections, including the northern Atlantic Ocean, southern Atlantic Ocean, eastern Indian Ocean, northwestern Pacific Ocean, and southwestern Pacific Ocean. The regression pattern of the preceding spring SST field with the PRI has also already shown that these SST anomalies develop and are maintained until summertime in these five key positive anomalous sections. However, only the southwestern Pacific Ocean SST fails to reach the 95% confidence level (figure not shown). Therefore, we define the detrended and normalized regional mean SST as the SST indices. A1 represents the North Atlantic SST (35°W–15°W, 40°N–50°N); A2 represents the South Atlantic SST (15°W–10°E, 5°S–5°N, 5°E–15°E, 18°S–5°S; because of its irregular shape, it is divided into two parts for calculation); A3 represents the northwestern Pacific SST (140°E–170°W, 25°N–35°N); A4 represents the southwestern Pacific SST (165°E–165°W, 40°S–25°S); and A5 represents the southeastern Indian SST (90°E–110°E, 20°S–10°S, 105°E–120°E, 40°S–20°S, same as the A2 area).

Figure 5.

Regression patterns of detrended SST fields with the PRI. (The light red and light blue shadings denote regions with correlations that are, respectively, positively or negatively significant at the 95% confidence level; the heavy red and heavy blue shadings denote regions with correlations that are, respectively, positively or negatively significant at the 99% confidence level.)

[14] Table 1 shows the correlation coefficients of these indices with the PRI, all of them at a 99% confidence level. Here, we use A1 + A2 as the Atlantic Ocean SST index, A3 + A3 + A5 as the Indo-Pacific Ocean SST index, and A1 + A2 + A3 + A3 + A5 as the five regions' mean index. The correlation coefficient between the southern Atlantic index and the PRI is 0.61, which is higher than that of the other individual regions, and the correlation coefficient between the mean index and the PRI is 0.72. Therefore, the SST anomaly index of the five key sections is more intimately related than the indices of the other regions to the PRI. To verify the primary effect on the Tibetan Plateau summer precipitation, the partial correlations of the Atlantic Ocean, Pacific Ocean, and Indian Ocean indices with the PRI were calculated; only the Atlantic signal exceeds the 95% confidence level. Thus, the dominant effect comes from the Atlantic Ocean. The ENSO signal, which comes from the Pacific Ocean, has an insignificant relationship with the southeastern Tibetan Plateau summer precipitation. Figure 6 shows the time series of the PRI and the A1 and A2 indices. This figure shows that in both the late 1980s and late 1990s, the North and South Atlantic Ocean indices do not match well with the PRI; however, the Atlantic Ocean indices are consistent with the PRI for the longer period.

Table 1. Correlation Coefficients of PRI With the Detrended and Normalized SST Key Sections' Mean Time Series, All Exceeding the 99% Confidence Level
Correlation CoefficientSST_A1SST_A2SST_A3SST_A4SST_A5A1 + A2A3 + A4 + A5Mean Index
PRI0.510.610.590.490.550.680.650.72
Figure 6.

PRI, detrended and normalized A1 region mean index, and A2 region mean index.

[15] To further understand how the positive Atlantic SST anomalies are associated with the atmospheric circulation patterns influencing the southeastern Tibetan Plateau summer precipitation, we plot the regression patterns of the detrended and normalized Atlantic SST and mean SST indices with the 500 hPa and 200 hPa height fields (Figure 7). Figures 7a and 7b show the regressions of the northern and southern Atlantic Ocean SST signals with the middle and upper levels of the troposphere, respectively. The figure displays a clear North Atlantic Ocean-Europe-Central-West Asia teleconnection pattern, which is consistent with Figures 3a and 3b. However, the regressions of the northwestern and southwestern Pacific Ocean signals with the middle and upper atmosphere prominently show the zonal circumglobal teleconnection patterns in the Northern Hemisphere over the middle latitudes and that the South Asian high and subtropical high signals over the North Pacific are overemphasized (figure not shown). Although the regression of the eastern Indian Ocean SST index with the 500 hPa and 200 hPa height fields is consistent with Figures 3a and 3b, which can obviously display the North Atlantic Ocean-Europe-Central-West Asia teleconnection patterns, these teleconnection patterns are much weaker (figure not shown). Figure 7c depicts the regression pattern of the mean index with the 500 hPa and 200 hPa height fields. This figure clearly shows the North Atlantic Ocean-Europe-Central-West Asia teleconnection pattern, with the weakened South Asian high and western Pacific subtropical high. Therefore, this pattern is well consistent with the pattern associated with the Atlantic Ocean SST, demonstrating that the signals mainly originate from the Atlantic Ocean for the year-to-year variability of the southeastern Tibetan Plateau summer precipitation.

