Climatic characteristics of mesoscale convective systems in the warm season in North China

In this study, a total of 339 mesoscale convective systems (MCSs) are obtained in North China using the temperature of brightness blackbody (TBB) data from the FY-2E in the warm season from 2010 to 2018. The number of meso-α-scale convective systems (MαCSs) is much more than that of meso-β-scale convective systems (MβCSs). The number of mesoscale elongated convective systems (MECSs) is more than that of mesoscale circular convective systems (MCCSs). Most MCSs occur in July and August, which have the widest influence range, the longest duration, and the strongest convection. The MαCS develops slowly and weakens rapidly. The diurnal variation of MαCSs presents a bimodal distribution, most of MαCSs form in the afternoon, while some of MαCSs form in the evening. The MCSs activities in the warm season of North China are concentrated in two belts, namely, the east–west-oriented belt along Henan Province, Shandong Province and the Yellow Sea, and the south–north-oriented belt along central-western Shandong, Tianjin City, the west of Bohai Sea and the northeast of Hebei Province. MCSs mainly move eastward, and only some MECSs move southwestward and northwestward. The easterly and northerly moving MCSs are mainly affected by the steering flow, while the southerly moving MCSs are mainly affected by storm propagation. The MCSs of North China mainly form in the high temperature, high humidity and high energy area, with favorable dynamic conditions, such as middle-level trough or vortex, low-level shear line, surface inverted trough or surface convergence line, and the terrain. Meanwhile, the MCS pregnant environment is often accompanied by low-level jet and relatively strong vertical wind shear.


Introduction
The sudden rainstorm, thunder and lightning, gale, hail and other severe convective weather events are all featured by large intensity and significant locality, and thus are difficult to be forecasted in operational business. These disastrous weather events pose a serious threat to the people's life and property, urban construction, port transportation and marine aquaculture industry. In fact, the mesoscale convective system (MCS) can often cause these severe convective weather events (Zong and Chen 2000;Zhao et al. 2017;Wang and Cui 2011). Therefore, the monitoring, forecasting and researching of MCSs have become a widely concerned topic.
The MCS often presents as a cumulonimbus, which can produce continuous precipitation in a wide area. Houze (1993) proposed a similar definition and further pointed out that the dynamics of MCSs is generally more complex than that of a single cumulonimbus or cumulonimbus line. When the cumulonimbus and cumulonimbus line combine, various cloud and precipitation structures will occur (Houze et al. 1989). There is a special category of MCS, namely, the mesoscale convective complex (MCC). Based on the satellite infrared images, Maddox (1980) defined the MCC using the area of the cold cloud cover with the temperature of brightness blackbody (TBB) below − 32 °C and − 52 °C. Later studies further proved that the criterion of − 52 °C can better reflect the development of MCC in the vertical direction (Augustine and Howard 1988;McAnelly and Cotton 1989;Zheng et al. 2008). MCC is a kind of circular MCS, and its occurrence frequency is far less than that of the mesoscale elongated convective system (MECS). Anderson and Arritt (1998) studied the permanent elongated convective system (PECS) and divided the large-scale MCS into MCC and PECS according to the eccentricity. Jirak et al. (2003) introduced the meso-βscale circular convective system (MβCCS) and the mesoβ-scale elongated convective system (MβECS) into the MCS categories. This classification method divides MCSs into four categories (MCC, PECS, MβCCS and MβECS), taking into account the scale, shape and duration of MCSs. According to the MCS classification of Jirak et al. (2003), there are significant differences between MCC and PECS in the spatio-temporal distribution and synoptic situation before the occurrence of convection (Anderson and Arritt 1998;Jirak et al. 2003).
