1. Introduction
Forecasting severe convective weather, which often accompanies short-term heavy rain, thunderstorms, or hailstorms, is a difficult problem and challenging task in the meteorological community (Wilson and Schreiber 1986; Choudhury et al. 2016; Zheng et al. 2016). Convective initiation (CI) is the initial formation of a convective storm characterized by obviously rapid development of lifting motion and is defined as the first occurrence of a radar echo of intensity greater than or equal to 35 dBZ from a growing cumulus cloud (Browning and Atlas 1965; Marshall and Radhakant 1978; Wilson and Schreiber 1986; Wilson and Mueller 1993; Mecikalski and Bedka 2006). CI is the most crucial and challenging aspect in operational convective nowcasting (0–2-h forecasting) because of the complicated interactions of multiscale atmospheric processes and underlying surface conditions. There are several tools available to nowcast CI: 1) analysis of surface observations to detect surface-based boundaries (Banacos and Schultz 2005; Madaus and Hakim 2016), 2) high-resolution numerical weather prediction models (Trentmann et al. 2009; Sun et al. 2014), 3) ground-based weather radar (Roberts and Rutledge 2003; Weckwerth et al. 2011), and 4) meteorological satellites (Mecikalski and Bedka 2006; Reif and Bluestein 2017). Surface-station observations and ground-based radar have proven to be extremely useful in nowcasting CI, but they are both limited in spatial extent, particularly over mountainous and ocean regions. Moreover, with numerical weather prediction models, it is very difficult to predict the timing and location of CI owing to the limitation of the data assimilation and not well understood of detailed initiation process. However, the high spatial and temporal resolution of satellite observation is one of the most effective means to monitor convective cloud clusters (CCs) over oceans and mountains regions.
Over the past several decades, statistical analysis of regional CI events has been essential for elucidating the mechanisms that control convective growth (Wilson and Schreiber 1986; Mercer and Richman 2007; Bai et al. 2020b). Climatological studies identified topographic features with a high likelihood of initiating convection (Banta and Schaaf 1987; Aoshima et al. 2008; Weckwerth et al. 2011; Chen et al. 2012; Huang et al. 2017). By detecting CI events in China, Chen et al. (2012) demonstrated that most storms initiate over the northwestern mountains of the contiguous North China in the afternoon as a result of solar heating. Huang et al. (2017) also found that CI signals peaks in the early afternoon and occur with high frequency in areas with remarkable terrain inhomogeneity over central eastern China. The CI occurrences in the coastal orography regions of South China also showed a primary peak in the early afternoon and a secondary peak at midnight (Bai et al. 2020a). In addition to topography, boundary layer convergence zones are important precursors to CIs, particularly over plain regions (Wilson and Schreiber 1986; Reif and Bluestein 2017; Stelten and Gallus 2017). Most storms initiating over the plains of eastern Colorado were found to be associated with convergence lines observed by radar (Wilson and Schreiber 1986). Reif and Bluestein (2017) documented that the nocturnal CI over the central Great Plains could occur on a surface boundary and on the cold side of a surface boundary, but could also occur without any nearby surface boundary. Huang et al. (2019) found that the sharp vegetation contrast over Hetao Irrigation District in China had a substantial impact on boundary activities and that 44% were associated with convection.
Synoptic-scale circulation often restricts the mesoscale environment for CI by affecting local thermal, dynamical, and moisture conditions (Brooks et al. 2018; Bai et al. 2021). Some synoptic circulation patterns, such as mei-yu front (He et al. 2018; Luo et al. 2018; Zhang et al. 2021), low-level jets (Reif and Bluestein 2017; Zhang et al. 2019; Du et al. 2020), the edge of the western Pacific subtropical high (WPSH) (Zhang et al. 2017; Liu et al. 2020; Zeng et al. 2022; Zhao et al. 2022), and tropical cyclones (TCs) (Chen et al. 2010; Meng and Wang 2016), may support the genesis of convection. Johns and Doswell (1992) concluded that there are three ingredients for CI: instability, moisture, and lifting mechanisms. Larger and more intense convective cells tend to occur in more unstable and humid low-level environmental conditions (Feng et al. 2022). The instability can be expressed in terms of convective available potential energy, which is a necessary condition for CI. Increasing moisture also increased the likelihood of CI by eliminating convective inhibition, lowering the level of free convection, and reducing the negative effects of dry entrainment on buoyancy (Bodine et al. 2010; Nelson et al. 2022). The deep convection initiation is generally suppressed in strong wind shear environment (Peters et al. 2022), however, this shear environment could govern the organization of updrafts and enables formation of severe storm (Smith et al. 2012). Generally, the environmental parameters can be considered as a predictor of initiation, and the mountain-induced or synoptic-scale lifting mechanism is also an important factor. Therefore, it is important for CI nowcasting to take into account the lifting mechanism and the complex local environmental conditions.
The middle reaches of the Yangtze River basin (YRB) is a typical climate used to study the climatology of CI owing to its diverse atmospheric convective environments and underlying surface conditions. As an economically developed region with dense population in East China, this area also suffers from frequent occurrence of severe convective weather. Some statistical studies of CI have been conducted in North China (Chen et al. 2012), central Eastern China (Huang et al. 2017), and South China (Bai et al. 2020a,b); however, few have focused on the middle reaches of the YRB in China. Additionally, some geostationary satellite-based CI studies have mainly relied on observations from GOES for North and South America, the Meteosat Second Generation for Europe, and Fengyun-2 (FY-2) for China. Here, we used the Chinese new-generation geostationary meteorological satellite Fengyun-4A (FY-4A), which has high spatial and temporal resolution to identify and track CIs.
In this study, we aimed to identify satellite-derived CIs using a hybrid cloud-object automatic tracking algorithm and to document the distribution and variability characteristics of CIs occurring during the warm season (May–September) of 2018–21. We also evaluated the dominant synoptic patterns and atmospheric environments for triggering convection over the middle reaches of YRB. Section 2 introduces the data and method. Section 3 presents spatial and temporal variations and cloud-top cooling rates of CIs and analyzes the duration and intensity of CI-related CCs. The CI-related composite circulations and atmospheric environments are described in section 4. A conclusion is given in section 5.
2. Data and method
a. Datasets
The FY-4A satellite, which was launched on 11 December 2016, carries four new instruments for monitoring rapidly changing weather systems and improving forecasting capabilities across a wide range of ocean, land, and atmosphere (Yang et al. 2017). In this study, blackbody brightness temperature (TBB) data from the FY-4A infrared channel centered at 10.8 μm of the Advanced Geosynchronous Radiation Imager (AGRI) was used to identify CIs. The spatial resolution of this data is 4 km at nadir, with 15 min for the full disk and 5 min for the China region. Because TBB data were available for dates after 12 March 2018, CIs over the study region (108°–119°E, 25.5°–35.7°N, the red box in Fig. 1) were identified during the warm seasons (May–September) from 2018 to 2021.
