STUDY ON SENSITIVITY OF WIND FIELD VARIATION TO STRUCTURE AND DEVELOPMENT OF CONVECTIVE STORMS

The environmental factors during the genesis of convective storms vary significantly^[1-4], which will, to some extent, affect the development of storms and the structural features of storms (i. e., shape, structure and intensity). If changes of environmental factors occur at different stages of the convective storms' development, the evolution of storms will be remarkably affected ^[8]. Studies on the physical mechanism associated with the development, organization and structural characteristics of convective storms have always been highly concerned ^[9]. The structures of the initial disturbance and different configurations of environmental fields have an obvious impact on the development of the convective systems ^[10]. The storm's development environment, including the distribution of Convective Available Potential Energy (CAPE), vertical wind shear, and water vapor, etc., is a considerable factor affecting storm structure and strength ^[11-15]. Among them, the distribution and variation of vertical wind shear play indispensable roles in the development of convective storms. These variations, including the wind shear strength and the shape of the wind vector rotating with height ^[16-17], determine the rotation in the storm to some extent ^[18]. Xu ^[19] finds that convection motion causes horizontal vorticity under vertical wind shear due to convergence and tilting, and the decrease of atmospheric density with height is one of the reasons why the center of vorticity maintained in the middle and lower part of convection motion. Large low-level wind shear is conducive to the development of severe convective clouds by extending their life cycles so as to promote precipitation ^[20]. The interaction between low-level wind shear and the ground cold pool also contributes to the development of linear convection ^[21-24]. CAPE is one of the important parameters to represent the environmental thermodynamic characteristics, which affects not only the spread and development of convective systems, but also the convection properties and intensity ^[25]. In a suitable environment for CAPE, the interaction between buoyancy and wind shear will promote the development of convective storms.

The above studies analyzed some possible impact mechanisms of storm's development environment on storm structure and intensity from different perspectives. Many involved environmental factors and the interactions among them make the impacts of environmental factors unclear. Moreover, the development of convective storms is of regional characteristics, and the characteristics of temperature, humidity, dynamic environment and underlying surface in different regions are quite different, so are the impacts of these environmental factors the same? With focus on the development of convective storms in a typical strong convection case in the Jianghuai area, numerical experiments were conducted in this study to discuss the possible effects of wind field variations on the structure and development of convective storms, and to further understand the physical formation and development processes of convective storms, so as to improve the predictability of such kind of weather.

The data used in the present study includes the observed hail data (WSB), sounding data, Doppler weather radar data and WRFv3.8.1 mesoscale numerical simulation data of Jiangsu Province on April 28, 2015. All the time referred to is Beijing Time unless otherwise specified.

At 08: 00 on April 28, two troughs and one ridge appeared at mid-high latitudes at 500 hPa of Asia (Fig. 1a). The ridge was located near Lake Baikal and had a trough on both the right and left side. A cut-off vortex, located in the North China region and matching the temperature trough, was formed at the south of the trough on the west side of the Sea of Okhotsk, and the vortex at 850 hPa and below matched up with the warm tongue extending eastwards (figure omitted), thereby forming an instability stratification with upper warmth and lower cooling. On the afternoon of April 28, as the cold vortex gradually moved to the south, the northerly airflow behind the vortex carrying cold air at 500 hPa enhanced and moved southward, which promoted the development of stratification instability and enhanced the vertical wind shear in the middle and lower troposphere. In this way, a favorable environment for the occurrence of convective weather was created. Associated with convergence line, the mesoscale low pressure, developing and moving on the ground in the Jianghuai area, triggered the initial convection.

Figure 1.  (a) 500 hPa height (solid line, unit: dagpm) and temperature field (dashed line, unit: ℃) at 08:00 on April 28, (b) hail observation on Aug 28, 2015.

