Physical Mechanism of Phased Variation of 2020 Extremely Heavy Meiyu in Middle and Lower Reaches of Yangtze River

Meiyu or plum rain is a kind of peculiar seasonal precipitation during the seasonal advancement of the East Asian summer monsoon (Tao and Chen ^[1]; Qian and Lee ^[2]). In the middle and lower reaches of the Yangtze River with dense population, and developed industry and agriculture, Meiyu is often characterized by concentrated periods of time and high intensity. It can cause rise in the water level of the main stream and tributaries of the Yangtze River and Dongting Lake, Panyang Lake, Taihu Lake and other rivers and lakes as well as urban waterlogging, leading to huge loss of people's lives and property. Over the years, Meiyu in the middle and lower reaches of the Yangtze River has been the focus of short-term and medium-range forecasts by the Central Meteorological Observatory (National Meteorological Center) and climate prediction by National Climate Center of China Meteorological Administration (CMA). Since the beginning of the 21st century, with the rise of the integrated "seamless" forecasting system for weather and climate prediction, 10-30 d extended-range forecasting has become another important domain and a very difficult point in meteorological science research. Undoubtedly, the extended-range forecasting technology of Meiyu has very important practical significance and challenge.

Meiyu is caused by East Asian summer monsoon. The major members of the summer monsoon such as WPSH, SAH and the southwest monsoon have significant effects on Meiyu. The abnormal changes of these members are important causes for the inter-annual difference of Meiyu and the drought and flood disasters in the middle and lower reaches of the Yangtze River (Sampe and Xie ^[3]; Liu et al. ^[4]; Ding et al. ^[5]). Therefore, for many years, scientists have done a lot of research to reveal the influence of various factors on summer monsoon members and the mechanisms. These factors include the ENSO cycle (Kosaka et al. ^[6]; Weng et al. ^[7]; Wang et al. ^[8]; Zhang et al. ^[9]; Yuan et al. ^[10]; Shan et al. ^[11]; Zhang et al. ^[12]; Hao et al. ^[13]), the Western Pacific Warm Pool (Huang et al. ^[14]; Li et at. ^[15]), the Indian Ocean SSTA (Wu et al. ^[16]; Hu et al. ^[17]; Kosaka et al. ^[6]; Wang et al. ^[8]; Liu et al. ^[18]), the thermal forcing (Wu et al. ^[19-21]; Lu et al. ^[22]) and mechanical forcing (Liu et al. ^[23]; Molnar et al. ^[24]; Chen and Bordoni ^[25]; Kong and Chiang ^[26]) of the Tibetan Plateau, and the Eurasian mid-high latitude atmospheric circulation (Hsu and Lin ^[27]; Liu et al. ^[28]; Wang et al. ^[29]), etc. However, most of the above-mentioned research aimed at the needs of climate prediction, using climatological statistical methods to analyze a certain main factor based on the analysis of inter-annual difference or making sensitive experiments. Their research were mostly on the influence of"climate factors"on the later circulation or the approaching flood and drought events.

In recent years, a large number of studies have confirmed that the 30-60d low-frequency oscillation and quasi-biweekly oscillation (WBWO) are the key factors for Meiyu (Chen et al. ^[30]; Song et al. ^[31]; Yuan et al. ^[10]; Ding et al. ^[5]), but the low-frequency analysis still has some limitations. It can be only used as a forecast reference or a method of forecasting weather process in combination with other statistical methods, mainly for qualitative forecasts. Therefore, it is far from meeting the requirements of the refined extended-range forecast.

