Mesoscale Factors Contributing to the Extreme Rainstorm on 20 July 2021 in Zhengzhou, China, as Revealed by Rapid Update 4DVar Analysis

Juanzhen Sun aNational Center for Atmospheric Research, Boulder, Colorado

Search for other papers by Juanzhen Sun in
Current site
Google Scholar
PubMed
Close
https://orcid.org/0000-0001-6048-6362
,
Rumeng Li bDepartment of Atmospheric and Oceanic Sciences, School of Physics, Peking University, Beijing, China

Search for other papers by Rumeng Li in
Current site
Google Scholar
PubMed
Close
,
Qinghong Zhang bDepartment of Atmospheric and Oceanic Sciences, School of Physics, Peking University, Beijing, China

Search for other papers by Qinghong Zhang in
Current site
Google Scholar
PubMed
Close
,
Stanley B. Trier aNational Center for Atmospheric Research, Boulder, Colorado

Search for other papers by Stanley B. Trier in
Current site
Google Scholar
PubMed
Close
,
Zhuming Ying aNational Center for Atmospheric Research, Boulder, Colorado

Search for other papers by Zhuming Ying in
Current site
Google Scholar
PubMed
Close
, and
Jun Xu cNational Meteorological Center of China, Beijing, China

Search for other papers by Jun Xu in
Current site
Google Scholar
PubMed
Close
Open access

Abstract

The purpose of this study is to diagnose mesoscale factors responsible for the formation and development of an extreme rainstorm that occurred on 20 July 2021 in Zhengzhou, China. The rainstorm produced 201.9 mm of rainfall in 1 h, breaking the record of mainland China for 1-h rainfall accumulation in the past 73 years. Using 2-km continuously cycled analyses with 6-min updates that were produced by assimilating observations from radar and dense surface networks with a four-dimensional variational (4DVar) data assimilation system, we illustrate that the modification of environmental easterlies by three mesoscale disturbances played a critical role in the development of the rainstorm. Among the three systems, a mesobeta-scale low pressure system (mesolow) that developed from an inverted trough southwest of Zhengzhou was key to the formation and intensification of the rainstorm. We show that the rainstorm formed via sequential merging of three convective cells, which initiated along the convergence bands in the mesolow. Further, we present evidence to suggest that the mesolow and two terrain-influenced flows near the Taihang Mountains north of Zhengzhou, including a barrier jet and a downslope flow, contributed to the local intensification of the rainstorm and the intense 1-h rainfall. The three mesoscale features coexisted near Zhengzhou in the several hours before the extreme 1-h rainfall and enhanced local wind convergence and moisture transport synergistically. Our analysis also indicated that the strong midlevel south/southwesterly winds from the mesolow along with the gravity-current-modified low-level northeasterly barrier jet enhanced the vertical wind shear, which provided favorable local environment supporting the severe rainstorm.

© 2023 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Qinghong Zhang, qzhang@pku.edu.cn

Abstract

The purpose of this study is to diagnose mesoscale factors responsible for the formation and development of an extreme rainstorm that occurred on 20 July 2021 in Zhengzhou, China. The rainstorm produced 201.9 mm of rainfall in 1 h, breaking the record of mainland China for 1-h rainfall accumulation in the past 73 years. Using 2-km continuously cycled analyses with 6-min updates that were produced by assimilating observations from radar and dense surface networks with a four-dimensional variational (4DVar) data assimilation system, we illustrate that the modification of environmental easterlies by three mesoscale disturbances played a critical role in the development of the rainstorm. Among the three systems, a mesobeta-scale low pressure system (mesolow) that developed from an inverted trough southwest of Zhengzhou was key to the formation and intensification of the rainstorm. We show that the rainstorm formed via sequential merging of three convective cells, which initiated along the convergence bands in the mesolow. Further, we present evidence to suggest that the mesolow and two terrain-influenced flows near the Taihang Mountains north of Zhengzhou, including a barrier jet and a downslope flow, contributed to the local intensification of the rainstorm and the intense 1-h rainfall. The three mesoscale features coexisted near Zhengzhou in the several hours before the extreme 1-h rainfall and enhanced local wind convergence and moisture transport synergistically. Our analysis also indicated that the strong midlevel south/southwesterly winds from the mesolow along with the gravity-current-modified low-level northeasterly barrier jet enhanced the vertical wind shear, which provided favorable local environment supporting the severe rainstorm.

© 2023 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author: Qinghong Zhang, qzhang@pku.edu.cn

1. Introduction

A multiday rainfall event that occurred in Henan province of China during 17–23 July 2021 caused catastrophic flooding and great loss of life and property. The most intense rainfall episode took place on 20 July in Zhengzhou, the capital city of Henan, with a 24-h accumulation of 624.1 mm. More strikingly an extreme 1-h rainfall accumulation of 201.9 mm was recorded between 0800 and 0900 UTC (LST = UTC + 8 h) in the afternoon from a rain gauge station, which broke mainland China’s 1-h rainfall record that had stood for 73 years. Three large rivers in the region all experienced severe floods with peak water levels exceeding the historically recorded maximum values. The event had significant societal impact with over 14 million people in Henan Province affected, more than 300 deaths, and a direct economic loss of ∼120 billion Yuan (Xu et al. 2022). Although the event had a 1-week duration, our focus in the current study is on the record-breaking rainfall that occurred in the city of Zhengzhou on 20 July. This episode will be referred to as the 7.20 extreme rainstorm or 7.20 Zhengzhou rainstorm in this paper.

The prolonged precipitation period over a broad region of Henan Province and the local nature of the extreme rainfall episode on 20 July suggested interactions of multiscale processes leading to the rainstorm. While current numerical prediction (NWP) models can predict synoptic-scale systems with reasonable accuracy, reliably resolving the mesoscale and convective-scale features remains a formidable challenge. Zhang et al. (2022) found that all global models from the major meteorological centers significantly underpredicted the 24-h rainfall amount on 20 July 2021. Furthermore, operational models, including convection-allowing regional models, significantly underpredicted the extreme 1-h rainfall in the city of Zhengzhou.

As has been affirmed through many previous studies, synoptic-scale weather systems mainly provide a setting in which severe convection develops (see, e.g., Newton 1963; Barnes and Newton 1983; Johns and Doswell 1992; Trapp 2013) through supplying convective instability and moisture. Mesoscale processes can further destabilize the environment, modify the vertical wind shear profiles, and provide convergence and upward motion, together with horizontal moisture advections, thereby setting the stage for triggering severe weather (Johnson and Mapes 2001; Doswell 1987). Examples of mesoscale processes associated with the initiation of severe weather are shortwave troughs, mesolows (see below for definition), topographically forced circulations, drylines, gravity waves, low-level jets, gust fronts, collision of convergence boundaries, and urban heat islands.

The study of effects of mesoscale processes on severe convection have traditionally relied mainly on enhanced observations from field campaigns and model simulations. Although knowledge has been gained through field campaigns and subsequent observational analysis, the approach is limited to weather events documented during short periods and over small regions. Extreme weather events, such as the 7.20 Zhengzhou rainstorm, are often captured only by operational networks that are typically not sufficient to depict important three-dimensional local circulations. This is due to either the lack of dense upper-air observations to complement often available surface networks or insufficient radar networks. In recent years, therefore, convection-allowing model (CAM) simulations that are initialized by reanalysis or real data have been applied to study convection initiation and subsequent upscale development (e.g., Zhang and Zhang 2012; Trier et al. 2015; Wang et al. 2016; Luo et al. 2018) and have provided valuable insights on the role of mesoscale circulations on severe convection. However, due to errors in model simulation and the limited predictability of mesoscale phenomena, the model simulations may not correctly resolve the mesoscale features leading to convective initiation, and could produce reasonable precipitation simulations, but for the wrong reasons (Zhang et al. 2013). An alternative approach for investigating mesoscale processes has been demonstrated in recent years. This approach employs rapid update 3D analyses that are obtained by constraining a CAM with observations of high spatial density and temporal frequency using advanced data assimilation (DA) techniques. By applying this approach, several recent studies have gained new knowledge on convective initiation by mesoscale motions (Marquis et al. 2016; Xiao et al. 2017, 2019, 2021; L. Zhang et al. 2021; Y. Wu et al. 2021; Y.-J. Wu et al. 2021). The approach takes advantages of both models and observations and is particularly useful for studying extreme or severe rainfall events that are sampled only with operational networks. These studies have shown that analyses produced by DA could identify critical features that could not be revealed by the observations or model alone.

Henan province in China has one of the best operational mesoscale observation networks in the world, featuring nine Doppler radars and a dense surface network in an area stretching 480 km from north to south and 560 km from east to west. By assimilating these observations using the Variational Doppler Radar Analysis System (VDRAS) (Sun and Crook 1997), we produced analyses that were updated every 6-min to investigate the main mesoscale contributors to the initiation and organization of the Zhengzhou 7.20 extreme rainstorm. VDRAS is a 4DVar DA system designed to create high-resolution and high-frequency atmospheric analyses by assimilating radar and surface observations. VDRAS can use small assimilation windows (6-min in this study) as compared with those of other 4DVar DA systems that assimilate convective-scale observations [e.g., WRF 4DVar (Wang et al. 2013)], and thereby better resolve mesoscale and convective-scale circulations.

For the study of mesoscale processes affecting the 7.20 rainstorm, there are two critical questions to be answered: 1) What were the initial triggering and development mechanisms that eventually led to the 7.20 extreme rainfall? 2) What local processes around Zhengzhou were responsible for rainfall intensification and its quasi-stationary maintenance that resulted in the 1-h record-breaking precipitation? Our goal is to reveal key mesoscale features that directly contributed to the initiation and intensification of the rainstorm, based on the 6-min update 3D VDRAS analyses. We illustrate in this paper that a low pressure system with a scale at the upper end of mesobeta-scale (2–200 km; Orlanski 1975) (mesolow hereinafter), which developed within an inverted trough southwest of Zhengzhou, played a key role in the initiation, early development, and local intensification of the 7.20 Zhengzhou rainstorm. We further show that a northeasterly barrier jet (Li and Chen 1998; Ke et al. 2019) and a downslope gravity current or katabatic flow (Poulos and Zhong 2008) north of Zhengzhou were also critical mesoscale factors that worked in synergy with the mesolow to intensify and maintain the quasi-stationary rainstorm near Zhengzhou.

