Calibration shifts in Scintrex CG‐3M gravimeters with an application to detection of microgravity changes at Iwo‐tou caldera,Japan

Calibration shift seriously influences gravity values measured using Scintrex CG‐3M gravimeters. We calibrated three Scintrex CG‐3M gravimeters three times (1999, 2003 and 2006) over eight years, using a calibration line with a gravity difference of 1.38 Gal. The scale factor correction coefficients (calibration factors) obtained here range from 0.9998–1.0005. The calibration factors vary with time by 89 ppm, ?102 ppm and ?126 ppm between the 1999–2003 surveys. The calibration shifts of two of the three gravimeters decreased to about 20 ppm or less in the second interval, the other remained about the same (142 ppm). The results indicate that they shifted at rates on the order of 10 ppm/year even several years after manufacturing. The large shift in calibration factors indicates that they must be corrected using calibrations done before and after the measurements to perform microgravity measurements when gravity differences between a reference gravity site and survey sites are on the order of a hundred milligals (mGal) or more. The results also indicate that the calibration factors change gradually with time, so their interpolation provides a good practical approximation for a specific survey time. We applied the time‐dependent calibration factors to microgravity monitoring at the Iwo‐tou caldera, Japan, where the gravity difference between the base site on the island and the reference site on Honshu (the mainland of Japan) is about 870 mGal mainly due to the 11° latitude difference. Gravity surveys were conducted every two years from 1998–2006. The correction of scale factors estimated from the repeated calibration surveys leads to satisfactory measurements, in which the average of the absolute differences between two Scintrex CG‐3M instruments in five surveys is reduced from 207μGal to 19μGal; for three of those surveys, it is less than 10 μGal. This result demonstrates the importance of repeated calibration surveys.