10:45 AM - 11:00 AM
[SVC33-07] The real-time correction method for the JMA Multi-GAS data and application to volcano monitoring and assessment
Keywords:Multi-GAS, real-time correction, volcanic gas b value, Azumayama, Kusatsu-Shiranesan
Introduction
Since 2015, the Japan Meteorological Agency has installed multi-component volcanic gas continuous observation system (hereinafter referred to Multi-GAS) at four volcanoes in Japan to monitor the concentration of volcanic gases (Seismology and Volcanology Department, JMA, 2022). This continuous and telemetric monitoring system was improved from the prototypal field-portable Multi-GAS system (Shonohara et al., 2005).
Issues in correction methods for changes in sensor sensitivity
Multi-GAS has gas sensors that use a controlled-potential electrolysis. This type of sensor experiences a decrease in voltage sensitivity with continued use, especially for hydrogen sulfide. To solve this problem, we replaced the sensor twice a year and calibrated it before and after the period of use. Assuming that the rate of change in sensitivity is constant during the period of use, we tried a method of linear correction (Kitagawa et al., 2022). However, this approach does not make the data available for monitoring until calibration is done. Therefore, development of a real-time correction method is desired.
Development of a real-time correction method
In order to realize real-time correction, it was necessary to fully understand the characteristics of the sensor under long-term continuous gas exposure. Therefore, we conducted experiments in which the sensor was artificially exposed under various conditions of exposure gas concentration and humidity, and grasped the temporal behavior of changes in sensor sensitivity. There are 16 combinations of exposure experiments: hydrogen sulfide exposure concentrations of 0, 5, 10, and 15 ppm, and mixed water vapor concentrations of 0.5, 1.0, 1.5, and 2.0%. The hydrogen sulfide sensor was exposed for approximately 120 hours. As a result, all sensitivity changes showed exponential behavior depending on duration of exposure. In addition, it was revealed that the higher the exposure concentration or the lower the humidity, the faster the sensitivity decreased. These exposure concentration dependence and humidity dependence were approximated by an exponential function. The coefficients of the function were determined by the least-squares method, and modeled as a sensitivity change function with exposure time, exposure gas concentration, and humidity as variables. With this model function, it became possible to calculate the sequential sensitivity change using the observed data, and the development of the real-time correction method was realized. By applying this method to a series of observation data, it was confirmed that the gap at the time of sensor replacement was almost eliminated. There was a prospect of real-time monitoring.
Possibility of new monitoring and evaluation method for volcanic activity using Multi-GAS continuous data
By using continuous volcanic gas concentration data, there is a possibility of a new technical method of monitoring and evaluation.
Evaluation of volcanic activity based on understanding of hydrothermal systems using concentration ratios of multiple components
It may be possible to evaluate volcanic activity while understanding volcanic activity by using not only the concentration ratio of two kinds of volcanic gases but also various concentration ratios. It is possible to show daily changes in the concentration ratios of the three gas components (CO2, H2S, SO2) on a triangular diagram (Stix and de Moor, 2018). This may provide a relative understanding of whether volcanic activity is magmatic or hydrothermal. The presentation will show examples of volcanic activities at Mt. Azumayama in 2019.
Proposed activity evaluation method using gas b-value based on frequency distribution for every concentration grade
The gas observed by Multu-GAS is relatively low in concentration because it is blown down by the wind. However, by analyzing the frequency gradient from the cumulative frequency distribution of the concentration obtained every second, it may be possible to know whether the current activity is during the period when the emission rate of high-concentration gas is dominant or not. This is similar to the b-value used as an index of seismic activity. Here we call it the gas b-value. Based on this idea, we will present a recent example of Mt. Kusatsu-Shiranesan.
Since 2015, the Japan Meteorological Agency has installed multi-component volcanic gas continuous observation system (hereinafter referred to Multi-GAS) at four volcanoes in Japan to monitor the concentration of volcanic gases (Seismology and Volcanology Department, JMA, 2022). This continuous and telemetric monitoring system was improved from the prototypal field-portable Multi-GAS system (Shonohara et al., 2005).
Issues in correction methods for changes in sensor sensitivity
Multi-GAS has gas sensors that use a controlled-potential electrolysis. This type of sensor experiences a decrease in voltage sensitivity with continued use, especially for hydrogen sulfide. To solve this problem, we replaced the sensor twice a year and calibrated it before and after the period of use. Assuming that the rate of change in sensitivity is constant during the period of use, we tried a method of linear correction (Kitagawa et al., 2022). However, this approach does not make the data available for monitoring until calibration is done. Therefore, development of a real-time correction method is desired.
Development of a real-time correction method
In order to realize real-time correction, it was necessary to fully understand the characteristics of the sensor under long-term continuous gas exposure. Therefore, we conducted experiments in which the sensor was artificially exposed under various conditions of exposure gas concentration and humidity, and grasped the temporal behavior of changes in sensor sensitivity. There are 16 combinations of exposure experiments: hydrogen sulfide exposure concentrations of 0, 5, 10, and 15 ppm, and mixed water vapor concentrations of 0.5, 1.0, 1.5, and 2.0%. The hydrogen sulfide sensor was exposed for approximately 120 hours. As a result, all sensitivity changes showed exponential behavior depending on duration of exposure. In addition, it was revealed that the higher the exposure concentration or the lower the humidity, the faster the sensitivity decreased. These exposure concentration dependence and humidity dependence were approximated by an exponential function. The coefficients of the function were determined by the least-squares method, and modeled as a sensitivity change function with exposure time, exposure gas concentration, and humidity as variables. With this model function, it became possible to calculate the sequential sensitivity change using the observed data, and the development of the real-time correction method was realized. By applying this method to a series of observation data, it was confirmed that the gap at the time of sensor replacement was almost eliminated. There was a prospect of real-time monitoring.
Possibility of new monitoring and evaluation method for volcanic activity using Multi-GAS continuous data
By using continuous volcanic gas concentration data, there is a possibility of a new technical method of monitoring and evaluation.
Evaluation of volcanic activity based on understanding of hydrothermal systems using concentration ratios of multiple components
It may be possible to evaluate volcanic activity while understanding volcanic activity by using not only the concentration ratio of two kinds of volcanic gases but also various concentration ratios. It is possible to show daily changes in the concentration ratios of the three gas components (CO2, H2S, SO2) on a triangular diagram (Stix and de Moor, 2018). This may provide a relative understanding of whether volcanic activity is magmatic or hydrothermal. The presentation will show examples of volcanic activities at Mt. Azumayama in 2019.
Proposed activity evaluation method using gas b-value based on frequency distribution for every concentration grade
The gas observed by Multu-GAS is relatively low in concentration because it is blown down by the wind. However, by analyzing the frequency gradient from the cumulative frequency distribution of the concentration obtained every second, it may be possible to know whether the current activity is during the period when the emission rate of high-concentration gas is dominant or not. This is similar to the b-value used as an index of seismic activity. Here we call it the gas b-value. Based on this idea, we will present a recent example of Mt. Kusatsu-Shiranesan.