J. Meteor. Res.  2019, Vol. 33 Issue (3): 478-490   PDF    
http://dx.doi.org/10.1007/s13351-019-8610-3
The Chinese Meteorological Society
0

Article Information

LI, Bo, Liu YANG, and Shihao TANG, 2019.
Intraseasonal Variations of Summer Convection over the Tibetan Plateau Revealed by Geostationary Satellite FY-2E in 2010–14 . 2019.
J. Meteor. Res., 33(3): 478-490
http://dx.doi.org/10.1007/s13351-019-8610-3

Article History

Received June 8, 2018
in final form March 18, 2019
Intraseasonal Variations of Summer Convection over the Tibetan Plateau Revealed by Geostationary Satellite FY-2E in 2010–14
Bo LI1,2, Liu YANG3, Shihao TANG1,2     
1. National Satellite Meteorological Center, China Meteorological Administration, Beijing 100081;
2. Key Laboratory of Radiometric Calibration and Validation for Environmental Satellites, China Meteorological Administration, Beijing 100081;
3. Meteorological Office of Longquanyi District of Chengdu, Sichuan Province, Chengdu 610100
ABSTRACT: Based on the infrared black body temperature (TBB) observed by the geostationary meteorological satellite FY-2E from 2010 to 2014, the seasonal migration, occurrence frequency, and intraseasonal variability of summer convection over the Tibetan Plateau (TP) and its surrounding areas are analyzed. The results show that in May, convection mainly occurs over the eastern edge of the TP; in June, following the onset of the Asian summer monsoon, the strongest (severe) convection occurs in the southeastern part of the TP; and in July–August, strong southwesterly winds transport abundant moisture to the eastern and central areas of the TP, leading to formation of an active convection belt over southeastern TP. The results also show that in the western TP, the area with convection frequency greater than 6% occupies the southern plateau around the 37th pentad, and gradually moves northward until the end of July; in the central plateau, convection (severe convection) becomes active since early (mid) June, and maintains through the entire late summer with three major northward movements until reaching 34°N; and in the eastern TP, the convection is relatively active since the beginning of May and its northward stretching is slightly later than that over the central plateau. Overall, summer convective activities are unevenly distributed over the TP, with frequency of convection decreasing from south to north; and they also exhibit considerable intraseasonal variability, the maximum of which is found over the middle reach of the Yarlung Zangbo River and the southeastern plateau. EOF analysis of summer convection frequency over the TP reveals two leading modes, with the first mode being a dipole variation pattern between the Indian monsoon region and the southeastern TP, and the second mode a tripole pattern over the western TP, the Indian continent west of 80°E, and the South Asian continent east of 80°E.
Key words: Tibetan Plateau     black body temperature (TBB)     severe convection     intraseasonal variation    
1 Introduction

Due to its unique geographic location and topography, the Tibetan Plateau (TP) has significant thermal and dynamic impacts on weather and climate in East Asia as well as across the globe (Ye et al., 1979; Wu et al., 2007). The dynamic effect of the TP on atmospheric circulation is to force the airflow to either climb up or flow around. The TP is located in the westerly belt in winter. Numerical experiments have revealed that under the dynamic forcing of TP, the upslope and bypass components of the airflow are equivalent in the wintertime (Qian et al., 1979). In summer, when the westerly belt shifts northward, the upslope component is smaller than the bypass component. As an uplifted heat source in the summer, the TP forced rising flow can affect the Pacific Ocean and North America to its east, and its influence can even reach the Southern Hemisphere (Li and Chen, 2003; Zhou et al., 2009). Intraseasonal and interannual variabilities of the atmosphere circulation surrounding the TP are greatly affected by the TP forcing (Li and Yanai, 1996; Zhang et al., 2006), which subsequently affect precipitation in East and Southeast Asia (Zhang et al., 2004; Chow et al., 2008).

The water vapor budget and precipitation over the TP directly influence the hydrological cycle in the TP and its surrounding regions. The northern TP is under the influence of the mid–high latitude atmospheric circulation systems, while the southern TP is mainly in control by the Asian monsoon system. Therefore, precipitation in the TP exhibits distinct regional features (Feng, 2011), and the low-frequency oscillation associated with precipitation also demonstrates significant regional differences (Duan et al., 2017). The dry and wet seasons in the TP are obvious, and precipitation is largely concentrated over May–September. There are two peaks of precipitation in the southern Himalayas and the Yarlung Zangbo River valley. In the southern Himalayas, the primary peak occurs in July–August and the secondary peak appears in February–March. In the Yarlung Zangbo River valley, the primary and secondary peaks appear in April and July, respectively. Other regions of the TP only experience a single precipitation peak that appears in July–August (Luo, 1992; Qiao and Zhang, 1994). Convective weather systems are active in summer due to the unique thermal and dynamic effects of the TP. Feng (2011) evaluated precipitation characteristics on multiple timescales in the TP based on station observations, satellite data, merged satellite data, high-resolution gridded data derived from station observations, and reanalysis products, and found that although these different datasets could all demonstrate the spatial pattern of summertime precipitation that decreases from southeast to northwest, the differences in spatial distribution, total amount, and frequency of precipitation, rainy days, etc., among these datasets cannot be ignored.

