Studies of Earth Dynamics with the Superconducting Gravimeter

I present investigations of Earth dynamics with the superconducting gravimeter GWR T020 at Metsahovi from 1994 to 2005. As a necessary background, the history and key technical details of the installation are given. Then a new data acquisition system (DAS) developed by the author is described. It has proved reliable and accurate. A pre-requisite for the study of weak phenomena is the removal of the tides from the gravity record and the careful treatment of spikes, offsets and drifts. I present the data processing methods and the development of the local tidal model at Metsahovi and show that improvements help in the modelling of long-term environmental effects. First, I have used the T020 as a long-period seismometer to study microseismicity and the Earth’s free oscillation. I present the annual variation, spectral distribution, amplitude and the sources of microseism at Metsahovi. High levels of microseism appear to be mainly generated by low-pressure areas (storms) in the Northern Atlantic. I have analyzed free oscillations excited by three large earthquakes: the spectra, attenuation and rotational splitting of the modes. An air pressure correction is required to bring down the noise level. The lowest modes of all different oscillation types are studied, i.e. the radial mode 0S0 , the “football mode” 0S2, and the toroidal mode 0T2. This last-mentioned mode is very weak in the vertical as the component is only generated through the Coriolis coupling. I have also detected the very low level (0.01 nms -1 ) incessant excitation of the Earth’s free oscillation with the T020. The recovery of global and regional variations in gravity with the superconducting gravimeter requires the modelling of local gravity effects. The most important of them is hydrology. The variation in the groundwater level at Metsahovi as measured in a borehole in the fractured bedrock correlates significantly (0.79) with gravity. The range of the gravity effect is 70 nms -2 and this phenomenon must be taken into account at an early stage of data processing. The influence of local precipitation, soil moisture and snow cover are detectable in the gravity record. A fortuitous experiment powerfully demonstrated the effect of snow on the laboratory roof, on that occasion -20 nms -2 . The gravity effect of the variation in atmospheric mass (range about 300 nms -2 ) and that of the non-tidal loading by the Baltic Sea (range about 50 nms -2 ) were investigated together, as sea level and air pressure are correlated. The effects were modelled both using regression on the local barometer and tide gauge, and with a Green’s function formalism based a detailed model of the load. In the regression approach, there is a trade-off between the coefficients. Using Green’s functions it was calculated that a 1 metre uniform layer of water in the Baltic Sea increases the gravity at Metsahovi by 31 nms -2 and the vertical deformation is –11 mm. Best results were obtained using the HIRLAM (High Resolution Limited Area Model) for the atmosphere. Regression on the sea level in Helsinki then decreases the RMS of gravity residuals by 16%, i.e. explains about 30% of their variance. The regression coefficient for sea level is 27 nms -2 m -1 , which is 87% of the uniform model. These studies are associated with temporal height variations using the GPS data of Metsahovi permanent station. Loading by air pressure and the Baltic Sea explains about 40% of the variance of daily GPS height solution.