The Lower Atmospheric Characteristics of Dust Storms Using Ground-Based Sensor Data: A Comparative Analysis of Two Cases in Jinan, China

Abstract

Two severe dust storm (DS) events (15–17 March and 28–29 March) hit northern China in 2021 consecutively. The lower atmospheric vertical dynamic and thermal structures during the two cases were compared using the ground-based sensor data from the microwave radiometer and radar wind profiler, combined with the environmental and meteorological observations data in Jinan, China. It was found that both cases occurred under the background of cold vortexes over northeastern China. The dust was transported through the cold air on the northwest route. During the dust period, 2–3 km was the west or northwest airflow, and below 2 km was the northeast wind. The variation in the dynamic structure determined the duration of the DS. During the DS maintenance phase, the vertical wind shear (VWS) below 3 km measured approximately 10 m∙(s∙km)⁻¹. The increased VWS during the dust intrusion period facilitated the transportation of dust. In contrast, the more significant VWS was not conducive to the maintenance of DS, and the shift to south wind control in the upper middle layer indicated the weakening of DS. In both cases, we observed a cliff-like decrease in relative humidity as a prominent indicator of dust outbreaks, occurring approximately 2–5 h beforehand. The diurnal difference between the vertical temperature and relative humidity during the dust maintenance period was found to be insignificant.

1. Introduction

Air pollution events such as dust storms have recently received increasing international attention. Their increasing frequency and severity pose significant challenges to human health, agricultural productivity, and the ecological environment [1]. Detailed studies of the lower atmospheric conditions using various observation methods have been performed to understand the formation and transmission mechanism of dust storms [2,3,4]. The impacts of dust storm (DS) events on the physical and chemical processes show that the particles in the DS could affect the chemical reaction and energy balance in the boundary layer, further causing harm to the environment and ecosystem [5,6]. Researchers also identified key factors that contribute to the occurrence and severity of DS, such as climate change, land use change, and anthropogenic activities [7,8]. Remote sensing data are widely used in the study of the origin and transport trajectories of the dust, such as MODIS (Modify Resolution Imaging Spectrometer), CALIPSO (Cloud-Aerosol Lidar Infrared Pathfinder Satellite Observation), and the Fengyun Satellite series, providing researchers with more data options [9,10,11,12]. In addition to satellite remote sensing data, ground-based remote sensing data are also broadly used to detect different weather. They have high temporal and vertical resolution and can conduct continuous profiles of the atmospheric layer under different weather conditions [13]. In recent years, as the China Meteorological Administration has successively started networking on new observation equipment, it is feasible and effective to explore the vertical structure of the atmosphere by using new data such as microwave radiometers and radar wind profilers [14,15,16,17].

In China, dust storms mainly occur in arid and semiarid areas such as the South Xinjiang Region, the Hexi Region, and their neighboring areas [18]. Based on data from 1980 to 2005, Zhang and Gao [19] found there were mainly three potential dust sources affecting northern China, and approximately 70% of the DS events impacting China originated from Mongolia. Bao et al. [20] investigated the spatiotemporal variations of border-crossing dust events between Mongolia and Inner Mongolia and supported the results that the sand and dust from Mongolia play a major contribution to the DS events in China. Previous investigations have been conducted on the vertical structure during the dust period in arid and semiarid areas of China. The vertical structure plays a crucial role in facilitating the mixing of dust aerosols [21]. By using the ground-based polarization Raman lidar, Zhang et al. [22] demonstrated that dust aerosols in the central Taklimakan Desert were regularly lifted over 6 km during the summer, while the majority of the dust remained restricted within 2 km. Based on the CALIPSO data analysis conducted in March 2021, Ye et al. [23] revealed that the dust conveyance height was mostly concentrated at 1–3 km and 3–10 km. By using microwave radiometer data from 2007 to 2013, Li et al. [24] discovered that before the occurrence of DS weather, there was a significant increase in temperature at the lower levels and a notable temperature decrease at higher altitudes in China’s semi-arid region. By utilizing atmospheric aerosol lidar, Yin [25] indicated that during the period of sand and dust pollution in May 2022 in Xinjiang Province, the trend of PM_2.5 and PM₁₀ was consistent, and there was heterogeneity in vertical height.