China has a vast territory and complex terrain. The unique geographical environment and monsoon climate lead to the special distributions of MCS and precipitation in China. For example, Zheng et al. (2008) pointed out that there are three MCS high-frequency regions in China and its surrounding areas, which are presented as three east-west-oriented belts. Besides, there are many other MCS survey studies in different regions of China, for example, in southern China (Li et al. 1989), in the Qinghai-Tibet Plateau (Yang and Tao 2005), in the Yellow Sea and its surrounding areas (Zheng et al. 1999(Zheng et al. , 2004, in the Huaihe River basin (Fei et al. 2005;Liu et al. 2015), in Yunnan Province and its surrounding areas (Duan et al. 2004), and in the whole China and its surrounding areas (Ma et al. 1997).
The terrain over North China is very complex, with the Taihangshan Mountains, the Yanshan Mountains and the Bohai Sea to its west, north and east, respectively (Fig. 1). Affected by the complex terrain and multi-scale circulation systems, the precipitation and convective storm over North China present large spatio-temporal variations (He and Zhang 2010;Chen et al. 2012Chen et al. , 2014, and thus are hard to be accurately forecasted. Zheng et al. (2008)

MCS classifications and the identifying method
Based on the MCS classification of Jirak et al. (2003), the MCS classification standards in this study have been improved. For the meso-β-scale convective system (MβCS), the limitation that the maximum area with the TBB ≤ − 52 °C must be above or equal to 50,000 km 2 has been removed. The specific classification criteria are shown in Table 1. The time when the scale condition is first satisfied is recorded as the formation time. The time of the maximum area is the maturity time. The last time when the scale condition is satisfied is recorded as the dissipation time.
To ensure the relative integrity of MCS area and other information in the study area, the MCS identification range is expanded by 4-6°. That is, the identification range is 105-130°E, 30-50°N (Fig. 1). Then, the MCS cases with the centroid position at the maturation time in the study area (110-126°E, 33-45°N) are selected to form the MCS data set. In this study, the image processing technology is used to identify and track the MCS, process the data and calculate the relevant characteristic parameters.
The processing steps are as follows.
(1) The TBB data are converted into binary images according to the threshold of − 52 °C, and the cloud clusters with an area of less than 30,000 km 2 are screened out using the recognition technology.
(2) The 8-domain method is used to extract the boundaries of all binary cloud clusters in each image, and the image processing method is used to extract the features, such as area, and the length of major axis and minor axis. Through the calculation, the characteristic values are obtained as follows.
The centroid position of the MCS is calculated as follows: where n is the number of grids with TBB ≤ − 52 °C within the MCS region. x i , y i and T i represent the longitude, latitude and TBB value of the grid point i, respectively.
(C) According to the area overlapping method and the closest movement of centroid, the convective cloud clusters in two consecutive time steps are matched. If the cloud cluster in the current time and several clouds in the next time are overlapped, the group with the largest overlapping area is selected. Therefore, this method ignores the merging and splitting processes of the convective systems.
(3)  Figure 2a shows the annual variation characteristics of MCS frequency. The annual number of MCS changes greatly from 2010 to 2018, showing a bimodal structure. The number of MCS is the largest in 2010 (60 cases), followed by 2016 (52 cases) and 2017 (45 cases). The number of MCS in these 3 years accounts for 46.3% of the total. There are only 14 cases in 2014, which is the least. In addition, the PECS accounts for the largest proportion every year, leading to the higher frequency of MαCS than that of MβCS. Meanwhile, the frequency of MECSs is significantly higher than that of MCCSs. Figure 2b shows the monthly variation of MCS frequency for all MCS categories. As can be seen, MCSs occur most frequently in July (102 cases), accounting for 30% of the total, followed by June and August (both with 62 cases). There are only 13 MCSs in September, which is the least.
