At least two (one in emergency during storms) of a set of 4 quartz thermally stabilized MOD. GMN-K gravimeters assembled in Russia by VNIIgeofisika were running during multibeam lines on medium or long range transits. The sensors were mounted on gyroscopic platforms at the sea level very close to ship's center of gravity (Tab.3). The instruments have a resolution of 0.04 mGal with an accuracy of 0.2 mgal. Damping by viscous siliceous-organic liquid is used to filter high frequency vertical accelerations. The gravimeter is a-static, having an elastic back connection, and the measuring principle is based on the return-to-horizontal of the system lever actuated by (a) detecting the movement by photoresistors, and (b) turning of the screw connected to the spring. The position of the screw is measured by two voltmeters, digitized and processed by PC every 15 sec during normal survey and every 30 sec during low speed survey, e.g. during dredging. Four gyroscopes connected in pairs along the diagonal, increase the moment of inertia of the pendulum platform, keeping it horizontal within few arcminutes.
The averages of two front-positioned instruments, were used to minimize cross-coupling. Drift of zero point was calculated from the on-board and portable measurements at the berth (see Tab.6). No station in anchor was performed. A few tie lines were run on the predominantly SW-NE multibeam acquisition pattern.
Filtered GPS positions, speed and course over the ground were used to calculate the Eötvös correction. A corrective smoothing procedure[Panteleev (1983)] was used to avoid time-delays and smooth the data. The methematical details can be found in the Appendix 7.1). This procedure can be used in place of the regularization technique. The filter can be built if the model of gravimeter as a dynamical system and the frequency characteristics of the desired signal are known. A first-order model with time constant of 300 sec and a low-pass filter with 1 Hz cut-off frequency were used.
Free-air gravity anomalies were obtained by removal of the GRF80 normal field [Moritz (1984),Torge (1989)], subtracting to the observed values the normal field and applying the Eötvös corrections.