Topographic map (color shading; m) over the middle reaches of the Yangtze River basin (red box) and its vicinity. The provincial boundaries and rivers are marked by black and blue lines, respectively. The rivers, mountains, and plains are labeled in gray, red, and purple fonts, respectively.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
Additionally, the 0.25° × 0.25° 1-h ECMWF ERA5 reanalysis dataset was used to analyze the composite circulations and atmospheric environments associated with CIs. To identify landscape variations, the topography data were derived from the 1-arc-min global relief model of Earth’s surface, which integrates land topography and ocean bathymetry (ETOPO1) produced by the National Geophysical Data Center of the National Oceanic and Atmospheric Administration. To classify the synoptic circulation affected by TCs, the western North Pacific TC historical track dataset was used. The historical composite radar reflectivity data over middle reaches of the YRB were used to evaluate the validity of CIs identified by satellite.
b. Method for identifying satellite-derived CIs
Roberts and Rutledge (2003) found that satellite-observed cloud tops characterized by subfreezing temperatures and large cooling rates occurred 30 min before the first detection of 35-dBZ radar echoes. Then, Mecikalski and Bedka (2006) developed a satellite convection analysis and tracking (SATCAST) algorithm using a set of multispectral CI “interest field,” but it produced a high false alarm ratio (FAR) of ∼69% (Mecikalski et al. 2008). The cloud object-tracking method was then developed to reduce the FAR (Walker et al. 2012). Therefore, this study used the cloud object-tracking method to identify CIs which consisted of three main steps: 1) CC tracking, 2) primary CI selection, and 3) cloud-top cooling rate calculation and CI determination. These steps are described below and illustrated in Fig. 2.
Flowchart of the identification of CIs. The three steps are presented using dashed boxes.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
1) CC tracking
In this study, in order to better identify CIs, we screened CCs using a TBB threshold below 273 K at the 10.8-μm band observed by FY-4A AGRI. The 273-K criterion has been widely used in previous studies to define CIs (Huang et al. 2017; Liu et al. 2019). The area threshold was set to as small as 50 km2 (approximately 2 pixels). For capturing smaller CCs, Huang et al. (2018) and Chen et al. (2019) proposed a hybrid objective tracking algorithm combining the conventional area overlapping (AOL) with the Kalman filter (KF) method. In this study, we modified the threshold of overlapping rates described by Chen et al. (2019) to track CCs using FY-4A data with higher temporal resolution. The main flowchart can be found in the upper dashed boxes of Fig. 2.
Box-and-whisker plots of the overlapping rates of CCs from two consecutive AGRI/FY-4A images (at intervals of 5 and 10 min) over the middle reaches of the YRB from May to September of 2018. The upper and lower edges of each box represent the rate of overlapping (ROL) at 75th and 25th percentiles, respectively, the orange solid lines indicate the medians, and the red triangles denote mean values. The whiskers extend from the box by 3.0 × the interquartile range, and the blue circles are abnormal ROLs, with the total percentage of abnormal values labeled for each group of CC size. Red lines and associated values denote thresholds used to distinguish normal values and outliers. The black numbers on the top denote the numbers of samples used to calculate the box-and-whisker plots.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
The KF method handled the scenario when the AOL method failed to track CCs smaller than 1000 km2 and when their ROLs were lower than 19%. The KF method could robustly track small and fast-moving systems by estimating the speed and direction of the target systems (Reid 1979; Xing et al. 2009). The steps of the KF tracking method were as follows: 1) setting the initial velocity variable by averaging the motion vectors of all the surrounding CCs in a 5° × 5° grid box; 2) predicting the positions of CCs based on the five equations reported by Huang et al. (2018); 3) matching two CCs if the centroid of the predicted CC had the shortest distance with the centroid of next observed CC over time but not matching two CCs if the shortest distance was larger than a given threshold of 30 km (i.e., suppose the maximum moving speed of the CC was 120 km h−1), resulting in termination of the trajectory; and 4) updating the state variable based on two matched CCs for the next iteration. One uncertainty of the KF method stems from the estimated position of the CCs, the larger the CCs, the more difficult it is to accurately determine its exact position. Moreover, larger CCs have higher ROLs. Therefore, we combined the AOL method with the KF method to track multiscale CCs.
2) Selection of primary CIs
The CCs with an extensive area in which the TBB was less than or equal to 273 K are usually clouds that have evolved into a certain stage rather than immature cumulus clouds during the CI period. Therefore, we specified that the contiguous TBB ≤ 273 K area had to be no more than 5000 km2 (approximately 200 pixels). The purpose of this area threshold was to reduce the influence of uncertainty from the observed image, and CCs with areas larger than 5000 km2 were most likely to be mature convective cloud systems or other cloud types (Huang et al. 2017; Sun et al. 2019).
3) Cloud-top cooling rates and CI determination
To confirm the validity of this method, we examined the satellite-derived and radar-identified CIs by using the probability of detection (POD, the ratio of correctly match radar number to the total number of radar-identified isolated CIs) and FAR (the ratio of the total satellite-derived CIs number minus the matched radar number to the total satellite-derived CIs) every day from 15 June to 15 August 2020. A radar-identified CI is determined when and where the convective cell first reaches 40 dBZ without a nearby (within 60 km in this study) preexisting (within 30 min in this study) storm and storm anvil (Bai et al. 2020b). A total of 705 radar-identified CIs, and the total number of satellite-derived CIs was 741. The mean POD over the study region was 70.5%, and the mean FAR was 33.06% (Table S1 in the online supplemental material). This mean POD was significantly higher than the result of Huang et al. (2017) of 50.3% without using cloud object-tracking method and based on 30-min interval FY-2E data. The mean FAR was comparable with the result of Sieglaff et al. (2011) of 34.8% in the central United States. However, PODs could reach 80.0% in some days, especially, in high incidence days of CIs such as 15 June, 15 July, 31 July, and so on (Table S1). The lowest cloud-shield brightness temperature (L-TEMP) was obtained by tracking the whole lifetime of CCs after initiation, indicating the intensity of the CCs. In Table 1, the intense CCs (L-TEMP ≤ 241 K) had lower FAR, while the weak CCs (L-TEMP > 241 K) had higher FAR because they do not grow far enough to eventually produce rainfall. Therefore, we selected only those CIs that could reach the L-TEMP ≤ 241 K, which is the TBB threshold of mesoscale convective complexes (MCCs) defined by Maddox (1980).
FY-4A count and FAR statistics over the middle reaches of the YRB from 15 Jun to 15 Aug 2020 within different L-TEMP ranges.
Figure 4 further showed some CI events captured based on the hybrid objective tracking algorithm using some continuous TBB images. For case 1 (Figs. 4a1–a4), a CI event was detected in the southeast region of Hubei Province in China using five continuous TBB images within 30 min; the mean cooling rate from 1134 to 1200 local standard time (LST = UTC + 8 h) was 23.44 K (15 min)−1, and the CI-related CCs developed into severe convective storms (SCSs) at 1334 LST 20 August 2020. For case 2 (Figs. 4b1–b4) and case 3 (Figs. 4c1–c4), one or more CI events occurred in Henan and Jiangxi Provinces, respectively, and these CIs with higher cloud-top cooling rates [19.84 and 20.81 K (15 min)−1, respectively] could develop into SCSs. These three cases of CI events were consistent with radar-observed isolated storms at the CI time (Figs. 4a1,b1,c1), further demonstrating the effectiveness of the hybrid cloud-object tracking method used in this study to identify smaller CIs.