In March 2021, a series of severe dust storms occurred in northern China. Many scholars have studied these DS cases in terms of the causes, transport, deposition, and radiation effects. Liang et al. [9] found that the “3.15” DS radiated from central and southern Mongolia to northern China. The significant contribution of Mongolian sand also received validation in the March 2021 case. Zhang et al. [26] studied the vertical diffusion mechanism of the boundary layer in Beijing during the transport of DS and pointed out that the eruption of turbulence under the influence of a cold front led to a sharp increase in the dust concentration on the surface, while the continuous turbulence caused the dust to dissipate gradually. By comparing the three dust events in March 2021, Wang et al. [27] revealed the effect of the vertical structure of the atmosphere on the dust concentration and found that the continuous strong downdraft led to the sharp rise of the dust concentration near the ground. Gui et al. [28] compared the optical characteristics of the dust processes of “3.15” and “3.27” and found that the direct forcing of short-wave radiation caused by the dust particles in the latter was stronger than that of the former. Jin et al. [29] simulated the dust events and found that 69% of the dust deposition in the North China Plain came from long-distance transport from Mongolia. Song et al. [30] investigated the carbonaceous components, inorganic elements, and water-soluble ions in dust storms, providing diverse evidence for dust source identification.

Based on the above investigations, research predominantly relies on conventional observational data for analysis. Existing studies primarily focus on individual DS case analyses employing a single observational instrument. There is relatively little research on the lower atmospheric characteristics of downstream areas in the dust case. The lower atmospheric vertical structure of DS remains unclear. Therefore, it is imperative to investigate the structural characteristics of the local vertical atmosphere in dust transport utilizing novel high-resolution data. Jinan (JN), the capital of Shandong Province, is at the lower reaches of the Beijing–Tianjin–Hebei region. It is one of the main downstream affected areas in the transmission process of dust storms. Affected by the dust transport in the upstream area, there were two large-scale dust events in Jinan on March 15–18 and March 28–29, 2021, which are only about 10 days apart. Although the circulation backgrounds affecting the two processes were similar, the intensity and duration of the two dust storms affecting Jinan were different. What were the influencing factors in the lower atmosphere?

This study explored the reasons mainly using high-resolution ground-based sensor data. In Section 2 of this paper, we introduce the high-resolution ground-based sensor data and methods. In Section 3, results are analyzed and discussed, including the comparisons of the ground meteorological elements, the intensity of cold air, and the vertical dynamic and thermodynamic structures. Lastly, the conclusions are summarized. Our ultimate purpose is to quantitatively understand the lower atmospheric vertical characteristics in the process of downstream dust invasion, maintenance, and dissipation.

2. Data and Methods

2.1. Ground-Based Remote Sensing Data

The two sources of ground-based remote sensing data are from the Chinese-made QFW-6000 microwave radiometer (MWR) and the CLC-11-F tropospheric radar wind profiler (RWP) installed in Pingyin, Jinan. The temporal resolution of temperature and relative humidity (RH) profiles of the MWR (116.41° E, 36.26° N) is 5 min, with a height resolution of 100 m (<2 km) and 250 m (2–10 km). The RWP (116.24° E, 36.15° N) provides the horizontal wind direction, horizontal wind speed, and vertical wind speed products with a time interval of 6 min. All the MWR and RWP data were under strict quality control for data consistency, continuity, and deviation based on the national standard of observation.

2.2. Environmental and Meteorological Data

The hourly meteorological observations were derived from the Chinese-made DZZ4 automatic weather device at JN National Meteorological Station. They include the surface temperature, pressure, relative humidity, wind direction, and wind speed. We also collected the 0.25° × 0.25° ERA5 reanalysis data from the ECMWF (European Center for Medium-Range Weather Forecasting) website to analyze the weather situation of the DS. The JN observational data of hourly PM₁₀ were observed by the Met One BAM-1020 particulate substance monitor from the National Air Quality Monitoring Network.

2.3. Methods

The vertical wind shear (VWS) and the vertical temperature lapse rate (VTLR) are important indicators of the lower atmosphere’s dynamics and thermodynamics, respectively [13,31]. We used the VWS to characterize the wind shear intensity between two heights. Hourly VWS was calculated with the resolution of 120 or 240 m, expressed as follows:

$$VWS=\frac{\sqrt{{(u_{z1}-u_{z2})}^2+{(v_{z1}-v_{z2})}^2}}{(z_1-z_2)}\times 1000$$

(1)

where VWS is the vertical wind shear (m·(s·km)⁻¹). $u_{z1}$ and $u_{z2}$ represent the zonal wind at z₁ and z₂, and $v_{z1}$ and $v_{z2}$ denote the meridional wind at z₁ and z₂, respectively (z₂ > z₁).