Following the development of satellite remote sensing techniques, more satellite data have been used in weather and climate studies of the TP. Due to the unique measurement method and advantages of spatial and temporal continuity, satellite data can effectively compensate for the large differences between precipitation observations and other cloud detection data (Feng, 2011; Wu et al., 2017), and thereby can be used to interpret the hydrological cycle in the TP from a different perspective.

Flohn (1968) proposed that there were 20–50 well developed cumulonimbus per 100,000 km2 on the plateau, indicating frequent local summer convective activities on the TP. Jiang and Fan (2002) found that areas of mesoscale convective activities can be divided into southeastern and southwestern subregions with the boundary along 95°E. Mesoscale convective activities are more frequent in the southwestern subregion, and only a few convective systems can propagate eastward and then affect precipitation in the Yangtze River valley. However, despite the small number of systems propagating eastward, they often result in heavy rainstorms in the Yangtze River valley. Zhuo et al. (2002) suggested that convective systems in the TP tend to move from the TP to the middle and lower reaches of the Yangtze River valley, and the time when the convective systems reach the Yangtze River valley is coincident with the occurrence time of heavy rainstorms there. Hu et al. (2010) statistically analyzed mesoscale convective systems (MCSs) originated in the TP during June–August of 1998–2001 based on the deep convection track dataset from ISCCP (International Satellite Cloud Climatology Project), and comprehensively described the origin, path, and precipitation characteristics of MCSs in the TP. Possible mechanisms for the impact of these systems on precipitation in China are summarized based on analysis of geopotential height and wind fields in abnormally strong and weak years of MCSs (Hu et al., 2008). Seasonal variations of the TP convective activity were also explored (Hu et al., 2016). Lu et al. (2016) compared the horizontal and vertical scales of cumulus and deep convective clouds over the TP, the land area to its east, the southern Indian Ocean, and the northwestern Pacific using CloudSat dataset. Their results indicate that the vertical scale of deep convective clouds in the TP is about 10 km, which is smaller than that in other areas. In addition, analysis of intensive observations during the Third Tibetan Plateau Atmospheric Science Experiment (TIPEX-III) indicates that the amount of total clouds, the top and amount of high clouds, and cloud depth, as well as microphysical features of clouds, all exhibit a distinct diurnal cycle (Liu et al., 2015; Zhao and Yuan, 2017).

Although many previous studies have investigated clouds and convection in the TP based on satellite datasets like CloudSat, CLIPSO (Cloud–Aerosol Lidar and Infrared Pathfinder Satellite Observations), and ISCCP, spatial and temporal resolutions of these datasets are relatively low. The black body temperature (TBB) measured by geostationary meteorological satellites represent the cloud top temperature in cloudy-sky condition and surface temperature in clear-sky condition. Thereby, TBB can reflect the genesis and development of a convective system as well as its intensity in a straightforward way, and well represent convective activities over the TP and moisture distribution in the atmosphere (Jiang and Fan, 2002; He et al., 2006). Using the geostationary meteorological satellite infrared data, Zhu and Chen (1999) statistically analyzed the weather and climatic features of summertime convective activities in the TP. Unfortunately, their study is based on a single year data of 1995. The study of Lin et al. (2006) on the spatial and temporal evolutions of MCSs in the TP is also based on one year data of 1998. In the present study, five-year TBB data from the Chinese geostationary satellite FY-2E are used to statistically analyze the convective systems in the TP and its surrounding areas and to explore the multi-scale features of summertime convection in this region.

2 Data

The data used in the present study are TBB from FY-2E, China’s domestically developed geostationary meteorological satellite, which was launched in June 2008 and positioned above the equator at 105°E since February 2009. It has drifted to 86.5°E since 1 July 2015 and continues to provide observational service. The main instrument of FY-2E is the infrared and visible-light spin-scanning radiator (VISSR). The TBB of the infrared window channel (IR1) of VISSR is used in the current study. The spectrum of the channel is 10.3–11.3μm, and the resolution at the sub-satellite point is 5 km. IR1 is mainly used to detect clouds in both daytime and nighttime, the temperature of the underlying surface, and the distinction between cloud and snow. The data cover the period of 2010–14.

The IR1 grayscale images at 3-h intervals (0000, 0300, 0600, 0900, 2100, 1500, 1900, and 2100 UTC) are transformed into TBB first, and then remapped to 0.1° × 0.1° grids. In the IR1 grayscale images, brighter hue indicates lower TBB, higher cloud top, and stronger convection. On the contrary, darker hue corresponds to higher TBB and inactive convective activity. Usually, the areas of TBB < –32°C(about 241 K)are convection active areas, while the areas of TBB < –52°C(about 221 K)are strong convection areas ( Yao et al., 2005).