In this context, a severe convective weather accompanied with hail (Fig. 1b), thunderstorm gale and short-term heavy precipitation (figure omitted) occurred in the Jianghuai area from the afternoon of April 28 to the first half of the night of April 29, 2015. According to the statistics from the civil affairs department, 610, 700 people in Jiangsu Province people suffered from the convective weather, five people died due to the disaster, and the direct economic loss reached 681, 959, 687 Yuan.

On the afternoon of April 28, the scattered and small-scale convective cells (figure omitted) in Jiaxiang County of Shandong Province near the surface mesoscale depression began to form. It developed rapidly and moved to the southeast; meanwhile, convective cells were constantly simulated along the convergence line extending southward from the mesoscale low-pressure center on the ground, and a strong convective echo strip gradually formed. At 16:00 (Fig. 2a), the convective echo strip was organized by several vigorous supercell storms. The reflectivity factor of the centers exceeded 60 dBZ, and the top height rose to 14-18 km. The storms were featured with strong echo center (greater than 60 dBZ) overhanging and echo pendency (Fig. 2b). Correspondingly, the velocity fields of storms were accompanied by a distinct mesocyclone (figure omitted). These supercell storms successively produced hail in many places at the northern border of Jiangsu and Anhui Provinces.

Figure 2.  (a) Reflectivity factor at 16:00, (b) vertical cross section of reflectivity factor at 16:13 along the black line in Fig. 2a, (c) reflectivity factor at 18:06, and (d) vertical cross section of reflectivity factor at 18:06 along the black line in Fig. 2c (unit: dBZ, the circle indicates the supercell storm causing the hail).

During the southward movement of convective echo, the range of echo area gradually expanded, and several well-developed convective storms were located at the right edge in the movement direction of the echo area. At 18:06, the supercell storm at the forefront of the echo area arrived at Luhe (Fig. 2c), and the storm top extended to 12 km, which was lower than before. The overhanging strong center (> 60 dBZ) in the storm fell down to the ground (Fig. 2d). At this time, Luhe suffered from 50 mm heavy hails, accompanied by a thunderstorm gale of 20 m s^-1. Afterwards, the convective storms continued to move southeast, passing through Yizheng, Zhenjiang, Danyang, Changzhou, Wuxi, Suzhou, and Wujiang, where it caused severe heavy hails successively. At 22:00, it moved to Shanghai and brought severe convective weather to the western part of Shanghai, and then gradually weakened and dissipated.

The following numerical tests for this severe convection process would be carried out to explore the possible impacts of wind field variations on the development of convective storms.

This severe convection occurred under the background of the cold vortex, accompanied by hail, thunderstorms, gale and short-term heavy precipitation. When the Jianghuai area was located at the rear of the cold vortex, the upper and middle troposphere prevailed a northerly airflow, which tended to form a relatively dry environment that is prone to produce hail and thunderstorms ^[26]. During the development of severe convection, the northerly wind in the upper and middle troposphere behind the cold vortex strengthened, and the gradual formation of a stronger northerly jet from the North China to Jianghuai area promoted the development of convection ^[27]. If the environmental wind field changes, does it affect the development of convection? If so, how does it affect?

The control experiment plan in this study used the non-hydrostatic mesoscale numerical model WRFv3.8.1, and took the reanalysis data of ERA-interim 0.75°×0.75° resolution of ECWMF at 18: 00 UTC on April 27 for every six hours as the initial and boundary conditions. The period of simulation time was 36 hours and was output hour by hour. A double nested grid (Fig. 3) with horizontal resolution of 9 km, 3 km and vertical stratification of 70 layers with inhomogeneity was used in the control experiment. MYJ boundary layer scheme, the scheme of the unified Noah land-surface model, the RRTM long-wave radiation scheme, and the Goddard short-wave radiation scheme were adopted in the model. On the basis of control experiment, numerical tests were carried out, and the experiment plan is shown in Table 1. Considering the spin-up time in the simulation process of mesoscale model, we selected the data six hours after the model integration for the present study.

Figure 3.  Numerical simulation of double nested areas.

