The geoid heights are needed for determining the orthometric heights from the Global Positioning System (GPS) ellipsoidal heights. There are several methods for geoid height determination. The paper presents a method employing the Artificial Neural Network (ANN) approximation together with the Least Squares Collocation (LSC). The surface obtained by the ANN approximation is used as a trend surface in the least squares collocation. In numerical examples four surfaces were compared: the global geopotential model (EGM96), the European gravimetric quasigeoid 1997 (EGG97), the surface approximated with minimum curvature splines in tension algorithm and the ANN surface approximation. The effectiveness of the ANN surface approximation depends on the number of control points. If the number of well distributed control points is sufficiently large, the results are better than those obtained by the minimum curvature algorithm and comparable to those obtained by the EGG97 model.

No abstract

Determination of the correlation between the energy-density changes in geomagnetic field and seismic events is a challenging scientific topic that allows the study of local tectonics using magnetometers. The magnetised properties of the crustal field of the Earth change due to natural sources and/or human activities that affect the environment. The latter can be avoided by setting up observatories in "geomagnetically-quiet" locations, while the natural sources, which describe the combined effects due to changes in the core, lithosphere, external or electromagnetically induced field, cannot be easily eliminated. This research focused on the investigation of local changes in the geomagnetic field in relation to evidently significant local tectonics in the vicinity of the PIA (Piran, Slovenia) geomagnetic observatory. It is obvious that geomagnetic measurements from PIA contain much higher levels of noise compared to the surrounding magnetometers in Italy and Croatia. According to previous geodynamic studies, the position of the PIA observatory is specific, since it is located at an Adriatic microplate that collides under the Eurasian plate. At this point it can be assumed that the reason for high-level noise in geomagnetic observations is due to the still ongoing Adria-Eurasia collisional process. Furthermore, the study of the earthquake on 1 November 2015 with a magnitude of 4.2 and its epicentre 150 km from PIA showed the correlation between higher energy density of the Earth's magnetic field and the earthquake occurrence. From the results acquired by the computational strategy described in the research, it is obvious that, as expected, a few days prior and after the earthquake higher Earth's magnetic field indicated some significant changes in the local geomagnetic field that might occur due to the increased stress in the Earth's crust in the northeastern part of the Adria-Eurasia collisional zone.

The geoid heights are needed for determining the orthometric heights from the Global Positioning System (GPS) ellipsoidal heights. There are several methods for geoid height determination. The paper presents a method employing the Artificial Neural Network (ANN) approximation together with the Least Squares Collocation (LSC). The surface obtained by the ANN approximation is used as a trend surface in the least squares collocation. In numerical examples four surfaces were compared: the global geopotential model (EGM96), the European gravimetric quasigeoid 1997 (EGG97), the surface approximated with minimum curvature splines in tension algorithm and the ANN surface approximation. The effectiveness of the ANN surface approximation depends on the number of control points. If the number of well distributed control points is sufficiently large, the results are better than those obtained by the minimum curvature algorithm and comparable to those obtained by the EGG97 model.

A B S T R A C TSignificant hysteresis effects after transport in the upright position are found during measurement for the purpose of calibration of Scintrex CG-3M gravimeters serial numbers 4372 and 4373. The effects amount to 0.60 and 0.15 μms −2 during the first 10 minutes and an additional 0.35 and 0.10 μms −2 during the next 10 minutes after transport for gravimeters 4372 and 4373, respectively. Such significant and not completely homogenous effects do not completely cancel out when forming gravity differences. An iterative MATLAB algorithm has been developed that models hysteresis effects using exponential functions that account for shifts in time and eliminates the hysteresis effects from the measurements. The method facilitates determination of an exponential function of hysteresis from multiple observation series with different unknown offsets and significant noise that prevents the application of conventional methods. Hysteresis elimination from measurements for the purpose of calibration resulted in (mostly significant) increases in the precision of adjusted calibration constants for both gravimeters. It is shown that, in order to obtain reasonable results, modelling the hysteresis effects should be based on approximately 30 minutes of readings. Finally, one can conclude that hysteresis effects surely do build up during transport, contrary to statements by Scintrex.