5:15 PM - 6:30 PM
[AOS10-P06] Characterization of long-term groundwater level changes associated with the 2016 Kumamoto Earthquake
Keywords:time series analysis
The recharge sources of groundwater are mainly rainfall and river water, and it is known that the groundwater level is affected by changes in the recharge rate from these sources, by atmospheric pressure and tides from the structure of observation wells, and by earthquakes and changes preceding earthquakes. A multivariate linear regression model that does not include the groundwater level itself as an explanatory variable is appropriate for detecting groundwater level changes due to earthquakes or changes preceding earthquakes (Matsumoto et al., 1989). In this study, we aim to extract the characteristics of long-term groundwater level fluctuations in the Kumamoto Plain using a multivariate linear regression model in order to clarify the relationship between long-term groundwater level fluctuations and seismic activity.
In Kumamoto City, groundwater levels are constantly observed at 33 groundwater level observation wells in 20 locations, and this information is available on the website. In this study, we obtained the data directly from the Water Conservation Division, Environment Promotion Department, Environment Bureau, and collected and compiled the data on the website. The daily total of precipitation and daily mean local pressure were obtained from the observation data measured and published by the Japan Meteorological Agency, and the global tide values were calculated using GOTIC2, which is available on the website of the Mizusawa Observatory of the National Astronomical Observatory of Japan. Using a multivariate linear regression model with groundwater level fluctuation as the objective variable and precipitation, atmospheric pressure, and earth tide as explanatory variables, we extracted the characteristics of long-term groundwater level fluctuation up to the Kumamoto earthquake by removing the effects of precipitation, atmospheric pressure, and earth tide.
In contrast to the results of the model calculations based on the 2007-2009 period, the measured values at Suizenji showed a decrease toward 2014 and an increase toward 2016. The strainer of the observation well at Suizenji is a porous Togawa lava formation, and groundwater fluctuations at this location may be influenced by subsurface pressure changes due to crustal movement.
In Kumamoto City, groundwater levels are constantly observed at 33 groundwater level observation wells in 20 locations, and this information is available on the website. In this study, we obtained the data directly from the Water Conservation Division, Environment Promotion Department, Environment Bureau, and collected and compiled the data on the website. The daily total of precipitation and daily mean local pressure were obtained from the observation data measured and published by the Japan Meteorological Agency, and the global tide values were calculated using GOTIC2, which is available on the website of the Mizusawa Observatory of the National Astronomical Observatory of Japan. Using a multivariate linear regression model with groundwater level fluctuation as the objective variable and precipitation, atmospheric pressure, and earth tide as explanatory variables, we extracted the characteristics of long-term groundwater level fluctuation up to the Kumamoto earthquake by removing the effects of precipitation, atmospheric pressure, and earth tide.
In contrast to the results of the model calculations based on the 2007-2009 period, the measured values at Suizenji showed a decrease toward 2014 and an increase toward 2016. The strainer of the observation well at Suizenji is a porous Togawa lava formation, and groundwater fluctuations at this location may be influenced by subsurface pressure changes due to crustal movement.