Wind Field Characteristics of the 13 June 2014 Downburst Event in Beijing Based on Meteorological Tower Records

Abstract

Understanding the characteristics of downburst wind fields is crucial for studying structural resistance to downbursts. Based on measured data from the 325 m meteorological tower in Beijing, this paper investigates the spatiotemporal evolution of mean and fluctuating winds during a non-stationary downburst. Key wind field parameters such as the mean wind speed, turbulence intensity, turbulence integral length scale, probability density function, power spectral density, evolutionary power spectral density, and gust factor are statistically analyzed. The results show that the wind speed of downburst undergoes rapid changes, with wind direction significantly influenced by outflow vortices at low altitudes and relatively stable at higher altitudes. When the event happens, the temperature decreases sharply. The mean wind speeds and turbulence integral length scale of the downburst exhibit pronounced “nose-shaped” profile characteristics at the moment when peak wind speed occurs. The turbulence intensity at lower altitudes predominantly exceeds that at higher altitudes. The probability density distribution function of the reduced fluctuating wind speed matches the standard Gaussian distribution curve. The fluctuating wind speeds of the downburst exhibit significant non-stationary characteristics, with their energy mainly distributing in the period of rapid change of wind speed in the time domain and concentrating in the vicinity of 0–0.1 Hz in the frequency domain. The gust factor reaches its maximum at the moment when the peak wind speed occurs.

1. Introduction

Downbursts are a type of damaging extreme wind, caused by rapidly descending air currents that impact the ground and characterized by intense radial flows and annular vortices [1,2,3,4,5,6]. They exhibit features such as short duration, an abrupt onset, high randomness, and high near-ground wind speeds, which result in severe damage to building structures. Furthermore, downbursts exhibit significant non-stationary and non-Gaussian characteristics [7,8,9,10]. Therefore, elucidating the wind field characteristics of downbursts is crucial for optimizing building design and reducing casualties.

Field measurements are the most direct and effective means of studying the characteristics of downburst wind fields. Currently, scholars from international communities have carried out some field studies on the characteristics of downburst wind fields. Meteorologist Goff [11] conducted over 20 preliminary thunderstorm detections at the Oklahoma Tower between 1971 and 1973, which investigated the wind speed profiles of gust fronts and which classified thunderstorm outflows into four types based on different stages of thunderstorm development. Fujita [12,13] made groundbreaking contributions to early field studies of thunderstorm winds through international projects such as NIMROD, JAWS, and MIST, which focused on downburst measurements, providing preliminary classifications based on their damage radii. Subsequently, Hjelmfelt [14] summarized the basic wind field characteristics of downbursts, including the initial diameter of the downflow, initial outflow height, and height of maximum horizontal wind speed, based on 26 field measurements obtained from the JAWS project. Corresponding schematic diagrams of downburst wind profiles were presented. The “nose-shaped” vertical wind speed profile of downburst, with higher speeds at the middle altitude and lower speeds at the lower and higher altitude, is the most distinguishing characteristic different from that of atmospheric boundary layer wind fields [4]. Canepa et al. [4] analyzed 10 downburst records obtained by lidar and found that the nose-shaped feature of the thunderstorm wind profile is mainly observed during the wind speed rise and peak segments, with maximum wind speeds usually occurring within the height range of 60 m to 250 m. Chen et al. [15] performed a comprehensive statistical analysis to delve into the temporal characteristics of wind speeds documented during two thunderstorm events at Texas Tech University. Building on this foundation, they introduced an empirical model tailored for simulating the extreme wind velocities associated with thunderstorm downbursts. Solari et al. [16] analyzed 93 sets of downburst field measurements and found that the mean turbulence intensity of downbursts was approximately 0.12, which is lower than that of synoptic winds ($I_v=0.17$), and turbulence integral length scale is relatively small as well. Zhang et al. [3] statistically analyzed over 200 downburst events in the northern Mediterranean and found that the turbulence intensity ranged from 0.08 to 0.18, with turbulence integral length scale between 25 m and 32 m, and both were unaffected by surface roughness. Subsequently, Shu et al. [17] analyzed six years of meteorological data from six weather stations in Hong Kong and found that the deviation of the gust factor correlates with terrain conditions and wind speed. Choi et al. [18,19] examined the wind speed profiles of more than 50 thunderstorm events recorded at Nanyang Technological University in Singapore and found that they are mainly influenced by the distance from the impact center, downburst intensity, and surface roughness. Gunter and Schroeder [20] analyzed the average and evolutionary characteristics of wind profiles for three typical downburst events and found that extreme wind speeds occurred above 100 m in all three events. Zhang et al.[21] analyzed 70 downburst wind speed records from a 325 m meteorological tower in Beijing and found that the turbulence intensity varied between 0.11 to 0.34, decreasing with height. In summary, current research on the characteristics of downburst wind fields primarily focuses on field measurements at a single height, while few studies examine downbursts at different heights. Additionally, because of their small spatial scale and short lifecycle, the available field measurement data for downbursts is limited, and further in-depth studies utilizing addition field data are required.