Figure 7.

Regression of the detrended and normalized A1–A2 region indices and the five regions' mean index with (a1, b1, c1) 500 hPa height and (a2, b2, c2) 200 hPa height fields. (The light (heavy) shading denotes regions with correlations that are positively (negatively) significant at the 95% confidence level.)

5 Further Analysis of the Associated Atmospheric Circulation Variability

[16] Figure 8 shows the regression patterns of the OLR and the 200 hPa wind field (the wind field is significant at the 95% confidence level) against the PRI. The negative OLR anomalies in the Indian Peninsula-southern Tibetan Plateau-southern China-East China Sea, the tropical Indian Ocean, and the Maritime Continent are shown to be associated more with the PRI. The anomalous positive SST over the tropical Atlantic Ocean corresponds with the anomalous negative OLR, so the convective activity there is strong and forms the cyclonic anomaly (Figures 2 and 3b). The air ascends over the tropical Atlantic SST and subsides over its north, the Arabian plateau and the Iranian plateau. Because of the strong convective activities over the Tibetan Plateau and the weak convections over the Iranian plateau and the Arabian plateau, a vertical circulation is formed that corresponds to a negative pressure anomaly over the Tibetan Plateau and a positive pressure anomaly over the Iranian plateau and the Arabian plateau. In addition, the descending air over the Iranian plateau and the Arabian plateau leads to less precipitation there. There are anomalous easterly winds at 200 hPa and anomalous westerly winds at 850 hPa over the tropical region (Figure 2). On the western Tibetan Plateau, the easterly wind anomalies bifurcate into two parts over the tropical region in the Northern Hemisphere: one part travels to the Atlantic Ocean, while the other travels over the western Mediterranean and shifts from an easterly to a westerly, forming an anticyclone over the Mediterranean. At the same time, the anomalous easterly winds, which come from the Arctic Ocean along the northern edge of Eurasia via the Laptev Sea-Kara Sea-Barents Sea-Norwegian Sea southward to 40°N, strengthen and converge with the westerlies over the northern Mediterranean, forming cyclonic anomalies in the upper troposphere.

Figure 8.

Regression patterns of detrended OLR and 200 hPa wind fields with the PRI. (The light shading (heavy) denotes the regions with correlations that are positively (negatively) significant at the 95% confidence level. The entire wind field denotes regions significant at the 95% confidence level.)

[17] In the upper troposphere, therefore, the anomalous easterly winds from the Arctic Ocean along the Eurasian coast to the northern Atlantic Ocean combine with the easterlies from the Tibetan Plateau to the Atlantic Ocean along the Northern Hemisphere tropical belt, forming a meridional North Atlantic teleconnection pattern in which the North Atlantic Ocean, the Mediterranean, and the Iranian plateau are anomalous anticyclone, cyclone, and anticyclone centers, respectively. Because a warm low pressure anomaly exists in the Tibetan Plateau and a high pressure anomaly exists in the Iranian plateau, the air rises more intensively in the vicinity of 90°E over the Tibetan Plateau because of stronger surface heating, inducing more rainfall. Thus, the primary impact of the southeastern Tibetan Plateau summer precipitation distribution is relevant to the North Atlantic Ocean teleconnections and the tropical Atlantic Ocean.

6 Conclusions and Discussions

[18] The summer precipitation interannual variability (May–September) over the Tibetan Plateau was studied based on monthly mean NCEP/NCAR reanalysis data during the period of 1979–2010. The results show that the large centers of summer precipitation over southeastern Tibetan Plateau are located in northern Burma, the northern Bay of Bengal, and Bangladesh. The first EOF mode of the southeastern Tibetan Plateau summer precipitation spatial distribution is consistent with its climatology in the spatial pattern, and the correlation coefficient between the Precipitation Index (PRI) and the first EOF principal component of the southeastern Tibetan Plateau summer precipitation is 0.97.

[19] Our study of the atmospheric circulation anomalies associated with the PRI indicates that a well-organized pattern originated from the Atlantic to the southeastern Tibetan Plateau. Thus, we analyzed the associated SST anomalies around the world's oceans and identified five key regions, with two in the Atlantic and three in the Indo-Pacific.