3 Result analysis 3.1 Climatic occurrence frequency of summertime convection over the TP

Monthly mean distributions of TBB (°C) over the TP and adjacent areas from May to August during 2010–14 are displayed in Fig. 1. Figure 1a shows that compared to other regions, the entire TP is covered by low TBB in May with TBB lower than –13°C over most areas. Over the Bay of Bengal and India, which are located to the south of the TP, TBB is higher than 10°C. In June (Fig. 1b), areas of low TBB are largely located in the central and eastern TP, where TBB is lower than –10°C over most areas, and TBB in the western TP is relatively high. As the Asian summer monsoon outbreaks and propagates northward, the Bay of Bengal and South China Sea are covered by large areas of clouds with lower TBB. In July (Fig. 1c), areas of lower TBB are mainly found in the central and southern TP, while TBB over the Bay of Bengal is still higher than that over the TP. The Bay of Bengal, the South China Sea, and the central–southern TP are covered by large-scale monsoon clouds with TBB lower than –10°C. In August (Fig. 1d), the distribution of low-TBB areas is similar to that in July, although the intensity of low TBB overall is weaker than that in July.

Figure 1 Monthly mean distributions of TBB (°C) over the Tibetan Plateau (TP) and adjacent areas in (a) May, (b) June, (c) July, and (d) August of 2010–14. The red solid line denotes the terrain height of 3000 m.

The occurrence frequency of convection in the TP is statistically analyzed by taking –32°C as the threshold. The results are shown in Fig. 2. In May (Fig. 2a), convective activities over the TP are mainly affected by the westerly wind belt. There are two areas of strong convective activity located near the northwestern and eastern edges of the TP respectively. The maximum values in the two areas are similar to those of occurrence frequency of convection lower than 12%. Although TBB over the TP is relatively low in May as shown in Fig. 1a, the occurrence frequency of convection is also relatively low (Fig. 2a). Lower TBB is mainly due to the high cloud top. In general, the occurrence frequency of convection in June decreases from southeast to northwest. Following the seasonal northward shift of the westerly wind belt and the outbreak of the Asian summer monsoon, southwesterly winds intensify to the south of the TP. The strong southwesterly flow carries abundant water vapor from the Bay of Bengal and climbs up the plateau along the topographical gap in the central and eastern plateau. The strongest convection occurs in the southeastern TP, where the occurrence frequency of convection is about 10%, and this area is coincident with the low TBB area shown in Fig. 1b. Over the Bay of Bengal, however, the occurrence frequency of convection can be larger than 20% due to the outbreak of the Asian summer monsoon.

Figure 2 Frequencies (%) of monthly convections (with TBB less than –32°C) occurring over the TP and its adjacent areas in (a) May, (b) June, (c) July, and (d) August of 2010–14. The red solid line denotes the terrain height of 3000 m.

In July and August, the occurrence frequency of convection over the TP overall decreases from south to north, and is independent of convective activities in South Asia. In July (Fig. 2c), due to the further intensifying of the east Asian summer monsoon, some southwesterly wind transports water vapor to the central TP and a convection-active zone forms in the southern TP, where the occurrence frequency of convection is greater than 12%. There are two convection-active areas located in the central–southern and southeastern TP, respectively. The central–southern area is centered at (31°N, 90°E), where the occurrence frequency of convection is above 20% and larger than that in the southeastern center. In addition, the Bay of Bengal to the southwest of the TP is the most active area of convection, where the area of the largest occurrence frequency gradually moves westward from 90°E in June to 80°E. In August (Fig. 2d), following the weakening and southward retreat of the Asian summer monsoon, convective activities become weaker than that in July. The major convective belt in the TP is still located in the central–southern area, where the occurrence frequency of convection reduces to around 12%. Convective activities in the Bay of Bengal have also weakened compared to those in July, and the occurrence frequency is only about 12%–20% in most areas. The center of convective activity slightly shifts eastward compared to that in July.

The occurrence frequency of severe convection in summer over the TP is obtained by using –52°C as the threshold. Results are shown in Fig. 3. Comparison of Figs. 2 and 3 indicates that the spatial distributions of occurrence frequencies of severe convection and convection over the TP and its surrounding regions are quite similar, i.e., strong convections always occur over areas of frequent convection. In May (Fig. 3a), there exists nearly no strong convection in the TP, while a few strong convections occur over the Bay of Bengal, northern India, and the Beibu Gulf, where the occurrence frequency of strong convection is mostly below 6%. In June (Fig. 3b), areas with occurrence frequency of strong convection higher than 2% are found in the southeastern TP and at the junction between Sichuan and Yunnan provinces. Severe convections vigorously develop over the Bay of Bengal, South Asia, and the Beibu Gulf, where occurrence frequency of strong convection higher than 4% has been found over a large area, especially in Bangladesh and the offshore areas near 90°E. In July (Fig. 3c), severe convection centers appear in the central–southern TP and the junction between the southeastern part of the TP and Sichuan Province, respectively, where the occurrence frequencies are approximately 3%–6%. In Fig. 3c, the strong convection center over the Bay of Bengal and South Asia at around 90°E extends westward and reaches 75°E, and the occurrence frequency is higher than 6% over large areas. In August (Fig. 3d), the occurrence frequency of severe convection decreases to lower than 3% in the southeastern TP and is within the range of 2.5%–6% in the central–southern TP. To the south of the TP, the occurrence frequency of severe convection over South Asia, the Bay of Bengal, and Beibu Gulf in August is around 4%, which is lower than that in July and is similar to that in June.