A set of numerical tests (Table 1) was designed to study the effects of changes in environmental dynamic characteristics on the occurrence and development of storms. The vertical wind shear conditions for the initial growth of the storm were changed by adjusting the middle - and low-level wind fields in the initial field of d02 region. The average zonal wind u and meridional wind v near the initial field around convection occurrence region, and the corresponding vertical distribution characteristics after adjustment are shown in Fig. 4a and Fig. 4b, respectively.

Figure 4.  Vertical distribution of (a) zonal wind u and (b) meridional wind v in the initial filed of control experiments and wind filed experiments (unit: m s^-1).

From the results of the control experiments, the initial period of convection is close to the actual situation, and at around 13:00 (figure omitted), an initial convective cells gradually form at the junction of Shandong, Jiangsu and Anhui Provinces, and then develop rapidly. At 15: 00 (Fig. 5a), the isolated convective cells along the convergence line develop and gradually connect with each other, and the reflectivity factor of the center on the echo strip reaches 55-60 dBZ, developing into a supercell storm. The intensity, location and shape of the convective storms are very similar to those in the real situation after one hour (Fig. 2a), but the narrower echo strips on the east side of the severe echo strips are not simulated. At 17:00 (Fig. 5b), the range of the convective echo area is enlarged, and the front-end convective storm develops fast and has moved to Luhe, which is very close to the actual echo shape and location of the storm after one hour (Fig. 2c), but the arousing convective cells around the convective storm are significantly stronger than the actual situation. Furthermore, a comparison of the formation of simulated convective storms at different development stages with the actual situation shows that the control experiment can basically simulate the formation and development process of convective storms, including the intensities and movement paths of convective storm at different stages of formation and development. However, the convective storm simulated in the control experiment develops more rapidly, and its development and lysis time are about one hour ahead of the that in the real situation. In view of the ideal effect of the model simulation, the following numerical experiments would be carried out on the basis of the control experiments.

Figure 5.  . Reflectivity factor in control experiments (unit: dBZ) at (a) 15: 00 and (b) 17:00 (the circle indicates the supercell storm causing hails).

The possible effects of wind field adjustment at different heights on the whole layer of ambient wind and its evolution at later stage were further analyzed. Fig. 6a1-6c1 shows the temporal evolution of wind field at different heights in the control experiment. From the wind field at middle and lower level (Fig. 6a1-6b1), we notice that the gradual formation and development of a mesoscale low vortex corresponded to the junction of Shandong, Jiangsu and Anhui Provinces. After integration for ten hours, the cyclonic circulation near the vortex become clearer, the vortex moves southward with time, and wind changes gradually from north to south in the eastern Jiangsu. At 500 hPa (Fig. 6c1), the northerly wind in western Jiangsu and its upstream areas of Shandong and Anhui Provinces significantly strengthens, which is consistent with the strengthening northerly wind in the rear part of the North China cold vortex moving southward in observation. When the low-level wind field strengthens, it can be seen from Fig. 6a2 that the wind field near the mesoscale vortex at the junction of Shandong, Jiangsu and Anhui Provinces becomes stronger than that in the control experiment, and the strengthening trend of the wind field is maintained over time, but the differences between the wind field changes at 700 hPa and 500 hPa level in this experiment and those in the control experiment are not significant (fig. 6b2-c2). With the increasing of the middle-level wind field (Fig. 6a3-c3), the wind field in each layer shows a significant strengthening trend. At 850 hPa (Fig. 6a3), the vortex moves southward faster, and the northerly wind behind the vortex moves southward rapidly. The reasons are as follows. On one hand, because of the difference of the vertical distribution of the observed wind field, the adjustment of the middle-level wind is bigger than that of the low-level wind. On the other hand, the mechanism of the momentum downward transmission from upper level may play a role. In the wind field weakening experiment (figure omitted), after the low-level wind weakens, the wind field at 850 hPa weakens only in 0-6 hours compared with that in the control experiment, while the difference of the subsequent changes in the low-level wind field, including the changes in the upper and middle-level wind field, are not significant between the two experiments. After the weakening of the mid-level wind field, the northerly wind near the vortex in the lower troposphere or in the western and upper reaches of Jiangsu Province at 500 hPa is obviously weaker than that in the control experiment.