Typical examples of satellite-derived CIs over the middle reaches of the YRB for the years of (a1)–(a4) 2020, (b1)–(b4) 2019, and (c1)–(c4) 2018. Each column presents at least one signal (blue circles) identified by the hybrid cloud-object tracking method. Radar-identified CIs corresponding to the satellite-derived CIs in (a2), (b2), and (c2) are denoted by the blue circles in radar mosaics in (a1), (b1), and (c1). The names of relevant provinces are marked in the center of each province.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
c. Hierarchical clustering analysis
The hierarchical clustering analysis (HCL) is used to classify the synoptic circulation patterns associated with CIs. The hourly geopotential height at 850 and 500 hPa covering our study region and the nearby region (15°–55°N, 105°–135°E) were chosen as the classification objective because the configuration of synoptic systems at these levels are important for CI based on previous studies (He et al. 2018; Luo et al. 2018; Zhang et al. 2017; Liu et al. 2020). The layers below 850 hPa are not considered as these levels may be influenced by the terrain elevation because the average altitude of terrain is ∼1200 m. This method starts with each cluster containing only one object, and merges the two most similar clusters with the smallest Ward’s distance until all objects are included in a single cluster (Ran et al. 2023). It was terminated objectively at a step where a significant “jump” in the smallest Ward’s distance (which is used to determine which two clusters should be merged at each step) occurred (Zhao et al. 2016, 2017a,b; Hu et al. 2019a,b). This is therefore the step when the actual merging procedure should be terminated. Finally, we documented the total number of clusters at this termination step.
This objective classification method is further investigated by comparing the other two methods, including PCT (obliquely rotated principal component analysis in T-mode) and SOM (self-organizing maps). Both HCL and SOM methods performed better separability between clusters than PCT, and HCL showed better consistency than SOM in Fig. S1 in the supplemental material. Based on the result, HCL method was the optimal choice in the current study.
3. Statistical characteristics of CIs
a. Spatial and temporal variations in CIs
Figure 5a presents the spatial distribution of CI numbers at the whole day identified using the above-described CI identification method. In total, 5215 CIs were identified in the study region from May to September of 2018–21. CIs primarily occurred over the southern part of the study region and the high-frequency regions (numbers > 12) associated with the mountains. Five notable high-frequency areas of CIs were found: Dabie Mountains (A), Mulianjiu Mountains (B), Xuefeng Mountains (C), Luoxiao Mountains (D), and Wuyi Mountains (E) (the names of the mountains were given in Fig. 1). In contrast to the high-frequency regions, the Jianghan, Jianghuai, and Poyang Lake plains (the names of the plains are given in Fig. 1) showed lowest CI frequencies. Figures 5b–e further show the spatial distributions of CI numbers at diurnal different periods in the study region from May to September of 2018–21; 3673 CIs during noon–afternoon (1100–1659 LST), accounting for about 72.2% of all CIs, were located in complex terrain regions that were induced by the diurnal cycle of local solar heating. The CIs during afternoon–nighttime (1700–2259 LST) dropped abruptly to 899, and their high-frequency areas moved to the foothills. The CIs during nighttime–morning (2300–1059 LST) continued to decrease, and their high-frequency areas were located in the Jianghan and Poyang Lake plains.
The 0.25° × 0.25° mosaic plots of CI numbers for (a) the whole day, (b) 1100–1659 LST, (c) 1700–2259 LST, (d) 2300–0459 LST, and (e) 0500–1059 LST over the middle reaches of the YRB from May to September of 2018–21. The black solid lines indicate the terrain at an altitude of 250 m. The high-frequency convection initiation regions, i.e., Dabie Mountains, Mulianjiu Mountains, Xuefeng Mountains, Luoxiao Mountains, and Wuyi Mountains, are labeled with A, B, C, D, and E, respectively.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
Figure 6 shows annual from May to September, monthly, and diurnal variations in CIs. The maximum total CI events occurred in 2018, accounting for about 39.6% of all CIs, and the minimum occurrence was in 2020, totaling almost half of that in 2018 (Fig. 6d). The low occurrence of CIs in 2020 may have been related to the extremely long duration of the mei-yu season from 1 June to 2 August (Ding et al. 2021), resulting in fewer instances of isolated convection over the study region. Analysis of monthly variations in CIs showed that the frequency was highest in August, followed by July, whereas that in May was the lowest (Fig. 6c). Huang et al. (2017) also found that satellite-derived CIs outbroke extensively over the whole central eastern China region in July and August. Diurnal variations in CIs throughout the study region exhibited a unimodal structure, with remarkable peak at noon (1200–1259 LST) (Fig. 6a). Nearly 81% of CIs were concentrated during noon–afternoon (1100–1859 LST), whereas the minority appeared at nighttime (Fig. 6a).
(a) Histogram of diurnal variation, (b) month–hour heatmap (shaded; numbers labeled on every mosaic), (c) histogram of monthly variability, and (d) histogram of annual variability in CI numbers over the middle reaches of the YRB from May to September of 2018–21.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
To further analyze temporal variations in CIs over different landscapes (i.e., plains and mountains), Fig. 7 shows the monthly and diurnal variations in CIs according to the elevation of the CI location. The results showed that the CI events in June were located over lower-elevation regions compared with those in other months (Fig. 7a). In contrast, the CIs forming in July and August occurred in regions with the higher elevation. This phenomenon was caused by the proportion of CIs during the noon–afternoon period was higher in July and August, 79.1% and 78.6%, respectively, whereas lowest in June of 65.7% (Fig. 6b). For the diurnal variations in Fig. 7b, CIs were most likely to occur at higher elevations (i.e., in the mountains) during noon–afternoon (1100–1859 LST) or at lower elevations (i.e., plains) during nighttime–morning (2000–1059 LST). Overall, CIs occurred frequently in the mountains in midsummer and during noon–afternoon, whereas the frequency of CIs in the plains was relatively low, peaking in June and during nighttime–morning.
