3D inversion of airborne electromagnetic data

Abstract

Time-domain airborne surveys gather hundreds of thousands of multichannel, multicomponent samples. The volume of data and other complications have made 1D inversions and transforms the only viable method to interpret these data, in spite of their limitations. We have developed a practical methodology to perform full 3D inversions of entire time-or frequency-domain airborne electromagnetic (AEM) surveys. Our methodology is based on the concept of a moving footprint that reduces the computation requirements by several orders of magnitude. The 3D AEM responses and sensitivities are computed using a frequency-domain total field integral equation technique. For time-domain AEM responses and sensitivities, the frequencydomain responses and sensitivities are transformed to the time domain via a cosine transform and convolution with the system waveform. We demonstrate the efficiency of our methodology with a model study relevant to the Abitibi greenstone belt and a case study from the Reid-Mahaffy test site in Ontario, Canada, which provided an excellent practical opportunity to compare 3D inversions for different AEM systems. In particular, we compared 3D inversions of VTEM-35 (time-domain helicopter), MEGATEM II (time-domain fixed-wing), and DIGHEM (frequency-domain helicopter) data. Our comparison showed that each system is able to image the conductive overburden and to varying degrees, detect and delineate the bedrock conductors, and, as expected, that the DIGHEM system best resolved the conductive overburden, whereas the time-domain systems most clearly delineated the bedrock conductors. Our comparisons of the helicopter and fixed-wing time-domain systems revealed that the often-cited disadvantages of a fixed-wing system (i.e., response asymmetry) are not inherent in the system, but rather reflect a limitation of the 1D interpretation methods used to date.

Joint 3D inversion of time- and frequency-domain airborne electromagnetic data

“…As described in Zhdanov (2009) and Cox et al (2010Cox et al ( , 2012, the forward modeling is based on the rigorous 3D contraction integral-equation method (Hursan and Zhdanov, 2002). For the modeling of a moving sensitivity domain, the volume of interest for a given source is defined as the subdomain of the 3D earth model encapsulated in the AEM system's footprint, and only sensitivities within the sensitivity domain need to be calculated.…”

“…A moving footprint takes advantage of the fact that the received signal of an AEM system is only influenced by a domain, termed the sensitivity domain, which is much smaller than the entire survey domain. Using the moving footprint approach, 3D inversion of entire AEM data sets suddenly became practical, reliable, rapid, and able to delineate deposit-scale features (Cox et al, 2010(Cox et al, , 2012. AEM surveys are conducted either in the frequency-domain (FD) or time-domain (TD).…”

Review on airborne electromagnetic inverse theory and applications

“…The footprint is defined as the lateral extent of the sensitivity of the AEM system (Liu and Becker, 1990;Beamish, 2003;Reid et al, 2006). Cox et al (2012) successfully extend this footprint-based approach to the time domain. Yin et al (2014b) give a more comprehensive definition for the EM footprint as the underground volume, in which the total contributions from the current in a half-space accounted for 90% of the secondary magnetic field.…”

Three‐dimensional inversion of towed streamer electromagnetic data

“…In the case of 3D inversion, this problem is exacerbated by the need to compute Fréchet derivatives and repeat the whole process for multiple iterations. Similar to airborne EM (Cox and Zhdanov ; Cox, Wilson and Zhdanov , ), we can exploit the fact that the towed streamer EM system's sensitivity domain is significantly less than the size of the survey area and we introduce the concept of 3D inversion with a moving sensitivity domain. That is, for a given transmitter‐receiver pair, the responses and Fréchet derivatives are computed from a 3D earth model that encapsulates the towed EM system's sensitivity.…”