Simultaneous calibration of instrument scale factor and drift rate in network adjustment for continental-scale gravity survey campaign

Abstract

SUMMARY The scale factor of each relative gravimeter must be calibrated both before and after the fieldwork of a terrestrial gravity survey, to reduce uncertainties and ensure high precision. Conventionally, such calibration is a time-consuming process performed following well-established baselines. We propose a new Bayesian method to estimate the scale factor in a hybrid gravity network that includes several absolute gravity observation stations. In this approach, the scale factor is estimated as a hyperparameter using the Akaike Bayesian information criterion and using known absolute gravity stations in the network or/and calibrated instruments as constraints. Testing the sensitivity of the gravity values and the residuals of the gravity difference between two successive stations to the change of the scale factor demonstrates the robustness of this method. We also test the sensitivity of the estimated scale factor in the presence of Gaussian noise and the non-linear instrumental drift rate. Moreover, if the maximum absolute gravity interval is greater than 60 per cent of the range of gravity values in the network, or if the known scale factors of calibrated gravimeters are well calibrated, this approach can provide reasonable estimates of the daily drift rate and the unknown scale factors, where the latter has an error of <3 × 10−5. We apply this approach to real gravity campaign data from Yunnan in China and use a cross-validation method to compare estimated gravity values and corresponding gravity values obtained from absolute gravity observations at the same stations, to validate how the proposed method improves estimation accuracy of the gravity value at each station.