Successful adaptation of 3D inversion methodologies for archaeological-scale, total field magnetic datasets

Despite the development of advanced processing and interpretation tools for magnetic datasets in the fields of mineral and hydrocarbon industries, these methods have not achieved similar levels of adoption for archaeological or very near surface surveys. Using a synthetic dataset we demonstrate that certain methodologies and assumptions used to successfully invert more regional-scale data can lead to large discrepancies between the true and recovered depths when applied to archaeological-type anomalies. We propose variations to the current approach, analysing the choice of the depthweighting function, mesh design and parameter constraints, to develop an appropriate technique for the 3D inversion of archaeological scale datasets. The results show a successful recovery of a synthetic scenario, as well as a case study of a Romano-Celtic temple in the UK. For the case study, the final susceptibility model is compared with two coincident ground penetrating radar surveys, showing a high correlation with the comparative depth slices. The new approach takes interpretation of archaeological datasets beyond a simple 2D visual interpretation based on pattern recognition.