Figure 6.  Temporal evolution of wind fields at different heights in each experiment (unit: m s^-1); Ctrl(a1-c1), Test-low-wind-200% (a2-c2) and Test-middle-wind-200% (a3-c3) (a. 850 hPa; b. 700 hPa; c. 500 hPa; black wind vector: 6-hour integration; red wind vector: 10-hour integration; blue wind vector: 18-hour integration).

The strong vertical wind shear helps to form a long-life well-organized convective system. During the development of the severe convection, the northerly wind at the rear of the cold vortex in the middle troposphere develops into a northerly jet over the convection zone, which promotes the rapid development and a long-term maintenance of convective storms. So, how does the development of convection differ when the wind fields in the middle and lower troposphere are changed?

Figure 7 shows the distribution of reflectivity factors during the initial and mature stages of convection development after the changes of wind fields in the middle and lower troposphere. It can be seen that the intensity and shape of convection development vary greatly in different experiments. While the low-level wind field weakens, compared with the results of control experiment(Fig. 5a), the convective cells at the northern junction of Jiangsu and Anhui Provinces are weaker in the initial stage of convection development (Fig. 7a). In the mature stage (Fig. 7e), the echo area extends and gradually expands southwards. The convective storm at the front end of the echo zone moves to the vicinity of Xuyi County, with its center intensity reaching 55 dBZ. The intensity of convective storm center is slightly weaker than that in the control experiment, but the range of the strong echo area is larger. When the low-level wind field strengthens (Fig. 7b, f), the development of convection in the initial and mature stages of the storm is obviously more intense than that in the weakening experiment of the low-level wind field, and also slightly stronger than that in the corresponding development stage of convection in the control experiment. After the weakening of the wind field in the middle-level troposphere, the initially formed convection is significantly weaker than that in the control experiment and other experiments (Fig. 7c). In the mature stage of convective storms (Fig. 7g), the convective storms move to the northern part of Xuyi County, which are weaker than those in the control experiment and other wind field experiments, and the echo area is smaller. After the enhancement of middle-level wind field, the convective cells generate one hour earlier (figure omitted), and the convection development is more intense than that in other experiments in the same period (Fig. 7d, h). The convective storms in the mature stage (Fig. 7h) are significantly different from the control experiment (Fig. 5b). Instead of forming a fast developed blocky supercell at the front end of the echo area, a linear convective system is formed in the central part of Jiangsu and Anhui Provinces. The reflectivity factor of echo center is greater than 60 dBZ, which is similar to the center intensity in the control experiment, but with a slightly larger high-value area. In addition, it is noteworthy that the convective storms move southward significantly faster after the enhancement of the middle-level wind field, and the formation and extinction time of storms in this experiment is about one hour earlier than that in the control experiment and other wind field experiments.

Figure 7.  Reflectivity factors of convective storms in the initial stage (a-d) and mature stage (e-h) in each wind field experiment (unit: dBZ). (a, e) Test-low-wind-50%, (b, f)Test-low-wind-200%, (c, g) Test-middle-wind-50%, and (d, h) Test-middle-wind-200%.

The results of above wind field experiments show that the development of convective storms is sensitive to the changes of wind fields in the middle and low level. After the enhancement of middle-level wind field, the initial convections form earlier, and the convective storms develop more rapidly and move faster. Furthermore, in the mature stage, the new convective storms on the right side of its moving direction gradually organize into a linear convective system. When the low-level wind field enhances, the convective storms in the initial stage are stronger than those in the control experiment. In the mature stage, the structures of convective storms are dense and the center intensities exceed 60 dBZ, which is similar to the control experiment. However, the range of the echo area at the rear part of the storm is smaller than that in the control experiment, and the speed of the convective storm moving southward is slower than that in the control experiment. After the weakening of middle- or low-level wind field, the intensity of storms in this group is weaker than that in the control experiment, and the storm is the weakest when the middle-level wind field weakens. The following parts would further analyze the differences in structural characteristics of different development stages of convective storms after the adjustments of the wind fields at different levels.