The VTLR (°C·hm⁻¹) is defined as:

$$\mathsf{\gamma }=-\frac{T_{z2}-T_{z1}}{z2-z1}\times 100$$

(2)

where γ is the vertical temperature lapse rate. $T_{z1}$ and $T_{z2}$ represent the temperature at z₁ and z₂, respectively (z₂ > z₁).

3. Results and Discussion

3.1. Surface Elements Changes

For the convenience of expression, we referred to the former dust process as “case 1” and the latter as “case 2”. According to the hourly variations of PM₁₀ concentrations, we divided the dust storm episode into three periods, corresponding to invasion (Period 1, P1), maintenance (Period 2, P2), and dissipation (Period 3, P3). In case 1 (Figure 1a), the minimum visibility was 1.0 km, and the PM₁₀ peak reached 1146.8 μg/m³. The intensity of case 2 (Figure 1b) was slightly weaker, but the peak value of PM₁₀ still reached 1101.2 μg/m³ with a minimum visibility of 1.9 km. Both cases show that RH appeared to have a cliff-like decrease during the P1. In case 1, as PM₁₀ started to rocket at 17:00 on the 15th, the RH decreased from 80% to 35% within 4 h. Meanwhile, in case 2, the RH decreased from 73% to 46% within 3 h from 11:00 on the 28th. Before the P1, the southerly wind dominated. A significant shift in the surface wind direction occurred during P1. As the cold air blew southward, the northeasterly wind dominated. The northeasterly wind gradually weakened, turning south during P3. The surface wind speed in P2 was slightly lower than that in P1. The cold air intrusion periods in the two cases were both afternoons, but the surface temperature before P1 in case 2 was 18.2 °C, which was 4 °C higher than in case 1.

3.2. Cold Air Intensity

The DS events we discussed were both affected by the transport and deposition of high-altitude dust. Both cases were formed under similar circulation patterns, as Figure 2 demonstrates. At 500 hPa, the cold vortex was located over the northeast (between 115° E–120° E, 45° N–50° N). The northwest airflow controlled JN at the base of the cold vortex. Case 1’s 500 hPa closed isotherm center was measured at −32 °C, but case 2’s cold center was slightly weaker at −28 °C. In each case, the temperature field trailed behind the height field, causing a blast of cold air to break through the North China Plain and head south. In comparison to case 2, case 1’s cold air influence range and intensity were noticeably larger. At 08:00 on March 15, there was a strong temperature front and an intense cold advection center (−10 × 10⁻⁴ °C/s) near 42° N, with northwest wind speeds ranging from 12–18 m/s. The isobars and isotherms at 500 hPa in the upstream region (Jing–Jin–Ji) were dense (Figure 2a). The northwest wind speed ranged from 8 to 12 m/s, and the corresponding surface cold advection center intensity was −7 × 10⁻⁴ °C/s (Figure 2c). At 8:00 on March 28, the highest cold advection on the ground was only −4 × 10⁻⁴ °C/s (Figure 2d), while the center of the cold advection at 850 hPa was −10 × 10⁻⁴ °C/s (Figure 2b). Compared to case 1, the surface wind speed in case 2 was approximately 2–4 m/s slower. In summary, the cold air intensity of case 1 was stronger than that in case 2. Later, a more thorough comparison and analysis of the lower atmospheric structure during the dust transmission is conducted.

3.3. Dynamic and Thermodynamic Structure in the Lower Atmosphere

In this section, we characterize the lower atmospheric characteristics of the three periods in the two cases using the RWP and MWR data. Under a similar cold vortex background, the wind field structure below 3 km in both cases presented the characteristics of “backflow”. In case 1, the northwest airflow was between 2 and 3 km, while the northeast wind was less than 1.5 km. The wind direction is reversed counterclockwise with heights (Figure 3a). In case 2, the characteristics of “backflow” are more typical, with west–northwest airflow at 2–3 km and northeast wind below 2 km. In case 1, a weak descending motion (0.8 m∙s⁻¹) occurred both before and during the P2, which was conducive to dust deposition. There was a clear northeast jet within 0–700 m from 16:00 on the 15th to the early hours of the following day, which was also helpful for the ongoing transport of dust.