Based on the wind speed data measured by ultrasonic anemometers at nine heights on the 325 m meteorological tower in Beijing, wind speed records on 13 June 2014 were selected in this paper. This event is extracted by Zhang et al. [21] and confirmed by the record of Beijing Meteorological Observatory. The spatiotemporal evolution characteristics of mean and fluctuating winds in non-stationary downbursts were investigated based on these data. The research encompasses key wind field parameters including the mean wind speed profile, turbulence intensity profile, turbulence integral length scale, probability density function, power spectral density and evolutionary power spectral density of fluctuating winds, and the gust factor of downbursts. The findings of this study provide a reference for a deeper understanding of the characteristics of downburst wind fields and provide both a data foundation and theoretical framework for further research on unified wind field modeling of downbursts and related load calculations and response analyses.

2. Overview of Field Measurements

The wind speed records measured in this study are obtained from the 325 m meteorological observation tower at the Institute of Atmospheric Physics, Chinese Academy of Sciences, as depicted in Figure 1. The tower is located in the northern part of Madian Bridge along the Northern Third Ring Road, in the Haidian District of Beijing, China, at the southwestern corner of Jiandemen Bridge on Beitucheng West Road (39°58′ N, 116°2′ E). Within a 1 km radius from the tower, there are no other significant high-rise buildings, except for some residential buildings approximately 60 m in height, located 100 m to the south of the Tower. According to the “Load Code for the Design of Building Structure” (GB 50009-2012) [22], the site surrounding the meteorological tower can be classified as Terrain Category C.

A total of 11 tri-axial ultrasonic anemometers were installed at heights of 8 m, 16 m, 32 m, 47 m, 64 m, 80 m, 140 m, 200 m, and 280 m on the tower. These anemometers are capable of simultaneously measuring three-dimensional instantaneous wind speed and temperature at a specific measurement point, with a sampling frequency of 10 Hz for all parameters. Additionally, to verify the accuracy of the measured fluctuating wind speeds, two ultrasonic anemometers at the same heights but in different directions (north and west) were installed at 32 m and 64 m, respectively, while the remaining anemometers were oriented towards the north.

The wind speeds collected by the WindMaster Pro ultrasonic anemometer are three-dimensional orthogonal instantaneous wind speeds, denoted as $U_i(i=N,E,W)$, where the positive directions correspond to south-to-north, east-to-west, and vertically upward, respectively. Prior to analyzing wind field characteristics, the raw data were preprocessed to remove outliers [23,24].

3. Downburst Wind Model

Instantaneous horizontal wind speed $U\left(t\right)$ of a downburst, the module of the two wind speed components U_E and U_N, can be modeled as the superposition of the time-varying mean wind speed term $\overline{U}\left(t\right)$ and the residual fluctuating wind speed term $u^{\prime }\left(t\right)$ [25,26], which is expressed as follows:

(1)$U\left(t\right)=\overline{U}+u^{\prime }\left(t\right)$

The time-varying mean wind speed $\overline{U}$ is a deterministic function that can be extracted using the moving average method. The moving time interval is a key factor influencing the accuracy of the processed results. Based on previous research, the basic time interval adopted in this paper is 30 s [25].

Furthermore, the residual fluctuating wind speed component $u^{\prime }\left(t\right)$ is a non-stationary random process, and its calculation formula can be expressed as:

(2)$u^{\prime }\left(t\right)={\sigma }_u\cdot {\tilde{u}}^{\prime }\left(t\right)$

where ${\sigma }_u$ is the time-varying standard deviation of the residual fluctuating wind speed. ${\tilde{u}}^{\prime }\left(t\right)$ is a stationary Gaussian random process with zero mean and unit standard deviation [2], defined as the reduced fluctuating wind speed.

Substituting Equation (2) into Equation (1), the original wind speed time history can be further expressed as:

(3)$U\left(t\right)=\overline{U}\cdot \left[1+I_U\cdot {\tilde{u}}^{\prime }\left(t\right)\right]$

where $I_U={\sigma }_u/\overline{U}$, representing the time-varying turbulence intensity.