The most intense convective storms in the mature stage in each experiment are selected as the research objects to compare their structural characteristics. Fig. 8a-8e shows the vertical structures of the convective storms in the mature stage in different wind fields. The structures and intensities of convective storms in these experiments are evidently distinct from each other. The convective storms in the mature stage in these experiments have commonalities: the tops of the storms all extend to around 200 hPa; the reflectivity factors of the storm centers exceed 50 dBZ, and are mostly located in the middle or upper part of the convective storms; the warm and humid inflow from the low level of the storm leads to the weak echo area in the middle and low level of the storm and the overhanging structure thereon. But it is obvious that the storm intensity increases by about 5-15 dBZ after the enhancement of middle- or low-level wind field, compared with the situation when the wind field weakens. In particular, when the middle-level wind is enhanced or weakened, the differences in storm intensity between the two experiments are more obvious: when the middle-level wind field weakens (Fig. 8d), the intensity of the storm center is 45-50 dBZ, which is the weakest among these experiments; when the middle-level wind field enhances, the convective storm develops to the most intense (Fig. 8e), and the intensity center is located in the middle of the storm with the intensity over 60 dBZ, and the area with the reflectivity factor exceeding 50 dBZ extends from the ground to 300 hPa.

Figure 8.  Flow field and vertical cross section of reflectivity factor along the black line in Fig. 7e-7h in the mature stage (shaded, unit: dBZ). (a) Ctrl, (b) Test-low-wind-50%, (c) Test-low-wind-200%, (d) Test-middle-wind-50%, and (e) Test-middle-wind-200%.

According to the structure of flow field in the convective storm at this moment, what is corresponded to the intensely developed convective storm in the environment with enhanced middle-level wind is a deep secondary circulation. The secondary circulation is very narrow, smaller in scale than the secondary circulation associated with convective storms in the control experiment. It only spans about 30 km from north to south. The updraft is located on the south side of the secondary circulation and is associated with the warm and moist air from the low level. This updraft flows near the top of the storm and begins to flow outward, forming a divergence at the top of the storm as well as a thin secondary circulation on the south side of the top of the storm. The downdraft on the north side of the secondary circulation accompanied by the convective storm merges with the downdraft of the convective storm on the north side to form a distinct northerly airflow near the ground, and then it intersects with the southerly warm and moist flow at front of the storm, which enhances the low-level convergence of the storm and is conducive to the enhancement of warm and moist inflows, thus leading to the intense development of convective storms. At the north side of the convective storm, there is another developing convective storm accompanied by a relatively weak secondary circulation, which is only located in the middle and lower troposphere. In the middle-level wind weakening experiment, the convective storm is also accompanied by a deep secondary circulation, with the updraft still located on the south side of the secondary circulation, and the ascending region is wider, but the intensity of ascending motion is weaker than that of the former (figure omitted), so do the development of convective storms. After the enhancement of low-level wind field, the secondary circulation associated with the convective storm is only located in the lower troposphere and the center of the circulation is below 800 hPa, though the updraft associated with the low-level warm and moist inflow develops to the vicinity of the storm top, resulting in the unbalanced development of the updraft and downdraft, with the downdraft only located in the middle and lower troposphere. When the low-level wind field weakens, the ascending branch of the secondary circulation associated with the convective storm moves to the north side of the circulation, so the ascending airflow in the storm dramatically weakens (figure omitted), and forms a weak and thin secondary circulation at the low level in the front of the storm.