The most crucial feature of extended-range weather forecast, which is between medium - and short-term forecast and climate prediction, is that not only the dynamic process of the atmosphere itself, but also the land surface processes including sea surface temperature (SST), the thermal effect of Tibetan Plateau, and other climatic factors influence the atmosphere. At present, the dynamical extended ensemble forecast that incorporates real-time SST and other land surface processes into the model is the main direction of extended-range forecast technology. The forecasting practice in recent years has shown that major ensemble models like ECMWF are somewhat capable of predicting the trends of planetaryscale fluctuations and large-scale weather systems such as major polar vortex, blocking high pressure, and WPSH. These models are very important in the extended-range forecasting of Meiyu. However, climate prediction and medium- and short-term forecasting have been divided into different operation systems. Very few meteorologists engaged in medium - and short-term forecasting analyzed the impact of Tibetan Plateau thermal effect and ENSO or other"climate factors"on weather processes, while, on the other hand, climate prediction experts have rarely analyzed the detailed weather systems and periodic changes of "climate factors". Therefore, there is a serious lack of research integrating synoptics and climatology on the complex mechanisms of extended-range weather processes, which results in the lack of forecasting ideas and also affects the judgment on the reliability of model prediction.

From June to July 2020, extremely severe precipitation in Meiyu season occurred in the Yangtze River and Huaihe River Basin. Some researches (Ding et al. ^[32]; Takaya et al. ^[33]; Wang ^[34]) deemed that the 2020 severe Meiyu was related to the quasi-biweekly oscillation (QBWO) or the persistent warming in the tropical Indian Ocean by creating an EAP / PJ-like teleconnection pattern over East Asia and intensified and southward displaced WPSH. In this paper, we are to use the daily dataset and the synoptic diagnostic method to analyze the detailed process of Meiyu rain belt in 2020 and investigate how the phased changes of atmospheric circulation affects the variation of rain belt. The influence and physical mechanism of the regional and phased changes of traditional"climate factors", such as the Tibetan Plateau thermal effects and SSTA, on the phased evolution of monsoon circulation and Meiyu rain belt will be analyzed as well. These jobs can provide some practical ideas for extended-range weather forecast.

The characteristics of periodic variation of Meiyu are analyzed using daily precipitation data from 2, 435 stations nationwide in China from June to July in 2020 and 1981-2010, provided by CMA National Meteorological Center. NCEP / NCAR (National Center for Environmental Prediction / National Center for Atmospheric Research) daily mean reanalysis data (2.5°× 2.5°), including wind, geo-potential height, temperature, specific humidity from June to July in 2020 and their multi-year average data are used to study the evolution features of atmospheric circulation and calculate the heat source and moisture sinks according to the formula provided by Yanai et al. ^[35]. Daily mean OLR and its climatic mean from June to July in 2020 and daily mean SSTA from May to July in 2020 are all provided by NOAA (National Oceanic and Atmospheric Administration).

According to the Meiyu monitoring standard of CMA, the 2020 Meiyu in the middle and lower reaches of the Yangtze River started on 9 June 2020, 8 days earlier than the climatic mean (17 June) and ended on 31 July, 20 days later than climatic mean (10 July), lasting for 52 days and extremely longer than normal.

Figure 1 is a latitude-time profile of the daily average rainfall along 105°-120°E in eastern China from June to July 2020. It shows that during the Meiyu season, the main rain belt in eastern China has never been intermittent, and the heavy precipitation processes occur very frequently, but the position of the rain belt is extremely unstable. It swings back and forth from Yellow River to southern Yangtze River Basin with a large north-south span, passing through the middle and lower reaches of the Yangtze River every time.

Figure 1.  Latitude-time profile of average rainfall (units: mm) along 105º-120ºE from June to July 2020.

During the Meiyu season, the accumulated rainfall over most parts of Yellow River and Huaihe River Basin, the middle and lower reaches of the Yangtze River and the Sichuan Basin amounted for 600-900 mm, which is 50-100% more than the climatic average, and even 1000-1400 mm in some areas, which is 150-200% more than the climatic mean (Fig. 2a). The average accumulated precipitation during this Meiyu season in the middle and lower reaches of the Yangtze River (east of 105° E, 27° - 32° N) ranked the first in the historical record since 1961 (figure omitted).

Figure 2.  Accumulated precipitation (contours, units: mm) and percentage of precipitation anomaly (shaded, units: %) (a, b, c, and d indicate the whole Meiyu season, the early, middle, and late phases of Meiyu season in 2020, respectively).