In addition, the gust factor is generally defined as:

(4)$G_u=\widehat{U}/{\overline{U}}_{max}$

where $\widehat{U}$ denotes the peak gust wind speed of a downburst, with a typical calculation time interval of 1 s, and ${\overline{U}}_{max}$ is the maximum value of the 30 s time-varying mean wind speed.

The autocorrelation coefficient method [9,27] determines the turbulence integral length scale by analyzing the autocorrelation of flow velocity data. The calculation formula is as follows:

(5)$L_u^t={\int }_0^{\mathrm{\infty }}{\rho }_{uu}\left(t\right)dt$

where ${\rho }_{uu}\left(t\right)$ is the autocorrelation coefficient of the fluctuating wind speed $u^{\prime }$.

Utilizing wind speed data collected from the 325-m meteorological tower located in Beijing, this research paper identifies and characterizes a typical downburst event that transpired on 13 June 2014, through meticulous analysis and screening procedures in accordance with the downburst identification methodology presented in literature [2]. To determine the typical temporal progression of the downburst, a comprehensive evaluation is conducted, which encompasses the assessment of the maximum 1-s gust wind speed, gust factor, temperature variation data, and a qualitative inspection of wind profile shapes. Consequently, Figure 2 illustrates the temporal variations of the instantaneous horizontal wind speed $U$, wind direction $\alpha$ (0° means north and 90° means west), and temperature $T$ at various heights during the aforementioned downburst event on 13 June 2014. This analysis contributes to the understanding of downburst characteristics and their impact on structural wind engineering, particularly in the context of high-rise structures and wind-sensitive designs.

From Figure 2, it can be observed that during the documented downburst event on the specified day, peak wind velocity attained a magnitude of 22.10 m/s. According to the national standard “Wind Scale”, promulgated in China in June 2012, this wind speed aligns with Beaufort Scale 9, indicating a capacity sufficient to inflict substantial damage upon structures, thereby highlighting the extreme peril associated with downburst phenomena. A meticulous examination of the temporal evolution plots reveals that the wind speed underwent rapid and pronounced fluctuations throughout the duration of the downburst event. As the wind speed of the downburst commenced its ascent, alterations in wind direction are particularly notable at lower elevations, whereas at higher altitudes, the wind direction underwent a gradual shift in tandem with the augmentation of wind speed. The results show that the wind direction at elevated heights is predominantly influenced by the velocity of the thunderstorm’s movement, whereas at lower altitudes, the wind direction is more significantly impacted by the dispersed outflow vortices resulting from the interaction between the downburst airflow and the ground surface. During the phase of rapid acceleration in downburst wind speed, the temperature decreased by 10 °C within an 8-min interval and subsequently stabilized. The abrupt change in both wind speed and temperature is the quintessential characteristic when a downburst occurs.

The downburst wind speed signal at a height of 47 m is decomposed using a moving average filter with a 30-s period. The variations of $U$, $\overline{U}$, $u^{\prime }$, ${\sigma }_u$, $I_u$, ${\tilde{u}}^{\prime }$ over time are illustrated in Figure 3, respectively. As observed from Figure 3, both the time-varying mean wind speed $\overline{U}$ and the instantaneous wind speed $U$ exhibit distinct wind speed mutation intervals. The maximum of the instantaneous horizontal wind speed $U$ occurs at 17:03, with a value of 14.49 m/s.

4. Characteristics of Downburst Wind Fields

4.1. Mean Wind Speed

Previous studies have shown that the mean wind speed profile of downbursts exhibits a “nose-shaped” characteristic [3,9,14,24,28]. The temporal evolution of the time-varying mean wind speed profile during the representative downburst case is explored in Figure 4.

An examination of Figure 4 elucidates that at 16:50 (Moment 1), 17:00 (Moment 2), and 17:01 (Moment 3), preceding the manifestation of the downburst event, minor variations occur in wind speed with respect to height, consistently maintaining around 3.15 m/s. When the downburst occurs, specifically at 17:03 (Moment 4), a clear “nose-shaped” profile is observed, with the maximum mean wind speed appearing at a height of 80 m. During the wind speed decline phase, at Moment 5, 6, 7, and 8, as the wind speed decreases, the “nose-shaped” characteristic of the wind speed profile also weakens.