Box-and-whisker plots of the (a) monthly and (b) diurnal variations in CI location elevations over the middle reaches of the YRB from May to September 2018–21. The box covers the 25th–75th percentiles, the horizontal lines in the boxes mark the medians, the triangles denote the mean values, the whiskers extend from the box by 1.5 × the interquartile range, and the black circles are those past the ends of the whiskers.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
b. Cooling rates of CIs
Roberts and Rutledge (2003) found that the largest values of cloud-top cooling rates corresponded to periods of intense cloud growth and were important indicators for discriminating between weakly precipitating storms (<35 dBZ) and vigorous convective storms (>35 dBZ). Several unique signatures, including rapid cloud-top cooling, overshooting tops, and above-anvil cirrus plumes, have been identified within satellite imagery of SCSs tops, and in particular, rapid cloud-top cooling is a widely used signature for forecasting SCSs (Dworak et al. 2012; Cintineo et al. 2013). In the current study, we calculated the average cloud-top cooling rates within 30 min and evaluated their relationships with the occurrences of SCSs. Figure 8 shows monthly and diurnal variations in average cooling rates within 30 min (
Box-and-whisker plots of the (a) monthly and (b) diurnal variations in average cooling rates within 30 min over the middle reaches of the YRB from May to September 2018–21. The lines, triangles, and circles have the same meanings as in Fig. 7.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
c. Duration and intensity of CI-related CCs
To further evaluate the development of the identified CI events, we analyzed the duration and intensity of CCs after triggering. Here, we used the CI-related trajectory described in section 2b. The duration of CCs was counted until dissipation or merging with other organized and large convective systems, and the lowest cloud-shield brightness temperature (L-TEMP) and maximum cloud-shield area (M-AREA) were calculated simultaneously during the duration. The L-TEMP indicated the intensity of the CCs, and TBBs lower than 221 K presented the best correlations with intensive convective rainfall (Goyens et al. 2012; Ai et al. 2016). Figure 9 presents the statistical characteristics of duration, L-TEMP, and M-AREA in the study region from May to September of 2018–21. More CCs for approximately 26% of the total CI events lasted for 30–60 min, the second peak is 120–180 min, and then the percent showed a decreasing trend as duration increased. The median duration of CCs was 173 min, with only 10 exceeding 900 min. The median duration of CCs over YRB was only half that of convective systems in South China (Bai et al. 2020b), which may be related to the greater moisture content over South China or the difference in tracking algorithms (Bai et al. 2021). L-TEMP occurred most often (35.7%) between 231 and 241 K and least often (7.0%) less than 201 K, and an L-TEMP less than 221 K accounted for a high percentage of cases (36.88%). Moreover, most CCs had an M-AREA between 1 and 2 × 103 km2, accounting for 27.5%, whereas 79 CCs reached an M-AREA greater than 50 × 103 km2. Overall, CCs with a duration of 30–60 min, L-TEMP between 231 and 241 K, and M-AREA between 1 and 2 × 103 km2 occurred most often, whereas 36.88% of CCs showed intense convection.
Histogram and line chart for (a) duration within different time ranges, (b) L-TEMP within different minimum temperature ranges, and (c) M-AREA within different maximum area ranges of CI-related CCs. The blue bars denote the numbers of CI-related CCs, and the red dots and curves represent the percent of CI-related CCs in every bar.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
4. Synoptic circulation patterns for noon–afternoon CIs
Because up to 81% of CIs occurred during noon–afternoon (1100–1859 LST) and there was a unimodal structure at noon, we focused on the noon–afternoon CIs in this section. The CI hours are defined as the previous hour when there was at least one CI occurred. In total, we got 1591 h for all 420 days from May to September of 2018–21. First, we used the quantitative criteria from Ren et al. (2006) and an TC historical best track dataset to identify TC-influenced CIs by selecting CIs for which the distance between the TC center and the CI location was 500–1100 km away from the TC center. There are 430 h affected by TCs, accounting for 27.02% of the total CI hours and 23.18% of all the noon–afternoon CI events. The TC pattern was not analyzed because of its diverse scales and paths, and it often interacted with the westerly trough or mei-yu front (Chen et al. 2010), making it difficult to clarify the dominant synoptic circulation on CIs. Then, we used the HCL method to classify the synoptic circulation patterns on 1161 selected hours without TCs. The HCL method is used to classify the synoptic circulation patterns associated with noon–afternoon CIs, four notable different synoptic circulations are obtained (Fig. 10). Four main synoptic circulations were also identified over Eastern China during summer precipitation in the previous study (Yang et al. 2021).
The composite synoptic circulation for the four patterns. The geopotential height (blue lines; dagpm) at 850 hPa and 588 dagpm (purple lines) at 500 hPa, wind field (wind barbs; 4 m s−1), and wind speed (color shading; m s−1) at 850 hPa. The red boxes denote the study region.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
Based on the main wind field at 850 hPa of the four patterns (Figs. 10a–d), we named them southwesterly flows (SW-Flows), southerly flows (S-Flows), westerly flows (W-Flows), and northwesterly flows (NW-Flows). The classification results (Fig. 11) showed that SW-Flows and S-Flows patterns accounting 37% and 36%, respectively, were the dominant circulations. The percentages of CI occurrence were 34% and 47% for SW-Flows and S-Flows, respectively, and S-Flows had the highest occurrence frequency (3.04 CIs per hour), followed by SW-Flows (2.09 CIs per hour), and the other two patterns were the lowest. The four patterns were related to seasonal movement of WPSH (purple lines in Fig. 10) from early to late summer, W-Flows primarily occurred in May, SW-Flows in June and July, S-Flows in July and August, and NW-Flows in September (Table 2). SW-Flows was the typical circulation of the mei-yu front, with the WPSH in the southeast of the study region and a strong southwesterly producing a maximum wind belt greater than 6 m s−1, whereas the S-Flows was under the control of the WPSH and weaker southerly winds affected the entire study region.
Histogram for the proportion of circulation samples and CI samples of the four patterns, and numbers on the top of each bar indicate the percentage of each pattern.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
Numbers of CI hours corresponding to the four synoptic circulation patterns in each month.
To investigate the thermodynamic and dynamic condition for CI hours, respectively. We calculated the most unstable convective available potential energy (MUCAPE) and the 0–3 km above ground level (AGL) bulk wind shear (SHR3) at the grids of all CI hours using ERA5 datasets and compared their differences for each circulation. According to the distribution of environmental parameters in Fig. 12, the average MUCAPE in the mountains were higher than that over flat regions, especially the higher-altitude Ta-pa and Wushan Mountains (the names were given in Fig. 1) possessed more MUCAPE, which could be attribute to the increased solar heating (Rasmussen and Houze 2016). Notably, the spatial distribution of higher MUCAPE in the mountains correlated well with the five high-frequency areas of CIs, indicating that thermodynamic conditions could be the main reason for the noon–afternoon CIs. For the two dominant circulation patterns, the higher MUCAPE areas were considerably broader than the other two patterns, which resulted in more CIs. For the dynamic condition, the strong SHR3 areas had distinct northward-movement characteristics from the W-Flows, SW-Flows, to S-Flows pattern, resulting in the S-Flows pattern having relatively weaker SHR3 in the southeastern mountains.
The average MUCAPE (shading; J kg−1) and SHR3 (vectors; m s−1) for the four patterns. The gray solid lines indicate the terrain at an altitude of 250 m. The purple dots in each panel denote the location of CIs.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
Figure 13 further showed the south–north variation of environmental parameters for the four patterns in the study region. The latitudinal average of the MUCAPE exhibited remarkable north–south regional differences, and the MUCAPE in the southern part was significantly higher than that of northern part. The hourly frequency of CI showed a decreasing trend from the southern to the northern part as MUCAPE decreased. The SW-Flows and S-Flows patterns had higher MUCAPE throughout the entire study region than the other two patterns, and their MUCAPE exceeded 1500 J kg−1 could move northward to 32.5°N. The northward movement characteristics of the strong SHR3 areas were further confirmed by the maximum value of the SHR3 curves, with the maximum value located in the southern part (27.5°–28.5°N), the central part (30.5°–31.5°N), and the northern part (32.5°–33.5°N) for W-Flows, SW-Flows, and S-Flows pattern, respectively. Although the strong SHR3 areas moved among the three patterns, the high-frequency areas of CIs did not correspond to a high shear, whereas correlated well with the weaker shear. Some studies (Peters et al. 2019; Nelson et al. 2022) indicate that CI is more inhibited when shear is relatively strong, compared to when shear is weak. It should also be noted that the SHR3 had the lowest of 4.12 m s−1 in the S-Flows pattern (Fig. 13b), with the highest frequency of CIs, confirming that the weaker shear environment increased the frequency of CIs.