As can be seen from Fig. 1, the Meiyu season in 2020 can be roughly divided into three stages: Stage 1, from 9 to 19 June, was the early period of Meiyu season, during which the heavy rainfall centers mainly occurred in the northward region from Yangtze River to Yellow River. The precipitation in most regions amounted to 100-250 mm, and some localized areas got 250-400 mm, which are 100-500% more than the climatic mean (Fig. 2b). In the second stage, from 20 June to 11 July, which was the mid-Meiyu season (close to the climatic mean Meiyu Season in terms of time), the heavy rainfall center mainly appeared near the middle and lower reaches of the Yangtze River, where the accumulated rainfall amounted to 400-800 mm, and some localized rainfall amount was about 1000 mm, generally less than climatic mean in most regions (Fig. 2c).The third stage or the late Meiyu season was from July 12 to 30, when the severe rainfall concentrated in almost the same regions as that at the first stage. The accumulated precipitation was generally 100-250 mm, with 250-400 mm in some areas, so most areas had the rainfall 100-300% more than the climatic mean (Fig. 2d). Therefore, it can be seen obviously that the main reasons for the extremely heavy Meiyu in 2020 are the early beginning, late retreat, and long duration of the rainy season.

Figure 3 shows the evolution of atmospheric circulation at the onset and the retreat of Meiyu in 2020. Before its onset, on 7 June (Fig. 3a), the western part of WPSH was weaker and southeasterly. There were three blocking highs at mid-high latitudes in Asia, located in eastern Urals to Lake Balkhash, Northeast China and the Sea of Okhotsk respectively. On 8-9 June (Fig. 3b-c), the blocking high over Northeast China weakened, being affected by the cold vortex moving eastward in front of the western blocking high. In the middle latitudes, SAH strengthened and its eastern section moved northward. In the low latitudes, the southwest monsoon strengthened and propagated eastward, converging with the easterly wind on the southwest side of WPSH and forming a tropical disturbance in the ocean near the Philippines. The western section of the WPSH was enhanced, stretching westward and northward under the combined effect of the in-phase superposition with the eastward high pressure ridge at the mid-high latitude, the strengthening of the SAH and the tropical disturbance (i.e. the baby stage of the second typhoon "Parrot" in 2020) near the Philippines. As a result, the unusual Meiyu season in 2020 broke out in the middle and lower reaches of the Yangtze River.

Figure 3.  The 500 hPa (black solid lines) and 200 hPa (shaded) geo-potential heights (units: dagpm) and 850 hPa wind vectors (units: m s^-1) on 7 (a), 8 (b), 9 (c) June and on 29 (d), 30 (e), 31(f) July 2020. Red dashed lines represent climatic mean 588 dagpm contours at 500 hPa, and red solid lines represent climatic mean potential height at 200 hPa.

The evolution of circulation situation in the retreating process of Meiyu is similar to that of the onset process. On 29-30 July (Fig. 3d-e), the blocking high in Northeast Asia weakened, the mid-high latitude circulation in East Asia became straight, the east section of the SAH developed northward, and a low pressure disturbance occurred over the ocean east to Luzon Island in the Philippines, causing WPSH to intensify and move northward. On 31 July (Fig. 3f), the low pressure disturbance entered the South China Sea and developed into a tropical cyclone, which was the third typhoon in 2020, named "Senrak". Finally, WPSH strengthened northward and the extreme Meiyu season retreated.