4.2. Turbulence Intensity

Turbulence intensity is the most important characteristic quantity that describes the characteristics of atmospheric turbulent motion. Prior to analyzing this intensity, data points deemed erroneous due to exceptionally low wind speeds, which induce significant and intense turbulence (specifically, with mean wind speeds $\overline{U}$ < 5 m/s and turbulence intensities $I_u$ > 0.2), were excluded from consideration in this study [9]. Figure 5 presents the profiles of turbulence intensity varying with height at different time instants during the downburst event.

Upon detailed examination of Figure 5, it is evident that during the occurrence of a downburst, the turbulence intensity at lower altitudes predominantly exceeds that at higher altitudes. Specifically, before the occurrence of the event, the overall turbulence intensity is relatively small fluctuating around a value of 0.1. Then, when the downburst occurs, turbulence intensity increases and exhibits a trend of an initial decrease, followed by an increase, and then a subsequent decrease as the altitude ascends.

4.3. Turbulence Integral Length Scale

Turbulence integral length scale, a critical parameter in structural wind load analysis, defines the average size of turbulent eddies in fluctuating winds and can be calculated using Equation (5). In this study, the turbulence integral length scales at different times are computed using the autocorrelation coefficient method with a 30-s moving average.

The profiles pertaining to the turbulence integral length scale related to the 10-min time interval centered with the analysis moment are depicted in Figure 6. At 16:43, 16:53, and 17:23, the turbulence integral length scale presents relatively small values and remains basically constant along the height. Notably, at 17:03, the turbulence integral length scales, representative of the vortex sizes induced by the downburst, initially increases with height, then decreases, and ultimately increases once again. During the period when the peak wind speed occurs, a pronounced “nose-shaped” profile emerges, with the apex of the nose positioned at a height of 80 m.

4.4. Probability Density Function

After evaluating and analyzing long-term wind speed data in central Beijing, the specific urban boundary layer structure within the Beijing area has been delineated, identifying below 80 m as the rough sublayer, with a height of 140 m located within the inertial sublayer; and 200 m and 280 m situated in the mixed layer [29]. Within the rough sublayer, the dynamic and thermal transfer processes of the atmosphere are directly influenced by the surrounding built environment, leading to a general reduction in atmospheric stability. The inertial sublayer, also known as the constant flux layer, exhibits uniform distribution of atmospheric turbulence flux and relatively stable turbulence transport within this altitude range, providing optimal conditions for wind field observations and turbulence analysis. In the urban mixed layer, the exchange of atmospheric matter and heat is significantly greater than the frictional force from the underlying rough elements, resulting in notable changes in turbulence movement patterns.

To investigate the Gaussian characteristics of reduced fluctuating wind speeds, Figure 7 presents the histograms of probability density distributions of reduced turbulent fluctuation and their fitted functions within a 10-min interval centered around the peak wind speed moment at the rough sublayer [18] (Figure 7a), the inertial sublayer (Figure 7b), and the mixed layer (Figure 7c). Additionally, the calculated statistical properties (mean $\mu$, standard deviation $\sigma$, skewness $\gamma$, kurtosis $\kappa$).

Figure 7 illustrates that the probability density distribution functions of reduced turbulent fluctuation basically coincide with the standard Gaussian distribution curve, with a better fit at 280 m compared to 47 m and 140 m. The means $\mu$ at all three heights are close to 0, and the skewness values are also small, indicating a balanced distribution of positive and negative wind speeds in the downburst samples. The standard deviations $\sigma$ are close to 1, suggesting a high degree of dispersion in the data and significant wind speed fluctuations. Except for the kurtosis $\kappa$ at 280 m being a little bit higher than 3, the kurtosis values at 47 m and 140 m are slightly lower than 3, indicating that the sample data distributions at these two heights are relatively flat, with a lower probability of extreme values compared to a normal distribution. This suggests that the reduced fluctuating wind speeds can be modeled as a stationary Gaussian random process with zero mean and unit standard deviation.

4.5. Power Spectral Density

The power spectral density of reduced turbulent fluctuation indicates the energy distribution of reduced fluctuating wind speed across different frequency ranges [30]. For this event, wind speed data within a 10-min interval centered around the peak wind speed moment were selected at three heights: the rough sublayer (47 m), the inertial sublayer (140 m), and the mixed layer (280 m). The fitting results for the power spectral density of reduced turbulent fluctuation are presented in Figure 8.