[20] The SSTs in these five regions display a comprehensive influence on the southeastern Tibetan Plateau summer precipitation distribution. However, the partial correlation analysis indicates that the dominant SST signals come from the Atlantic Ocean. We then investigated why the SST in the Atlantic can affect the precipitation in the southeastern Tibetan Plateau. The evident North Atlantic Ocean-Europe-Central-Western Asia teleconnection pattern is found to be associated with the Atlantic SST, resulting in the South Asian high being more westward than normal, with a stronger upward motion in the southeastern Tibetan Plateau, stronger southwesterly low-level winds in the Northern Indian Ocean and the Bay of Bengal, and more water vapor transport to the southeastern Tibetan Plateau. This teleconnection pattern is well consistent with the Rossby wave propagation from the Atlantic to South Asia found in the previous studies by Hoskins and Ambrizzi [1993] and Ding and Wang [2005]. Hence, there are two possible ways that the anomalous positive Atlantic SST may influence the southeastern Tibetan Plateau summer precipitation. Previous studies examined the relationship between wintertime SST anomalies in the northwestern Atlantic and atmospheric circulation fluctuations over the area between 60°W–40°W and 40°N–50°N [Palmer and Sun, 1985; Wallace and Jiang, 1987; Wallace et al., 1990; Peng, 1993; Peng et al., 1995; Ting and Peng, 1995; Peng and Fyfe, 1996]. A positive height anomaly was found in perpetual November experiments in response to a warm SST anomaly over the northwestern Atlantic, in which transient eddies serve as an important damping mechanism [Ting and Peng, 1995; Peng et al., 1995]. In our study, a barotropic positive potential height anomaly exists in the north in response to a warm SST anomaly over the northeast Atlantic Ocean via air-sea interaction. The anomalous anticyclonic circulation over the North Atlantic influences the southeastern Tibetan Plateau summer precipitation through the North Atlantic Ocean-Europe-Asia teleconnection. Moreover, the positive tropical Atlantic SST anomaly generates an anomalous cyclonic circulation there through the convection activity (Figures 2 and 3b). The negative OLR anomaly over the tropical Atlantic corresponds to the anomalous strong activity, so the ascending motion forms over the tropical Atlantic, and subsidence forms over its north (Figure 8). Hence, the tropical Atlantic SST can influence the southeastern Tibetan Plateau summer precipitation via the zonal convection activities. Furthermore, the dynamics of the southeastern Tibetan Plateau summer precipitation are related to stationary wave activity bridging the North Atlantic and the Eurasian continent (Figure 3a), which shows that the stationary wave from the North Atlantic Ocean to the Tibetan Plateau is discontinuous at the Iranian plateau. There are two parts of energy reaching the southeastern Tibetan Plateau. One portion is from the North Atlantic Ocean to the south of Mediterranean Sea and the Iranian plateau via the teleconnection; the other portion is from Arabia and the Iranian plateau to the Tibetan Plateau. Sun et al. [2009] found that the warmer tropical Atlantic SST can impact the summer North Atlantic Oscillation through the meridional wave-like train which may provide the possible connection between the North and tropical Atlantic Ocean. Thus, both the North Atlantic Ocean-Europe-Asia teleconnection and the tropical Atlantic zonal wave can affect the SET_PR.

[21] In summary, the three-dimensional atmospheric circulation associated with the Atlantic SST and the southeastern Tibetan Plateau summer precipitation is comprehensively analyzed in this work to document the associated dynamical processes.

[22] The Indo-Pacific SST has comparatively less influence on the southeastern Tibetan Plateau summer precipitation. This finding may be related to the Indo-Pacific SST being more influential on the western Pacific subtropical high (WPSH) [Gao and Wang, 2012] and the East Asian summer monsoon [Chang et al., 2000; Zhou et al., 2009] and to the WPSH having a moderate impact on the southeastern Tibetan Plateau precipitation.

[23] Previous studies have illustrated large rainfall biases in the southwestern Tibetan Plateau in the state-of-the-art climate models. Whether the primary cause is the regional simulation biases of the summer rainfall over the southwestern Tibetan Plateau or the inaccurate depiction of the connection between the Atlantic Ocean and the southwestern Tibetan Plateau summer rainfall is unclear. Moreover, although Li et al. [2012] found a North Atlantic-Mediterranean Sea wave train in the upper level in response to a warm SST in the North Atlantic in a sensitive experiment during summertime, most previous works have focused on the northwestern Atlantic Ocean during wintertime. How the north and tropical Atlantic Ocean interacts with the atmosphere to affect the southeastern Tibetan Plateau climate during summertime at synoptic and intraseasonal scales needs further investigation.

Acknowledgments

[24] This research was supported by the Major State Basic Research Development Program of China (973 Program, grant 2009CB421406) and the National Natural Science Foundation of China (grants 41130103 and 41210007).

Ancillary