With reference to the radar data analysis, the storms in the mature stage show a typical super-cell structure, and the storms are accompanied by mesocyclone. Does the variations in the wind field affect rotation in the convective storms? What are the differences in the rotational characteristics associated with the storms in different environmental wind fields? The vertical distribution characteristics of the internal vorticity of the storms in the mature stage in the wind field experiments will be used to analyze the differences in the rotational characteristics of the storms in different wind fields (Fig. 9a-e). After the enhancement of middle-level wind field, the convective storms develop most intensely. The vorticity characteristics of the developed convective storms in this experiment (Fig. 9e) show that a strong and narrow vertical developing cyclonic vorticity area is formed near the middle level of the storm, which lies on the north side of the strong updraft, and two centers emerge in the cyclonic vorticity region. The stronger center is located at 700 hPa with a value of 5 10^-3 s^-1, significantly stronger than vorticity in the control experiment, while the weaker center is located near the top of the storm. Next to the north side of cyclonic vorticity area is the anti-cyclonic vorticity area, lying between 400 to 500 hPa with a center value of -1.5 10^-3 s^-1, and its intensity is weaker than that in the cyclonic vorticity area. Apparently, the rotation in the middle level of the convective storm is stronger than that in the low level. To some extent, the existence of the rotational flow in the middle level helps to maintain the steady coexistence of the updraft and the downdraft in the storm, promoting the further development of the storm. Another pair of cyclonic and anti-cyclonic vorticity areas appears in another convective storm on the north side of this convective storm, but their locations are basically below 700 hPa and the intensities are obviously weaker. In the middle level wind field weakening experiment, the updraft area lying on the front side of the storm correspond to the cyclonic vorticity area (Fig. 9d). The cyclonic vorticity slopes slightly southward with height. It has two vorticity centers, lying between 600-900 hPa and 400-600 hPa, respectively, and the center of the latter is slightly southward. The corresponding downdraft area at the backside of the storm is the anti-cyclonic vorticity area. Compared with the middle-level wind field enhancement experiment, the rotation in the storm weakens in this environment, and the differences in the rotation intensity between the middle and low levels are reduced. In the low-level wind field enhancement experiment (Fig. 9c), the vorticity at the middle level of the storm has not yet developed significantly. The corresponding vorticity center appears near the middle part of the storm with a center value of 1.5 10^-3 s^-1. It is noteworthy that the cyclonic vorticity at the middle and lower levels in front of the convective storm has obviously developed, which may be associated with the enhancement of warm and moist inflows in the boundary layer. The convection intensity weakens as the low-level wind field weakened (Fig. 9b). The rotation in the storm is mainly located in the upper and middle part of the storm, and the cyclonic vorticity area is located on the rear side of the storm, with the center near 500 hPa and a center value of about 1.5 10^-3 s ^-1. The anti-cyclonic vorticity area is located in front of the storm, with the relative large value area between 400 and 650 hPa.

Figure 9.  The vertical cross section of the vorticity (contours, unit: 10^-3 s^-1) and reflectivity factor (shaded, unit: dBZ) along the black line in Fig. 7e-7h in the mature stage of storms. (a) Ctrl, (b)Test-low-wind-50%, (c) Test-low-wind-200%, (d) Test-middle-wind-50%, and (e) Test-middle-wind-200%.

To sum up, the results of the above wind field adjustment experiments show that there are obvious differences in the structure and development of convective storms after the change of the wind field in the middle and lower troposphere. The enhancement of the middle-level wind field is conducive to the reduction of the convective scale of the storm, but the stronger development of convection and the more intense rotation at the middle level of the storm caused by it, are conductive to the formation of a larger negative pressure, thereby promoting the development of the vertical circulation in the storm, and leading to the intense development of convective storms. The development of low-level wind field facilitates the development of updraft on the inflow side of the storm, which promotes the development of the cyclonic vorticity on the inflow side of the storm at the middle and lower level. The secondary circulation accompanied by the storm is formed in the middle and low troposphere. When the wind field weakens, the development of convection weakens in middle or lower layer, especially in the middle layer. The scale of the secondary circulation accompanied by the convective storm increases, but the intensity and the vorticity in the middle layer of the storm decrease compared to other experiments.