It can be observed that the downburst satisfies Kolmogorov’s “−5/3 power law” in the high-frequency range. The fitted results show poor agreement between the measured power spectrum and the Davenport spectrum [30], but better agreement with both the Von Karman spectrum [31] and the Solari spectrum [32]. The Von Karman spectrum accurately reflects the energy distribution of reduced turbulent fluctuation in both the mid- and high-frequency ranges, but it is slightly higher than the measured values in the low-frequency range. In contrast, the Solari spectrum is lower than the measured values in the mid-frequency range but higher in the low-frequency range.

As illustrated in Figure 8, the downburst satisfies Kolmogorov’s “−5/3 power law” in the high-frequency range. The measured power spectrum exhibits good fitting with both the Davenport spectrum [30] and the Solari spectrum [31], whereas the fitting with the Von Karman spectrum [32] is relatively poor. The Davenport spectrum accurately reflects the energy distribution of fluctuating wind speeds in both the mid- and high-frequency ranges, but it is slightly higher than the measured values in the low-frequency range. The Solari spectrum is lower than the measured values in the mid-frequency range but higher in the low-frequency range.

4.6. Evolutionary Power Spectral Density

The evolutionary power spectral density (EPSD) is utilized to characterize the non-stationary characteristics of thunderstorm winds. Among them, the EPSD based on the Priestley method possesses a clear physical meaning, namely, it can describe the evolution process of spectral energy in the time-frequency domain [28,33]. The window function and weight function are chosen as follows:

(6)$g\left(u\right)=\left\{\begin{array}{c}1/\left(2\sqrt{h\pi }\right)\left|u\right|\le h;\\ 0\left|u\right|>h.\end{array}\right.$

(7)$W_{T^{\prime }}\left(t\right)=\left\{\begin{array}{c}1/T^{\prime }\left|t\right|\le {\displaystyle \frac{1}{2}}{T}^{\prime };\\ 0\left|t\right|>{\displaystyle \frac{1}{2}}{T}^{\prime }.\end{array}\right.$

In the equation, the parameters $h$ and $T\prime$ are set to 7 and 200, respectively [24,34].

The evolutionary power spectral density (EPSD) and its normalized estimates (NPSD) of the fluctuating wind speed at the 47 m height during the downburst event are depicted in Figure 9. As shown in Figure 9a, during this event, the energy demonstrates pronounced non-stationary characteristics. Notably, the energy is predominantly concentrated within the frequency band of 0 to 0.1 Hz and during the time interval encompassing approximately 1200 to 1400 s. It is suggested that, in the context of structural design for resistance to downbursts, particular emphasis should be given to the dynamic response characteristics of structures possessing natural frequencies within this specified range. The estimated evolutionary power spectral density (EPSD) is normalized using time-varying variance, as shown in Figure 9b. It is observed that the distribution of the NPSD is non-stationary.

Figure 10 presents estimates of the time-varying standard deviation obtained using the Evolutionary Power Spectral Density (EPSD) integration method and the moving average method, respectively, along with the error values between the two methods (characterized by Mean Absolute Error (MAE) and Root Mean Square Error (RMSE)). Comparative analysis reveals that the MAE and RMSE of the two methods are approximately 0.17 and 0.25, respectively, and the correlation coefficient (CC) between the two methods is 0.91. This indicates that the results estimated by the two methods are relatively consistent and exhibit a strong positive correlation. Furthermore, the time-varying standard deviation obtained by the EPSD integration method is smoother, which validates the feasibility of employing the EPSD integration method for modeling downburst wind fields and calculating structural responses.

4.7. Gust Factor

The gust factor $G_u$ can be expressed by Equation (4), where ${\overline{U}}_{max}$ represents the maximum of the time-varying mean wind speeds. The peak gust wind speed is typically calculated using a time interval of 1 s given the instantaneous nature of downbursts. Table 1 presents the variation of gust factors with the height and time, the same definition with that in Figure 6.

As illustrated in Table 1, the gust factor attains its zenith at 17:03 when the peak wind speed occurs. As the height ascends, the gust factor undergoes a zigzag tendency. This intricate behavior may be attributed to the distinctive vertical wind profile characteristics of the downburst.

5. Conclusions

This paper utilizes the measured data of a downburst obtained from the Beijing Meteorological Tower on 13 June 2014 to investigate the wind field characteristics of the downburst. The conclusions are as follows:

• (1). During the occurrence of the downburst, the wind speed undergoes rapid changes. The wind direction varies significantly at low altitudes due to the influence of outflow vortices, while it remains relatively stable at higher altitudes. The temperature decreases sharply with increasing wind speed. The mean wind speed profile of the downburst differs significantly from that of synoptic winds, exhibiting a unique “nose-shaped” profile at the moment when peak wind speed occurs.

