DAMPED LEAST‐SQUARES INVERSION OF TIME‐DOMAIN AIRBORNE EM DATA BASED ON SINGULAR VALUE DECOMPOSITION1

Abstract

AbstractAirborne electromagnetic (AEM) methods are increasingly being used as tools of geological mapping, groundwater exploration and prospecting for coal and lignite. In such applications, quantitative interpretation is commonly based on the layered‐earth model. A new approach, a damped least‐squares inversion with singular value decomposition, is proposed for interpretation of time‐domain, towed‐bird AEM data. Studies using theoretical and field AEM data indicate that inversion techniques are dependable and provide fast converging solutions. An analysis has been made of the accuracy of model parameter determination, which depends on resistivity and thickness distribution. In the common case of conductive overburden, upper‐layer resistivity and thickness are usually well determined, although situations exist where their separation becomes difficult. In the case of a resistive layer overlying a conductive basement, the layer thickness is the best‐determined parameter. In both cases, estimates of basement resistivity are the least reliable. Field data obtained with the Chinese‐made M‐l AEM system in Dongling, Anhui Province, China, were processed using the described inversion algorithm. The survey area comprised fluvial Cenozoic clays and weathered Mesozoic sediments. Inversion of AEM data resulted in accurate depth‐to‐bedrock sections and realistic estimates of the resistivities of overburden and bedrock which agree with the results of drilling and resistivity sounding.