The variations of environmental wind field have an evident impact on the intensity and structure of storms. How does wind field variation affect the occurrence and development of the storm? Will this variation be transmitted in other ways directly or indirectly affecting the development and structure of the storms? A further discussion will be presented in the following sections.

Changes in wind fields at different levels have changed the dynamic environment for the storm growth. In addition, due to the different nature of the air carried by the wind field: cooling, warming, drying and wetting, the intensity and distribution of temperature and humidity advection in the convection area will be affected, which may give rise to changes in the structure of atmospheric stratification in the convection area. Fig. 10a-10e show that the average potential pseudo-equivalent temperature θ_se changes over time and height in the generating area of convection (30-35°N, 117-120° E) in the control and wind filed experiments. It is obvious that θ_se gradually decreases with height below 500 hPa in the middle and lower troposphere with feature of stratification instability, and all these instabilities develop further in the afternoon. In the control experiments, from 12:00 to 14:00, the θ_se center with a value of 328 K appears near the ground. When the low-level wind weakens (Fig. 10b), the formation time of the warm and moist center near the ground is close to that in the control experiment, but the center value reduces. In the course of severe convection, the southerly in the boundary layer is below 850 hPa, so the weakening of the low-level wind field may lead to the weakening of the warm and moist advection in the convection area, thus inhibiting the development of instability stratification. After the enhancement of low-level wind field, the warm and moist center near the ground appears slightly earlier than that in the control experiment, but the central value has no remarkable change. In the low-level wind field adjustment experiment (Fig. 10b-c), the low-level thermal environment has changed, but the thermal stratification above the boundary layer is almost identical to that in the control experiment (Fig. 10a). After the weakening of middle-level environmental wind field (Fig. 10d), the relatively dry and cold zone (value of 316 K) at the middle level shrinks upwards, especially in the early stage of convection. This will reduce the θ_se vertical gradient in the middle and lower levels of the convection zone, which is not conducive to the development of stratification instability. When the middle-level wind field enhances (Fig. 10e), the thermal stratification in the convective area changes significantly, and the middle-level dry-cold area gradually expands downwards. In the later stage of convection development, a dry-cold center with a value of 314 K appears at 600-700 hPa. The main period of the unstable stratification development is slightly earlier than that in the control experiment and other wind field experiments, which may be one of the reasons for the rapid development of convective storm in the early stage. Another evident change is presented in the middle and later stages of convection development, and that is due to the obvious enhancement of the northerly wind in the middle level. The relatively dry-cold area of the middle level gradually extends downwards, and the stratification in the convection area tends to be stable, which may result in earlier extinction of convective storms than that in other experiments. It can be seen from the above analysis that changes in wind fields at different heights will affect the environmental thermal stratification and may indirectly affect the development of storms.

Figure 10.  Average potential pseudo-equivalent temperature θ_se with time-height change (unit: K) in convection area (30-35°N, 117-120°E) in control and wind field experiments. (a) Ctrl, (b) Test-low-wind-50%, (c) Test-low-wind-200%, (d) Test-middle-wind-50%, and (e) Test-middle-wind-200%.

Figure 11a-d shows that the ground deviation winds in wind field experiments relative to the control experiment and the distribution of reflectivity factors at the same moments in the mature stage of the convective storms. It can be seen from the figures above (Fig. 11a-b) that the ground wind field in the low-level wind field experiments has little difference with that in the control experiment; in the deviation wind field, the cyclonic circulation associated with the mesoscale low pressure at the junction with Shandong, Jiangsu and Anhui Provinces is very weak; the shape and location of echoes in the two groups of experiments are quite different. In the wind field weakening experiments (Fig. 11a), the convection spreads southward along the convergence line, shaping as a strip. In the enhancement experiments of low-level wind field (Fig. 11b), the convective storms develop near the low-pressure center, whose shape is closer to an isolated supercell storm. The downdraft in the storm forms an evident divergence area of deviation wind. In the middle-level wind field enhancement experiments, the ground deviation wind field is large compared with the control experiment, and the ground wind field deviation tends to be apparent after three hours of integration (figure omitted). Corresponding to the location of the northerly jet in the middle troposphere, the northerly draft on the right rear of the convective storm enhances in the ground wind field, which is more intense than the northerly wind in this area in the control experiments, so the deviation wind field is dominated by the northerly wind. This enhanced northerly draft meets the southerly wind in front of the storm movement, enhancing convergence on the ground and forming an obvious east-west convergence line. The convections along the convergence line is gradually organized into a thriving linear convective system (Fig. 11d). After the middle-level wind field weakens, it can be seen that the deviation wind in the right rear of the convective storm is southerly (Fig. 11c), while the ground wind fields of this experiment and the control experiment are dominated by the northerly wind (figure omitted), which indicates that the surface northerly wind also weakens after the middle-level wind is abated.

