9:45 AM - 10:00 AM
[MTT39-04] Characteristics of pressure disturbances observed in an ultra high-resolution 100 point microbarometor network
Keywords:micro-pressure fluctuation, wind measurement, infrasound, turbulence, boundary layer
1. Background
Imada and Nakajima (2022)[1] developed an inexpensive micro-pressure observation system that can be installed in large numbers. The system measures pressure every 35 ms with an accuracy of 0.5 Pa, and may be able to detect fluctuations caused by thunders, turbulence, etc. Hiramine et al. (2023)[2] made 100 of such micro-barometers and conducted narrow-area concentrated field observations. Hiramine et al. (2024)[3] investigated the spatiotemporal spectrum and the relationship between the speed of isolated disturbances and wind. Here we will add spatiotemporal correlation analysis and discuss the relationship between wind and pressure disturbances comprehensively.
2. Field observation
The observation was performed at a flat square in Kyushu University from 15:12 to 17:20 on April 8, 2023. Micro-barometers were placed at 10-times-10 grid points with 0.295 m intervals. Each micro-pressure gauge consists of a capacitive MEMS pressure sensor, DPS310, a microcomputer, M5Stack-ATOM Lite. Data was collected on a Debian GNU/Linux PC via a Wi-Fi-connected UDP. An ultrasonic anemometer was placed at about 2.8 m from the observation area to measure the wind at 1.5 m above the ground. The following shows the results of the analysis of 15:12~15:23, 15:32~15:44, 16:10~16:44, and 17:10~17:20, for which data were obtained every 35 ms for all 100 units.
3. Results
3.1 Spatiotemporal spectrum
The power spectrum was created by performing a three-dimensional Fourier transform, rotating the horizontal wavenumber axis to east-west and north-south, and integrating the results for the east-west/north-east wavenumber. In general, the power was higher at lower wavenumbers and frequencies, but the power was biased toward the component corresponding to the mean wind direction (Fig. 1), and the bias shifted toward higher frequencies as the wind speed increased.
3.2 Tracking of isolated disturbances
At each time, the location of minimum or maximum pressure field is searched as a candidate of localized disturbance. The center positions of disturbances were estimated by the center of gravity of the 9 upper/lower pressure points. Examining the temporal continuity, we identified 64 isolated disturbances that lasted longer than 0.5 seconds in the area. 47 of these were low pressure, suggesting the existence of vortices with a diameter around 1 m (Fig. 2). The velocity of each disturbance was obtained by least-squares fitting its trajectory to a linear function of time. Compared with the observation by the ultrasonic anemometer, the direction of disturbance movement generally corresponded to the wind direction, but the magnitude of the disturbance speed was about half that of the wind (Fig.3).
3.3 Spatiotemporal correlation
For short and long-period variations whose characteristic frequency bands are 0.5-1 Hz and 0.05-0.1 Hz, respectively, two-point lag correlations were calculated for one point near the center of the area (reference point) and each of the other points (Fig. 4a). The distribution of delay times of the maximum correlation coefficient for each point showed a plane wave-like structure, and the propagation speed was about twice that of the wind. The direction was generally consistent with the wind, but there were times when the direction differed by nearly 90 degrees in the long-period variation. The horizontal scale of the disturbance, estimated as the product of the correlation time in autocorrelation function (Fig. 4b) and the propagation speed, was about 6 to 10 (60 to 160) meters for short(long)-period.
4. Discussion
The movement of the disturbances obtained in each analysis did not necessarily correspond to the observed winds (Fig. 5). Bearing the advective nature of vorticity in mind, the movement of isolated disturbances can possibly correspond to wind speeds at heights much closer to the ground. Conversely, the characteristic velocity found in the spatiotemporal correlation may correspond to a wind field much taller than the anemometer height because their estimated horizontal scales, which presumably are comparable to their vertical scales, were much larger than the observation network.
Reference
[1] E. Imada and K. Nakajima, MTT45-P04, Japan Geoscience Union Annual Meeting 2022.
[2] T. Hiramine, E. Imada and K. Nakajima, MTT37-P03, Japan Geoscience Union Annual Meeting 2023.
[3] T. Hiramine, E. Imada and K. Nakajima, MTT38-P03, Japan Geoscience Union Annual Meeting 2024.