Considering different categories of MCS, MCC, PECS and MβECS all occur most frequently in July, while MβCCS occurs most frequently in August. This is because the rainy season in North China is from the middle 10 days of July to the last 10 days of August every year, and the interaction among the weather systems of the westerly belt, subtropics and tropics is frequent, which is conducive to the generation and development of MCSs. Table 2 shows the duration of all kinds of MCSs. The development stage represents the period from formation to maturity, the weakening stage represents the period from maturity to dissipation, and the total length represents the total time from formation to dissipation. In general, the average duration of MCS in the warm season is 8.5 h. The average duration of MCC is 10 h, which is 0.6 h shorter than that of PECS. The average duration of MβCCS is 3.7 h, which is shorter than that of MβECS. It is revealed that the duration of MECSs is longer than that of MCCSs with the same scale. The average durations of the development stage for MCC and PECS are 6.0 h and 6.2 h, respectively. The average durations of the weakening stage for MCC and PECS are 5.0 h and 5.5 h, respectively, which are both less than those of the development stage. This result indicates that the MCSs in North China develop slowly and weaken rapidly. This is different from the MCSs in the lower reaches of the Yellow River (Zhuo et al. 2012). However, one conclusion is still consistent with them. That is, the characteristics of the PECS are more obvious than those of the MCC, and MCC and PECS also show the characteristics of slow development and rapid weakening in most months. The average duration By analyzing the distribution characteristics of MCS duration in each month (the MCC in September is not taken into consideration due to its small number), we find that the MCC duration is the longest in July, and the PECS duration is the longest in May, which are both followed by August. The durations of MβCCS and MβECS are the longest in August, followed by July. This is because in July and August the western Pacific subtropical high (WPSH) moves northward (Liu et al. 2007), and the frequency of low-level jet increases, which favors the establishment of unstable stratification. Thus, MCS can maintain for a long time in July and August.

Duration and diurnal variation characteristics
The durations of MCC and PECS in the development and weakening stages present a unimodal structure (Fig. 1). In general, the development stage lasts about 4-5 h, and the weakening stage lasts about 4 h. Most of these systems can maintain about 6-7 h. In addition, the duration of some cases is about 8-12 h. The time of the development and weakening stages of the MβCCS and MβECS also show a unimodal structure. The peak appears at about 2 h, and the duration is mainly 3-7 h. Meanwhile, the number of MβCS is less than that of MαCS. Therefore, the average duration of MCS is mainly determined by MαCS. Figure 3 gives the diurnal variations of the formation, maturity and dissipation times of MCSs. It can be seen that the MCS formation time presents a bimodal structure. The peak period is mostly from afternoon to evening. The first peak is mainly at 1500-1700 BJT (Beijing Time), and the second peak is at 0300-0400 BJT. The peaks of maturity time are at 0500-0600 BJT and 2000-2100 BJT, and the peaks of dissipation time are at 0400-0500 BJT and 0000-0100 BJT. It is mentioned above that the mean duration of MCS is mainly determined by the MCC and PECS. Similarly, the diurnal variation characteristics of MαCS (especially PECS) are similar to those of MCS (Fig. 3b). The peaks of MβCS formation are mainly at 1600-1700 BJT and 0200-0300 BJT, which is similar to those of MαCS (Fig. 3c). However, there are many peaks of the maturity time and dissipation time, which have no obvious concentration period.