While the anomalous large-scale circulation during the week of 17–23 July 2021 has been studied in several recent papers (X. Zhang et al. 2021; Ran et al. 2021; Xu et al. 2022; Cai et al. 2022), studies on mesoscale features have been relatively few. Two recent studies based on WRF-ARW model simulations attempted to answer the second question raised above. Yin et al. (2023) presented a WRF model simulation that reproduced heavy rainfall amounts on 20 July 2021 close to what were observed. Based on their simulation, they suggested that an arc-shaped updraft in the model simulation caused by the southwesterly flow of the Huang-Huai cyclone west of Zhengzhou was responsible for the quasi-stationary extreme rainfall. Wei et al. (2023) suggested directional convergence of the horizontal wind toward Zhengzhou was mainly caused by two factors: the north–south convergence between the southward wind component resulting from the northeasterly barrier jet parallel to Taihang Mountains and the southerly winds associated with a mesobeta-scale vortex, and the west–east convergence between the environmental easterly flows and a westerly flow due to a low-level cold pool. While analyses from both studies revealed the low-level flows converged from multiple directions toward Zhengzhou and provided a partial explanation of the causes, there were other important mesoscale processes that have yet to be explained. Moreover, the initiation processes responsible for the early development of the rainstorm have yet to be investigated or reported.

The remainder of the paper is organized as follows. VDRAS and its validation are described in section 2. Section 3 provides an overview of the large-scale environment and the convective precipitation evolution. The local environment and moisture transport in the Zhengzhou area are described in section 4, which is followed by section 5 in which the mesoscale factors contributing to the initiation and intensification of the 7.20 Zhengzhou rainstorm are diagnosed. A summary and conclusions are given in section 6.

2. Description of VDRAS and analysis validation

VDRAS is a convective-scale analysis system based on the four‐dimensional variational assimilation (4DVar) technique. The 4DVar is constrained by a cloud model with the prognostic variables of three-dimensional (3D) wind components, liquid water potential temperature, total water mixing ratio, and rainwater/snow mixing ratio. The microphysics parameterization used in VDRAS is a Kessler-type warm rain and a simple ice scheme (Chang et al. 2016). The 4DVar assimilation window length for each cycle is typically set to 5–15 min, allowing at least two scanning radar volumes and two time-levels of surface AWS (Automated Weather Station) measurements to be assimilated in each DA cycle. The assimilation variables from radar are radial velocity and rain and snow mixing ratios converted from reflectivity following Gao and Stensrud (2012), and those from the surface network are x- and y-wind components, temperature, and specific humidity. Recently lightning data assimilation has been added to VDRAS (Xiao et al. 2022), but it was not used in the current study. In addition to the observational data, hourly WRF simulations are used to provide the first guess background and boundary conditions, which contain information about the large-scale environment that is not usually contained in radar and surface observations. The output from VDRAS consists of 3D wind, temperature, humidity, microphysics variables (rainwater, snow, cloud water, and dry ice), and many diagnosed fields including vertical shear, horizontal divergence and vorticity, updraft helicity, CAPE (convective available potential energy), CIN (convective inhibition).

VDRAS has been applied to several extreme/severe rainfall and flooding cases to investigate mesoscale processes that led to heavy precipitation (e.g., Gochis et al. 2015; Xiao et al. 2017, 2019, 2021; L. Zhang et al. 2021; Y.-J. Wu et al. 2021). These studies all pointed to the important role played by mesoscale circulations associated with local topography in convection initiation and heavy precipitation, such as local terrain variations, coastal bays, and urban heat islands. The 7.20 rainstorm also occurred in a region with complex terrain. Therefore, it is reasonable to assume that the local topography may have impacted the event.

The latest release of VDRAS was used in the present study. This version has several major upgrades including the addition of a terrain scheme based on the immersed boundary method (IBM) (Tseng and Ferziger 2003), the improvement of surface DA, and a stepwise refinement analysis scheme. The IBM terrain scheme, described in detail in Tai et al. (2017), simulates the topographic effect in a height coordinate with the underground grid points treated as “ghost cell” following Tseng and Ferziger (2003). The IBM terrain scheme has an advantage in high-resolution simulations of steep terrain over the terrain-following scheme, and thus has been applied for simulations of flows between buildings (Lundquist et al. 2012; Tai et al. 2017). The surface DA was improved by interpolating surface temperature and humidity observations to the nearest model level using their vertical distributions in the background fields (i.e., from WRF), rather than assuming the standard atmosphere (Chen et al. 2016). The stepwise refinement analysis scheme enables multiple VDRAS analysis passes in which refined horizontal and vertical resolutions on a smaller domain are used in each new pass and the analysis from the previous pass is used as the first guess.

In the current study, VDRAS was configured with two analysis passes on a Δx = 4 km domain D1 and a Δx = 2 km domain D2, respectively (Fig. 1a), covering most of Henan province. The 4DVar analysis was conducted every 6 min with a 7-min assimilation window length, starting from 0600 UTC 19 July in D1 and 1700 UTC 19 July in D2, and ending at 1000 UTC 20 July in both domains. The analysis cycles were continuous with the analysis from last cycle used as the first guess after blending with the WRF simulation (details given later) to start the current cycle. The total number of analysis datasets produced during the whole analysis period was 170 for D2 and 280 for D1. The radial velocity and reflectivity from 17 radars were assimilated (see Fig. 1a for their names and locations). These radars all have the same scanning mode with nine elevation angles (0.5°, 1.5°, 2.5°, 3.4°, 4.4°, 6.1°, 9.9°, 14.6°, and 19.6°) and a volume scan rate of ∼6 min. Since the radars are synchronized with a time offset of less than 1 min, in the 7-min 4DVar assimilation window two volumes from each radar were assimilated in each cycle. Nearly 2850 surface stations were available in the inner D2 domain (Fig. 1b), and ∼60% of them had measurements of the state variables of wind, temperature, dewpoint temperature, and pressure. Most of the surface observations had an update interval of 5 min, meaning two measurements for each variable from each station were assimilated in each analysis cycle. The surface data were extrapolated to the first VDRAS model level based on the slope from the WRF simulation at each grid point.

Fig. 1.
Fig. 1.

(a) VDRAS 4-km D1 domain (outlined by the thick black square) and 2-km D2 domain (outlined by the blue square) and locations of 17 Doppler radars. The color shading represents the terrain height. The major mountain ranges—Funiu, Taihang, and Song—are labeled in purple. (b) Surface stations within the VDRAS domain. The red dots represent stations with measurements of wind, temperature, dewpoint, and pressure, and the green dots represent the stations with fewer observed variables. The border of Zhengzhou is shown by the black contour. The city center of Zhengzhou1 is marked by the black star.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The VDRAS domain was centered at 39.4°N, 113.2°E, and comprised an area of 400 km × 400 km with 2-km horizontal spacing and 150-m vertical spacing in D2, and areas of 560 km × 560 km with 4-km horizontal and 300-m vertical grid spacings in D1 (Fig. 1a). The model top is 12.6 km. The WRF 4-km forecast valid at 3, 4, or 5 h, whichever is closest to the analysis time, was used to drive the VDRAS D1 Δx = 4 km analysis by providing boundary conditions and first guess fields after blending with VDRAS analysis from the previous cycle. The WRF forecasts were initialized by NCEP GFS (Global Forecasting System) analysis every 3 h. The VDRAS D1 analysis results were then used to drive the D2 Δx = 2 km analysis. The use of the blended first guess is critical for maintaining temporal consistency of the 6-min updated analyses and to avoid the divergence of 4DVar analysis from the large-scale mean. Similar practices have been used in other high-resolution data assimilation systems. For example, WRF data assimilation systems have applied a blending of WRF simulations with the driving GFS forecasts (Feng et al. 2020; Schwartz et al. 2022).

The accuracy of VDRAS analysis has been verified against various data sources—for example, surface observation (Sun et al. 2010; L. Zhang et al. 2021), dual-Doppler radar synthesis (Crook and Sun 2004), radiosonde (Tai et al. 2011), and aircraft observations (Sun and Crook 1998). The convective parameters (CAPE, CIN, and vertical wind shear) diagnosed from VDRAS analysis were validated against those derived from reanalysis and radiosonde data in Xiao et al. (2017). A practical approach to validate VDRAS analyses has also been used (X. Zhang et al. 2021) by computing the RMSE between the assimilated observations and the DA analysis in the observation space, that is, observation − analysis (O-A). By comparing O-A with observation − background (O-B), the amount of error reduced by DA can be evaluated. Figure 2 compares RMSEs of radial velocity computed for all 170 VDRAS D2 analyses (O-A; Fig. 2a) over the whole analysis period and for their WRF first guess counterparts (O-B; Fig. 2b) with observed radial velocity from seven radars within D2. The maximum of mean error averaged over all radars is reduced from ∼3.4 m s−1 in the WRF background to 1.9 m s−1 for the VDRAS analysis in the PBL and the maximum of the mean errors above the PBL are reduced from ∼2.8 m s−1 (Fig. 2b) to ∼1.5 m s−1 (Fig. 2a). The RMSEs for u, υ, T, and qυ at the lowest VDRAS model level2 (75 m AGL) against surface observations range from 0.8 to 1.25 m s−1 for u and υ (Figs. 3a,b), from 0.6° to 0.8°C for temperature (Fig. 3c), and from 1.0 to 1.3 g kg−1 for specific humidity (Fig. 3d). Note that the errors increase with cycle time due to the growth of convection.

Fig. 2.
Fig. 2.

RMSEs of radial velocity computed from (a) VDRAS analysis and (b) WRF background against observed radial velocity over all analysis times for seven radars in D2 (Fig. 1a). The mean RMSE over all radars is shown by the black line.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

Fig. 3.
Fig. 3.

RMSEs of VDRAS-analyzed 10-m velocity components (a) u and (b) υ, (c) 2-m temperature, and (d) water vapor mixing ratio against surface observations with respect to analysis time (UTC).

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The above validation confirmed that the VDRAS analysis fitted well to radial velocity in the 3D domain and to surface wind, temperature, and humidity. With the optimal fitting to the radar and surface observations and the large-scale background information from WRF simulations, together with the constraint of the convection-allowing numerical model in VDRAS, dynamically balanced 3D analysis of wind, temperature, humidity, and microphysical variables were obtained and provided valuable data for our study of mesoscale features.

3. Overview of large-scale environment and precipitation evolution

a. Large-scale environment

Henan has a diverse landscape with mountains in the west and floodplains in the east as part of the North China Plain (Fig. 1a, and Fig. 4c for a view of larger region). The capital city Zhengzhou is at the east of Mount Song, which connects to an extensive network of mountains and plateaus via the Funiu Mountains (Fig. 1a). To the northwest of Zhengzhou is the Taihang Mountain range that extends more than 400 km northward. During the week-long extreme rainfall period, Henan was situated at the southwest periphery of the northward extension of the west Pacific Subtropical High (WPSH) (Figs. 4a–c) with weak temperature gradient (Fig. 4a). Together with the WPSH, Typhoon In-Fa over the East China Sea to the southeast of Henan and Typhoon Cempaka to its south over the South China Sea modulated the environmental flows and moisture transport to the region of interest. The 700-hPa relative humidity field (Fig. 4b) indicates that the near saturated environment in the mid-atmosphere in Henan had spread northward from the South China Sea, whereas the moist air in the boundary layer was transported from the East China Sea by the strong southeasterly flow between the WPSH and Typhoon In-Fa (Fig. 4d). The winds at both 700 and 500 hPa were dominated by southwesterlies and southeasterlies in most part of Henan corresponding to an inverted trough on the southwest periphery of WPSH and the mesolow within the trough in western Henan. The mesolow can be well identified by closed geopotential height contours to the left of the star symbol in Fig. 4b. From comparing Fig. 4a with Fig. 4b, it can be inferred that the mesolow developed between 1200 UTC 19 July and 0000 UTC 20 July. In section 4, we will illustrate that the mesolow played a pivotal role in the initiation, development, and intensification of the 7.20 extreme rainstorm using the rapid update VDRAS analyses.