Figure 3 Frequencies (%) of monthly severe convections (with TBB less than –52°C) occurring over the TP and its adjacent areas in (a) May, (b) June, (c) July, and (d) August of 2010–14. The red solid line denotes the terrain height of 3000 m.
3.2 Intraseasonal variability of summer convection

The seasonal variability of monsoon activities contains multiple sub-seasonal variabilities with distinct regional features, especially in complex terrain area like the TP (Lau et al., 1988; Kang et al., 2002), where convective activities are closely linked with the East Asian summer monsoon and demonstrate significant temporal variabilities.

Large topographic differences can be found between the western and eastern TP. Due to the presence of the Himalayas in the southwestern TP, the terrain elevation is above 7000 m there, while in the southeastern TP, there are many canyons distributed along the north–south direction. The northward advancement of the summer monsoon rainbelt exhibits different features in the western and eastern TP (Feng, 2011). For this reason, time–latitude cross-sections of pentad mean TBB (°C) averaged over the western TP (75°–80°E), central TP (80°–95°E), and eastern TP (95°–105°E) during 2010–14 are shown in Fig. 4. In the western TP (Fig. 4a), the southern border of the TP is located at around 30°N on average. Summer TBB lower than –2°C occurs at the 38th pentad, disappears in the western TP at the 47th pentad, and can reach the northernmost latitude of 38°N. In addition, in the western TP between 34°–40°N, a belt of low TBB maintains from the 25th pentad (early May) until the mid summer. Combined with Figs. 12, it can be found that this low TBB belt actually represents the cloudy area that occurs from early May to June over the northern TP centered along 36°N. In the central TP (Fig. 4b), a belt of low TBB centered along 35°N maintains from late May to mid June, while TBB lower than –2°C appears in the central TP at about 34th pentad (mid June) and remains until to the end of August. This low TBB is centered along 30°N and can reach the northernmost latitude of 37°N. Lower TBB centered at about 33°N occurs over the eastern TP in early May (Fig. 4c), while TBB lower than –2°C also appears in another area centered around 32°N in the eastern TP at the 30th pentad (end of May), and remains there until the end of August with northernmost latitude of 37°N.

Figure 4 The time–latitude cross-sections of pentad mean TBB (°C) averaged over (a) the western TP (75°–80°E), (b) central TP (80°–95°E), and (c) eastern TP (95°–105°E) during 2010–14. The abscissa axis indicates time (pentad), and the average position of the southern foothills of the plateau is represented by the red solid line.

The time–latitude cross sections of pentad mean occurrence frequency of convection from May to August over the western, central and eastern TP during 2010–14 are displayed in Fig. 5, which shows that the propagations of occurrence frequency of convection are distinctly different in the above three regions of TP. In the western TP, convections are active in the northwest (Fig. 5a) with the center located over 36°–38°N, where the maximum value is about 12%. In the central and southern area of the western TP, convective activities are relatively weak. The occurrence frequency larger than 6% appears in the southern TP at around the 37th pentad and ends at the 48th pentad, and the northernmost region is around 33°N. Areas of active convection can be found in the southern TP and Indian Ocean, where the occurrence frequency of convection is larger than 20% from mid July to early August.

Figure 5 As in Fig. 4, but for the pentad mean occurrence frequency (%) of summer convection over the TP.

In the central TP (Fig. 5b), occurrence frequency of convection higher than 4% appears in the north part from May to early June. In its central–southern portion, the beginning time when the occurrence frequency is greater than 6% is slightly earlier and ends later than that in its western portion. Active convections maintain throughout the midsummer with the frequency larger than 15% in the center. In addition, high occurrence frequency of convection experiences three periods of northward propagation in the central TP during 34th–37th pentads, 39th–42nd pentads, and 43rd–46th pentads respectively, and the northernmost area is around 34°N. Although the occurrence frequency of convection is high (> 15%) in the Indian monsoon region to the south of TP, water vapor cannot be transported across the entire plateau due to the blocking of steep topography in the southern TP. Thereby, relatively low convective activities can be found in the southern foothill of TP (around 27°N)and convections on the north and south of the southern foothill of TP are independent of each other. Convections can develop further in the eastern and central TP as a result of the water vapor being carried to the plateau along rugged gaps in the mountains. In the eastern TP (Fig. 5c), the terrain is relatively flat with lower elevation, and convection starts in early May and is active throughout the entire summer. The high frequency area advances northward during 36th–38th, 40th–41st, and 43rd–46th pentads, slightly later than that in the central TP. Convective activities are the most active during 36th–40th pentads with the center located at 33°N, where the frequency is higher than 20% (from late June to mid July).