During the 2020 Meiyu, blocking highs were active in the middle and high latitudes of Eurasia. Ding ^[32] pointed out that the atmospheric circulation in the 2020 Meiyu (June-July) showed a "two ridges and one trough"pattern on average. The daily analysis shows that the blocking situation was constantly changing (figure omitted). In the early stage of Meiyu, the blocking high was in a double-blocking pattern, respectively located to the west of Ural Mountains and to the east of the Sea of Okhotsk (Fig. 4a). In the middle stage of Meiyu, the blocking situation over the mid-high latitudes was of single mid-blocking pattern, lying near Lake Baikal (Fig. 4b). Then, in the late stage of Meiyu, it became a double-blocking pattern again, and the blockings were located to the east of the Ural Mountains and over Northeast Asia respectively (Fig. 4c). The three types of blockings all led to flat westerly at the midlatitudes from Lake Balkhash to East Asia. The high-latitude cold air guided by the northwest airflow in the front of the blocking high or other high pressures moved southward into the middle and lower reaches of the Yangtze River along with the westerly fluctuation. During the early period of Meiyu, the two blocking high pressures were far away from China, and the western blocking high was inclined westward. Therefore, the westerly and the cold air were the most northerly. The blocking highs in the middle and late Meiyu periods were connected to the polar high-pressure dam. Therefore, the cold air was very strong, and the westly was significantly southward than that in the early Meiyu stage. The appearance and maintenance of blocking high and polar high dam in the middle and late July is one of the important reasons for the extension of the Meiyu.

Figure 4.  Mean potential height at 500 hPa (black solid lines, units: dagpm) and the anomalies (shaded) during the three stages of Meiyu in 2020 (a, b, and c are for the early, middle, and late stages, respectively); d, e, and f are the corresponding potential heights at 200 hPa (shaded, units: dagpm) and wind anomaly at 850 hPa (vector, units: m s^-1), respectively. Red dashed lines indicate climatic mean potential height at 500 hPa in a-c and 200 hPa in d-f.

During the three stages of Meiyu season in 2020, the monsoon circulation had the following outstanding characteristics: WPSH was significantly stronger and more westward than usual (Fig. 4a-c), the SAH was stronger and more eastward, the Somali jet stream was also much stronger, but the southwest monsoon from the Bay of Bengal to South China Sea was remarkably weaker (Fig. 4d-f). The ridge lines of WPSH and SAH were significantly more northerly than climatic mean in the early Meiyu period, being roughly the same as climatic mean in the middle Meiyu period, but significantly more southward than climatic mean in the late Meiyu period. As the strong WPSH continuously guided the southwest monsoon into the south of the Yangtze River and South China and nearby, the southwest monsoon over this region became stronger than usual, converging with the southward cold air in the middle and lower reaches of the Yangtze River. Then, with the help of high-level divergence in the northeast quadrant of the SAH, heavy rainfall events occurred frequently over this region.

During the Meiyu period, the north-south fluctuations of the ridge lines of WPSH and SAH were closely correlated to the fluctuation of circulation in mid-high latitudes (Fig. 5a-b). Because the cold air moved southward one flow after another with the flat westerly fluctuation, the cold air evolved constantly into weakening period and strengthening period alternatively. While the cold trough deepened, the cold air moved southward, causing the WPSH and SAH to retreat southward. And the in-phase superposition of westerly fluctuation wave ridge with WPSH and SAH induced the strengthening and northward lifting of the later. As a result, the WPSH, SAH and the rain belt swayed north and south frequently

Figure 5.  Latitude-time profile for mean potential height at 500 hPa along 110º-130ºE (a, black lines) and mean potential height at 200 hPa along 110º-120ºE (b, black lines). Red solid line and dashed line indicate zero lines of mean zonal wind along 110º-130ºE at 500 hPa for the year of 2020 and the climatic mean, respectively, regarded as the ridge of WPSH; blue solid line and dashed line indicate zero lines of mean zonal wind along 110º-120ºE at 200hPa for the year of 2020 and the multi-year average, respectively, regarded as the eastern ridge of SAH.