Fig. 4.
Fig. 4.

ERA5 reanalysis fields showing the large-scale environment: (a) temperature (shading), geopotential height (contour), and wind at 700 hPa at 1200 UTC 19 Jul; (b) as in (a), but at 0000 UTC 20 Jul and the shaded field is relative humidity; (c) as in (b), but at 500 hPa and the shaded field is topography; and (d) as in (b), but at 975 hPa. The city center of Zhangzhou is marked by the red star, and the coastline and the border of Henan province are shown by dark-blue and purple contours, respectively. The Taihang and Funiu Mountain ranges are labeled in (c), and Typhoons In-Fa and Cempaka are labeled by the cyan letters “I” and “C,” respectively, in (b).

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

b. Precipitation evolution

Under the favorable large-scale environment, widespread heavy precipitation occurred in the northern part of Henan province beginning on 17 July 2021. On 19–20 July, the precipitation area moved southward and was centered near Mount Song and its adjacent eastern plains area. The precipitation episode leading to the Zhengzhou rainstorm started around 1900 UTC 19 July and reached a domain-averaged maximum above 180 mm at 0600 UTC and a station maximum of 201.9 mm at 0900 UTC (Fig. 5). The area-averaged hourly rainfall was computed by averaging over all gauge stations in the VDRAS analysis domain D2 (Fig. 1b).

Fig. 5.
Fig. 5.

(a) Accumulated hourly rainfall vs UTC time at Zhengzhou weather station (blue bar; left vertical axis) and areal average hourly rainfall over all stations in the VDRAS analysis domain D2 (red curve; right vertical axis). Also shown is the (b) 12-h accumulated rainfall during 0000–1200 UTC and (c) 3-h rainfall during 0700–1200 UTC 20 Jul 2021. The center of Zhengzhou city is marked by a white star in (b) and (c).

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The 12- (Fig. 5b) and 3-h (Fig. 5c) precipitation accumulations from a gauge adjusted radar QPE (qualitative precipitation estimate) product from VDRAS indicate that intense rainfall occurred during the 3 h from 0700 to 1000 UTC and was concentrated only in a small area near Zhengzhou (Fig. 5c), suggesting the local nature of the extreme rainstorm. Although the 12-h accumulated precipitation is more widespread, the pattern of precipitation near Zhengzhou is like that of the 3-h rainfall because the most intense rainfall occurred in the later hours (Figs. 5a and 6).

Fig. 6.
Fig. 6.

Observed column maximum reflectivity at nine selected times from a mosaic of all radars in D2. The white line is the 300-m terrain elevation. The city center of Zhengzhou is marked by the black star. The areas above 42 dBZ are outlined by black contours.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The evolution of reflectivity (Fig. 6) indicates that scattered convective cells were embedded within a widespread precipitation system (Figs. 6a–c) before two distinct convective cells were organized near Zhengzhou at 0524 UTC (Fig. 6d). These two cells then merged into one system and strengthened in the next 1–2 h (Figs. 6e,f). The intensified system remained over Zhengzhou (Figs. 6f–h) for more than 2 h until it broke into smaller cells (Fig. 6i) and dissipated.

To view the convective evolution in more detail, storm swaths of observed reflectivity greater than 42 dBZ between 0254 and 0630 UTC are displayed in Fig. 7a, and the quasi-stationary locations of the rainstorm at later hours are shown by contours with different colors in Fig. 7b. The three cells labeled C1, C2, and C3 each moved toward Zhengzhou (black star symbol in Fig. 7b). The cells C1 and C2 initiated in the southeast foothills of Mount Song and merged into one system west of Zhengzhou (Fig. 7a). Cell C3, which initiated in the plains southeast of Zhengzhou, moved northward while dissipating but later reintensified at ∼0524 UTC northeast of Zhengzhou and propagated northward. From 0630 to 0700 UTC, the second merging process occurred (Fig. 7b), resulting in the larger, stronger single cell (orange contours and blue contours in Fig. 7b). This storm moved slowly through central Zhengzhou during the next 2–3 h before dissipating at 1000 UTC and was responsible for the large rainfall produced from 0700 to 1000 UTC and the record-breaking 1-h rainfall recorded at the Zhengzhou station.

Fig. 7.
Fig. 7.

(a) Observed reflectivity swaths between 0254 and 0630 UTC (stage I) produced by plotting reflectivity greater than 42 dBZ every 12 min before 0454 UTC and every 24 min thereafter. The interval of gray shades is 2 dBZ. The three cells C1, C2, and C3 that initiated around 0254 UTC and contributed to the organized Zhengzhou extreme rainstorm are all labeled. Note that C3 reintensified at 0524 UTC to the northeast of Zhengzhou (black star). (b) Contour plots of reflectivity (stage II) at 0630 (green) and 0648 (orange) UTC (>40 dBZ with 4-dBZ interval), as well as at 0700 (blue), 0800 (red), 0900 (cyan), and 1000 (magenta) UTC (>45 dBZ with 2-dBZ interval).

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The precipitation evolution suggested two stages of convective development. Stage I covers the period of the initiation and early development of the convective cells but before the second merging occurred at 0630 UTC (Fig. 7a). Stage II represents the period of local intensification and quasi-stationary maintenance of the storm. In section 5, we will present key mesoscale features in both stages contributing to the 7.20 rainstorm based on the VDRAS 4DVar analysis.

4. Local environmental conditions

It was shown in the last section that after the second merging of the two convective cells near Zhengzhou, the strengthened storm was maintained locally for multiple hours. In this section, we will examine the local environment that was supportive of the severe local storm and its maintenance. We will also discuss the convergent flow and moisture transport near Zhengzhou in the few hours leading to the extreme rainfall. The mesoscale disturbances that contributed to the favorable local environment and horizontal convergence and moisture transports will then be diagnosed in section 5.

At 0600 UTC 19 July, before any significant mesoscale disturbances developed, the Zhengzhou area had cooler temperatures than the adjacent east plains (Fig. 8a). The surface winds in the area consist of weak northeasterlies parallel to the slopes of Taihang Mountains north and northwest of Zhengzhou, and easterlies from the east plains. Above the surface a southeasterly jet that was part of the anticyclonic circulation near the southwest periphery of the WPSH had maximum speeds at 2 km MSL, and extended from southeast to northwest across Zhengzhou as shown by VDRAS analysis in Fig. 8b.3

Fig. 8.
Fig. 8.

(a) Surface 2-m temperature analysis overlaid by surface 10-m wind vectors valid at 0600 UTC 19 Jul 2021, produced from observations of surface stations using the Barnes (1964) objective analysis technique. (b) VDRAS wind speed overlaid by wind vectors at Z = 2 km MSL valid at the same time as (a). The white contour represents the 300-m terrain height, and the city center of Zhengzhou is marked by the white star. Also shown are observed radiosonde soundings at the Zhengzhou station [black dot in (a)] valid at 0600 UTC (c) 19 and (d) 20 Jul 2021.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The observed sounding at 0600 UTC 19 July from the Zhengzhou radiosonde station (Fig. 8c) indicates that the southeasterlies extended up to nearly 500 hPa and the wind is northeasterly near the surface, consistent with the VDRAS analysis in Figs. 8a and 8b. A deep nearly saturated layer in the lower to middle troposphere was associated with the southeasterlies and overlaid the near surface air with CAPE of 1313 J kg−1 on 0600 UTC 19 July (Fig. 8a), which was reduced to 157 J kg−1 24 h later (Fig. 8b) as a result of ∼1°C surface cooling.

The horizontal distribution of CAPE from the VDRAS analysis at 0600 UTC 20 July (Fig. 9a) indicates that the atmosphere northwest of Zhengzhou had CAPE values less than 200 J kg−1 due to recent precipitation over the mountain and valley region (Fig. 6). However, the atmosphere southeast of Zhengzhou was unstable with CAPE > 1000 J kg−1 resulting in a horizontal CAPE gradient across the Zhengzhou area. The mean CAPE (averaged over the orange box in Fig. 9a) increased rapidly between 0430 and 0600 UTC (Fig. 9c), which is not surprising given the rapid transport of warm and moist air by the strong environmental southeasterlies in the early afternoon hours and a likely local concentration of the unstable air by local motions, which will be elucidated later.

Fig. 9.
Fig. 9.

(a) Near-surface CAPE that was computed by lifting parcels at 75 m AGL (one-half of VDRAS vertical spacing) overlaid by wind vectors at 75 m AGL valid at 0600 UTC 20 Jul from VDRAS. (b) As in (a), but for 0–3-km shear magnitude overlaid by the shear vectors. (c) As in (a), but for water vapor mixing ratio. (d) Time series of mean 0–3-km shear (black solid line), y component of the shear vector (black dotted line), and CAPE by lifting parcel from 75 m (red line), averaged over the orange square shown in (a) and (b). The city center of Zhengzhou is marked by the black star and the 300-m terrain elevation is shown by the green contour in (a)–(c).

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The larger near-surface water vapor mixing ratio (Fig. 9c) in the southeast part of the domain is shown by Fig. 9c, and appears to have been a critical moisture source. The southeast moist air exhibits a clear contrast with the drier air in the mountainous northeast part including the valley. It is noted that large water vapor mixing ratios were also present in the area directly north of Zhengzhou. It will be shown later in this section that this relatively moist region also played an important role in the moisture transport to Zhengzhou.

The 0–3-km bulk vertical shear (hereinafter referred to as shear) is large over an area extending from the southern part of the valley to east of Zhengzhou (Fig. 9b) before the occurrence of the extreme rainfall. The time series of the area mean of the 0–3-km shear suggests a rapid increase between ∼0230 and ∼0600 UTC, reaching an area-averaged value of nearly 19 m s−1 at ∼0600 UTC (Fig. 9d). It should be noted that 0–6-km bulk vertical wind shears were also examined, and they exhibited similar local patterns and temporal variations, but with a higher magnitude, and were dominated by the northward y component like the 0–3-km shear (Fig. 10b). The high shear provided a favorable environment for the rainstorm to persist.

Fig. 10.
Fig. 10.