The time–latitude cross-sections of pentad mean occurrence frequency of severe convection from May to August over the western, central and eastern TP during 2010–14 are displayed in Fig. 6. It is seen that the propagation of occurrence frequency of strong convection over the three portions of the TP are similar to that of convection. In the western TP (Fig. 6a), severe convections begin to develop at around the 35th pentad and end at the 47th pentad, and the occurrence frequency is relatively low. In the central TP (Fig. 6b), strong convections first appear at around the 34th pentad and maintain throughout the midsummer with the frequency over the central convection area larger than 4%. In addition, high occurrence frequency of strong convection experiences three northward propagations during 36th–38th, 40th–42nd, and 43rd–45th pentads respectively, and can reach the northernmost latitude of 35°N. Although the occurrence frequency of severe convection is large (> 11%) in the Indian monsoon region south of the TP, water vapor cannot be transported across the entire plateau due to the blocking of steep topography in the southern TP. Thereby, an area of relatively low convective activities is found near 27°–28°N. In the eastern TP (Fig. 6c), strong convections are the most active during 36th–40th pentads, and severe convections advance northward significantly during the 36th–38th and 43rd–46th pentads.

Figure 6 As in Fig. 4, but for the pentad mean occurrence frequency (%) of summer severe convection over the TP.

In order to investigate the evolution of precipitation intensity accompanied with convective activities, time–latitude cross-sections of TRMM 3B42 pentad mean precipitation intensity during May–August of 2010–14 are shown in Fig. 7. Although high occurrence frequency of convection appears from early May to June on the TP, the intensity of corresponding precipitation is weak at the same time. Based on the threshold of precipitation larger than 2.5 mm day–1, rainy season in the western TP begins from the 31st pentad (Fig. 7a) and ends at late August, and the rainy area can reach the northernmost latitude of 33°N, corresponding to the northernmost location of convection activities. Precipitation intensity in the Indian monsoon region is greater than 8 mm day–1; however, due to the blocking of water vapor transport by steep topography in the southern TP, heavy rainfall cannot be observed over the western TP. This is consistent with the result shown in Fig. 5a. In the central TP (Fig. 7b), the rainy season begins at the 34th pentad, which agrees well with the time of sudden increase in occurrence frequency of convection. However, the area of large precipitation located to the north of the central area where the most frequent convections appear. This result actually reflects the effect of the non-convective precipitation. In the eastern TP (Fig. 7c), precipitation intensity larger than 2.5 mm day–1 appears as early as the end of May, while mean precipitation intensity larger than 5 mm day–1 appears at the 34th pentad and ends at late August. This is consistent with the beginning (the 34th pentad) and ending time of high occurrence frequency of convection over the eastern TP. During the period of 34th–40th pentads, convective precipitation makes a large contribution to the total rainfall amount over the eastern TP, while the center of convective activities is coincident with the precipitation center. Since the 41st pentad, the area of large precipitation is located to the south of the center of active convections (33°N), which reflects the contribution of non-convective rainfall in late summer over the eastern TP.

Figure 7 As in Fig. 4, but for the pentad mean precipitation (mm day–1) over the TP.

As shown in Fig. 5, convections on the southern and northern sides of topography in the southern TP are out of phase due to the topographic blocking of water vapor transport. Figure 8 displays time–longitude cross-sections of pentad mean occurrence frequency of convection during May–August of 2010–14 over the areas of active convection in the main body of the TP (28°–35°N) and to the south of the TP (18°–26°N). As shown in Fig. 8a, convections over the main body of the TP suddenly intensify simultaneously along 93°E and 102°E at the 33rd pentad (late June), and two active convection centers gradually form and propagate westward. The western center experiences two active periods, i.e., during the 33rd–35th pentads when the active convection center slightly shifts westward, and during the 36th–43rd pentads when the active convection center propagates from 93°E to near 85°E and the occurrence frequency of convection is higher than 15% per pentad. Then during the 43rd–46th pentads, the occurrence frequency reduces to 12%–15% per pentad. The eastern area of active convection is centered along 102°E and mainly located on the southeastern part of the TP, and gradually shifts westward during the 33rd–38th pentads, while the active convection area expands westward to near 94°E. The occurrence frequency of convection in this center is higher than 20%, and gradually weakens after the 40th pentad. Figure 8a also shows that there exists a relatively low convection area near 110°E downstream of the TP, while the active convection area to the east of 115°E is largely affected by the East Asian summer monsoon and demonstrates different characters compared to convections originated in the TP.

Figure 8 Time–longitude cross-sections of the pentad mean occurrence frequency (%) of convection averaged over (a) the main body of the TP (28°–35°N) and (b) the south of the TP (18°–26°N) during 2010–14. The vertical axis indicates time (pentad).

Figure 8b shows that the convections related to the summer monsoon suddenly strengthen at the 28th pentad over India and the Bay of Bengal, and propagate westward, reaching 83°E at the 31st pentad. In this stage, convections are most active near 92°E. At the 33rd pentad, convections in this area intensify again and propagate westward; by the 34th pentad, active convections cover most of the area to the west of 95°E. The convective center is located near 91°E from late June to early July, and shifts to about 78°E in mid and late July. The occurrence frequency is higher than 30% in the center. Since early August, the active convection center slightly moves eastward and is located at 84°E at the 46th pentad.

