How remote can the far remote reference site for magnetotelluric measurements be?

Remote reference (RR) magnetotelluric (MT) measurements are made to reduce the bias caused by noise in electric field E and magnetic field H at a local site. RR only works when noise at a local and remote sites are uncorrelated. A study has been undertaken to find the farthest distance of a far remote site for crustal study which maintains the effectiveness of the RR technique. From theoretical studies the conditions for valid RR estimates have been obtained. The study shows that the remote site can be kept at a considerably large distance from the local site. In a field experiment a fixed local site and several remote sites have been selected at distances of 80, 115, and 215 km away in the frequency range of 30 Hz (∼0.03 s) to 0.00055 Hz (∼1800 s) at midlatitudes where the wavelengths of the source magnetic fields are long compared to the site separation. The remote sites are distributed over diverse geological settings. All the data of the fixed local site have been remote reference processed with the corresponding remote sites. The study reveals that using a remote site located at as large a distance as 215 km results in unbiased observations and remains effective in improving the MT data quality for all frequency ranges. If the data were acquired at long periods and or in the high latitudes, then to extract the stable uniform field estimates of the impedance one has to carry out robust processing technique.