The (a) water vapor difference and (b) υ-wind difference computed between 0600 and 0300 UTC at 0.7 km MSL, overlaid by the wind vectors [black arrows in (a)] at 0600 UTC and the difference wind vectors [white arrows in (b)]. The white line shows the 300-m terrain elevation, and the green line outlines the areas of reflectivity greater than 42 dBZ. The orange-outlined square marks the area (same as in Fig. 9a) used for the computation of moisture transports shown in (c) and (d). (c) Temporal variations of integrated vapor transports (IVT) to the column within the orange square in (a) through the north (dashed red), south (solid red), east (dashed green), and west (solid green) sides. The IVT in the low level (<1.5 km) through the north side (dotted red) is also shown. (d) Temporal variations of net moisture transports through the north–south sides (red) and east–west sides (green), and the net moisture transport to the column through all four sides (black). (e) Backward air trajectories reaching Zhengzhou at 0400 UTC, beginning at 0000 UTC, at the levels 0.3, 0.6, 1.0, 2.0, 3.0, 4.0, and 5.0 km. (f) As in (e), but for the air mass reaching Zhengzhou at 0700 UTC, beginning at 0400 UTC. The terrain map is shown by gray shadings.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The low-level winds are increasingly confluent in the Zhengzhou region (Fig. 10a). This general flow pattern started as early as ∼2200 UTC 19 July (not shown), but extreme 1-h rainfall amounts did not occur until after 0600 UTC 20 July (Fig. 5a). Hence, it is reasonable to speculate that significant mesoscale motions developed in the few hours prior to the extreme rainfall and enhanced the low-level convergence and water vapor transport.

The y component of wind difference and the difference vectors between 0600 and 0300 UTC (Fig. 10b) both indicate a distinct area of southward change of wind north of Zhengzhou with a magnitude greater than 8 m s−1, along with an area of northward change of wind at Zhengzhou and to its east and southeast. These wind changes contributed to the enhancement of local wind and moisture transport in the meridional direction. As shown in Fig. 10a, large local increases of moisture near Zhengzhou were nearly 4 g kg−1 during the short period of only 3 h.

To quantify from which direction the moisture transport was greatest, we computed the vertically integrated water vapor transport (IVT) over all model levels through the four sides of the orange square in Figs. 10a and 10b. Figure 10c shows that the east and south sides had positive (inward) transport during the entire period from 0000 to 1000 UTC while the north and west sides maintained negative (outward) transport. However, it is interesting to note that the north side had positive transport below 1.5 km, and the low-level IVT increased noticeably from ∼0300 to ∼0600 UTC, which is consistent with the southward turning of wind in the area north of Zhengzhou (Fig. 10b). It is further shown in Fig. 10d that the net IVT in the north–south direction is much larger than that in the east–west direction (Fig. 10d), and the IVT in the north–south direction as well as the total IVT increased markedly between 0330 and 0600 UTC, which could have played a critical role in setting the stage for the extreme local precipitation in the next 2–3 h.

Backward trajectories were calculated by tracing backward in time a point near Zhengzhou using 6-min updated winds from VDRAS analysis. When comparing the trajectories during 0000–0400 UTC and 0400–0700 UTC at different vertical levels, it is found that low level trajectories between 0.3 and 0.6 km turned anticlockwise associated with the enhancement of northerly winds (Fig. 10b), and those in the 3.0–5.0-km layer turned clockwise associated with the increase of the midlevel southerly winds, which is consistent with the increased IVT in the north–south direction (Fig. 10d). In the next section, we will provide analysis of the possible mesoscale features responsible for the rapid increase of wind and horizontal moisture transports in the few hours before the extreme rains occurred at Zhengzhou.

5. Mesoscale factors contributing to the development of 7.20 Zhengzhou rainstorm

It was shown in section 3 (Fig. 7) that the initiation of the rainstorm occurred in the southeast foothills of Mount Song before intensifying a few hours later near Zhengzhou. In this section we will first describe the mechanisms for the initiation and early development of the rainstorm (stage I), and then the local processes near Zhengzhou that contributed to its intensification (stage II).

a. The mesolow

1) The evolution of the mesolow

The mesolow centered in western Henan formed when the midtropospheric WPSH intruded southwestward (Fig. 4c) as mentioned in section 3. Here we examine the evolution of the mesolow in more detail using VDRAS analyzed pressure perturbation and wind fields (Fig. 11). The perturbation pressure is relative to the horizontal mean pressure averaged over D2 (Fig. 1) at 1800 UTC 19 July. During the formation process of the inverted trough, the prevailing southeasterlies in the region to the southeast of Funiu Mountains and Mount Song (Fig. 8b) gradually turned to southerlies (Fig. 11a). A cyclonic circulation centered at Funiu Mountains became identifiable at 2100 UTC (Fig. 11b) when the flow along the southeast slopes of Funiu Mountains turned to southwesterly. From 1800 to 0000 UTC, the central pressure decreased, and the minimum perturbation pressure became <−4 hPa at 0000 UTC as the center of the cyclonic circulation moved northeastward (Figs. 11a–c). Consistent with the motion of the mesolow, the winds both south of Zhengzhou and along the eastern part of Mount Song evolved from southeasterly (Fig. 11a) to southerly, and the southwesterly winds in the eastern part of Funiu Mountains strengthened (Fig. 11c).

Fig. 11.
Fig. 11.

Evolution of pressure perturbation (difference from the horizontal mean over D2 at 1800 UTC 19 Jul) and wind vectors at Z = 3 km at (a) 1800, (b) 2100, and (c) 0000 UTC 19 and (d) 0300, (e) 0600, and (f) 0800 UTC 20 Jul 2021. The green line represents the 300-m terrain height. The city center of Zhengzhou is marked by the red star. The orange target symbol marks the center of the mesolow circulation.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

During the next few hours, the mesolow center continued moving northeastward toward the valley as the negative pressure perturbation weakened by about 1 hPa at 0300 UTC (Fig. 11d). Thereafter, the mesolow center did not change position significantly, but the area with the perturbation pressure less than −2.5 hPa continued to move northeastward. Consequently, the southerly winds near Zhengzhou became stronger and the southwesterly winds extended farther northeast over the Mount Song area (Fig. 11e). The magnitude of the southerly winds south of Zhengzhou strengthened to above 16 m s−1 (Figs. 12a–d) and maintained that strength for a few hours (Figs. 12d–f) up to ∼0900 UTC. The depth of the strong southerly winds (>12 m s−1) extended from the height of ∼600 m to ∼6 km above the surface (Fig. 12h). The southwesterly winds had a similar depth, but a higher base and were slightly weaker (not shown). We will show later in this section that both the southerly and southwesterly airstreams associated with the mesolow played important roles in development of extreme rainfall. It is noted from Fig. 12h that a shallow low-level northerly flow extended ∼80 km from point A, which will also be discussed later in this section.

Fig. 12.
Fig. 12.

Evolution of y component of wind and wind vectors at the horizontal plane of Z = 3.0 km at (a) 1800 and (b) 2100 19 Jul 2021 and (c) 0000, (d) 0400, (e) 0600, (f) 0800, (g) 1000 UTC 20 Jul 2021 and at (h) the vertical plane along the black line A–B at 0600 UTC. The green line represents the 300-m terrain height. The city center of Zhengzhou is marked by the black star.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

2) The roles of the mesolow in stage I of the rainstorm

During the development and movement of the mesolow, convection initiated and organized. We first discuss the role of the mesolow in initiating the three convective cells (C1, C2, and C3 in Fig. 7a) in stage I that later merged to form the Zhengzhou rainstorm, and then explain how the mesolow contributed to the local intensification in stage II.

All three convective cells C1, C2, and C3 (labeled by the blue letters and arrows) were located within convergence bands (Figs. 13a,b) northeast of the mesolow center. The spiral convergence bands that produced C1 and C2 occurred directly within the mesolow circulation, whereas the convergence band that triggered C3 initially formed along a wind shift in the environmental southeasterly flow, but later also became part of the mesolow circulation. As the mesolow moved northeastward, the rain cells C1 and C2 merged (Figs. 13c,d) when the two inner convergence bands merged. In contrast, rain cell C3 developed further as it moved north-northwestward in the southeasterly flow (Figs. 13a–d). By 0524 UTC, the two distinct storms C1 + C2 and C3 grew to larger sizes near Zhengzhou. In the next 2 h, these two storms merged into a single larger storm over Zhengzhou and intensified.

Fig. 13.
Fig. 13.

Convergence field (as positive values) overlaid by wind vectors and observed 42-dBZ reflectivity (black contour) at z = 1.6 km. The three convective cells that eventually merged to form the 20 Jul (i.e., 7.20) extreme rainstorm are labeled by blue letters and arrows. The green line indicates the 300-m terrain height. The city of Zhengzhou is marked by the black star.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The merging process of C1 + C2 and C3 is illustrated further by the low-level (Z = 0.4 km) convergence field (Fig. 14). The two convergence bands where these two storms were located moved closer to each other between 0524 and 0624 UTC (Figs. 14a,b) before colliding by 0724 UTC (Fig. 14c). The vertical structure of the colliding convergence bands is shown in Figs. 14e and 14f, where enhanced near-surface northeasterlies can be seen entering the storms. The storm C1 + C2 moved with the midlevel southwesterly winds on the southeast side of the mesolow and then encountered the low-level northeasterly flow, resulting in enhanced convergence, updraft, and a stronger merged storm C1 + C2 + C3. The northeasterly flow could have resulted from the orographic effect of the Taihang Mountains, which will be discussed next.

Fig. 14.
Fig. 14.

(a)–(d) Convergence field (positive values for convergence) overlaid by wind vectors and observed 42-dBZ reflectivity (blue contour) at Z = 0.4 km on a subdomain of Fig. 13. The three convective cells that eventually merged to form the 7.20 extreme rainstorm are labeled by blue letters and arrows. The green line indicates the 300-m terrain height. The city of Zhengzhou is marked by the black star. (e) A vertical cross section of convergence field at 0524 UTC along the black straight-line A–B in (a) overlaid by reflectivity contours of 42, 44, and 48 dBZ (green). (f) As in (e), but at 0724 UTC along the black straight-line C–D in (c).

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

Given the proximity of the initiated convective cells to the east slopes of Mount Song at 0254 UTC (Fig. 13a), it is reasonable to speculate that terrain effects may have also influenced the convection initiation in addition to the convergence bands produced by the mesolow. Furthermore, terrain effects may have contributed to the formation of the mesolow, in which case they may indirectly contribute to the convection initiation.

Other than directly contributing to the initiation of the convective cells that produced the Zhengzhou rainstorm, the northeast quarter of the cyclonically circulating winds in the midtropospheric mesolow also played an important role in modifying the local environment by increasing the vertical wind shear near Zhengzhou (Figs. 9b,d). When the center of the mesolow circulation reached the west end of the valley (Fig. 11d), the Z = 3 km southerly flow around Zhengzhou was strengthened (Fig. 12), which was one of the two major factors responsible for the enhanced shear. Other factors associated with the enhancement of the low-level winds will be explained next in this section.