The most fundamental reason for the onset and maintenance of monsoon is the land-sea thermal contrast (Zhou and Zhou ^[36]; Ding et al. ^[37]). Fig. 6 reveals the average temperature and its anomaly at the mid-high levels (300-500 hPa) during the three stages of Meiyu season in 2020. It can be seen that in all the three stages, affected by the cold advection of northwest airflow, the temperature anomaly in the northwestern part of Tibetan Plateau was negative, while the eastern and southern parts of Tibetan Plateau and the middle and lower reaches of the Yangtze River had positive temperature anomaly. The atmospheric warming center in the Asia-Pacific region was located in the southeast of Tibetan Plateau, with central value above - 12℃, which was higher than the climatic mean (- 13 to - 14℃). This is particularly beneficial for formations of the stronger and eastward SAH and the stronger and westward WPSH. In the early Meiyu period, the positive temperature anomaly from the eastern part of Tibetan Plateau to the middle and lower reaches of the Yangtze River was the strongest, reaching 3-5℃, and its location was also the most northern, which was very favorable to maintain the WPSH northward. While in the late period, the situation became different. The distribution of wind and temperature fields show that the positive temperature anomalies in the eastern and southern parts of Tibetan Plateau and the middle and lower reaches of the Yangtze River were related to the warm advection transport of the SAH anticyclone. In addition, they were also related to the heat sources in Tibetan Plateau and the release of latent heat from the condensation of heavy rainfalls of Meiyu.

Figure 6.  300-500 hPa mean temperature (black solid line, units: ℃) and the anomalies (shaded), mean winds (vector, units: m s^-1). a, b, and c are for the early, middle and late stage of Meiyu, respectively. Red lines are climatic mean temperatures of -14℃, -13℃ and -12℃, respectively.

Figure 7a displays the daily evolution curves of the apparent heat source and apparent water vapor sink of Tibetan Plateau from June to July 2020. Seen from the figure, the variation trends of the apparent heat source and apparent water vapor sink are similar, and most of the time they are both stronger than climatic mean. This suggests that the stronger latent heat of convective condensation has an important contribution to the stronger heat source. In addition, the phase and amplitude changes of the apparent heat source and apparent water vapor sinks are basically the same as those of the north-south swings of SAH and WPSH, especially in the middle and late stages of the Meiyu season.

Figure 7.  Apparent heat sources (red solid line, units: W m^-2) and apparent water vapor sinks (black solid line, units: W m^-2) over the Qinghai-Tibet Plateau (27.5°- 37.5°N, 70°- 102.5°E) (a, red and black dashed lines represent the climatic mean of apparent heat sources and apparent water vapor sinks, respectively), latitude-time profile of OLR (black solid line, units: W m^-2) and the anomalies (shaded, units: W m^-2) along 85°-102.5°E (b, the red solid line and dashed line represent the western ridge of WPSH and the eastern ridge of SAH, respectively) during June-July 2020.

Daily changes of convection and heat sources (figure omitted) show that the evolution of heat sources in Tibetan Plateau are significantly affected by convection during the Meiyu season, especially in the eastern part. Fig. 7b presents the latitude-time profile of OLR and its anomaly along eastern part of the Qinghai-Tibet Plateau (85°-102.5°E). It shows that the convection and its anomaly changed almost synchronously with apparent heat sources and apparent water vapor sinks and their anomalies, especially in the central and northern areas of eastern Tibetan Plateau. This reflects that the dry and cold air brought about by the fluctuations of the mid-latitude flat westerly converged with the warm and humid air over Tibetan Plateau and formed precipitation. Then, the latent heat of condensation was released, strengthening the heat sources in the eastern part of Tibetan Plateau, where an obvious positive temperature anomaly was formed and thus the SAH was enhanced, moving more eastward. These activities were conducive to the development of stronger and more westward WPSH and regulated its north-south shifting.

Ju and Slingo et al. ^[38] and Huang et al. ^[39] have shown that tropical convective heating may be a key factor linking Asian monsoon and ENSO. Huang and Sun ^[14] found that convective activities of the western Pacific warm pool have an important influence on the anomalous activity of WPSH. In mid - and short-term forecasts, whether tropical convection or typhoon is active is also a significant basis for judging the development trend of WPSH. Bao ^[40] found that temporal and spatial evolutions of real-time SSTA were the key factors for the similarities and differences between summer monsoon circulation and Meiyu in the middle and lower reaches of the Yangtze River in 2016 and 1998. This research result was used successfully in the medium-term and extended-range operational forecasts of National Meteorological Centre, CMA, in July 2020. The trends of WPSH and Meiyu rain belt, and the precipitation trends of seven major rivers including the Haihe River, Yellow River, Yangtze River and Huaihe River were successfully predicted through analyzing the real-time SSTA.

