11:00 AM - 1:00 PM
[MTT45-P04] Development of an atmospheric pressure observation system for observing Lamb waves excited by ground motion
Keywords:M5Stack, Lamb waves, tsunamis, Tonga volcano, microbarometer, early warning
Introduction
Lamb waves are surface trapped atmospheric waves that propagate almost non-dispersively at the sound speed averaged in the lower atmosphere. Lamb waves (about 311m/s) is faster than tsunami (about 200m/s) and they retain the shape of source due to their almost non-dispersive properties, as was demonstrated in the case of the Lamb waves associated with the great Tohoku earthquake and tsunamis (Arai et al. 2011). Therefore, the development of an extensive pressure measurement network for the observation of Lamb waves will be useful for the early warning of tsunamis. In this study, we developed an inexpensive and massively installable atmospheric pressure observation system to detect Lamb waves, and verified its accuracy.
Outline of the developed system
We choose infenion DPS310 as the pressure sensor. It is connected to a microcomputer, M5Stack Basic or ATOMLite, which sends measurement commands to the DPS310 at regular intervals via I2C communication. The collected raw values are converted to pressure data in SI units with temperature correction by the microcomputer and sent via UDP communication to a Linux server which stores the data as text files. The clock in the microcomputer is synchronized with the NTP server of the National Institute of Information and Communications Technology (NICT) every hour. The microcomputer automatically reboots every month to prevent the internal timer overflow arising in a long period of operation. The physical dimension is about 10cm, and the installation cost per unit is less than 10,000 yen.
The DPS310 barometric pressure sensor has an oversampling (OS) function. The OS number must be chosen with compromise; larger OS number results in more accurate but less frequent measurements. Further, when OS number is small, the number of data bits is small, so that the quantization error is large. Based on the results of a number of trials, we set the pressure OS to 8times, the temperature OS to 4times, and the measurement frequency to every 35ms as optimal settings for Lamb wave observations.
The verification of accuracies
The time accuracy was verified by setting eight of the developed devices on a table in a room whose door is open/close to cause pressure fluctuation. By comparing the pressure signals observed by the devices, the relative time accuracy of the system was found to be within 20ms. The accuracy of the barometric pressure measurements was estimated in the same setting as the difference between the data obtained from each devices and the ensemble average of the data obtained from all devices. The results showed that the accuracy of the pressure measurements was about 0.5Pa. This indicates that pressure fluctuations of several tens of Pa, such as Lamb waves excited by tsunami, can be easily detected. Since the installation cost per unit is low, multiple units can be installed in a single location to increase the accuracy.
The case of Hunga Tonga-Hunga Ha’apai volcano
On January 15, at around 13:10 Japan time, Hunga Tonga-Hunga Ha’apai volcano near the Tonga Islands erupted, and the lamb wave generated by the eruption was observed by the installed system. The details will be presented at the conference.
Lamb waves are surface trapped atmospheric waves that propagate almost non-dispersively at the sound speed averaged in the lower atmosphere. Lamb waves (about 311m/s) is faster than tsunami (about 200m/s) and they retain the shape of source due to their almost non-dispersive properties, as was demonstrated in the case of the Lamb waves associated with the great Tohoku earthquake and tsunamis (Arai et al. 2011). Therefore, the development of an extensive pressure measurement network for the observation of Lamb waves will be useful for the early warning of tsunamis. In this study, we developed an inexpensive and massively installable atmospheric pressure observation system to detect Lamb waves, and verified its accuracy.
Outline of the developed system
We choose infenion DPS310 as the pressure sensor. It is connected to a microcomputer, M5Stack Basic or ATOMLite, which sends measurement commands to the DPS310 at regular intervals via I2C communication. The collected raw values are converted to pressure data in SI units with temperature correction by the microcomputer and sent via UDP communication to a Linux server which stores the data as text files. The clock in the microcomputer is synchronized with the NTP server of the National Institute of Information and Communications Technology (NICT) every hour. The microcomputer automatically reboots every month to prevent the internal timer overflow arising in a long period of operation. The physical dimension is about 10cm, and the installation cost per unit is less than 10,000 yen.
The DPS310 barometric pressure sensor has an oversampling (OS) function. The OS number must be chosen with compromise; larger OS number results in more accurate but less frequent measurements. Further, when OS number is small, the number of data bits is small, so that the quantization error is large. Based on the results of a number of trials, we set the pressure OS to 8times, the temperature OS to 4times, and the measurement frequency to every 35ms as optimal settings for Lamb wave observations.
The verification of accuracies
The time accuracy was verified by setting eight of the developed devices on a table in a room whose door is open/close to cause pressure fluctuation. By comparing the pressure signals observed by the devices, the relative time accuracy of the system was found to be within 20ms. The accuracy of the barometric pressure measurements was estimated in the same setting as the difference between the data obtained from each devices and the ensemble average of the data obtained from all devices. The results showed that the accuracy of the pressure measurements was about 0.5Pa. This indicates that pressure fluctuations of several tens of Pa, such as Lamb waves excited by tsunami, can be easily detected. Since the installation cost per unit is low, multiple units can be installed in a single location to increase the accuracy.
The case of Hunga Tonga-Hunga Ha’apai volcano
On January 15, at around 13:10 Japan time, Hunga Tonga-Hunga Ha’apai volcano near the Tonga Islands erupted, and the lamb wave generated by the eruption was observed by the installed system. The details will be presented at the conference.