The evolution of the y component of wind indicates that the increased southerly winds by the mesolow (Fig. 12) was the main reason for the northward wind change south of Zhengzhou between 0300 and 0600 UTC (Fig. 10b). The strengthened southerly winds contributed to the formation of the Zhengzhou rainstorm by producing enhanced moisture transports and favorably influencing horizontal convergence in the north–south direction through its juxtaposition with the enhanced northerly flow to its north, which will be discussed in sections 5b and 5c.

To summarize, the mesolow contributed to the 7.20 extreme rainstorm via these four factors: 1) production of spiral convergence bands that triggered the initial convective cells, 2) strengthening of the vertical wind shear near Zhengzhou by enhancement of the southerly winds over a deep vertical layer, 3) contribution to the merging of the storms C1 + C2 and C3 as the southwesterly flow from the mesolow encountered the low-level northeasterly flow, and 4) enhancing the y component of winds south of Zhengzhou that increased the convergence with the enhanced northerly flow north of Zhengzhou.

b. The northeasterly barrier jet

The area north of Zhengzhou, stretching from the valley in the western foothills of the Taihang Mountains in the east, is prone to the influences of terrain including flow channeling or blocking. On 20 July a northeasterly flow along the slopes of the Taihang Mountains influenced the area, in addition to the mesolow circulation. Based on the location, the height (Figs. 16e,f), and the parallel wind direction to the elevation isoline, the low-level northeasterlies appear to be enhanced by a barrier jet near the Taihang Mountains. The barrier jet and its contribution to the 7.20 Zhengzhou rainstorm was noted earlier by Wei et al. (2023).

The temporal variation of the northeasterly jet and its connection to the 7.20 Zhengzhou rainstorm are revealed in Fig. 15. At 0100 UTC (Fig. 15a), the valley area was dominated by flows associated with the mesolow circulation with easterlies turning to northeasterlies except for the foothills to its northeast where a northeasterly flow was present. During the next several hours, the northeasterly flow grew stronger (Fig. 15b) and spread southward toward Zhengzhou (Figs. 15c,d).

Fig. 15.
Fig. 15.

Barrier jet wind speed (shading), computed by projecting full winds to the direction parallel to the slopes of Tainhang Mountains (about 56° from the north), overlaid by the full wind vectors at the horizontal plane of Z = 0.7 km at (a) 0100, (b) 0400, (c) 0600, and (d) 0800 UTC 20 Jul, and at the vertical cross sections along the black line (e) A–B in (c) and (f) C–D in (d). The storms C1 + C2 and C3 are marked by magenta letters and arrows in (c). Note that only the speed of northeasterly winds (with directions from 180° to 270°) was shown. The black contour represents the 300-m terrain height, and the green contour is 42-dBZ observed reflectivity [also 46 and 52 dBZ in (e) and (f)]. The city center of Zhengzhou is marked by the red star.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

One of the contributions of the barrier flow to extreme rainfall was to provide lifting as this airstream converged with the low-level southwesterlies from the mesolow (Figs. 14e,f).

This northeasterly barrier jet played a critical role in the merging of cells C1 + C2 and C3 (Fig. 15c) as it extended southward toward Zhengzhou. Without the southward extension of the barrier jet, the merging and the intensification of the storm might not have occurred. Hence, an important question examined in the next subsection is what process(es) caused the southward extension of the barrier jet?

c. The downslope gravity current

A gravity current moving down the slopes of Taihang Mountains is the third factor that contributed to the Zhengzhou rainstorm. Our analysis indicated that this current was the main reason for the northerly wind increases north of Zhengzhou (Fig. 10b, blue shadings north of star symbol) and for the southward extension of the northeasterly flow, which played a critical role in the re-intensification of cell C3 and its subsequent merging with cell C1 + C2 (Figs. 15c,d). Below we examine the evolution of the gravity current using the analysis of surface observations and VDRAS cross sections. The surface analysis is a product of VDRAS that is performed on surface observations using the objective analysis method of Barnes (1964) before 4DVar data assimilation.

A northerly flow (in the oval in Fig. 16a) from the high peaks (>1500 m) of Taihang Mountains to the south slopes persisted during the morning hours until 0100 UTC, accompanied by a negative temperature perturbation of near −3.0°C (Figs. 16a–e), which could have been produced by higher longwave radiation loss over the mountain top due to drier air (not shown). The cold air mass along the slopes was then advanced southwestward as the flow was subsequently channeled by the northeasterly barrier jet. The southwestward advance of the cold air mass halted in the next few hours (Fig. 16f) as the gravity current weakened, likely due to the increased radiative heating of the slope in the morning, which might be expected based on the enhanced warming southeast of Zhengzhou. Around noontime, a secondary gravity current (in the green oval in Fig. 16f) developed from the area west of the Taihang Mountain ridge (an elevated valley) and both currents persisted in the next few hours (Figs. 16g–i). By tracing the cold air mass that accompanied the secondary gravity current backward in time in vertical cross-sections across the west slopes of Taihang ridge (not shown), we found that this cold air mass was also originated from the high peaks of the Taihang and accumulated in the valley by downslope winds. The cold air mass over the higher elevation of Taihang Mountains was maintained up to the late morning, likely because of the cloudy sky that prevented radiative heating (Fig. 6). After noontime, the cold air mass over the slopes began to advance south-southeastward toward Zhengzhou, as a result of increased surface pressure in the cold air mass (not shown). In addition to the effect of these downslope currents, the strengthening of the cold air mass from 0500 to 0800 UTC may have also influenced by the evaporative cooling of the persistent light precipitation in the valley and over the lower slopes in the early afternoon hours. However, given the largely near-saturated condition in the area, the latter may have played a smaller role.

Fig. 16.
Fig. 16.

Gridded two-dimensional analysis of temperature perturbation overlaid by surface wind analysis at (a) 2148, (b) 2242, and (c) 2318 UTC 19 and (d) 0000, (e) 0100, (f) 0400, (g) 0500, (h) 0600, and (i) 0800 UTC 20 Jul 2021. The temperature perturbation is relative to the horizontal mean temperature in D2 at 1800 UTC 19 Jul. The green ovals in (a) and (f) mark the areas of the northerly gravity currents descending from the peaks of the Taihang Mountains. The city center of Zhengzhou is marked by the red star. The 300- and 1100-m terrain heights are represented by the red contours. The black line A–B in (b) indicates the cross section shown in Fig. 17.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

The vertical structure of the downslope gravity current at different times is shown in Fig. 17 within a north–south vertical cross section across Zhengzhou (marked in Fig. 16b). A cold air mass on top of the Taihang Mountains started to increase in depth and intensity, and had a minimum perturbation of <−2.0°C at 2030 UTC with very weak near-surface winds (Fig. 17a).

Fig. 17.
Fig. 17.

Vertical cross sections of potential temperature perturbation overlaid with in-plane winds along the black line A–B in Fig. 16b at (a) 2030, (b) 2206, and (c) 2318 UTC 19 and (d) 0000, (e) 0100, (f) 0400, (g) 0500, (h) 0600, and (i) 0800 UTC 20 Jul 2021. The potential temperature perturbation was obtained by subtracting the horizontal mean potential temperature averaged over the VDRAS domain D2 at 1800 UTC 19 Jul 2021. The green contour lines in (h) and (i) are observed reflectivity above 42 dBZ with a 2-dBZ interval.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

VDRAS analyzed humidity (not shown), indicated that the air at mountaintop was subsaturated and had relative humidities that were much lower than those over the lower slopes and the valley, which was favorable for nighttime radiative cooling. As the cold air mass intensified during the next few hours, the near surface winds became downslope but were still quite weak (Figs. 17b–d), which corresponded to the period of the southwestward propagation of the cold air mass shown in Figs. 16a–d. After 0100 UTC, however, stronger southward advancement occurred, and the northerly wind component increased to ∼8 m s−1 (Figs. 17e–i). Between 0600 and 0800 UTC, these northerlies associated with the cold air mass met the southerly winds from the mesolow and resulted in substantial ascending motion and rapid convective development (Figs. 17h,i).

6. Summary and conclusions

In this study, we investigated the mesoscale factors that led to the initiation, development, and intensification of the 7.20 Zhengzhou rainstorm that produced extreme 1-h rainfall amounts on 20 July 2021. The investigation was based on VDRAS 2-km analyses with 6-min updates obtained by assimilating radar radial velocity and reflectivity observations from an operational network consisting of 17 Doppler radars, and wind, temperature, and humidity measurements from a dense surface network. The 4DVar data assimilation system VDRAS produced 170 continuously cycled 2-km analyses at a 6-min interval over a period of 17 h with the boundary conditions and a first guess background provided by WRF 4-km simulations.

Our analysis of precipitation evolution indicated that the 7.20 Zhengzhou rainstorm was formed by sequential merging of three convective cells. The storm lifetime can be characterized by two stages. Stage I covers the period when the three convective cells (C1, C2, and C3) that eventually formed the rainstorm initiated and developed, and the first two cells merged (C1 + C2), before they reached Zhengzhou. Stage II includes the period of local intensification and quasi-stationary maintenance during which the second merging of the cells C1 + C2 and C3 occurred. We further illustrated that a lower-tropospheric mesolow that developed west of Henan within a larger-scale inverted trough played an important role in the initiation and development of the rainstorm. Two other factors, a barrier jet due to the blocking effect of the Taihang Mountains and a downslope gravity current from the Taihang Mountains, also played crucial roles in the intensification and maintenance of the rainstorm during stage II. The specific contributions from each of these mesoscale systems are summarized below with the aid of the schematic diagram (Fig. 18) illustrating the lower-tropospheric flow structure in stage II.

  1. The mesolow contributed to the 7.20 Zhengzhou extreme rainstorm in the following ways: (i) by supporting convergence bands within its circulation, along which convective cells initiated and later merged as the bands collided, (ii) by enhancing the southerly wind component, which resulted in large 0–3-km vertical shear (which helps storms to persist) when the southerlies overlaid surface northeasterlies (Fig. 18a), (iii) by promoting near-surface convergence as its southwesterly flow met low-level northeasterly flow, and (iv) by facilitating deep moisture transport from the south toward the region of persistent convection.

  2. The barrier jet contributed to the rainstorm by enhancing the vertical shear in conjunction with the midlevel southerly and southwesterly flows from the mesolow. In addition, the northeasterly barrier flow was also one of the main contributors to the low-level wind convergence and lifting.

  3. The downslope gravity current and the cold air mass acted as a barrier to extend the northeasterly flow toward Zhengzhou, by displacing it southward, and thereby strengthening the low-level convergence as the modified northeasterly flow encountered the southerly flow from the mesolow. The downslope gravity current also played an important role in further stabilizing the low-level atmosphere in the area north of Zhengzhou, which may have also contributed to the strengthening of the northeasterly barrier flow.

Fig. 18.
Fig. 18.