Figure 9 presents time–longitude cross-sections of TRMM 3B42 pentad mean precipitation over the convection active area in the main body of the TP (Fig. 9a) and the southern TP (Fig. 9b), respectively. By comparing Fig. 9a with Fig. 8a, it can be found that in the main body of the TP, the center of large occurrence frequency of convection does not always correspond to the precipitation center, indicating the notable contribution of non-convective precipitation to the total summer rainfall amount over the TP. Figure 9b shows that the strongest precipitation centers in South Asia and the adjacent oceans are basically consistent with the center of large occurrence frequency of convection shown in Fig. 8b. This result reflects the great contribution of convective precipitation to the total monsoon rainfall in these regions.

Figure 9 As in Fig. 8, but for the pentad mean precipitation (mm day–1).

The occurrence frequency of convection could be high during some periods and low in other periods because convection indeed occurs unevenly during various periods of the summer. Therefore, for the study of climatic features of convection, it is important to include not only the occurrence frequency of convection, but also its intraseasonal distribution, i.e., the intraseasonal variability of convection occurrence. In this study, intraseasonal variability of convection (strong convection) is investigated based on the standard deviations of the pentad occurrence frequency of convection during June–August of 2010–14. Figure 10a shows that the standard deviations of the occurrence frequency of convective activities overall decreases from south to north. Large intraseasonal variability of convection can be found in two areas in the TP, which are respectively located over the middle reaches of the Yarlung Zangbo River in the central–southern TP and the junction of Tibet Region, Qinghai Region, and Sichuan Province in the southeastern TP. The variability is larger in the former area with the value greater than 5%. Note that the distributions of intraseasonal variability in these areas are extremely uneven and prone to flood and drought disasters. The standard deviation of convective frequency is small in the northern part of the TP, and the values are mostly below 2%.

Figure 10 Standard deviations of the pentad mean occurrence frequency (%) of (a) convection and (b) severe convection during June–August of 2010–14.

As shown in Fig. 10b, the standard deviation of occurrence frequency of severe convection also decreases from south to north. Large intraseasonal variability of severe convection can be found in two areas on the TP, i.e., the middle reaches of the Yarlung Zangbo River in the central–southern TP and the junction of Tibet, Qinghai, and Sichuan provinces in the southeastern TP. The intraseasonal variability of severe convection is comparable beween the above two areas, both with a standard deviation of more than 1.5% over the central areas, which is different from 5% of the convection case.

3.3 Principal modes of pentad occurrence frequency of summertime convection

The above results indicate that convections are most active in July–August over the TP. For this reason, EOF analysis is applied to the pentad occurrence of convection frequency during July–August (mid and late summer) of 2010–14 over the TP and its surrounding areas. The first two leading modes of variability for summertime convection are derived based on the North Rule Test (North, et al., 1982). Their spatial patterns and corresponding principal components (PCs) are shown in Figs. 1112. The first two leading modes account for 14.9% and 10.5% of the total variance, respectively. As shown in Fig. 11a, the first leading mode (EOF1) demonstrates opposite patterns between the southeastern TP centered at (30°N,100°E) and the area from India to the Bay of Bengal. Combined with PC1 (Fig. 12a), it can be found that this mode reflects the high occurrence frequency of convection in the southeastern TP and low frequency in the northwestern TP, India, and the Bay of Bengal in early and mid July and in middle and late August. From mid July to mid-August, the occurrence frequency decreases in the southeastern TP but increases in the northwestern TP and India. Figure 5c shows that late June to early July is the active convection period for the eastern TP, while the occurrence frequency decreases subsequently but increases again since mid August and maintains there until the end of August.

Figure 11 The first two independent dominant modes (a) EOF1 and (b) EOF2 of convection frequency during July–August of 2010–14. The red solid line denotes the terrain height of 3000 m.
Figure 12 The first two principal components of convection occurence frequency during July–August of 2010–14: (a) PC1 and (b) PC2, which respectively account for 14.9% and 10.5% of the total variance.

The second leading mode (EOF2; Fig. 11b) shows a tripole pattern of variability for occurrence frequency over the western TP, the area to the west of 80°E in India, and the area to the east of South Asia. PC2 (Fig. 12b) indicates that when the time coefficient is positive from late July to early August in most of the years, the occurrence frequency of convection increases over areas to the west of 90°E in the TP and to the west of 80°E in India, corresponding to the western center of active convection in the main body of the TP and South Asia shown in Fig. 8. Meanwhile, the occurrence frequency of convection is low in the area to the east of 80°E in South Asia. In 2011, this mode prevails over the entire July and late August.

4 Conclusions

Based on the infrared TBB of the geostationary meteorological satellite FY-2E from 2010 to 2014, the climatic characteristics of summer convection over the TP and its surrounding areas are analyzed. Intraseasonal variability of the summer convection and the leading modes of occurrence frequency of convection are explored. Major conclusions are as follows.