Latitudinal average of the MUCAPE (red dotted lines; J kg−1), SHR3 (blue dotted lines; m s−1), and hourly frequency of CI (green dotted lines) for the four patterns. The black dashed lines denote the north–south division of the study region.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
The local environmental conditions at the CI points between the CI hour and other times at the same hour of the day but without CI (non-CI hour) in each circulation pattern was also investigated. The CI points at non-CI hour chosen in this study should satisfy two criteria: 1) no CI events occurred within a 100-km radius at the same CI hour of the other days to avoid repeated CI event at the same time and location, 2) no CCs were covered in the CI points to avoid the influence of the convection. The corresponding statistics are shown in Fig. 14. For the CI hours, the mean values of MUCAPE in the two dominant patterns were higher than the other two patterns, indicating a highly unstable atmospheric environment. The mean values of the lifted index (LI) in these two dominant patterns were both −4.8 K, lower than the other patterns, which confirms this conclusion. The mean SHR3 and 0–6 km AGL bulk wind shear (SHR6) in S-Flows pattern were both significantly lower than the other three patterns, and the mean SHR3 and SHR6 in W-Flows pattern were the highest of the four patterns. Comparing CI hours and non-CI hours in Figs. 14a and 14b, it was found that the mean values of MUCAPE (LI) at CI hours were significantly higher (lower) than those of non-CI hours in each pattern, and more than 50% of non-CI hours were concentrated in the environment with MUCAPE below 1300 J kg−1, but more than 50% of CI hours were higher than 1900 J kg−1, indicating a more thermally unstable environment at CI hours. However, the mean values of SHR3 and SHR6 at CI hours were equal to or lower than those of non-CI hours in each pattern (Figs. 14c,d), indicating a weaker dynamic environment at CI hours. Therefore, the local environmental conditions at the points between the CI hours and non-CI hours suggested that noon–afternoon CIs were more controlled by thermally unstable environments.
Violin plots of the (a) MUCAPE (J kg−1), (b) LI (K), (c) SHR3 (m s−1), and (d) SHR6 (m s−1) for the four patterns. The pink (blue) shaded color of the violin represents the environmental parameters at the CI points of the CI (non-CI) hours. The outline of the violin represents the probability density. The upper and lower edges and the whiskers of the box inside the violin have the same meaning as in Fig. 7. The green line (white dot) inside the box represents the mean value (median), and the mean value is marked next to it.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
To investigate the moisture condition that contributed to CIs under different circulation patterns, we have calculated the vertically integrated moisture flux convergence (VIMFC) from surface to 300 hPa. Figure 15 showed that a remarked moisture flux convergence (negative VIMFC) occurred in the mountain regions for the four patterns, indicating that upslope flows near mountains may aid in the moisture pooling. Higher-altitude Ta-pa and Wushan Mountains had fewer CIs due to their weaker moisture convergences compared to the eastern mountains, although their average MUCAPEs were higher. The moisture flux convergences were stronger for SW-Flows and W-Flows patterns due to their stronger low-level jets for transporting moisture. The vertically integrated moisture flux (VIMF) also presented that the moisture was mainly from the southern part of the study region for SW-Flows and W-Flows patterns, whereas the moisture was localized for the S-Flows pattern. Note that CIs under the S-Flows pattern had narrow moisture pooling, but their local moisture convergences were stronger.
The average of the vertically integrated moisture flux convergence (VIMFC; shading; 10−4 kg m−2 s−1), and the vertically integrated moisture flux (VIMF; vectors; kg m−1 s−1) for the four patterns. The gray solid lines indicate the terrain at an altitude of 250 m. The purple dots in each panel denote the location of CIs.
Citation: Journal of Hydrometeorology 25, 3; 10.1175/JHM-D-23-0157.1
5. Conclusions
A high spatial and temporal resolution TBB dataset from the Chinese new-generation geostationary meteorological satellite FY-4A was used to recognize and track CIs over the middle reaches of the YRB from May to September of 2018–21. These CIs were identified using a hybrid objective tracking algorithm combining the conventional AOL with the KF method. Then, CI datasets were used to analyze the spatial and temporal variations and cloud-top cooling rates of CIs. To evaluate the development of the identified CI events, we further analyzed the duration and intensity of CI-related CCs. Finally, 1161 h without TCs occurring during noon–afternoon were selected and classified into four synoptic patterns, and their composite circulations and atmospheric environments were analyzed using ERA5 reanalysis data. The main findings are as follows.
In total, 5215 CIs were identified in the study region from May to September of 2018–21. Analysis of monthly variations showed that the frequency of CIs was highest in August and lowest in May. Nearly 81% of CIs occurred at noon–afternoon (1100–1859 LST) in the mountains, with the highest frequency in the southern mountains of the study region. Five high-frequency areas were located: Dabie Mountains, Mulianjiu Mountains, Xuefeng Mountains, Luoxiao Mountains, and Wuyi Mountains. By contrast, the frequency of CIs in the plains was relatively low. The diurnal variation of CIs throughout the study region showed a unimodal structure, with a peak appearing at noon (1200–1259 LST). The CIs during noon–afternoon were mainly located in the mountains, whereas the high-frequency areas moved to the plains from afternoon to morning.
The CIs in July and August had faster cooling rates than those in May and June, and CIs at noon–afternoon had faster cooling rates than those at nighttime–morning. Most CI-related CCs had duration of 30–60 min, a cloud brightness temperature between 231 and 241 K, and a cloud area between 1 and 2 × 103 km2; however, 36.88% of CCs showed intense convection.
The synoptic circulation during noon–afternoon CI hours, without the influence of TCs, was classified into four patterns using the HCL algorithm. The SW-Flows and S-Flows were the dominant circulation patterns, consistent with the typical circulation of the mei-yu front and the weak southerly under the control of the WPSH, respectively. The S-Flows pattern had the highest occurrence frequency of CIs, followed by the SW-Flows pattern, and the other two patterns were the lowest.
Analysis of the environmental conditions for CI hours in the four patterns suggests that 1) two dominant patterns had broader areas of higher MUCAPE, and stronger SHR3 areas had distinct northward movement characteristics from the W-Flows, SW-Flows, to S-Flows pattern, whereas the SHR3 was the weakest in the S-Flows pattern; 2) the high-frequency areas of CIs were most likely to occur in stronger MUCAPE and weaker SHR3 environments. Comparing the environmental conditions of the CI points between the CI hours and the non-CI hours, CIs were more controlled by thermally unstable environments. Meanwhile, CIs tend to occur in more unstable and moisture flux convergence areas.
The current study primarily focused on the synoptic circulation patterns and atmospheric environments of noon–afternoon CIs over the middle reaches of the YRB. Future investigation efforts should be directed toward evaluation of nocturnal convection.
Acknowledgments.