Schematic diagram summarizing the major processes that contributed to the formation of the 7.20 extreme rainstorm at Zhengzhou (b) in stage II at 0600 UTC 20 Jul 2021, as compared with (a) the undisturbed environment at 0600 UTC 19 Jul. The dark-blue contour lines stand for the mesolow pressure perturbation, the dashed brown line represents the 300-m terrain elevation that separates the western mountainous and eastern plain regions, and the triangle represents the location of Zhengzhou. The red, magenta, green, and blue arrows stand for the winds from the large-scale environment, the winds from the mesolow, the barrier flow, and the downslope gravity current. The black dashed line illustrates the position of the cold air front associated with the gravity current.

Citation: Monthly Weather Review 151, 8; 10.1175/MWR-D-22-0337.1

Overall, the mesolow played a critical role in the initiation and early development of the first two convective cells and their merging during stage I southwest of Zhengzhou and it also provided a favorable local environment for the extreme rainstorm by enhancing the northward moisture transport within the supportive large-scale environment. The interaction between the downslope gravity current and the barrier flow played critical roles in the second merging process in stage II by enhancing wind convergence near Zhengzhou and increasing near-surface southward moisture transport toward Zhengzhou.

Wei et al. (2023) also examined the mesoscale systems that contributed to the extreme rainstorm at stage II based on a high-resolution WRF simulation. Our analysis confirmed the following findings from their study. 1) The extreme rainstorm was maintained by the dynamical lifting of low-level converging flows from the south, north, and the east of the storm. 2) The low-level northerly flow that originated from the barrier jet along the slopes of the Taihang Mountains played an important role by converging with the southerly winds from the mesolow. However, there are also distinct differences between the two studies. The current study revealed the important role of the downslope gravity current in extending and veering the barrier flow toward Zhengzhou, and thereby strengthening the low-level convergence as the modified flow encountered the southerly flow from the mesolow (Fig. 15). Although Wei et al. (2023)’s explanation of the expanded barrier jet by the Rossby radius of deformation was reasonable, it could not explain the southward veering of the barrier flow toward Zhengzhou, which was essential and a major contributor to the increased north–south moisture convergence (Fig. 10). As a result of the missing gravity current, the rainstorm in the WRF simulation was located to the south of Zhengzhou (Fig. 3 in Wei et al. 2023) and failed to reproduce the reintensification of the cell C3 north of Zhengzhou and its subsequent merging with C1 + C2 (Figs. 13c–e). Furthermore, although the mesolow was successfully simulated in Wei et al. (2023), its center location was offset westward, which resulted in spurious precipitations in the west part of the valley and an unrealistic cold pool northwest of Zhengzhou that enhanced the west–moueast convergence together with the easterly environmental flow. Our analysis (Figs. 10b–d), however, suggested that the north–south convergence played a dominant role in intensifying and maintaining the 7.20 rainstorm.

The main purpose of the current study was to identify the main mesoscale factors that led to the formation of the 7.20 Zhengzhou rainstorm. Further studies using a subkilometer NWP model and sensitivity experiments are necessary to better understand the interaction of these mesoscale features. For example, interesting questions to investigate would be whether the northeasterly barrier flow could have expanded southeastward even if there were no downslope gravity current? and whether the Zhengzhou rainstorm would occur without the southeastward expansion of the barrier flow?

Revealing the mesoscale factors responsible for this extreme rainstorm is an important first step toward improving their numerical prediction. Future research will examine how and to what extent these factors contributed to the extreme rainfall as well as the predictability of these mesoscale features. Using high-resolution ensemble simulations in combination with sensitivity studies of model physics and data assimilation, we hope to investigate the capability and deficiencies of the current convection-allowing NWP models in predicting the mesoscale processes and the timing and location of the heavy rainfall event.

1

The entire area of Zhengzhou City extends ∼80 km north–south and ∼165 km east–west as shown by the red border line in Fig. 1b. For convenience, however, we will refer to the city center (Fig. 1b, star symbol) as Zhengzhou in this paper unless otherwise specified.

2

Since VDRAS uses a height coordinate, the model levels are relative to mean sea surface (MSL). The results on the planes above ground level (AGL) were interpolated from the original height coordinate.

3

All following figures in sections 4 and 5 are produced from VDRAS analysis, unless otherwise specified.

Acknowledgments.

The authors thank Dr. Yuan Ho for his help in using IDV for some of the figures. The National Center for Atmospheric Research (NCAR) is sponsored by the National Science Foundation. This work was performed as part of NCAR’s Short Term Explicit Prediction (STEP) program, which is supported by National Science Foundation funds for the U.S. Weather Research Program (USWRP). This work was also supported by the National Natural Science Foundation of China (Grant 42030607).

Data availability statement.

Please contact the corresponding author Qinghong Zhang for permission to access the data.

REFERENCES

  • Barnes, S. L., 1964: A technique for maximizing details in numerical weather map analysis. J. Appl. Meteor., 3, 396409, https://doi.org/10.1175/1520-0450(1964)003<0396:ATFMDI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., and C. W. Newton, 1983: Thunderstorms in the synoptic setting. Thunderstorm Morphology and Dynamics, E. Kessler, Ed., University of Oklahoma, 75–112.

  • Cai, X., T. Chen, Y. Chen, J. Fu, and N. Hu, 2022: Dynamic impact of upper tropospheric cold low on persistent extreme rainstorm of Henan during 17–22 July 2021 (in Chinese). Meteor. Mon., 48, 545555, https://doi.org/10.7519/j.issn.1000-0526.2022.021802.

    • Search Google Scholar
    • Export Citation
  • Chang, S.-F., Y.-C. Liou, J. Sun, and S.-L. Tai, 2016: The implementation of the ice-phase microphysical process into a four-dimensional Variational Doppler Radar Analysis System (VDRAS) and its impact on parameter retrieval and quantitative precipitation nowcasting. J. Atmos. Sci., 73, 10151038, https://doi.org/10.1175/JAS-D-15-0184.1.

    • Search Google Scholar
    • Export Citation
  • Chen, X., K. Zhao, J. Sun, B. Zhou, and W.-C. Lee, 2016: Assimilating surface observations in a four-dimensional variational Doppler radar data assimilation system to improve the analysis and forecast of a squall line case. Adv. Atmos. Sci., 33, 11061119, https://doi.org/10.1007/s00376-016-5290-0.

    • Search Google Scholar
    • Export Citation
  • Crook, N. A., and J. Sun, 2004: Analysis and forecasting of the low-level wind during the Sydney 2000 forecast demonstration project. Wea. Forecasting, 19, 151167, https://doi.org/10.1175/1520-0434(2004)019<0151:AAFOTL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., III, 1987: The distinction between large-scale and mesoscale contribution to severe convection: A case study example. Wea. Forecasting, 2, 316, https://doi.org/10.1175/1520-0434(1987)002<0003:TDBLSA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Feng, J., J. Sun, and Y. Zhang, 2020: A dynamic blending scheme to mitigate large‐scale bias in regional models. J. Adv. Model. Earth Syst., 12, e2019MS001754, https://doi.org/10.1029/2019MS001754.

    • Search Google Scholar
    • Export Citation
  • Gao, J., and D. J. Stensrud, 2012: Assimilation of reflectivity data in a convective-scale, cycled 3DVAR framework with hydrometeor classification. J. Atmos. Sci., 69, 10541065, https://doi.org/10.1175/JAS-D-11-0162.1.

    • Search Google Scholar
    • Export Citation
  • Gochis, D., and Coauthors, 2015: The great Colorado flood of September 2015. Bull. Amer. Meteor. Soc., 96, 14611487, https://doi.org/10.1175/BAMS-D-13-00241.1.

    • Search Google Scholar
    • Export Citation
  • Johns, R. H., and C. A. Doswell III, 1992: Severe local storms forecasting. Wea. Forecasting, 7, 588612, https://doi.org/10.1175/1520-0434(1992)007<0588:SLSF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., and B. E. Mapes, 2001: Mesoscale processes and severe convective weather. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 71–122, https://doi.org/10.1175/0065-9401-28.50.71.

  • Ke, C.-Y., K.-S. Chung, T.-C. C. Wang, and Y.-C. Liou, 2019: Analysis of heavy rainfall and barrier-jet evolution during Mei-Yu season using multiple Doppler radar retrievals: A case study on 11 June 2012. Tellus, 71A, 1571369, https://doi.org/10.1080/16000870.2019.1571369.

    • Search Google Scholar
    • Export Citation
  • Li, J., and Y.-L. Chen, 1998: Barrier jets during TAMEX. Mon. Wea. Rev., 126, 959971, https://doi.org/10.1175/1520-0493(1998)126<0959:BJDT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lundquist, K. A., F. K. Chow, and J. K. Lundquist, 2012: An immersed boundary method enabling large-eddy simulations of flow over complex terrain in the WRF Model. Mon. Wea. Rev., 140, 39363955, https://doi.org/10.1175/MWR-D-11-00311.1.

    • Search Google Scholar
    • Export Citation
  • Luo, L., M. Xue, K. Zhu, and B. Zhou, 2018: Explicit prediction of hail in a long-lasting multicellular convective system in eastern China using multimoment microphysics schemes. J. Atmos. Sci., 75, 31153137, https://doi.org/10.1175/JAS-D-17-0302.1.

    • Search Google Scholar
    • Export Citation
  • Marquis, J., Y. Richardson, P. Markowski, J. Wurman, K. Kosiba, and P. Robinson, 2016: An investigation of the Goshen County, Wyoming, tornadic supercell of 5 June 2009 using EnKF assimilation of mobile mesonet and radar observations collected during VORTEX2. Part II: Mesocyclone-scale processes affecting tornado formation, maintenance, and decay. Mon. Wea. Rev., 144, 34413463, https://doi.org/10.1175/MWR-D-15-0411.1.

    • Search Google Scholar
    • Export Citation
  • Newton, C. W., 1963: Dynamics of severe convective storms. Severe Local Storms, Meteor. Monogr., No. 27, Amer. Meteor. Soc., 33–58.

  • Orlanski, I., 1975: A rational subdivision of scales for atmospheric processes. Bull. Amer. Meteor. Soc., 56, 527530, https://www.jstor.org/stable/26216020.

    • Search Google Scholar
    • Export Citation
  • Poulos, G., and S. Zhong, 2008: An observational history of small-scale katabatic winds in mid-latitudes. Geogr. Compass, 2, 17981821, https://doi.org/10.1111/j.1749-8198.2008.00166.x.

    • Search Google Scholar
    • Export Citation
  • Ran, L., and Coauthors, 2021: Observational analysis of the dynamic, thermal, and water vapor characteristics of the “7.20” extreme rainstorm event in Henan Province, 2021. Chin. J. Atmos. Sci., 45, 13661383, https://doi.org/10.3878/j.issn.1006-9895.2109.21160.

    • Search Google Scholar
    • Export Citation
  • Schwartz, C. S., J. Poterjoy, J. R. Carley, D. C. Dowell, G. S. Romine, and K. Ide, 2022: Comparing partial and continuously cycling ensemble Kalman filter data assimilation systems for convection-allowing ensemble forecast initialization. Wea. Forecasting, 37, 85112, https://doi.org/10.1175/WAF-D-21-0069.1.