In summary, combined with the duration characteristics of MCS (Table 2), it can be found that there are two typical lifecycles of the MCSs in North China. The first type of MCSs form in the afternoon, mature in the evening and dissipate in the early morning. The second type of MCSs form in the evening, mature in the early morning and dissipate in the morning or at noon. This result is basically consistent with the conclusions in the lower reaches of the Yellow River (Zhuo et al. 2012) and the conclusions of diurnal variation of convective activity in North China (Chen et al. 2012).  Some studies have shown that the MCS forming in the afternoon is not only related to the large-scale dynamic forcing, but also relates to the underlying surface and local thermal circulation (Yu et al. 2009). However, the MCS forming at night and maturing in the early morning may be related to the diurnal variation of low-level jet and the topographic effect, such as the mountain-valley circulation and the sealand thermal difference (He and Zhang 2010;Zheng et al. 2013). Using composite reflectivity mosaic (CRM) data to study the diurnal variation of convective storm activity in North China during the warm season, Chen et al. (2012) found that the peak frequency of storms at about 0200 BJT at night is related to the nighttime convection on the plain. As the convection propagates downstream, it evolves from vertical convection to inclined convection, so that strong storms are formed at favorable locations downstream. He and Zhang (2010) showed that the diurnal cycle of convective precipitation over North China is related to the diurnal variation of the mountain-plains solenoid (MPS) induced by differential heating between mountains and the plains. There are at least two possible mechanisms responsible for the MPS-related nocturnal convective precipitation over the plains. First, in the afternoon the convective precipitation initiates or intensifies over the mountains or slopes, and then steered by the middle-level mean flow the convective precipitation propagates southeastward to the plains. Second, the upward branch of MPS over the plain can promote the initiation or enhancement of the convection and, meanwhile, 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

Minimum TBB of MCSs, the area and eccentricity at the MCS maturity time
The minimum TBB of MCSs, the average area and eccentricity at the maturity time for different MCS categories are shown in Table 3. It can be seen that the average area of MCSs at the maturity time is about 1.49 × 10 5 km 2 , which is smaller than those in North America (Jirak et al. 2003) and the lower reaches of the Yellow River (Zhuo et al. 2012). The average area of MβCCS and MβECS are similar, with the value being 0.43 × 10 5 km 2 . The average area of PECSs is the largest with the value being 2.03 × 10 5 km 2 . A large part of PECSs develop from the merging of different convective clouds, and thus the average area of PECSs is greater than that of MCCs. The average eccentricities of MCC and PECS at the maturity time in North China are 0.80 and 0.43, respectively, which are slightly smaller than those in the United States (0.83 and 0.50, respectively) (Jirak et al. 2003). Similarly, the eccentricities of MβCCS and MβECS are 0.80 and 0.48, respectively, which are also slightly smaller than those in the United States (0.84 and 0.53, respectively). The average minimum TBB of MCCs in North China in the warm season is − 72 °C, which is about 1 °C lower than that of PECS and close to that of MCC in the lower reaches of the Yellow River in summer (Zhuo et al. 2012). It is about 8 °C higher than the average minimum TBB of MCC (− 80 °C) in the southern China (Xiang and Jiang 1995). Meanwhile, there is a MCC case with the minimum TBB of − 88 °C. It is 2 °C higher than that of the case in summer in the lower reaches of the Yellow River (Zhuo et al. 2012), and about 4 °C lower than that of the existing case in the southern China.
The monthly variation of MCSs (Table 3) show that the mean area of MCS at the maturity time decreases gradually from April to August. The eccentricity and minimum TBB show a unimodal structure, with the maximum values of 0.54 and − 71 °C in July, respectively. It also can be seen from Table 3 that the minimum TBB of MCC in June and August is lower than that of PECS, and the two are almost equal in other months. In addition, the average eccentricities of MCC and PECS are the largest in July, the average area of MCC at the maturity time in July is larger than that of the PECS in the same scale, and the minimum TBBs of the two are almost equal. The minimum TBB of all MCS categories in August is the lowest compared with in other months. That is, the convection develops more vigorously from July to August. These characteristics of MCC and PECS may be related to the northward movement of the western Pacific subtropical high (WPSH) in July and August. The North China is located on the edge of the WPSH, and the warmwet southwesterly airflow continuously provides sufficient water-vapor and momentum to North China, which is conducive to the development of MCSs (Zhao and Sun 1990).