• (2). The turbulence intensity and turbulence integral length scale increases when the downburst occurs. The turbulence intensity at lower altitudes predominantly exceeds that at higher altitudes. The “nose-shaped” turbulence integral length scale profile can be observed at the moment when the peak wind speed occurs, with the nasal tip appearing at a height of 80 m similar with the mean wind profile.

• (3). The probability density distribution function of the reduced fluctuating wind speed during the downburst basically coincides with the standard Gaussian distribution curve. The measured power spectral density of the downburst satisfies Kolmogorov’s “−5/3 power law” in the high-frequency range, with energy gradually increasing as the altitude rises, and it aligns well with the Davenport spectrum.

• (4). The residual fluctuating wind speed of the downburst exhibits significant non-stationary characteristics. Its energy is mainly distributed in the period of rapid change of wind speed in the time domain and concentrating in the vicinity of 0–0.1 Hz in the frequency domain. The time-varying standard deviation estimated by the moving average method and the evolutionary power spectral integral method are in good agreement and the latter is smoother. This validates the feasibility of employing spectral methods for downburst wind modeling and structural response calculations. The gust factor attains its maximum value when the peak wind speed occurs.

Author Contributions

Conceptualization, S.Z. and Z.Q.; methodology, S.Z. and Y.W.; software, Y.W. and X.X.; validation, Z.Q., D.Z., Q.C. and X.X.; formal analysis, Y.W. and K.G.; data curation, S.Z. and K.G.; writing—original draft, S.Z., Y.W. and K.G.; writing—review and editing, Z.Q., D.Z. and Q.C. supervision, Z.Q., D.Z., Q.C. and X.X. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

For privacy reasons, data cannot be made public.

Conflicts of Interest

Author Xiaoda Xu was employed by the company Central Research Institute of Building and Construction Co., Ltd. MCC. Authors Zengzhi Qian, Daxing Zhou and Qing Cao were employed by the company China Railway Construction Group Co., Ltd. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as potential conflicts of interest.

Footnotes

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Figures and Table

Figure 1. The picture of Beijing 325 m meteorological tower.

Figure 2. Typical downburst recorded on 13 June 2014.

Figure 3. Wind velocity decomposition of a typical downburst record.

Figure 4. Evolution process of time-varying mean wind speed profiles.

Figure 5. Evolution process of turbulence intensity profiles.

Figure 6. Evolution process of turbulence integral length scale profiles.

Figure 7. Probability density function of reduced turbulent fluctuation.

Figure 7. Probability density function of reduced turbulent fluctuation.

Figure 8. Power spectral density of reduced turbulent fluctuation.

Figure 9. Estimation of EPSD and NPSD.

Figure 10. Estimation of time-varying standard deviations.

Table 1

Gust factors $G_u$ at different heights.

Height (m)	16:43	16:53	17:03	17:13	17:23
8	1.32	1.64	1.87	1.79	1.65
16	1.49	1.59	1.89	1.80	1.65
32	1.61	1.76	1.92	1.81	1.59
47	1.69	1.78	2.14	1.71	1.54
64	1.61	1.73	1.94	1.69	1.69
80	1.80	1.81	2.03	1.80	1.52
140	1.61	1.64	2.14	1.71	1.55
200	1.81	1.86	1.90	1.69	1.27
280	1.94	1.98	2.05	1.81	1.31

References

1. Canepa, F.; Burlando, M.; Romanic, D.; Solari, G.; Hangan, H. Experimental investigation of the near-surface flow dynamics in downburst-like impinging jets. Environ. Fluid Mech.; 2022; 22, pp. 921-954. [DOI: https://dx.doi.org/10.1007/s10652-022-09870-5]

2. Zhang, S.; Solari, G.; Burlando, M.; Yang, Q.S. Directional decomposition and properties of thunderstorm outflows. J. Wind Eng. Ind. Aerodyn.; 2019; 189, pp. 71-90. [DOI: https://dx.doi.org/10.1016/j.jweia.2019.03.014]

3. Zhang, S.; Solari, G.; De Gaetano, P.; Burlando, M.; Repetto, M.P. A refined analysis of thunderstorm outflow characteristics relevant to the wind loading of structures. Probabilistic Eng. Mech.; 2017; 54, pp. 9-24. [DOI: https://dx.doi.org/10.1016/j.probengmech.2017.06.003]