The SSTA distribution in the three stages of Meiyu in 2020 is shown in Fig. 8a-c. It can be seen that during the early Meiyu period (Fig. 8a), the South China Sea, the sea east of south-central Philippines, and the Indonesian islands had obvious positive SSTA, and the corresponding convection (ITCZ) was also stronger than normal (Fig. 9a). The northern boundary of ITCZ was also more northerly than climatic mean (Fig. omitted). Due to the release of latent heat of convective condensation, a strong and huge heat source belt with more northerly boundary was formed (Fig. 9d), which was very beneficial for the activity of stronger, more westward and northerly WPSH (Liu et al. ^[41]).

Figure 8.  SSTA during the Meiyu season in 2020 (a, b, and c are for the early, middle and late stages of the Meiyu season, respectively; units: ℃).

Figure 9.  OLR and its anomalies during the Meiyu season in 2020 (a, b, and c are for the early, middle, and late stages of the Meiyu season, respectively; units: ℃; d, e, and f are for the corresponding apparent heat sources and the anomalies).

In the middle stage of Meiyu, the positive SSTA in northern South China Sea and the sea east to northcentral part of the Philippine did not change much. However, the positive SSTA in southern South China Sea, southern Philippines, and the Indonesian islands got enhanced (Fig. 8b). As the positive SSTA center moved southward to the southern Philippines and near the equator, ITCZ and the stronger convection area (Fig. 9b) as well as the corresponding strong heat sources also moved southward together (Fig. 9e). As a result, WPSH was still strong and westward, but obviously retreated southward. Accordingly, the Meiyu rain belt also shifted southward to the middle and lower reaches of the Yangtze River.

In the late stage of the Meiyu season, the overall pattern of SSTA in the South China Sea and the tropical western Pacific remained unchanged. The positive SSTA in the north-central part of South China Sea and the sea east to the central part of the Philippines strengthened together with the SSTA near the equator (Fig. 8c). Therefore, the ITCZ area was enlarged and intensified, and the north boundary was more northern than that in the middle stage. However, as the consequence of the strong convection and heat source near the equator (Fig. 9c and Fig. 9f), WPSH remained southerly, and the Meiyu continued.

From the above analysis, it can be concluded that from June to July 2020, the distribution characteristics and phased development of SSTA in the South China Sea and the tropical western Pacific region have a profound impact on WPSH and Meiyu through tropical convection and heat sources. This is also one critical reason for the success of medium-term and extendedrange forecasts of Meiyu made by the National Meteorological Centre.

It can be seen from Fig. 8 that during the Meiyu period, most of the Indian Ocean north to 15° S has positive SSTA, but positive SSTA near the equator are not prominent, especially in the area with longitude of 40-60°E, SSTA is relatively low, and there is no obvious abnormal convection and heat sources (Fig. 9). The upper and middle atmosphere temperature is also showing a positive anomaly weaker than that in the northern part of the Arabian Sea and Tibetan Plateau. The positive temperature anomaly with positive SSTA in the central equatorial Indian Ocean is also weaker than that of the Tibetan Plateau, and so the cross-equatorial airflow (the Somali jets) is stronger than climatic mean. However, due to strong WPSH, the southwest monsoon in South Asia and Southeast Asia is still weaker than usual.

The latitude-time profile of mean SSTA along 110°-140°E (Fig. 10a) denotes that there was an obvious 10-day positive SSTA increasing process in the South China Sea and ocean east to the Philippines from the 6th pentad of May to the 2nd pentad of June. The maximum value of positive SSTA reached 1℃, exceeding the SSTA in the equatorial area of the same longitude. It was conducive to the development of convection. From 6 June, the convection in this area gradually developed (Fig. 10b). On 8 June, a tropical disturbance appeared in the northwestern Pacific east to Samar island of the Philippines. Later, it developed into the typhoon "Parrot", and then moved to northwest, causing severe building of convection and heat source (Fig. 10b and c), which drove WPSH to move sharply northward, initiating the extremely severe Meiyu process.