(1) In May, convective activities in the TP are mainly affected by the westerlies. Major areas of convection are located at eastern edges of the TP. In June, following the onset of the Asian summer monsoon and the intensification of the southwesterlies, the strongest convections (severe convections) occur over the southeastern part of the TP. During July–August, strong southwesterly winds rise along the topographic gaps in the southwestern TP and bring about abundant moisture to the eastern and central TP. As a result, an active convection belt forms over the southeastern TP and two active convection (severe convection) centers are embedded in the belt.

(2) The zonal variations of convection over the western, central, and eastern parts of the TP are quite different as time revolves. In the western TP, convective activities are generally weak; the area with convection frequency greater than 6% reaches the southern foothill of the plateau at about the 37th pentad, and reaches the northernmost position at the end of July and early August. In the central plateau, convection (severe convection) becomes active slightly earlier than in the western TP, and maintains there over the entire mid and late summer. The occurrence frequency experiences three major northward movements and reaches the northernmost latitude of 34°N. Convection in the eastern part of the TP has been active since the beginning of May, while convective activities (severe convective activities) experience three (two) times of northward movement.

(3) The standard deviation of convection (severe convection) occurrence frequency in general decreases from south to north. Large intraseasonal variability of convection can be found in two areas of the TP, which are respectively located at the middle reaches of the Yarlung Zangbo River in the central–southern TP and the junction of Tibet, Qinghai, and Sichuan provinces of the southeastern TP.

(4) The first two leading EOF modes of pentad occurrence frequency of summer convection over the TP are independent modes. EOF1 demonstrates the opposite patterns between the southeastern TP centered at (30°N, 100°E) and the area from India to the Bay of Bengal, while EOF2 shows a tripole pattern of convection occurrence frequency over the western TP, the area to the west of 80°E in India, and the area to the east of 80°E in South Asia.

Convective activities in the TP are complicated. Convective cells with different intensities and various life spans are often embedded in the cloud systems over the TP in summer. For this reason, data with finer resolutions on both spatial and temporal scales are needed to comprehensively describe the summertime convections in the TP. The new generation of geostationary satellites, Himawari-8 and FY-4, can provide observations that are spatially and temporally dense with high quality and more observation channels over the TP. In the future, we will further analyze the characteristics of summertime convections in the TP based on the above new satellite data.

Acknowledgments. We acknowledge Professor Ping Zhao of the Chinese Academy of Meteorological Sciences for his many valuable suggestions on this work.