4. Canepa, F.; Burlando, M.; Solari, G. Vertical profile characteristics of thunderstorm outflows. J. Wind Eng. Ind. Aerodyn.; 2020; 206, 104332. [DOI: https://dx.doi.org/10.1016/j.jweia.2020.104332]

5. Zhang, S.; Guo, K.; Yang, Q.; Xu, X. Review of Wind Field Characteristics of Downbursts and Wind Effects on Structures under Their Action. Buildings; 2024; 14, 2653. [DOI: https://dx.doi.org/10.3390/buildings14092653]

6. Lombardo, T.F.; Zickar, S.A. Characteristics of measured extreme thunderstorm near-surface wind gusts in the United States. J. Wind Eng. Ind. Aerodyn.; 2019; 193, 103961. [DOI: https://dx.doi.org/10.1016/j.jweia.2019.103961]

7. Burlando, M.; Zhang, S.; Solari, G. Monitoring, cataloguing and weather scenarios of thunderstorm outflows in the Northern Mediterranean. Nat. Hazards Earth Syst. Sci.; 2018; 18, pp. 2309-2330. [DOI: https://dx.doi.org/10.5194/nhess-18-2309-2018]

8. Fang, Z.; Wang, Z.; Li, Z.; Yan, J.; Huang, H. Wind field characteristics of stationary and moving downbursts based on the test of impinging jet with a movable nozzle. J. Wind Eng. Ind. Aerodyn.; 2023; 232, 105266. [DOI: https://dx.doi.org/10.1016/j.jweia.2022.105266]

9. Zhang, A.L.; Zhang, S.; Xu, X.D.; Zhong, H.B.; Li, B. Variation characteristics of the wind field in a typical thunderstorm event in Beijing. Appl. Sci.; 2022; 12, 12036. [DOI: https://dx.doi.org/10.3390/app122312036]

10. Burlando, M.; Romanić, D.; Solari, G.; Hangan, H.; Zhang, S. Field data analysis and weather scenario of a downburst event in Livorno, Italy, on 1 October 2012. Mon. Weather Rev.; 2017; 145, pp. 3507-3527. [DOI: https://dx.doi.org/10.1175/MWR-D-17-0018.1]

11. Byers, H.R.; Braham, R.R. The Thunderstorm: Final Report of the Thunderstorm Project; US Government Printing Office: Washington, DC, USA, 1949.

12. Fujita, T.T.; Wakimoto, R.M. Five scales of airflow associated with a series of downbursts on 16 July 1980. Mon. Weather Rev.; 1981; 109, pp. 1438-1456. [DOI: https://dx.doi.org/10.1175/1520-0493(1981)109<1438:FSOAAW>2.0.CO;2]

13. Fujita, T.T. Downbursts: Meteorological features and wind field characteristics. J. Wind Eng. Ind. Aerodyn.; 1990; 36, pp. 75-86. [DOI: https://dx.doi.org/10.1016/0167-6105(90)90294-M]

14. Hjelmfelt, M.R. Structure and life cycle of microburst outflows observed in Colorado. J. Appl. Meteorol. Climatol.; 1988; 27, pp. 900-927. [DOI: https://dx.doi.org/10.1175/1520-0450(1988)027<0900:SALCOM>2.0.CO;2]

15. Chen, L.; Letchford, C.W. Numerical simulation of extreme winds from thunderstorm downbursts. J. Wind Eng. Ind. Aerodyn.; 2007; 95, pp. 977-990. [DOI: https://dx.doi.org/10.1016/j.jweia.2007.01.021]

16. Solari, G.; Burlando, M.; DE Gaetano, P.; Pia Repetto, M. Characteristics of thunderstorms relevant to the wind loading of structures. Wind Struct.; 2015; 20, pp. 763-791. [DOI: https://dx.doi.org/10.12989/was.2015.20.6.763]

17. Shu, Z.R.; Li, Q.S.; He, Y.C.; Chan, P.W. Gust factors for tropical cyclone, monsoon and thunderstorm winds. J. Wind Eng. Ind. Aerodyn.; 2015; 142, pp. 1-14. [DOI: https://dx.doi.org/10.1016/j.jweia.2015.02.003]

18. Choi, E.C. Extreme wind characteristics over Singapore–an area in the equatorial belt. J. Wind Eng. Ind. Aerodyn.; 1999; 83, pp. 61-69. [DOI: https://dx.doi.org/10.1016/S0167-6105(99)00061-6]

19. Choi, E.C. Wind characteristics of tropical thunderstorms. J. Wind Eng. Ind. Aerodyn.; 2000; 84, pp. 215-226. [DOI: https://dx.doi.org/10.1016/S0167-6105(99)00054-9]

