Objectives and Methods  Based on the mobile gravity data from 2015 to 2016 in the southern segment of the North-South Seismic Belt, this paper conducts in-depth comparative analysis of classical adjustment and quasi-stable adjustment, and provides a method for evaluating the coefficient of one degree term of relative gravimeter. Two quantitative parameterized indexes, P and R, are defined with the gravity change and error, providing a quantitative basis for analyzing the significance level of the gravity changes in the network.

  Results  Results indicate that accuracy of gravity values is closely related to the shape of gravity network, and long branches should be connected into closed ring as much as possible before data preprocessing. The overall shape of gravity change from 2015 to 2016 calculated by quasi-stable adjustment is in consistent with that by classical adjustment, but the index (P=0.55) is lower than that of classical adjustment (P=1.70). The starting benchmark of classical adjustment is of great importance to the gravity change in the network. Under the same error of gravity change, the significance level of the gravity changes is reduced with fixed gravity value of benchmark compared to that with changing gravity value, with the index P reduced from 1.70 to 0.96.

  Conclusions  The quasi-stable point is crucial to the result of the quasi-stable adjustment, which determines the authenticity of the gravity changes in the network. These results can provide reference for the adjustment calculation of mobile gravity network.