This work was supported by the National Natural Science Foundation of China (Grants U2142202, 42105012, 41975058, 42230612, and 41975057), the Open Project Fund of China Meteorological Administration Basin Heavy Rainfall Key Laboratory (Grant 2023BHR-Z01), the Key Scientific and Technological Project of Hubei Meteorological Bureau (Grants 2022Z02 and 2022Q06), the Fengyun Application Pioneering Project (Grant FY-APP-2022.0104), the Natural Science Foundation of Hubei Province (Grant 2022CFB025), and the Joint supported by Hubei Provincial Natural Science Foundation and the Meteorological Innovation and Development Project of China (2023AFD096). We thank Jianping Guo at the Chinese Academy of Meteorological Sciences for his suggestions on the hybrid objective-tracking algorithm combining the conventional area overlapping (AOL) method with the Kalman filter (KF) method used in this paper.
Data availability statement.
The data used in this manuscript were obtained from publicly available data repositories: blackbody brightness temperature data at the 10.8-μm band observed by FY-4A AGRI were obtained from the National Satellite Meteorological Center (NSMC) of China website at http://satellite.nsmc.org.cn/. ERA5 data were obtained from the Climate Data Store (https://cds.climate.copernicus.eu/cdsapp#!/home). ETOPO1 data were available at https://www.ncei.noaa.gov/access/metadata/landing-page/bin/iso?id=gov.noaa.ngdc.mgg.dem:316#Description. The western North Pacific (WNP) TC historical track dataset was obtained from the Shanghai Typhoon Institute of the China Meteorological Administration (https://tcdata.typhoon.org.cn/zjljsjj_zlhq.html). The historical composite radar reflectivity data over middle reaches of the YRB were provided by the National Meteorological Information Center of the China Meteorological Administration (registration required).
REFERENCES
Ai, Y., W. Li, Z. Meng, and J. Li, 2016: Life cycle characteristics of MCSs in middle east China tracked by geostationary satellite and precipitation estimates. Mon. Wea. Rev., 144, 2517–2530, https://doi.org/10.1175/MWR-D-15-0197.1.
Aoshima, F., A. Behrendt, H.-S. Behrendt, and V. Wulfmeyer, 2008: Statistics of convection initiation by use of Meteosat rapid scan data during the Convective and Orographically-Induced Precipitation Study (COPS). Meteor. Z., 17, 921–930, https://doi.org/10.1127/0941-2948/2008/0337.
Bai, L., G. Chen, and L. Huang, 2020a: Convection initiation in monsoon coastal areas (South China). Geophys. Res. Lett., 47, e2020GL087035, https://doi.org/10.1029/2020GL087035.
Bai, L., G. Chen, and L. Huang, 2020b: Image processing of radar mosaics for the climatology of convection initiation in South China. J. Appl. Meteor. Climatol., 59, 65–81, https://doi.org/10.1175/JAMC-D-19-0081.1.
Bai, L., G. Chen, Y. Huang, and Z. Meng, 2021: Convection initiation at a coastal rainfall hotpot in South China: Synoptic patterns and orographic effects. J. Geophys. Res. Atmos., 126, e2021JD034642, https://doi.org/10.1029/2021JD034642.
Banacos, P. C., and D. M. Schultz, 2005: The use of moisture flux convergence in forecasting convective initiation: Historical and operational perspectives. Wea. Forecasting, 20, 351–366, https://doi.org/10.1175/WAF858.1.
Banta, R. M., and C. B. Schaaf, 1987: Thunderstorm genesis zones in the Colorado Rocky Mountains as determined by traceback of geosynchronous satellite images. Mon. Wea. Rev., 115, 463–476, https://doi.org/10.1175/1520-0493(1987)115<0463:TGZITC>2.0.CO;2.
Bodine, D., P. L. Heinselman, B. L. Heinselman, R. D. Palmer, and D. Michaud, 2010: A case study on the impact of moisture variability on convection initiation using radar refractivity retrievals. J. Appl. Meteor. Climatol., 49, 1766–1778, https://doi.org/10.1175/2010JAMC2360.1.
Brooks, H. E., and Coauthors, 2018: A century of progress in severe convective storm research and forecasting. A Century of Progress in Atmospheric and Related Sciences: Celebrating the American Meteorological Society Centennial, Meteor. Monogr., No. 59, Amer. Meteor. Soc., https://doi.org/10.1175/AMSMONOGRAPHS-D-18-0026.1.
Browning, K. A., and D. Atlas, 1965: Initiation of precipitation in vigorous convective clouds. J. Atmos. Sci., 22, 678–683, https://doi.org/10.1175/1520-0469(1965)022<0678:IOPIVC>2.0.CO;2.
Chen, D. D., and Coauthors, 2019: Mesoscale convective systems in the Asian monsoon region from Advanced Himawari Imager: Algorithms and preliminary results. J. Geophys. Res. Atmos., 124, 2210–2234, https://doi.org/10.1029/2018JD029707.
Chen, L., Y. Li, and Z. Cheng, 2010: An overview of research and forecasting on rainfall associated with landfalling tropical cyclones. Adv. Atmos. Sci., 27, 967–976, https://doi.org/10.1007/s00376-010-8171-y.
Chen, M., Y. Wang, F. Gao, and X. Xiao, 2012: Diurnal variations in convective storm activity over contiguous North China during the warm season based on radar mosaic climatology. J. Geophys. Res., 117, D20115, https://doi.org/10.1029/2012JD018158.
Choudhury, H. P., P. Roy, S. Kalita, and S. Sharma, 2016: Spatio-temporal variability of the properties of mesoscale convective systems over a complex terrain as observed by TRMM sensors. Int. J. Climatol., 36, 2615–2632, https://doi.org/10.1002/joc.4516.
Cintineo, J. L., M. J. Pavolonis, J. M. Sieglaff, and A. K. Heidinger, 2013: Evolution of severe and nonsevere convection inferred from GOES-derived cloud properties. J. Appl. Meteor. Climatol., 52, 2009–2023, https://doi.org/10.1175/JAMC-D-12-0330.1.
Ding, Y., Y. Liu, and Z.-Z. Hu, 2021: The record-breaking Meiyu in 2020 and associated atmospheric circulation and tropical SST anomalies. Adv. Atmos. Sci., 38, 1980–1993, https://doi.org/10.1007/s00376-021-0361-2.
Du, Y., G. Chen, B. Han, C. Mai, L. Bai, and M. Li, 2020: Convection initiation and growth at the coast of South China. Part I: Effect of the marine boundary layer jet. Mon. Wea. Rev., 148, 3847–3869, https://doi.org/10.1175/MWR-D-20-0089.1.
Dworak, R., K. Bedka, J. Brunner, and W. Feltz, 2012: Comparison between GOES-12 overshooting-top detections, WSR-88D radar reflectivity, and severe storm reports. Wea. Forecasting, 27, 684–699, https://doi.org/10.1175/WAF-D-11-00070.1.
Feng, Z., A. Varble, J. Hardin, J. Marquis, A. Hunzinger, Z. Zhang, and M. Thieman, 2022: Deep convection initiation, growth, and environments in the complex terrain of central Argentina during CACTI. Mon. Wea. Rev., 150, 1135–1155, https://doi.org/10.1175/MWR-D-21-0237.1.
