Hitem – A high moment / high production TEM system

During the nineties the transient electromagnetic method (Nabighian and Macnae, 1991) became the method preferred for mapping the extension of aquifers in the sedimentary environments in Denmark. The most commonly used system has been the Geonics PROTEM 47 system with a 40 x 40 m2 transmitter loop. More than 45,000 soundings have been carried out by use of this system. However, for many tasks its penetration depth is insufficient and a wish rose to develop a high moment TEM system (HiTEM) with the same model resolution capabilities and similar production rates as the 40 x 40 m2 system. The development of the HiTEM system to its present state took about 3 years. The system is capable of transmitting up to 75 A in a 30 x 30 m2 transmitter loop. This offers a magnetic moment of the system about 15 times larger than the conventional 40 x 40 m2 system and a penetration depth up to 300 m compared to the 120-150 m of the PROTEM 47 system. Since spring 2001 more than 2,000 HiTEM soundings have been carried out. Several array configurations were tested, as significant technical difficulties arise when high currents are transmitted in a small loop and the Earth response is measured in the centre of the loop. For this reason and because of expected IP-effects (Flis et al., 1989), the receiver coil was moved outside the loop. This configuration is more complex to handle in the data interpretation process than the central loop configuration, and it took some time to gain knowledge of the strengths and limitations of the configuration. Thorough 3-D modelling of various arrays and model scenarios (Toft, 2001) showed that the offset configuration is extremely sensible to small-scale, near-surface inhomogeneities at early times. In addition, the offset of the receiver coil has to be known with an impossible accuracy in a practical field situation. All this led to the characteristic HiTEM sounding which includes both a central and an offset loop measurement. Early times are measured in the central loop configuration, and late times are measured in the offset loop configuration at 2.4 A and 75 A, respectively. Both sensitivities and locations of the arrays differ, but they are interpretatively the same sounding. A traditional joint inversion of the two data sets often results in a poor data fit due to the model inconsistencies between the data sets. The solution to this has been the Mutually Constrained Inversion (MCI) approach. In a MCI inversion two models, one for each data set, are constrained to each other, and the MCI algorithm inverts the two data sets concurrently and produces two similar models. The hardness of the constraint serves as a handle which the interpreter can use to meet certain requirements: the harder the constraint, the more similar the out coming models, but the more consistent data sets are asked to provide a satisfying fit to data.