Temperature Estimation in the Geothermal Areas by Incontact Electromagnetic Geothermometer

A new method of incontact sub-surface temperature estimation developed in (Spichak et al., 2007a) is tested by real data collected in the northern Tien Shan. The technique proposed is based on magnetotelluric (MT) sounding from the Earth’s surface followed by teaching artificial neural network (ANN) by the correspondence between the electrical conductivity revealed from the MT data and the temperature well logs available. It is shown that practical application of this method enables, first, to refine the temperature estimates in cases when the amount of temperature logs available is insufficient and, second, to perform more precise temperature estimation in extrapolation mode.