Acknowledgements
This research was supported by a grant from JSPS Scientific Research Grants 22K18872.
Imada and Nakajima (2022)[1] developed an inexpensive micro-pressure observation system that can be installed in large numbers. The system measures pressure every 35 ms with an accuracy of 0.5 Pa, and may be able to detect fluctuations caused by thunders, turbulence, etc. Hiramine et al. (2023)[2] made 100 of such micro-barometers and conducted narrow-area concentrated field observations. Hiramine et al. (2024)[3] investigated the spatiotemporal spectrum and the relationship between the speed of isolated disturbances and wind. Here we will add spatiotemporal correlation analysis and discuss the relationship between wind and pressure disturbances comprehensively.
2. Field observation
The observation was performed at a flat square in Kyushu University from 15:12 to 17:20 on April 8, 2023. Micro-barometers were placed at 10-times-10 grid points with 0.295 m intervals. Each micro-pressure gauge consists of a capacitive MEMS pressure sensor, DPS310, a microcomputer, M5Stack-ATOM Lite. Data was collected on a Debian GNU/Linux PC via a Wi-Fi-connected UDP. An ultrasonic anemometer was placed at about 2.8 m from the observation area to measure the wind at 1.5 m above the ground. The following shows the results of the analysis of 15:12~15:23, 15:32~15:44, 16:10~16:44, and 17:10~17:20, for which data were obtained every 35 ms for all 100 units.
3. Results
3.1 Spatiotemporal spectrum
The power spectrum was created by performing a three-dimensional Fourier transform, rotating the horizontal wavenumber axis to east-west and north-south, and integrating the results for the east-west/north-east wavenumber. In general, the power was higher at lower wavenumbers and frequencies, but the power was biased toward the component corresponding to the mean wind direction (Fig. 1), and the bias shifted toward higher frequencies as the wind speed increased.
3.2 Tracking of isolated disturbances
At each time, the location of minimum or maximum pressure field is searched as a candidate of localized disturbance. The center positions of disturbances were estimated by the center of gravity of the 9 upper/lower pressure points. Examining the temporal continuity, we identified 64 isolated disturbances that lasted longer than 0.5 seconds in the area. 47 of these were low pressure, suggesting the existence of vortices with a diameter around 1 m (Fig. 2). The velocity of each disturbance was obtained by least-squares fitting its trajectory to a linear function of time. Compared with the observation by the ultrasonic anemometer, the direction of disturbance movement generally corresponded to the wind direction, but the magnitude of the disturbance speed was about half that of the wind (Fig.3).
3.3 Spatiotemporal correlation
For short and long-period variations whose characteristic frequency bands are 0.5-1 Hz and 0.05-0.1 Hz, respectively, two-point lag correlations were calculated for one point near the center of the area (reference point) and each of the other points (Fig. 4a). The distribution of delay times of the maximum correlation coefficient for each point showed a plane wave-like structure, and the propagation speed was about twice that of the wind. The direction was generally consistent with the wind, but there were times when the direction differed by nearly 90 degrees in the long-period variation. The horizontal scale of the disturbance, estimated as the product of the correlation time in autocorrelation function (Fig. 4b) and the propagation speed, was about 6 to 10 (60 to 160) meters for short(long)-period.
4. Discussion
The movement of the disturbances obtained in each analysis did not necessarily correspond to the observed winds (Fig. 5). Bearing the advective nature of vorticity in mind, the movement of isolated disturbances can possibly correspond to wind speeds at heights much closer to the ground. Conversely, the characteristic velocity found in the spatiotemporal correlation may correspond to a wind field much taller than the anemometer height because their estimated horizontal scales, which presumably are comparable to their vertical scales, were much larger than the observation network.
Reference
[1] E. Imada and K. Nakajima, MTT45-P04, Japan Geoscience Union Annual Meeting 2022.
[2] T. Hiramine, E. Imada and K. Nakajima, MTT37-P03, Japan Geoscience Union Annual Meeting 2023.
[3] T. Hiramine, E. Imada and K. Nakajima, MTT38-P03, Japan Geoscience Union Annual Meeting 2024.
Acknowledgements
This research was supported by a grant from JSPS Scientific Research Grants 22K18872.