    • Search Google Scholar
    • Export Citation
  • Sun, J., and N. A. Crook, 1997: Dynamical and microphysical retrieval from Doppler radar observations using a cloud model and its adjoint. Part I: Model development and simulated data experiments. J. Atmos. Sci., 54, 16421661, https://doi.org/10.1175/1520-0469(1997)054<1642:DAMRFD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sun, J., and N. A. Crook, 1998: Dynamical and microphysical retrieval from Doppler radar observations using a cloud model and its adjoint. Part II: Retrieval experiments of an observed Florida convective storm. J. Atmos. Sci., 55, 835852, https://doi.org/10.1175/1520-0469(1998)055<0835:DAMRFD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sun, J., M. Chen, and Y. Wang, 2010: A frequent-updating analysis system based on radar, surface, and mesoscale model data for the Beijing 2008 forecast demonstration project. Wea. Forecasting, 25, 17151735, https://doi.org/10.1175/2010WAF2222336.1.

    • Search Google Scholar
    • Export Citation
  • Tai, S.-L., Y.-C. Liou, J. Sun, S.-F. Chang, and M.-C. Kuo, 2011: Precipitation forecasting using Doppler radar data, a cloud model with adjoint, and the Weather Research and Forecasting Model: Real case studies during SoWMEX in Taiwan. Wea. Forecasting, 26, 975992, https://doi.org/10.1175/WAF-D-11-00019.1.

    • Search Google Scholar
    • Export Citation
  • Tai, S.-L., Y.-C. Liou, J. Sun, and S.-F. Chang, 2017: The development of a terrain-resolving scheme for the forward model and its adjoint in the four-dimensional Variational Doppler Radar Analysis System (VDRAS). Mon. Wea. Rev., 145, 289306, https://doi.org/10.1175/MWR-D-16-0092.1.

    • Search Google Scholar
    • Export Citation
  • Trapp, R. J., 2013: Mesoscale-Convective Processes in the Atmosphere. Cambridge University Press, 346 pp.

  • Trier, S. B., G. Romine, D. A. Ahijevych, R. J. Trapp, R. S. Schumacher, M. C. Coniglio, and D. J. Stensrud, 2015: Mesoscale thermodynamic influences on convection initiation near a surface dryline in a convection-permitting ensemble. Mon. Wea. Rev., 143, 37263753, https://doi.org/10.1175/MWR-D-15-0133.1.

    • Search Google Scholar
    • Export Citation
  • Tseng, Y.-H., and J. H. Ferziger, 2003: A ghost-cell immersed boundary method for flow in complex geometry. J. Comput. Phys., 192, 593623, https://doi.org/10.1016/j.jcp.2003.07.024.

    • Search Google Scholar
    • Export Citation
  • Wang, H., J. Sun, X. Zhang, X.-Y. Huang, and T. Auligne, 2013: Radar data assimilation with WRF-4DVAR. Part I: System development and preliminary testing. Mon. Wea. Rev., 141, 22242244, https://doi.org/10.1175/MWR-D-12-00168.1.

    • Search Google Scholar
    • Export Citation
  • Wang, Q., M. Xue, and Z. Tan, 2016: Convective initiation by topographically induced convergence forcing over the Dabie Mountains on 24 June 2010. Adv. Atmos. Sci., 33, 11201136, https://doi.org/10.1007/s00376-016-6024-z.

    • Search Google Scholar
    • Export Citation
  • Wei, P., and Coauthors, 2023: On the key dynamical processes supporting the 21.7 Zhengzhou record-breaking hourly rainfall in China. Adv. Atmos. Sci., 40, 337349, https://doi.org/10.1007/s00376-022-2061-y.

    • Search Google Scholar
    • Export Citation
  • Wu, Y., J. Sun, Z. Ying, L. Xue, D. Chen, and W. Lin, 2021: Effects of local-scale orography and urban heat island on the initiation of a record-breaking rainfall event. J. Geophys. Res. Atmos., 126, e2021JD034839, https://doi.org/10.1029/2021JD034839.

    • Search Google Scholar
    • Export Citation
  • Wu, Y.-J., Y.-C. Liou, Y.-C. Lo, S.-L. Tai, S.-F. Chang, and J. Sun, 2021: Precipitation processes of a thunderstorm occurred on 19 August 2014 in northern Taiwan documented by using a high resolution 4DVar data assimilation system. J. Atmos. Soc. Japan, 99, 10231044, https://doi.org/10.2151/jmsj.2021-049.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, M. Chen, Y. Wang, and Z. Ying, 2017: The characteristics of weakly forced mountain-to-plain precipitation systems based on radar observations and high-resolution reanalysis. J. Geophys. Res. Atmos., 122, 31933213, https://doi.org/10.1002/2016JD025914.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, M. Chen, X. Qie, Z. Ying, Y. Wang, and L. Ji, 2019: A comparison of environmental and mesoscale characteristics between two types of mountain-to-plain precipitation systems in the Beijing Region, China. J. Geophys. Res. Atmos., 124, 68566872, https://doi.org/10.1029/2018JD029896.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, X. Qie, L. Ji, Z. Ying, M. Chen, and L. Zhang, 2021: Lightning data assimilation scheme in a 4DVAR system and its impact on very-short-term convective forecasting. Mon. Wea. Rev., 149, 353373, https://doi.org/10.1175/MWR-D-19-0396.1.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, L. Ji, L. Zhang, Z. Ying, Z. Chen, M. Chen, and C. Xu, 2022: A study on local-scale thermal and dynamical mechanisms in the initiation of a squall line under weak forcing. J. Geophys. Res. Atmos., 127, e2021JD035561, https://doi.org/10.1029/2021JD035561.

    • Search Google Scholar
    • Export Citation
  • Xu, J., R. Li, Q. Zhang, Y. Chen, X. Liang, and X. Gu, 2022: Extreme large-scale atmospheric circulation associated with the “21·7” Henan flood. Sci. China Earth Sci., 65, 18471860, https://doi.org/10.1007/s11430-022-9975-0.

    • Search Google Scholar
    • Export Citation
  • Yin, L., F. Ping, J. Miao, and S. Jin, 2023: Analysis on precipitation efficiency of the “21.7” Henan extremely heavy rainfall event. Adv. Atmos. Sci., 40, 374392, https://doi.org/10.1007/s00376-022-2054-x.

    • Search Google Scholar
    • Export Citation
  • Zhang, D.-L., Y. Lin, P. Zhao, X. Yu, S. Wang, H. Kang, and Y. Ding, 2013: The Beijing extreme rainfall of 21 July 2012: “Right results” but for wrong reasons. Geophys. Res. Lett., 40, 14261431, https://doi.org/10.1002/grl.50304.

    • Search Google Scholar
    • Export Citation
  • Zhang, L., J. Sun, Z. Ying, and X. Xiao, 2021: Initiation and development of a squall line crossing Hangzhou Bay. J. Geophys. Res. Atmos., 126, e2020JD032504, https://doi.org/10.1029/2020JD032504.

    • Search Google Scholar
    • Export Citation
  • Zhang, M., and D.-L. Zhang, 2012: Subkilometer simulation of a torrential-rain-producing mesoscale convective system in East China. Part I: Model verification and convective organization. Mon. Wea. Rev., 140, 184201, https://doi.org/10.1175/MWR-D-11-00029.1.

    • Search Google Scholar
    • Export Citation
  • Zhang, X., H. Wang, X. Wang, L. Shen, D. Wang, and H. Li, 2021: The character of extreme precipitation and its associate circulation in Henan July 2021. Trans. Atmos. Sci., 44, 672687.

    • Search Google Scholar
    • Export Citation
  • Zhang, Y., H. Yu, M. Zhang, Y. Yang, and Z. Meng, 2022: Uncertainties and error growth in forecasting the record-breaking rainfall in Zhengzhou, Henan on 19–20 July 2021. Sci. China Earth Sci., 65, 19031920, https://doi.org/10.1007/s11430-022-9991-4.

    • Search Google Scholar
    • Export Citation
Save
  • Barnes, S. L., 1964: A technique for maximizing details in numerical weather map analysis. J. Appl. Meteor., 3, 396409, https://doi.org/10.1175/1520-0450(1964)003<0396:ATFMDI>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Barnes, S. L., and C. W. Newton, 1983: Thunderstorms in the synoptic setting. Thunderstorm Morphology and Dynamics, E. Kessler, Ed., University of Oklahoma, 75–112.

  • Cai, X., T. Chen, Y. Chen, J. Fu, and N. Hu, 2022: Dynamic impact of upper tropospheric cold low on persistent extreme rainstorm of Henan during 17–22 July 2021 (in Chinese). Meteor. Mon., 48, 545555, https://doi.org/10.7519/j.issn.1000-0526.2022.021802.

    • Search Google Scholar
    • Export Citation
  • Chang, S.-F., Y.-C. Liou, J. Sun, and S.-L. Tai, 2016: The implementation of the ice-phase microphysical process into a four-dimensional Variational Doppler Radar Analysis System (VDRAS) and its impact on parameter retrieval and quantitative precipitation nowcasting. J. Atmos. Sci., 73, 10151038, https://doi.org/10.1175/JAS-D-15-0184.1.

    • Search Google Scholar
    • Export Citation
  • Chen, X., K. Zhao, J. Sun, B. Zhou, and W.-C. Lee, 2016: Assimilating surface observations in a four-dimensional variational Doppler radar data assimilation system to improve the analysis and forecast of a squall line case. Adv. Atmos. Sci., 33, 11061119, https://doi.org/10.1007/s00376-016-5290-0.

    • Search Google Scholar
    • Export Citation
  • Crook, N. A., and J. Sun, 2004: Analysis and forecasting of the low-level wind during the Sydney 2000 forecast demonstration project. Wea. Forecasting, 19, 151167, https://doi.org/10.1175/1520-0434(2004)019<0151:AAFOTL>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Doswell, C. A., III, 1987: The distinction between large-scale and mesoscale contribution to severe convection: A case study example. Wea. Forecasting, 2, 316, https://doi.org/10.1175/1520-0434(1987)002<0003:TDBLSA>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Feng, J., J. Sun, and Y. Zhang, 2020: A dynamic blending scheme to mitigate large‐scale bias in regional models. J. Adv. Model. Earth Syst., 12, e2019MS001754, https://doi.org/10.1029/2019MS001754.

    • Search Google Scholar
    • Export Citation
  • Gao, J., and D. J. Stensrud, 2012: Assimilation of reflectivity data in a convective-scale, cycled 3DVAR framework with hydrometeor classification. J. Atmos. Sci., 69, 10541065, https://doi.org/10.1175/JAS-D-11-0162.1.

    • Search Google Scholar
    • Export Citation
  • Gochis, D., and Coauthors, 2015: The great Colorado flood of September 2015. Bull. Amer. Meteor. Soc., 96, 14611487, https://doi.org/10.1175/BAMS-D-13-00241.1.

