On the sensitivity analysis of separated-loop MRS data

In this study we investigate the sensitivity analysis of separated-loop magnetic resonance sounding (MRS) data and, in light of deploying a separate MRS receiver system from the transmitter system, compare the parameter determination of the separated-loop with the conventional coincident-loop MRS data. MRS has emerged as a promising surface-based geophysical technique for groundwater investigations, as it provides a direct estimate of the water content. The method works based on the physical principle of NMR during which a large volume of protons of the water molecules in the subsurface is excited at the specific Larmor frequency. The measurement consists of a large wire loop (typically 25 - 100 m in side length/diameter) deployed on the surface which typically acts as both a transmitter and a receiver, the so-called coincident-loop configuration. An alternating current is passed through the loop deployed and the superposition of signals from all precessing protons within the investigated volume is measured in a receiver loop; a decaying NMR signal called Free Induction Decay (FID). To provide depth information, the FID signal is measured for a series of pulse moments (Q; product of current amplitude and transmitting pulse length) during which different earth volumes are excited. One of the main and inevitable limitations of MRS measurements is a relatively long measurement dead time, i.e. a non-zero time between the end of the energizing pulse and the beginning of the measurement, which makes it difficult, and in some places impossible, to record SNMR signal from fine-grained geologic units and limits the application of advanced pulse sequences. Therefore, one of the current research activities is the idea of building separate receiver units, which will diminish the dead time. In light of that, the aims of this study are twofold: 1) Using a forward modeling approach, the sensitivity kernels of different separated-loop MRS soundings are studied and compared with that of the conventional coincident-loop sounding. 2) The posterior parameter determination of central-loop MRS data is studied in a joint MRS and TEM data analysis scheme. The MRS kernel is the function that describes the spatial distribution of the sensitivity of the MRS measurement and, for given site specifications, depends on Q. In the typical 1D earth parametrization, a complete MRS measurement forms the 1D MRS kernel as a function of depth and Q; here referred to as a 1D kernel structure. For the conventional coincident-loop configuration, the 1D kernel structure covers the excited earth's volume throughout the applied Qs. As a result, the shallower parts of the subsurface are mainly sampled using smaller Qs and the deeper parts are mainly sampled using higher Qs. Here we also study the sensitivity kernel of the separated-loop MRS configuration, i.e. when different loops are used for transmitting the current and receiving the signal, and highlight the cases where an increased excited earth's volume is sampled throughout the Qs. We examined the optimal geometry of the receiver loop and the conditions under which a separate receiver system can be used are addressed. The results of this study suggest that it can be beneficial to develop MRS instrumentation where the receiver system is separated from the transmitter system.