20. Gunter, W.S.; Schroeder, J.L. High-resolution full-scale measurements of thunderstorm outflow winds. J. Wind Eng. Ind. Aerodyn.; 2015; 138, pp. 13-26. [DOI: https://dx.doi.org/10.1016/j.jweia.2014.12.005]

21. Zhang, S.; Yang, Q.S.; Solari, G.; Li, B.; Huang, G.Q. Characteristics of thunderstorm outflows in Beijing urban area. J. Wind Eng. Ind. Aerodyn.; 2019; 195, 104011. [DOI: https://dx.doi.org/10.1016/j.jweia.2019.104011]

22. GB 50009-2012 Code for Design Loads of Buildings and Other Civil Engineering Structures; China Architecture & Building Press: Beijing, China, 2012.

23. DE Gaetano, P.; Repetto, M.P.; Repetto, T.; Solari, G. Separation and classification of extreme wind events from anemometric records. J. Wind Eng. Ind. Aerodyn.; 2014; 126, pp. 132-143. [DOI: https://dx.doi.org/10.1016/j.jweia.2014.01.006]

24. Huang, G.; Jiang, Y.; Peng, L.; Solari, G.; Liao, H.; Li, M. Characteristics of intense winds in mountain area based on field measurement: Focusing on thunderstorm winds. J. Wind Eng. Ind. Aerodyn.; 2019; 190, pp. 166-182. [DOI: https://dx.doi.org/10.1016/j.jweia.2019.04.020]

25. Abd-Elaal, E.S.; Mills, J.E.; Ma, X. An analytical model for simulating steady state flows of downburst. J. Wind Eng. Ind. Aerodyn.; 2013; 115, pp. 53-64. [DOI: https://dx.doi.org/10.1016/j.jweia.2013.01.005]

26. Peng, L.; Huang, G.; Chen, X.; Yang, Q. Evolutionary spectra-based time-varying coherence function and application in structural response analysis to downburst winds. J. Struct. Eng.; 2018; 144, 04018078. [DOI: https://dx.doi.org/10.1061/(ASCE)ST.1943-541X.0002066]

27. Flay, R.G.J.; Stevenson, D.C. Integral length scales in strong winds below 20 m. J. Wind Eng. Ind. Aerodyn.; 1988; 28, pp. 21-30. [DOI: https://dx.doi.org/10.1016/0167-6105(88)90098-0]

28. Xu, X.; Zhang, S. Property of a typical urban thunderstorm outflow relevant to wind load on structures. IOP Conf. Ser. Earth Environ. Sci.; 2019; 218, 012086. [DOI: https://dx.doi.org/10.1088/1755-1315/218/1/012086]

29. Li, B.; Li, C.; Yang, Q.; Tian, Y.; Zhang, X. Full-scale wind speed spectra of 5 Year time series in urban boundary layer observed on a 325 m meteorological tower. J. Wind Eng. Ind. Aerodyn.; 2021; 218, 104791. [DOI: https://dx.doi.org/10.1016/j.jweia.2021.104791]

30. Davenport, A.G. The Spectrum of Horizontal Gustiness Near the Ground in High Winds. Q. R. Met Soc.; 1961; 87, pp. 194-211. [DOI: https://dx.doi.org/10.1002/qj.49708737208]

31. Roncallo, L.; Solari, G. An evolutionary power spectral density model of thunderstorm outflows consistent with real-scale time-history records. J. Wind Eng. Ind. Aerodyn.; 2020; 203, 104204. [DOI: https://dx.doi.org/10.1016/j.jweia.2020.104204]

32. Morfiadakis, E.E.; Glinou, G.L.; Koulouvari, M.J. The suitability of the von Karman spectrum for the structure of turbulence in a complex terrain wind farm. J. Wind Eng. Ind. Aerodyn.; 1996; 62, pp. 237-257. [DOI: https://dx.doi.org/10.1016/S0167-6105(96)00059-1]

33. Priestley, M.B. Evolutionary spectra and non-stationary processes. J. R. Stat. Soc. Ser. B (Methodol.); 1965; 27, pp. 204-229. [DOI: https://dx.doi.org/10.1111/j.2517-6161.1965.tb01488.x]

34. Chen, L.; Letchford, W.C. A deterministic–stochastic hybrid model of downbursts and its impact on a cantilevered structure. Eng. Struct.; 2003; 26, pp. 619-629. [DOI: https://dx.doi.org/10.1016/j.engstruct.2003.12.009]