11:00 AM - 11:15 AM
[SGD01-08] Reading value dependence of scale factors for LaCoste & Romberg relative gravimeters
Keywords:gravity, scale factor, relative gravimeter, absolute gravimeter, volcano
Campaign relative gravity measurement is one of the most useful methods to understand spatiotemporal mass variations in volcanoes. The relative gravity value obtained with a spring-type portable relative gravimeter can be expressed as g = f(x), where x and f indicate the reading value (corresponding to the elongation of the spring) and the conversion function provided by the gravimeter’s manufacturer. For the accurate estimation of the relative gravity values, the conversion function has been calibrated using a scale factor (SF; denoted by S) to be g = S * f(x). The SF value can be calculated as the ratio of the absolute gravity difference to the relative one, which are both measured between any two gravity points.
SF has been conventionally regarded as a constant value for each gravimeter, but several studies have been discussing the existence of the time variation in SF these days. For example, Onizawa (2019) pointed out that the SF value of a Scintrex CG5 gravimeter depends on its reading value linearly (i.e., S = S(x)). He also found the time variation in the SF associated with the instrumental drift of the reading value (i.e., S = S(x(t))). However, SF values have not been calibrated for LaCoste & Romberg gravimeters sufficiently, even though they are often used in and around volcanic areas. In order to observe actual spatiotemporal gravity variations due to volcanic activity, S(x) should be determined for each LaCoste gravimeter in advance and its effect should be corrected from the collected gravity data accurately.
Therefore, we calibrated SF values for spring-type relative gravimeters belonging to Kyoto University, to discuss the reading value dependence of the SF values. We first collected the reading values at four gravity points (Ishioka, Kyoto, Aso and Sakurajima) using three LaCoste gravimeters (LC-G534, LC-G605 and LC-G680) and one Scintrex gravimeter (CG5-150241330) in September 2020. We then converted the reading values to the gravity values using the conversion functions provided for the corresponding gravimeters, and subtracted the effects of short-period tide and instrumental height/drift from the obtained gravity values. We finally calculated the SF value as the ratio of the absolute gravity difference to the relative one for each gravimeter and each pair of the gravity points; we here utilized the absolute gravity values at four gravity points measured in 2018 and 2020.
The SF values of four gravimeters were calculated to be in the range of 0.9991 to 1.0003. We also found that the SF values of each gravimeter vary linearly to the reading value of the corresponding gravimeter. By fitting a regression line to the distribution of the SF value to the reading value, the rate of the SF variation to the reading value was estimated for each gravimeter as follows. LC-G534: -2.15 +/- 0.91, LC-G605: -7.09 +/- 0.93, LC-G680: -7.20 +/- 0.92, CG5-150241330: +10.82 +/- 0.90 [E-7 /mGal]. In the case of the LC-G534 and LC-G680 gravimeters, the SF values expected from the regression lines were found to be consistent with those determined in 2018 (Kazama et al., 2019). From these respects, we succeeded in quantifying the linear dependence of the SF on the reading value for the LaCoste gravimeters. In contrast, the past SF value for the CG5 gravimeter (Kazama et al., 2019) differed from the regression line obtained in this study, and the difference exceeded the 95% confidence interval of the regression line. This result may imply the existence of the SF’s higher-order dependence on the reading value in the case of the CG5 gravimeter, so we should determine the SF variation with the wider range of the reading value in our future study.
SF has been conventionally regarded as a constant value for each gravimeter, but several studies have been discussing the existence of the time variation in SF these days. For example, Onizawa (2019) pointed out that the SF value of a Scintrex CG5 gravimeter depends on its reading value linearly (i.e., S = S(x)). He also found the time variation in the SF associated with the instrumental drift of the reading value (i.e., S = S(x(t))). However, SF values have not been calibrated for LaCoste & Romberg gravimeters sufficiently, even though they are often used in and around volcanic areas. In order to observe actual spatiotemporal gravity variations due to volcanic activity, S(x) should be determined for each LaCoste gravimeter in advance and its effect should be corrected from the collected gravity data accurately.
Therefore, we calibrated SF values for spring-type relative gravimeters belonging to Kyoto University, to discuss the reading value dependence of the SF values. We first collected the reading values at four gravity points (Ishioka, Kyoto, Aso and Sakurajima) using three LaCoste gravimeters (LC-G534, LC-G605 and LC-G680) and one Scintrex gravimeter (CG5-150241330) in September 2020. We then converted the reading values to the gravity values using the conversion functions provided for the corresponding gravimeters, and subtracted the effects of short-period tide and instrumental height/drift from the obtained gravity values. We finally calculated the SF value as the ratio of the absolute gravity difference to the relative one for each gravimeter and each pair of the gravity points; we here utilized the absolute gravity values at four gravity points measured in 2018 and 2020.
The SF values of four gravimeters were calculated to be in the range of 0.9991 to 1.0003. We also found that the SF values of each gravimeter vary linearly to the reading value of the corresponding gravimeter. By fitting a regression line to the distribution of the SF value to the reading value, the rate of the SF variation to the reading value was estimated for each gravimeter as follows. LC-G534: -2.15 +/- 0.91, LC-G605: -7.09 +/- 0.93, LC-G680: -7.20 +/- 0.92, CG5-150241330: +10.82 +/- 0.90 [E-7 /mGal]. In the case of the LC-G534 and LC-G680 gravimeters, the SF values expected from the regression lines were found to be consistent with those determined in 2018 (Kazama et al., 2019). From these respects, we succeeded in quantifying the linear dependence of the SF on the reading value for the LaCoste gravimeters. In contrast, the past SF value for the CG5 gravimeter (Kazama et al., 2019) differed from the regression line obtained in this study, and the difference exceeded the 95% confidence interval of the regression line. This result may imply the existence of the SF’s higher-order dependence on the reading value in the case of the CG5 gravimeter, so we should determine the SF variation with the wider range of the reading value in our future study.