Figure 11.  Deviation of ground wind filed between different wind field experiments and the control experiment, and reflectivity factor (shaded, unit: dBZ) at 16:00. (a) Test-low-wind-50%, (b) Test-low-wind-200%, (c) Test-middle-wind-50%, and (d) Test-middle-wind-200%.

It is concluded from the differences of ground wind field between wind field experiments and the control experiments that the changes of the middle and lower level wind fields would affect the ground wind field to some extent, which leads to changes in ground convergence and may affect the convection development. The comparison between experiments shows that the variation of the middle-level wind field has the most significant effects on the ground wind field. Of course, this may be correlated to the large difference in wind speed between the middle and lower troposphere in the process of severe convection.

Numerical experiments on wind field adjustment in the course of a typical severe convection under the cold vortex over the Jianghuai area are conducted to discuss the possible effects of wind field variations on storm structure and development process. The conclusions are drawn as follows.

(1) Variations of wind fields in the middle or lower troposphere have significant impacts on the development of convective storms. Compared with the low-level wind field, the differences in the development speed and intensity of the convective storms caused by the adjustment of the middle-level wind field are more obvious. When the middle - or low-level wind field weakens, the intensity of storms development decreases. The weakening of low-level wind field presents as the weaker convection intensity at the initial stage of the storm, while the weakening of middle-level wind field shows the weaker convection intensity at different stages of its development. The convective storms develop more vigorously when the wind field in the middle - or low-level is strengthened, especially in the middle-level wind field. In the middle-level wind field enhancement experiments, the convective storms develop most rapidly, and the formation and extinction of convections are about one hour ahead of those in other experiments, and the storm movement speed up.

(2) In different wind field experiments, the structural characteristics of convective storm at different stages of the development are evidently different. It can be seen from the structural characteristics of storms at the mature stage in each experiment that the convective storms develop rapidly when the middle - or low-level wind field enhances. Particularly, the enhancement of the middle-level wind field is conducive to the strengthening of rotation in middle level of the convective storm; the storm scale is reduced, and the convection develops more intensely. The enhancement of the low-level wind field is more favorable to the development of secondary circulation in the low level of storms and the middle - and low-level cyclonic vorticity on the inflow side. When the wind field weakens, regardless of the middle level or the low level, the convective storms weaken. Particularly, when the middle-level wind weakens, the convection distinctly abates. The scale of secondary circulation associated with the storm increases, but the corresponding updraft weakens.

(3) Wind field adjustment not only changes the dynamic environment of storms, but may change the thermal environment of storms to some extent, thus affecting the development of convective storms. Moreover, variations of wind fields in the middle and lower troposphere will also affect wind field on the ground and give rise to changes of convergence on the ground, resulting in differences in the development of the storms.

In this paper, the formation, development and structure of convective storms in different wind fields were studied through different numerical schemes, and the possible impacts of wind field variations on storms development were analyzed. The above work is of certain significance for understanding the role of the wind field in the formation and development of convective storms. The conclusions drawn from the individual cases in this paper require additional verification and further in-depth investigation, because the research done in this paper is preliminary, and the interaction among numerous meteorological factors that affect the formation and development of convective storms makes this process more complicated.