Error propagation in electromagnetic transfer functions: what role for the magnetotelluric method in detecting earthquake precursors?

Abstract

Listen

Translate article

SUMMARY Whether electromagnetic precursors exist is a contentious issue that should be considered in the context of the focal depth and epicentral distance at which earthquakes nucleate, and with due regard to externally driven changes in the electromagnetic fields recorded at the Earth’s surface. If high-frequency electromagnetic signals are generated by earthquakes, then these signals may not be detectable because of attenuation occurring between the focal depth of the earthquake and the Earth’s surface. This is particularly true in seismotectonic regions (e.g. subduction zones) where saline fluids (including seawater), raised heat flow and partial melts often lead to enhanced conductivities. A technique for examining temporal variations in magnetotelluric (MT) spectra is presented, which: (i) considers both non-inductive and inductive parts of the MT transfer function; (ii) allows electromagnetic skin depths and adjustment distances to be calculated; (iii) takes account of error propagation in the transfer functions and; and (iv) acknowledges possible nonearthquake sources of observed variations in transfer functions and their errors. Data processing tends to eliminate non-stationary electromagnetic disturbances of the source field. Hence the method is more suited to detecting hypothesized changes in subsurface conductivity than short-lived electromagnetic bursts. The technique is applied to MT data from seismotectonic regions in New Zealand (Mt. Ruapehu), the Andes (western Cordillera) and the Mediterranean (Montecristo island) and to a culturally quiet, control site situated on a stable craton in central Australia. We demonstrate that the MT method can resolve apparent conductivity changes of the order of a few percent over 48 hr time intervals and give thresholds for other interval lengths. Comparison of the results obtained from seismotectonic regions with those from central Australia, and consideration of external source field effects, highlight how temporal variations of electromagnetic fields might be speciously interpreted as earthquake precursors.