Geographical distribution of MCSs
The geographical distribution of MCSs in North China is given in Fig. 4. The formation area of MCCSs (MCC and MβCCS) is relatively concentrated, with two concentration belts. One is the east-west-oriented concentration zone along the Henan Province, the Shandong Province and the Yellow Sea, which is consistent with Zheng et al. (2008). The other is along the central and western Shandong, the Tianjin City, the west of Bohai Sea and the northeast of Hebei Province, which is similar to the spatial characteristics of warm-season convective storms (Su et al. 2011;Chen et al. 2012) and precipitation (He and Zhang 2010) in North China. At the formation time of MECSs (PECS and MβECS), the distribution of the centroid positions is relatively scattered; however, it can be found that they are also distributed very densely in the same area as MCCSs. To analyze the influence area and frequency of MCSs, the frequency distribution of all centroids is counted at the 1.0° × 1.0° grids. As shown in Fig. 5a, the regions with the most frequent influence of MCSs in the warm season are mainly located in the northeast of Hebei Province (southern Chengde, Tangshan, Qinhuangdao and Cangzhou), the central and western Shandong, the western Bohai Sea and the Yellow Sea, with the occurrence frequency of more than 20. In the south of Chengde and west of Shandong Province, the occurrence frequency can reach the maximum of 32-33. In terms of the monthly distribution, the MCS is most frequent in summer from June to August (Fig. 5b-d). The area with the most frequent influence in June is located in the southern-central Chengde in Hebei Province and the central Shandong Province with the occurrence frequency of 7-12. The area with the most frequent influence in August concentrates near Tianjin City, Cangzhou in Hebei Province, the western coastal area of the Bohai Sea and the southwest of Shandong Province with the occurrence frequency of 9-13. In July the area influenced by MCSs and the occurrence frequency is the largest. The influence area in July is mainly located in the south of Chengde in Hebei Province, the west of the Bohai Sea, the central and southern Shandong Province and the Yellow Sea, with the occurrence frequency of 10-18. Liu et al. (2007) showed that the northward movement of the WPSH in July and August can influence the location and distribution of heavy rainfall over North China. Chen et al. (2012) found that long-lived and organized convective systems such as MCS occur much more frequently on the southeastern plains, especially in July.
In general, the high-frequency area of MCS activity in North China from June to August is consistent with the MCS formation area and the high-value area of summer precipitation around the Bohai Sea (Duan et al. 2013). There are at least two reasons for the MCS concentration belts. First, the convection initiates locally in the concentration belt and then are strengthened into MCSs. Second, the convection initiates in the afternoon near the top of the mountain ranges, and then propagates downslope and southeastward to the central North China Plain around midnight and the early morning. Later, the convection can reach the Bohai Sea (He and Zhang 2010;Chen et al. 2012Chen et al. , 2014.

Moving characteristics of MCSs
MCSs often bring much convective precipitation. Therefore, the research on the moving characteristics of MCSs is helpful to improve the accuracy of precipitation forecast, especially the short-term nowcasting. Zeng et al. (2013) found that the main moving path of the MECSs is from southwest to northeast, while the MCCSs tend to move from northwest to southeast. Wang et al. (2013) surveyed the MCSs under the background of the northern cold vortex and found that there are two main moving paths for the PECS. One is from the west to southeast and the other is from the west to northeast. Moreover, there are a few MCCs moving from the northeast to southwest and from the east to west.
