Noise reduction in magnetotelluric time-series with a new signal–noise separation method and its application to a field experiment in the Saxonian Granulite Massif

摘要: SUMMARY In the presence of large and continuous correlated noise signals in measured electric magnetic time-series, even robust remote-reference methods give erroneous estimates MT transfer functions. If clean remote time-series are available, it is possible to separate derive unbiased functions with signal–noise separation method (SNS) (Larsen et al. 1996). practice, time series also contain some results can be improved by using a second data set SNS-remote-reference technique. We tested this from Saxonian Granulite Massif (SGM), Germany, where strong were detected. used which recorded 350 km away and, for short periods, sites profile across SGM itself (distance 5 km). To show efficiency we first determined ‘true’ function low level. step reproduced processing very noisy sections. able determine useful when variations have less than 10 per cent share variations. identified dominant pipelines trains.