Goyens, C., D. Lauwaet, M. Schröder, M. Demuzere, and N. P. M. Van Lipzig, 2012: Tracking mesoscale convective systems in the Sahel: Relation between cloud parameters and precipitation. Int. J. Climatol., 32, 1921–1934, https://doi.org/10.1002/joc.2407.
He, Z., Q. Zhang, K. Zhao, and H. Hu, 2018: Initiation and evolution of elevated convection in a nocturnal squall line along the Meiyu Front. J. Geophys. Res. Atmos., 123, 7292–7310, https://doi.org/10.1029/2018JD028511.
Hu, Y., Y. Deng, Z. Zhou, C. Cui, and X. Dong, 2019a: A statistical and dynamical characterization of large-scale circulation patterns associated with summer extreme precipitation over the middle reached of Yangtze River. Climate Dyn., 52, 6213–6228, https://doi.org/10.1007/s00382-018-4501-z.
Hu, Y., Y. Deng, Z. Zhou, H. Li, C. Cui, and X. Dong, 2019b: A synoptic assessment of the summer extreme rainfall over the middle reaches of Yangtze River in CMIP5 models. Climate Dyn., 53, 2133–2146, https://doi.org/10.1007/s00382-019-04803-3.
Huang, X., C. Hu, X. Huang, Y. Chu, Y.-h. Tseng, G. J. Zhang, and Y. Lin, 2018: A long-term tropical mesoscale convective systems dataset based on a novel objective automatic tracking algorithm. Climate Dyn., 51, 3145–3159, https://doi.org/10.1007/s00382-018-4071-0.
Huang, Y., Z. Meng, J. Li, W. Li, L. Bai, M. Zhang, and X. Wang, 2017: Distribution and variability of satellite-derived signals of isolated convection initiation events over central eastern China. J. Geophys. Res. Atmos., 122, 11 357–11 373, https://doi.org/10.1002/2017JD026946.
Huang, Y., Z. Meng, W. Li, L. Bai, and X. Meng, 2019: General features of radar-observed boundary layer convergence lines and their associated convection over a sharp vegetation-contrast area. Geophys. Res. Lett., 46, 2865–2873, https://doi.org/10.1029/2018GL081714.
Johns, R. H., and C. A. Doswell III, 1992: Severe local storms forecasting. Wea. Forecasting, 7, 588–612, https://doi.org/10.1175/1520-0434(1992)007<0588:SLSF>2.0.CO;2.
Liu, J., J. Zhang, F. Liu, and L. Yang, 2020: Drivers of a sudden mesoscale rainstorm in arid and semi-arid regions at the edge of the western Pacific subtropical high. Meteor. Appl., 27, e1884, https://doi.org/10.1002/met.1884.
Liu, Z., and Coauthors, 2019: Local severe storm tracking and warning in pre-convection stage from the new generation geostationary weather satellite measurements. Remote Sens., 11, 383, https://doi.org/10.3390/rs11040383.
Luo, X., M. Xue, and J. Fei, 2018: Simulation and analysis of the initiation of a squall line within a Meiyu frontal system in East China. Atmosphere, 9, 183, https://doi.org/10.3390/atmos9050183.
Madaus, L. E., and G. J. Hakim, 2016: Observable surface anomalies preceding simulated isolated convective initiation. Mon. Wea. Rev., 144, 2265–2284, https://doi.org/10.1175/MWR-D-15-0332.1.
Maddox, R. A., 1980: Mesoscale convective complexes. Bull. Amer. Meteor. Soc., 61, 1374–1400, https://doi.org/10.1175/1520-0477(1980)061<1374:MCC>2.0.CO;2.
Marshall, J. S., and S. Radhakant, 1978: Radar precipitation maps as lightning indicators. J. Appl. Meteor., 17, 206–212, https://doi.org/10.1175/1520-0450(1978)017<0206:RPMALI>2.0.CO;2.
Mecikalski, J. R., and K. M. Bedka, 2006: Forecasting convective initiation by monitoring the evolution of moving cumulus in daytime GOES imagery. Mon. Wea. Rev., 134, 49–78, https://doi.org/10.1175/MWR3062.1.
Mecikalski, J. R., K. M. Bedka, S. J. Paech, and L. A. Litten, 2008: A statistical evaluation of GOES cloud-top properties for predicting convective initiation. Mon. Wea. Rev., 136, 4899–4914, https://doi.org/10.1175/2008MWR2352.1.
Mecikalski, J. R., W. M. MacKenzie Jr., M. Koenig, and S. Muller, 2010: Cloud-top properties of growing cumulus prior to convective initiation as measured by Meteosat Second Generation. Part I: Infrared fields. J. Appl. Meteor. Climatol., 49, 521–534, https://doi.org/10.1175/2009JAMC2344.1.
Meng, W., and Y. Wang, 2016: A diagnostic study on heavy rainfall induced by landfalling Typhoon Utor (2013) in South China: 2. Postlandfall rainfall. J. Geophys. Res. Atmos., 121, 12 803–12 819, https://doi.org/10.1002/2015JD024647.
Mercer, A. E., and M. B. Richman, 2007: Statistical differences of quasigeostrophic variables, stability, and moisture profiles in North American storm tracks. Mon. Wea. Rev., 135, 2312–2338, https://doi.org/10.1175/MWR3395.1.
Nelson, T. C., J. Marquis, J. M. Peters, and K. Friedrich, 2022: Environmental controls on simulated deep moist convection initiation occurring during RELAMPAGO-CACTI. J. Atmos. Sci., 79, 1941–1964, https://doi.org/10.1175/JAS-D-21-0226.1.
Peters, J. M., W. Hannah, and H. Morrison, 2019: The influence of vertical wind shear on moist thermals. J. Atmos. Sci., 76, 1645–1659, https://doi.org/10.1175/JAS-D-18-0296.1.
Peters, J. M., H. Morrison, T. C. Nelson, J. N. Marquis, J. P. Mulholland, and C. J. Nowotarski, 2022: The influence of shear on deep convection initiation. Part I: Theory. J. Atmos. Sci., 79, 1669–1690, https://doi.org/10.1175/JAS-D-21-0145.1.
Ran, X., Y. Xi, Y. Lu, X. Wang, and Z. Lu, 2023: Comprehensive survey on hierarchical clustering algorithms and the recent developments. Artif. Intell. Rev., 56, 8219–8264, https://doi.org/10.1007/s10462-022-10366-3.
Rasmussen, K. L., and R. A. Houze Jr., 2016: Convective initiation near the Andes in subtropical South America. Mon. Wea. Rev., 144, 2351–2374, https://doi.org/10.1175/MWR-D-15-0058.1.
Reid, D., 1979: An algorithm for tracking multiple targets. IEEE Trans. Automat. Contr., 24, 843–854, https://doi.org/10.1109/TAC.1979.1102177.
Reif, D. W., and H. B. Bluestein, 2017: A 20-year climatology of nocturnal convection initiation over the central and southern Great Plain during the warm season. Mon. Wea. Rev., 145, 1615–1639, https://doi.org/10.1175/MWR-D-16-0340.1.