References
Chow, K. C., J. C. Chan, X. L. Shi, et al., 2008: Time-lagged effects of spring Tibetan Plateau soil moisture on the monsoon over China in early summer. Int. J. Climatol., 28, 55–67. DOI:10.1002/joc.1511
Duan, L. J., A. M. Duan, W. T. Hu, et al., 2017: Low frequency oscillation of precipitation and daily variation characteristic of air–land process at Shiquanhe station and Linzhi station in Tibetan Plateau in the summer of 2014. Chinese J. Atmos. Sci., 41, 767–783. DOI:10.3878/j.issn.1006-9895.1701.16205
Feng, L., 2011: Dignostic and simulation analyses on the summer precipitation and associated water vapor transport over the Tibetan Plateau. Master degree dissertation, The Graduate School of Chinese Academy of Sciences, Beijing, 190 pp. (in Chinese)
Flohn, H., 1968: Contributions to a meteorology of the Tibetan highlands. Atmospheric science paper, No. 130, Colorado State University, Fort Collins, 120 pp.
He, Y. H., C. Y. Li, Q. Jin, et al., 2006: Relationship between the low frequency oscillation of TBB in summer of Qinghai–Tibet Plateau and drought/flood in central China. Plateau Meteor., 25, 658–664. DOI:10.3321/j.issn:1000-0534.2006.04.013
Hu, L., Y. D. Li, R. Fu, et al., 2008: The relationship between mobile mesoscale convective systems over Tibetan Plateau and the rainfall over eastern China in summer. Plateau Meteor., 27, 301–309.
Hu, L., Y. D. Li, and S. Yang, 2010: Observation and analysis of mesoscale convective systems derived from the Qinghai–Tibet Plateau. Part I: Origin, path, development and characteristics of ground precipitation. Proceedings of 2010 Symposium on Meteorological Science and Technology between the Two Sides of the Taiwan Straits, Chinese Meteorological Society, Beijing, 14 pp. (in Chinese)
Hu, L., D. F. Deng, S. T. Gao, et al., 2016: The seasonal variation of Tibetan convective systems: Satellite observation. J. Geophys. Res. Atmos., 121, 5512–5525. DOI:10.1002/2015JD024390
Jiang, J. X., and M. Z. Fan, 2002: A primary study of the relationship between TBB fields and water vapor distribution over Qinghai–Tibet Plateau in summer. Plateau Meteor., 21, 20–24. DOI:10.3321/j.issn:1000-0534.2002.01.004
Kang, I. S., K. Jin, B. Wang, et al., 2002: Intercomparison of the climatological variations of Asian summer monsoon precipitation simulated by 10 GCMs. Climate Dyn., 19, 383–395. DOI:10.1007/s00382-002-0245-9
Lau, K. M., G. J. Yang, and S. Shen, 1998: Seasonal and intraseasonal climatology of summer monsoon rainfall over East Asia. Mon. Wea. Rev., 116, 18–37. DOI:10.1175/1520-0493(1988)116<0018:SAICOS>2.0.CO;2
Li, C. F., and M. Yanai, 1996: The onset and interannual variability of the Asian summer monsoon in relation to land–sea thermal contrast. J. Climate, 9, 358–375. DOI:10.1175/1520-0442(1996)009<0358:TOAIVO>2.0.CO;2
Li, W., and L. X. Chen, 2003: Characteristics of the seasonal variation of the surface total heating over the Tibetan Plateau and its surrounding area in summer 1998 and its relationship with the convection over the subtropical area of the western Pacific. Adv. Atmos. Sci., 20, 343–348. DOI:10.1007/BF02690792
Lin, H., J. X. Jiang, Y. B. Yang, et al., 2006: Spatial–temporal evolvement trends of mesoscale convective systems over Qinghai–Tibet Plateau. Geomatics. Inf. Sci. Wuhan Univ., 31, 576–581.
Liu, L. P., J. F. Zheng, Z. Ruan, et al., 2015: Comprehensive radar observations of clouds and precipitation over the Tibetan Plateau and preliminary analysis of cloud properties. J. Meteor. Res., 29, 546–561. DOI:10.1007/s13351-015-4208-6
Lu, Z. X., Y. Y. Li, and L. X. Fang, 2016: Horizontal and vertical scales of convective clouds over China and the surrounding oceans. Acta Meteor. Sinica, 74, 935–946. DOI:10.11676/qxxb2016.074
Luo, S. W., 1992: Study on Some Kinds of Weather Systems over and around the Qinghai–Tibet Plateau. Science Press, Beijing, 162 pp. (in Chinese)
North, G. R., T. L. Bell, R. F. Cahalan, et al., 1982: Sampling errors in the estimation of Empirical Orthogonal Functions. Mon. Wea. Rev., 110, 699–706. DOI:10.1175/1520-0493(1982)110<0699:SEITEO>2.0.CO;2
Qiao, Q. M., and Y. G. Zhang, 1994: Synoptic Study of the Qinghai–Tibet Plateau. China Meteorological Press, Beijing, 250 pp. (in Chinese)
Qian, Y. F., Q. Q. Wang, and D. Z. Ye, 1979: Tibetan Plateau Meteorology. Science Press, Beijing, 232–249. (in Chinese)
Wu, C., L. P. Liu, and X. C. Zhai, 2017: The comparison of cloud base observations with Ka-band solid-state transmitter-based millimeter wave cloud radar and ceilometer in summer over Tibetan Plateau. Chinese J. Atmos. Sci., 41, 659–672. DOI:10.3878/j.issn.1006-9895.1701.16170
Wu, G. X., Y. M. Liu, Q. Zhang, et al., 2007: The influence of mechanical and thermal forcing by the Tibetan Plateau on Asian climate. J. Hydrometeor., 8, 770–789. DOI:10.1175/JHM609.1
Yao, X. P., Y. B. Yu, and B. K. Zhao, 2005: Structural characteristic of Meiyu frontal cloud system and its probable causes. Plateau Meteor., 24, 1002–1011. DOI:10.3321/j.issn:1000-0534.2005.06.022
Ye, D. Z., Y. X. Gao, and M. Y. Zhou, 1979: Qinghai–Tibet Plateau Meteorology. Science Press, Beijing, 120–126. (in Chinese)
Zhang, Q. Y., Z. H. Jin, and J. B. Peng, 2006: The relationships between convection over the Tibetan Plateau and circulation over East Asia. Chinese J. Atmos. Sci., 30, 802–812. DOI:10.3878/j.issn.1006-9895.2006.05.09
Zhang, Y. S., T. Li, B. Wang, 2004: Decadal change of the spring snow depth over the Tibetan Plateau: The associated circulation and influence on the East Asian summer monsoon. J. Climate, 17, 2780–2793. DOI:10.1175/1520-0442(2004)017<2780:DCOTSS>2.0.CO;2
Zhao, P., and Y. Yuan, 2017: Characteristics of a plateau vortex precipitation event on 14 July 2014. J. Appl. Meteor. Sci., 28, 532–543. DOI:10.11898/1001-7313.20170502
Zhou, X. J., P. Zhao, J. M. Chen, et al., 2009: Impacts of thermodynamic processes over the Tibetan Plateau on the Northern Hemispheric climate. Sci. China Earth Sci., 52, 1679–1693.
Zhu, G. F., and S. J. Chen, 1999: Convective activities over the Qinghai–Tibet Plateau and adjacent regions in summer of 1995. Plateau Meteor., 18, 9–19. DOI:10.3321/j.issn:1000-0534.1999.01.002
Zhuo, G., X. D. Xu, and L. S. Chen, 2002: Instability of eastward movement and development of convective cloud clusters over Tibetan Plateau. J. Appl. Meteor. Sci., 13, 448–456. DOI:10.3969/j.issn.1001-7313.2002.04.008