Electric field decomposition Data reduction Data reduction

Noise evaluation

Time series recorded during the 96 hour break in transmission between tows allowed background noise levels on the instruments to be obtained by applying the same least squares fitting procedure as was used to extract data amplitudes. Although there is an implicit assumption that the noise level remained constant over the four days of the CSEM experiment, this approach allows direct comparison with the data. Noise levels are plotted in Figure. 2 at frequencies chosen to cover the full range transmitted during the experiment. The noise values determined for each receiver have been normalised by the source dipole moment to provide estimates that can be compared directly with the data amplitudes shown in later figures.

Sources of noise can be divided into two categories: external and instrumental. External sources include electromagnetic fields from ionospheric and cultural sources, and motionally induced fields caused by microseisms, water currents and tectonic activity. Electric fields decay rapidly in the conductive seawater layer, with the result that ionospheric noise at frequencies greater than 0.03 Hz is effectively screened by the water depths encountered on the Reykjanes Ridge. Similarly, high frequency noise from cultural sources is also removed from the spectrum. Between 0.1 Hz and 2 Hz the dominant sources of external noise are microseisms generated by the non-linear interference of sea surface gravity wave trains (Webb & Cox 1986). These produce pressure fluctuations which can propagate to the seafloor even in deep water, resulting in water and seafloor motions which generate electromagnetic fields. In addition, tectonically generated seismic activity can cause ground and water motions which also induce electromagnetic fields (Webb & Cox, 1982; Webb & Cox, 1986).

Noise generated in the instrument itself by the amplifiers and electrodes is independent of the receiver dipole length and so signal to noise ratio increases with increasing receiver dipole length. The ELFs were substantially noisier than the LEMs due to their shorter receiver dipole length, although the factor of 30 difference in antenna length suggests that the difference between ELF and LEM noise levels should be larger if the observed values were due only to instrumental noise. Kermit was nearly an order of magnitude more noisy than the other ELF instruments because of contamination by noise from an MT magnetometer deployed in tandem with the instrument. Quail, which was re-deployed after releasing prematurely, was much noisier during its second deployment than its first. This could be due either to an external noise source associated with the change in location or to instrumental noise introduced between deployments (it was modified to record MT as well as CSEM data).

Electric field decomposition Data reduction Data reduction

Go to first page