    • Search Google Scholar
    • Export Citation
  • Johns, R. H., and C. A. Doswell III, 1992: Severe local storms forecasting. Wea. Forecasting, 7, 588612, https://doi.org/10.1175/1520-0434(1992)007<0588:SLSF>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Johnson, R. H., and B. E. Mapes, 2001: Mesoscale processes and severe convective weather. Severe Convective Storms, Meteor. Monogr., No. 50, Amer. Meteor. Soc., 71–122, https://doi.org/10.1175/0065-9401-28.50.71.

  • Ke, C.-Y., K.-S. Chung, T.-C. C. Wang, and Y.-C. Liou, 2019: Analysis of heavy rainfall and barrier-jet evolution during Mei-Yu season using multiple Doppler radar retrievals: A case study on 11 June 2012. Tellus, 71A, 1571369, https://doi.org/10.1080/16000870.2019.1571369.

    • Search Google Scholar
    • Export Citation
  • Li, J., and Y.-L. Chen, 1998: Barrier jets during TAMEX. Mon. Wea. Rev., 126, 959971, https://doi.org/10.1175/1520-0493(1998)126<0959:BJDT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Lundquist, K. A., F. K. Chow, and J. K. Lundquist, 2012: An immersed boundary method enabling large-eddy simulations of flow over complex terrain in the WRF Model. Mon. Wea. Rev., 140, 39363955, https://doi.org/10.1175/MWR-D-11-00311.1.

    • Search Google Scholar
    • Export Citation
  • Luo, L., M. Xue, K. Zhu, and B. Zhou, 2018: Explicit prediction of hail in a long-lasting multicellular convective system in eastern China using multimoment microphysics schemes. J. Atmos. Sci., 75, 31153137, https://doi.org/10.1175/JAS-D-17-0302.1.

    • Search Google Scholar
    • Export Citation
  • Marquis, J., Y. Richardson, P. Markowski, J. Wurman, K. Kosiba, and P. Robinson, 2016: An investigation of the Goshen County, Wyoming, tornadic supercell of 5 June 2009 using EnKF assimilation of mobile mesonet and radar observations collected during VORTEX2. Part II: Mesocyclone-scale processes affecting tornado formation, maintenance, and decay. Mon. Wea. Rev., 144, 34413463, https://doi.org/10.1175/MWR-D-15-0411.1.

    • Search Google Scholar
    • Export Citation
  • Newton, C. W., 1963: Dynamics of severe convective storms. Severe Local Storms, Meteor. Monogr., No. 27, Amer. Meteor. Soc., 33–58.

  • Orlanski, I., 1975: A rational subdivision of scales for atmospheric processes. Bull. Amer. Meteor. Soc., 56, 527530, https://www.jstor.org/stable/26216020.

    • Search Google Scholar
    • Export Citation
  • Poulos, G., and S. Zhong, 2008: An observational history of small-scale katabatic winds in mid-latitudes. Geogr. Compass, 2, 17981821, https://doi.org/10.1111/j.1749-8198.2008.00166.x.

    • Search Google Scholar
    • Export Citation
  • Ran, L., and Coauthors, 2021: Observational analysis of the dynamic, thermal, and water vapor characteristics of the “7.20” extreme rainstorm event in Henan Province, 2021. Chin. J. Atmos. Sci., 45, 13661383, https://doi.org/10.3878/j.issn.1006-9895.2109.21160.

    • Search Google Scholar
    • Export Citation
  • Schwartz, C. S., J. Poterjoy, J. R. Carley, D. C. Dowell, G. S. Romine, and K. Ide, 2022: Comparing partial and continuously cycling ensemble Kalman filter data assimilation systems for convection-allowing ensemble forecast initialization. Wea. Forecasting, 37, 85112, https://doi.org/10.1175/WAF-D-21-0069.1.

    • Search Google Scholar
    • Export Citation
  • Sun, J., and N. A. Crook, 1997: Dynamical and microphysical retrieval from Doppler radar observations using a cloud model and its adjoint. Part I: Model development and simulated data experiments. J. Atmos. Sci., 54, 16421661, https://doi.org/10.1175/1520-0469(1997)054<1642:DAMRFD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sun, J., and N. A. Crook, 1998: Dynamical and microphysical retrieval from Doppler radar observations using a cloud model and its adjoint. Part II: Retrieval experiments of an observed Florida convective storm. J. Atmos. Sci., 55, 835852, https://doi.org/10.1175/1520-0469(1998)055<0835:DAMRFD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sun, J., M. Chen, and Y. Wang, 2010: A frequent-updating analysis system based on radar, surface, and mesoscale model data for the Beijing 2008 forecast demonstration project. Wea. Forecasting, 25, 17151735, https://doi.org/10.1175/2010WAF2222336.1.

    • Search Google Scholar
    • Export Citation
  • Tai, S.-L., Y.-C. Liou, J. Sun, S.-F. Chang, and M.-C. Kuo, 2011: Precipitation forecasting using Doppler radar data, a cloud model with adjoint, and the Weather Research and Forecasting Model: Real case studies during SoWMEX in Taiwan. Wea. Forecasting, 26, 975992, https://doi.org/10.1175/WAF-D-11-00019.1.

    • Search Google Scholar
    • Export Citation
  • Tai, S.-L., Y.-C. Liou, J. Sun, and S.-F. Chang, 2017: The development of a terrain-resolving scheme for the forward model and its adjoint in the four-dimensional Variational Doppler Radar Analysis System (VDRAS). Mon. Wea. Rev., 145, 289306, https://doi.org/10.1175/MWR-D-16-0092.1.

    • Search Google Scholar
    • Export Citation
  • Trapp, R. J., 2013: Mesoscale-Convective Processes in the Atmosphere. Cambridge University Press, 346 pp.

  • Trier, S. B., G. Romine, D. A. Ahijevych, R. J. Trapp, R. S. Schumacher, M. C. Coniglio, and D. J. Stensrud, 2015: Mesoscale thermodynamic influences on convection initiation near a surface dryline in a convection-permitting ensemble. Mon. Wea. Rev., 143, 37263753, https://doi.org/10.1175/MWR-D-15-0133.1.

    • Search Google Scholar
    • Export Citation
  • Tseng, Y.-H., and J. H. Ferziger, 2003: A ghost-cell immersed boundary method for flow in complex geometry. J. Comput. Phys., 192, 593623, https://doi.org/10.1016/j.jcp.2003.07.024.

    • Search Google Scholar
    • Export Citation
  • Wang, H., J. Sun, X. Zhang, X.-Y. Huang, and T. Auligne, 2013: Radar data assimilation with WRF-4DVAR. Part I: System development and preliminary testing. Mon. Wea. Rev., 141, 22242244, https://doi.org/10.1175/MWR-D-12-00168.1.

    • Search Google Scholar
    • Export Citation
  • Wang, Q., M. Xue, and Z. Tan, 2016: Convective initiation by topographically induced convergence forcing over the Dabie Mountains on 24 June 2010. Adv. Atmos. Sci., 33, 11201136, https://doi.org/10.1007/s00376-016-6024-z.

    • Search Google Scholar
    • Export Citation
  • Wei, P., and Coauthors, 2023: On the key dynamical processes supporting the 21.7 Zhengzhou record-breaking hourly rainfall in China. Adv. Atmos. Sci., 40, 337349, https://doi.org/10.1007/s00376-022-2061-y.

    • Search Google Scholar
    • Export Citation
  • Wu, Y., J. Sun, Z. Ying, L. Xue, D. Chen, and W. Lin, 2021: Effects of local-scale orography and urban heat island on the initiation of a record-breaking rainfall event. J. Geophys. Res. Atmos., 126, e2021JD034839, https://doi.org/10.1029/2021JD034839.

    • Search Google Scholar
    • Export Citation
  • Wu, Y.-J., Y.-C. Liou, Y.-C. Lo, S.-L. Tai, S.-F. Chang, and J. Sun, 2021: Precipitation processes of a thunderstorm occurred on 19 August 2014 in northern Taiwan documented by using a high resolution 4DVar data assimilation system. J. Atmos. Soc. Japan, 99, 10231044, https://doi.org/10.2151/jmsj.2021-049.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, M. Chen, Y. Wang, and Z. Ying, 2017: The characteristics of weakly forced mountain-to-plain precipitation systems based on radar observations and high-resolution reanalysis. J. Geophys. Res. Atmos., 122, 31933213, https://doi.org/10.1002/2016JD025914.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, M. Chen, X. Qie, Z. Ying, Y. Wang, and L. Ji, 2019: A comparison of environmental and mesoscale characteristics between two types of mountain-to-plain precipitation systems in the Beijing Region, China. J. Geophys. Res. Atmos., 124, 68566872, https://doi.org/10.1029/2018JD029896.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, X. Qie, L. Ji, Z. Ying, M. Chen, and L. Zhang, 2021: Lightning data assimilation scheme in a 4DVAR system and its impact on very-short-term convective forecasting. Mon. Wea. Rev., 149, 353373, https://doi.org/10.1175/MWR-D-19-0396.1.

    • Search Google Scholar
    • Export Citation
  • Xiao, X., J. Sun, L. Ji, L. Zhang, Z. Ying, Z. Chen, M. Chen, and C. Xu, 2022: A study on local-scale thermal and dynamical mechanisms in the initiation of a squall line under weak forcing. J. Geophys. Res. Atmos., 127, e2021JD035561, https://doi.org/10.1029/2021JD035561.

    • Search Google Scholar
    • Export Citation
  • Xu, J., R. Li, Q. Zhang, Y. Chen, X. Liang, and X. Gu, 2022: Extreme large-scale atmospheric circulation associated with the “21·7” Henan flood. Sci. China Earth Sci., 65, 18471860, https://doi.org/10.1007/s11430-022-9975-0.

    • Search Google Scholar
    • Export Citation
  • Yin, L., F. Ping, J. Miao, and S. Jin, 2023: Analysis on precipitation efficiency of the “21.7” Henan extremely heavy rainfall event. Adv. Atmos. Sci., 40, 374392, https://doi.org/10.1007/s00376-022-2054-x.

    • Search Google Scholar
    • Export Citation
  • Zhang, D.-L., Y. Lin, P. Zhao, X. Yu, S. Wang, H. Kang, and Y. Ding, 2013: The Beijing extreme rainfall of 21 July 2012: “Right results” but for wrong reasons. Geophys. Res. Lett., 40, 14261431, https://doi.org/10.1002/grl.50304.

    • Search Google Scholar
    • Export Citation
  • Zhang, L., J. Sun, Z. Ying, and X. Xiao, 2021: Initiation and development of a squall line crossing Hangzhou Bay. J. Geophys. Res. Atmos., 126, e2020JD032504, https://doi.org/10.1029/2020JD032504.

    • Search Google Scholar
    • Export Citation
  • Zhang