To further explore the moving characteristics of different categories of MCSs in North China in the warm season, in this study, the due north is regarded as 0°, the due east is considered as 90°, and the moving direction is divided into 16 directions. The moving direction and distance of the cloud cluster are calculated according to the position of the cloud clusters at different times. Figure 4a1-a4 shows the moving tracks of different MCSs. Figure 4b1-4b2 shows the moving distance and direction by taking the location of MCS formation as the origin. Table 4 shows the detailed statistical results. Figure 4 and the statistical results in Table 4 show that from 2010 to 2018 there are 101 MCSs moving to the due east and 97 MCSs moving to the east-northeast, accounting for 29.8% and 28.6%, respectively. The MCCs moving to the east-southeast and northeast are relative less, accounting for 11.5% (39) and 8.0% (27), respectively. The cases with these four moving directions accounts for 77.9% of the total MCS, which is related to the fact that the westerly belt system is dominant in the middle latitudes over North China for a long time. The comparison of various categories of MCSs ( 3 2 3 5 2 3 4 1 1 4 3 6 6 6 2 1 2 1 1 3 6 3 6 5 5 7 9 6 5 6 9 13 5 7 3 1 2 4 6 6 10 6 2 5 7 11 14 13 9 9 2 2 1 1 1 1 1 2 3 9 3 4 6 3 6 4 7 3 7 12 5 22 7  1 2 5 1 2 3 3 2 1 4 1 3 1 1 2 2 4 9 7 10 2 6 2 2 1 2 2 2 3 3 6 5 11 12 11 6 8 4 3 3 2 3 1 1 2 1 2 1 4 3 3 2 3 3 2 2 3 1 1 2 1 4 1 5 4 5 2 3 2 1 1 2 1 2 1 6 8 4 4 2 1 1 1 1 1 1 3 2 2 1 4 2 7 11 2 1 5 5 1 1 2 2 3 3 2 1 1 1 3 4 6 2 1 2 4 2 2 5 1 1 3 3 3 1 6 1 2 3 2 2 3 4 3 2 1 2 2 5 3 2 5 5 1 2 There are also a few MCCs with the moving direction of east-southeast. The MECS mainly moves in the northeast-east direction, accounting for about 65%. Different from the MCCSs, there is a part of MECSs moving to the southwest and northwest. Section 4.3 will discuss the MCS movement in detail from the perspective of advection and storm propagation (Table 5). Through the statistics on the moving distance of MCSs ( Fig. 4b and Table 5), it is found that the average moving distance of MCSs is 440 km. Considering different categories of MCSs, the average moving distance of PECS is 580 km, and is 140 km larger than that of MCC. The average moving distances of MβCCS and MβEC are the same (160 km), which is approximately one-third of that of MαCS. There are 39 quasi-stationary MCS cases with the moving distance below 100 km, accounting for 11.5% of the total MCS, and most of them are MβECS. There are 30 MCSs with the moving distance above 1000 km and almost all of them are PECSs, which is related to the continuous initiation and merging of convective clouds in front of the PECSs. Moreover, it can be found that the moving distance of MCC is mainly within 200-500 km, accounting for more than a half of the total. There are a significant number of PECSs with the moving distance between 200-500 km and 500-1000 km, both accounting for more than 1/3 of the total. The moving distance of MβCSs is almost within 100-500 km, mainly between 100 and 200 km.

Environmental conditions
The motion of a convective system can be considered as the vector sum of (1) an advection component approximated by the direction and magnitude of the mean cloudlayer wind, which is represented by the average wind of 300, 500, 700 and 850 hPa, and (2) a propagation component governed by the location of new cell (Newton and Katz 1958;Corfidi 2003). MCS propagation can be influenced by many factors, such as the distribution of convective available potential energy (CAPE), convective inhibition (CIN), outflow boundaries, gravity waves and orographic effects. Past studies have shown that the propagation component is directly proportional but opposite in direction to the LLJ (Corfidi et al. 1996;Corfidi 2003). As mentioned above, most MCSs in North China move eastward, they have the longest moving distance, and their average moving direction is consistent with the advection component ( Fig. 6a-1). The MCSs moving northward are formed in the southeast quadrant of the mid-high latitude vortex, their moving distance is longer and the moving direction is also affected obviously by advection ( Fig. 6b-1). Figure 6c-1 shows that MCSs moving southward have a larger moving angle to the advection direction, which means that advection component has less influence and the propagation component perhaps has greater influence. The column (3) of Fig. 6 shows that MCS is mainly generated in the baroclinic zone of high energy tongue (high CAPE) (Maddox 1983;Laing and Fritsch 2000). Its occurrence and development can be affected by the westerly trough or cold vortex in the middle troposphere, the 850 hPa shear line, the surface inverted trough and the surface convergence line (Fig. 6). The WPSH peripheral southerly LLJ transports warm and humid air, and the left front and the heading of LLJ bring dynamic uplift (Maddox 1983;Wilson et al. 2007). The column (1) of Fig. 6 shows that the MCS formation location mostly corresponds to the region with 0-6 km vertical wind shear larger than 10-12 m/s, where it is conducive to the strengthening of MCSs (Maddox 1983;Wilson et al. 2007). From Fig. 6b-1, c-1, it is found that the high-frequency centers of MCS formation are located at the southern slope of Yanshan Mountain and the eastern slope of Taishan Mountain. Considering the topographic amplification effect, the southerly or southeasterly wind in the boundary layer is forced to rise on the windward slope, which is conducive to the formation or enhancement of convective storms (Sun 2005;Chen et al. 2012;Zhang et al. 2013). Yuan et al. (2014) showed that topography is very important for the spatiotemporal distribution of short-term precipitation in the warm season in North China. Figure 7 shows the conception models of main environmental characteristics at the formation time of easterly and northerly moving MCSs.