Ren, F., G. Wu, W. Dong, X. Wang, Y. Wang, W. Ai, and W. Li, 2006: Changes in tropical cyclone precipitation over China. Geophys. Res. Lett., 33, L20702, https://doi.org/10.1029/2006GL027951.
Roberts, R. D., and S. Rutledge, 2003: Nowcasting storm initiation and growth using GOES-8 and WSR-88D data. Wea. Forecasting, 18, 562–584, https://doi.org/10.1175/1520-0434(2003)018<0562:NSIAGU>2.0.CO;2.
Sieglaff, J. M., L. M. Cronce, W. F. Feltz, K. M. Bedka, M. J. Pavolonis, and A. K. Heidinger, 2011: Nowcasting convective storm initiation using satellite-based box-averaged cloud-top cooling and cloud-type trends. J. Appl. Meteor. Climatol., 50, 110–126, https://doi.org/10.1175/2010JAMC2496.1.
Smith, B. T., R. L. Thompson, J. S. Grams, C. Broyles, and H. E. Brooks, 2012: Convective modes for significant severe thunderstorms in the contiguous United States. Part I: Storm classification and climatology. Wea. Forecasting, 27, 1114–1135, https://doi.org/10.1175/WAF-D-11-00115.1.
Stelten, S., and W. A. Gallus Jr., 2017: Pristine nocturnal convective initiation: A climatology and preliminary examination of predictability. Wea. Forecasting, 32, 1613–1635, https://doi.org/10.1175/WAF-D-16-0222.1.
Sun, F., D. Qin, M. Min, B. Li, and F. Wang, 2019: Convective initiation nowcasting over China from Fengyun-4A measurements based on TV-L1 optical flow and BP_Adaboost neural network algorithms. IEEE J. Sel. Top. Appl. Earth Obs. Remote Sens., 12, 4284–4296, https://doi.org/10.1109/JSTARS.2019.2952976.
Sun, J., and Coauthors, 2014: Use of NWP for nowcasting convective precipitation: Recent progress and challenges. Bull. Amer. Meteor. Soc., 95, 409–426, https://doi.org/10.1175/BAMS-D-11-00263.1.
Trentmann, J., and Coauthors, 2009: Multi-model simulations of a convective situation in low-mountain terrain in central Europe. Meteor. Atmos. Phys., 103, 95–103, https://doi.org/10.1007/s00703-008-0323-6.
Walker, J. R., W. M. MacKenzie Jr., J. R. Mecikalski, and C. P. Jewett, 2012: An enhanced geostationary satellite-based convective initiation algorithm for 0–2 h nowcasting with object tracking. J. Appl. Meteor. Climatol., 51, 1931–1949, https://doi.org/10.1175/JAMC-D-11-0246.1.
Weckwerth, T. M., J. W. Wilson, M. Hagen, T. J. Emerson, J. O. Pinto, D. L. Rife, and L. Grebe, 2011: Radar climatology of the COPS region. Quart. J. Roy. Meteor. Soc., 137 (S1), 31–41, https://doi.org/10.1002/qj.747.
Wilson, J. W., and W. E. Schreiber, 1986: Initiation of convective storms at radar-observed boundary-layer convergence lines. Mon. Wea. Rev., 114, 2516–2536, https://doi.org/10.1175/1520-0493(1986)114<2516:IOCSAR>2.0.CO;2.
Wilson, J. W., and C. K. Mueller, 1993: Nowcasts of thunderstorm initiation and evolution. Wea. Forecasting, 8, 113–131, https://doi.org/10.1175/1520-0434(1993)008<0113:NOTIAE>2.0.CO;2.
Xing, J., H. Ai, and S. Lao, 2009: Multi-object tracking through occlusions by local tracklets filtering and global tracklets association with detection responses. 2009 IEEE Conf. on Computer Vision and Pattern Recognition, Miami, FL, Institute of Electrical and Electronics Engineers, 1200–1207, 1200–1207, https://doi.org/10.1109/CVPR.2009.5206745.
Yang, J., Z. Zhang, C. Wei, F. Lu, and Q. Guo, 2017: Introducing the new generation of Chinese geostationary weather satellites, Fengyun-4. Bull. Amer. Meteor. Soc., 98, 1637–1658, https://doi.org/10.1175/BAMS-D-16-0065.1.
Yang, Y., R. Wang, F. Chen, C. Liu, X. Bi, and M. Huang, 2021: Synoptic weather patterns modulate the frequency, type and vertical structure of summer precipitation over eastern China: A perspective from GPM observation. Atmos. Res., 249, 105342, https://doi.org/10.1016/j.atmosres.2020.105342.
Zeng, W., G. Chen, L. Bai, Q. Liu, and Z. Wen, 2022: Multiscale processed of heavy rainfall over East Asia in summer 2020: Diurnal cycle in response to synoptic disturbances. Mon. Wea. Rev., 150, 1355–1376, https://doi.org/10.1175/MWR-D-21-0308.1.
Zhang, F., Q. Zhang, and J. Sun, 2021: Initiation of an elevated mesoscale convective system with the influence of complex terrain during Meiyu season. J. Geophys. Res. Atmos., 126, e2020JD033416, https://doi.org/10.1029/2020JD033416.
Zhang, M., Z. Meng, Y. Huang, and D. Wang, 2019: The mechanism and predictability of an elevated convection initiation event in a weak-lifting environment in central-eastern China. Mon. Wea. Rev., 147, 1823–1841, https://doi.org/10.1175/MWR-D-18-0400.1.
Zhang, S. M., and Coauthors, 2017: Analysis on two severe convective weather events occurred near the subtropical high ridge in Jiangsu Province (in Chinese). Torrential Rain Dis., 36, 422–430.
Zhao, Q., X. Chen, N. Wang, and L. Peng, 2022: Environmental characteristics and trigger mechanism of severe convective under the influence of subtropical high in Guanzhong area of Shaanxi Province (in Chinese). Meteor. Mon., 48, 28–43, https://doi.org/10.7519/j.issn.1000-0526.2021.061001.
Zhao, S., Y. Deng, and R. X. Black, 2016: Warm season dry spells in the central and eastern United States: Diverging skill in climate model representation. J. Climate, 29, 5617–5624, https://doi.org/10.1175/JCLI-D-16-0321.1.
Zhao, S., Y. Deng, and R. X. Black, 2017a: A dynamical and statistical characterization of U.S. extreme precipitation events and their associated large-scale meteorological patterns. J. Climate, 30, 1307–1326, https://doi.org/10.1175/JCLI-D-15-0910.1.
Zhao, S., Y. Deng, and R. X. Black, 2017b: Observed and simulated spring and summer dryness in the United States: The impact of the Pacific sea surface temperature and beyond. J. Geophys. Res. Atmos., 122, 12 713–12 731, https://doi.org/10.1002/2017JD027279.
Zheng, Y., M. Xue, B. Li, J. Chen, and Z. Tao, 2016: Spatial characteristics of extreme rainfall over China with hourly through 24-hour accumulation periods based on national-level hourly rain gauge data. Adv. Atmos. Sci., 33, 1218–1232, https://doi.org/10.1007/s00376-016-6128-5.
