GJI Geomagnetism, rock magnetism and palaeomagnetism Decomposition analysis of the BEAR magnetotelluric data: implications for the upper mantle conductivity in the Fennoscandian Shield

SUMMARY Long-period magnetotelluric (MT) and geomagnetic depth sounding data (GDS) have been acquired on the Fennoscandian Shield under the framework of the Baltic Electromagnetic Array Research (BEAR) project. The field campaign was carried out in the summer of 1998 when variations of the natural electromagnetic field were recorded simultaneously at 46 MT and 20 GDS stations. The key targets of the project are to investigate the electrical properties of the upper mantle and to determine the depth to the lithosphere‐asthenosphere boundary in the Fennoscandian craton. A challenging task emerges from the fact that numerous highly conductive crustal bodies and local conductivity contrasts generate galvanic and inductive distortions to the calculated transfer functions in the research area. We present here a systematic decomposition and dimensionality analysis of the BEAR data and use the results of this analysis to verify regions for which 1-D inversion is justified. We argue that most of the BEAR data represent regional 2-D and 3-D structures with local galvanic distortion. The decomposition of the long-period (T > 3000 s) MT impedance tensors yield a set of smoothly varying regional strike directions. Yet strike angles vary significantly in the scale of the BEAR array and have abrupt regional changes in some areas. The spatial behaviour of strike angles cannot be connected with largescale geological units. Moreover, strong variation of strike azimuths over the BEAR array convincingly shows that the strike angles cannot be associated with present day plate motion or mantle convection, because that would require a consistent strike azimuth over the whole array. Observed long-period strike angles indicate mainly upper mantle 2-D and 3-D structures or frozen in anisotropy induced by several Palaeoproterozoic and Archaean events. The dimensionality analysis of the BEAR data shows that in the northeastern part of the array the regional structure is approximately 1-D. 1-D inversion of selected data from the western Lapland-Kola Domain reveals a conducting layer in the middle crust. An increase of conductivity is required also at depths greater than 170 km providing a minimum estimate of the lithosphere thickness beneath the target area. Partial melts or dissolved water in olivine are most plausible sources for increased conductivity at such depths.