In case 2 (Figure 3b), there was a significant downward momentum transportation during the P1. At approximately 12:00 on March 28, a maximum wind speed center of 16–20 m∙s⁻¹ was observed at an altitude of 2.5–3 km. The center rapidly decreased over time, accompanied by a gradual grounding and rotation of wind direction. This was conducive to dust transport and deposition of the dust. Four hours later (at 16:00 on March 28), the PM10 concentration on the ground reached the maximum (Figure 1b). However, in case 1, the downward momentum transportation was weak during the dust invasion period (Figure 3a). In case 2, the ascending and descending movements alternated in the dust-influenced stage, which was conducive to dust maintenance. In addition, the north wind component (red line) also played an important role in dust transportation. The north wind component of the two cases was maintained between 4 and 8 m∙s⁻¹ in the dust invasion and maintenance (P1 and P2). Before and after the two cases dissipated, the wind above 2 km first turned to the south (the north wind component was positive), signifying that the influence of cold air transmission had ended. This indicated that the influence of the dust storm tended to weaken or terminate.

VWS signifies the state of vertical dynamic mixing in the atmosphere, which plays a vital role in both pollutant accumulation and diffusion. To analyze the vertical characteristics of the dynamic structure in the lower atmosphere during various periods of the DS, we calculated the mean profiles of VWS and north wind components, as depicted in Figure 4. The VWS and northerly wind components of the two cases exhibited distinct characteristics with heights. Both P1 and P2 were affected by northerly wind transport (Figure 4d,e). Although the cold air intensity of case 1 was stronger (Section 3.2), the northerly wind component was weaker than that of case 2. In P3, the upper-middle level (1–3 km) wind first turned to the south (Figure 4f). There was little difference between the VWS in P1 and P2 of case 1. However, the VWS of P1 in case 2 was significantly larger than that in P2, especially when the VWS increased to about 15 m∙(s∙km)⁻¹ at the height of 1.5–2.5 km, resulting in explosive growth of dust concentrations. The VWS in P2 of both processes was about 10 m∙(s∙km)⁻¹. In the P3, the VWS over 2 km increased to 10–20 m∙(s∙km)⁻¹ in both cases (Figure 4c). Additionally, there was an observed increase in the southerly wind component over 2 km (Figure 4f, the northerly wind component turned positive). This indicated that the wind direction shifting to the south and large VWS over 2 km was conducive to the dust dissipation.

By comparing the vertical structural changes of RH and temperature retrieved by MWR in the two cases (Figure 5), a common feature can be found: the RH in the lower atmosphere before the dust intrusion showed a cliff-like decline. As can be seen from Figure 1, the PM₁₀ concentration peaked at 23:00 on March 15 and 16:00 on March 28, respectively, while the RH profiles showed an obvious decline at 18:00 on March 15 (Figure 5a) and around 14:00 on March 28 (Figure 5b). Meanwhile, the maximum value area of the humidity gradient first appeared near 1.5–2.5 km, which was at least 2–5 h earlier than the peak dust concentration on the ground. Thus, a cliff-like decrease in RH about 2 km, occurring 2–5 h beforehand, serves as an indicator of an abrupt increase in dust concentrations. During these two cases, the height of dust transport in the upstream Beijing area varied from 1 to 2.5 km [32]. Affected by the northwest cold air transmission, the downstream area of Jinan will inevitably experience dust transportation followed by deposition. The height of approximately 2 km may be where the sand and dust first invaded, so this altitude range will first appear as a cliff-like drop in RH. During case 1, the isotherms (red line) exhibited a continuous decrease, whereas in case 2, they displayed a minor peak. The explanation for this phenomenon is as follows. In case 1, the solar radiation warming near the ground was weakened with time, and the temperature in the lower atmosphere gradually decreased due to the influence of cold air transmission. However, during case 2, a weak cold air mass combined with surface heat absorption at noon to produce a small temperature peak at approximately 15:00.