Figure 10.  Latitude-time profile of mean SSTA along 110°-140°E during May-July 2020 (a), OLR and its anomaly (b), and apparent heat source and its anomaly (c) along 110°-140°E during 1 Jun-5 August 2020.

From the third pentad in June, the positive SSTA east to the Philippines began to decrease. However, positive SSTA in the equatorial region south of 5° N increased. The ITCZ and the heat source belt retreated southward noticeably, and so was it with WPSH. From the beginning of the third pentad in July, positive SSTA east to the Philippines slowly increased, with the center of positive anomaly gradually edging northward from the area near equator. Around 24 July, the centers of positive SSTA and ITCZ shifted northward to the ocean east to the Philippines when convection began to form in this region. On 28 July, a tropical disturbance was born here. It marched northwestward and gradually evolved into the third typhoon"Senrak", which caused WPSH to develop drastically and moved northward in large span, finally stopping the extremely heavy Meiyu in 2020.

From the above analysis, we can conclude that real-time SSTA in Indian Ocean has no obviously direct effect on WPSH and Meiyu. The distribution characteristics and phased development of real-time SST anomalies in the South China Sea and the tropical western Pacific region had a profound impact on WPSH and Meiyu through tropical convection and heat sources. This is also one of the important reasons for the success of the Central Meteorological Observatory's extendedrange forecast of Meiyu.

This paper has discussed the physical mechanisms of the Meiyu season in 2020 by analyzing the mediumterm and extended-range changes of the atmospheric circulation, thermal effect of the Qinghai-Tibet Plateau, SSTA and the tropical convection. The conclusions are as follows:

(1) The 2020 Meiyu can be roughly divided into three stages: early, middle and late periods. Throughout the whole Meiyu season, WPSH was stronger and westward, SAH was stronger and eastward, and blocking highs were very active. All these factors made the situation very beneficial to the continuous heavy precipitation in Meiyu season.

(2) The blocking high situations during the three stages of the 2020 Meiyu were different, but they all formed flat westerly winds in the middle latitudes from the east of Balkhash Lake to East Asia. The westerly wave formed by blocking highs interacted with WPSH and SAH, jointly created procedural precipitation and heat sources increasing over the northeastern Qinghai-Tibet Plateau, and caused the ridges of WPSH and SAH and the rain belt to move drastically north and south.

(3) Two reasons should be responsible for the stronger and westward WPSH and the stronger and eastward SAH. One was the higher temperatures, produced by the warm advection transport, the heat sources in Tibetan Plateau, in the upper and middle atmosphere over the eastern and southern parts of Tibetan Plateau and nearby. The other was the strong positive SSTA in equatorial western Pacific producing strong convection and heat sources. The appearance and maintenance of blocking high and polar high dam, and the maintenance of strong positive SSTA in the western equatorial Pacific were the critical contributors to the southward persistence of WPSH and the prolonged Meiyu season in 2020. Real-time SSTA in the Indian Ocean has no obviously direct effect on WPSH and Meiyu.

(4) Both the early onset of Meiyu and the late end of Meiyu in 2020 were associated with the northward movement of WPSH, which was induced by the flattening of circulation at mid-high latitudes due to the blocking high weakening and the phased strengthening of positive SSTA in the ocean east of the Philippines, resulting in the intensification of tropical convection and the generation of typhoons.

It can be seen from the above conclusions that the phased evolutions of mid-high latitude circulation, the thermal effect of Tibetan Plateau, and real-time SST have an extremely important impact on the Meiyu season in China through circulation interactions of different latitudes and levels and air-sea interactions. For climate prediction, it is difficult to grasp such details, which, in fact, are significant factors affecting the accuracy of climate prediction. At the same time, these details are actually the focus and key to forecasting the changes of major rain belt and continuous heavy precipitation in Meiyu season when we do the mediumterm and extended-range forecasts. Therefore, they should be analyzed in greater detail for the forecasting operation to form medium-term forecast and extendedrange forecast that integrate synoptics and climatology. This is crucial to medium-term forecast and extendedrange forecasting.