Conclusions
In this study, the hourly TBB data from the FY- There are significant diurnal variations for the MαCSs in North China, and the curves of MαCS diurnal variations show a bimodal structure. Many MαCSs form in the afternoon, mature in the evening and dissipate in the early morning. Some others form in the evening, mature in the early morning and dissipate in the morning or at noon.
The high-frequency areas of MCS formation and MCS activity in the warm season in North China are overlapped. There are two MCS concentration belts, namely, the east-west-oriented belt along Henan Province, Shandong Province and the Yellow Sea, and the south-north-oriented belt along central-western Shandong, Tianjin City, the west of Bohai Sea and the northeast of Hebei Province. Most of MCSs occur in the summer from June to August. MCSs have the widest influence range and the highest frequency in July, followed by August.
The average area of MCSs at the maturity time in the warm season in North China is about 1.49 × 10 5 km 2 , which is smaller than that in North America and the lower reaches of the Yellow River. The mean eccentricities of the four categories of MCSs at the maturity time are all smaller than those in the United States. The mean minimum TBB of MCCs in the warm season in North China is − 72 °C, which is close to that in the lower reaches of the Yellow River in summer and about 8 °C higher than that in the southern China. The convection develops much more vigorously from July to August. The average eccentricity of MCC and PECS is the largest in July, and the average area of MCC at the maturity time is larger than that of PECS in the same scale. The minimum TBB of all categories of MCSs in August is the lowest in all months.
In the warm season from 2010 to 2018, 58.4% of the MCSs move to the due east and east-northeast in North China, and the MCSs moving to the east-southeast and northeast account for 19.5%. Different from the MCCSs, a part of MECSs move to the southwest and northwest. The easterly and northerly moving MCSs are obviously affected by the advection component, and their moving distance is longer. The MCSs moving southerly have the shortest moving distance, and the influence of advection component is smaller than the storm propagation component. The average moving distance of MCSs is 440 km. The average moving distance of MβCS is between 100 and 200 km, which is one-third of that of MαCS. The moving distance of MCC is mainly 200-500 km, and that of PECS is between 200 and 1000 km.
The MCSs in North China are mainly formed in the high temperature, high humidity and high energy area. The westerly trough or vortex in the middle troposphere, the lowlevel shear line, the surface inverted trough or the surface convergence line are favorable dynamic systems for MCSs. The MCS initiation is often accompanied by the LLJ and the relatively strong 0-6 km vertical wind shear. Terrain effect is also one of the important factors for MCS formation.
Data availability In this study, the temperature of brightness blackbody data from FY-2E geostationary satellite can be found at the service website of the National Satellite Meteorological Center of China (http:// satel lite. nsmc. org. cn/ porta lsite/ defau lt. aspx), the ERA5 reanalysis data can be found online at the following link (https:// apps. ecmwf. int/ datas ets/ data/ inter im-full-daily). The output data of identified MCS cases can be shared if requierd by anyone.

Declarations
Conflict of interest There are no relevant financial or non-financial interests to disclose.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.