The maintenance period (P2) of the two dust cases both went through day and night. Therefore, the thermal structure at 0–3 km is represented by the mean profiles of the RH and VTLR in day and night, respectively, as shown in Figure 6. The times 08:00 and 20:00 are taken as the dividing time points of day and night. In Figure 6a,b, the RH of the two cases had a large difference below 1.5 km and a small difference above 1.5 km. The difference between the mean RH values at each height during the day was at most 20%, while the difference at night was at most 15%. In the same case, the RH of day and night showed a similar shape with heights. The maximum was concentrated below 1.5 km, and the difference between day and night at each height was not large (the maximum is about 15%). However, by comparing with the 25–75% percentile value, we found that the fluctuation range of the RH during the day was larger than that at night. During the dust maintenance period, the RH in the lower layer decreased and remained at a low value, so the diurnal difference in the humidity profile was small. During the day, as the solar radiation first increased and then weakened, the RH also fluctuated, resulting in a wider range of changes in the daytime. As shown in Figure 6c,d, there was little difference between day and night in the VTLR in the dust maintenance period. It was maintained at approximately 0.5 °C∙hm−1 and oscillated more dramatically below 2 km. In summary, during the dust maintenance periods, the vertical temperature and relative humidity changes during the day and night were significantly small.

4. Discussion and Conclusions

According to statistics provided by the China Meteorological Administration, a total of 37 large-scale and intense dust storm (DS) events occurred in northern China between 2000 and 2023, averaging approximately 1.5 occurrences per year. These events were predominantly observed from March to May. The most severe dust storm event in Jinan was notably observed in March 2021, subsequent to the year 2010. Our research utilized the MWR and RWP data to compare the vertical dynamic and thermal structure of the lower atmospheric layer in two dust events that occurred in March 2021 in Jinan, China. The two cases we studied during this period (March 15–18 and March 28–29) are representative. Gao et al. [33] have suggested that the dust storm frequency is tightly linked to the atmospheric general circulations, and the cyclones are believed to play a dominant role in causing the DS events in northern China. We found a similar pattern casing these two DS events in this study. Both cases occurred under the influence of the northeastern cold vortex.

Then, we examined the impacts of the dynamic structure within the lower atmosphere during the DS events. The change of wind direction is a critical point for the different periods of the DS. As the most important dynamic condition, strong northerly wind has a significant positive correlation with the spring dust storm frequency [34]. The dust weather was generated through the transmission of cold air under the influence of the northwest airflow. The west–northwest airflow is 2–3 km, and the northeasterly wind is below 2 km. In both cases, the initial shift to the south wind control above 2 km was observed at the beginning of the dissipation period, indicating the dust would weaken. The VWS increases rapidly in both the invasion and dissipation periods of the dust. The VWS in the dust maintenance stage of both cases was about 10 m∙(s∙km)⁻¹. The typical feature of case 1 was that there was a boundary layer jet with a wind speed of approximately 8 m∙s⁻¹ at 0–700 m. This phenomenon was conducive to dust transport. Case 2 was characterized by a noticeable downward momentum at 2.5–3 km, which was favorable for the sharp increase in dust concentration. In addition, this finding was in line with the study conducted by Wang et al. [27], which examined a typical dust storm event in Beijing and proposed that the intensity of vertical downdraft directly influenced the amount of dust deposition on the ground.

The MWR is an effective method to detect the vertical thermodynamic characteristics at different heights in DS events. Our analysis suggested the change features of relative humidity before the dust concentration sharply rises, consistent with Hao et al. [35]. They found that the overall atmospheric humidity showed a steep decline trend before the particulate concentration on the ground increased dramatically in Xingtai, Hebei Province. However, we conducted further investigation into the indicative characteristics of height and time, and the results may have broader implications and significance. The maximum value of the humidity gradient first appeared near 1.5–2.5 km. If the relative humidity in this height range shows a cliff-like decline, it signifies that the dust concentration will rise sharply within 2–5 h. In addition, during the dust maintenance period, the diurnal difference in the vertical structure of temperature and humidity in the lower atmospheric layer was found to be small. However, the fluctuation range of relative humidity in the daytime was larger than that at night.

Overall, based on the two ground-sensor data sources, this study reveals the structural characteristics of the lower atmosphere in the downstream region during the dust transport process. It is important to note that our conclusions are specific to these two cases and should not be generalized. Therefore, we plan to carry out further investigations to explore more dust events by employing a wider range of ground sensor instruments or numerical simulations, thereby enhancing our understanding of the dust transport mechanism.