Imaging of groundwater with nuclear magnetic resonance

This review has provided information about the various aspects of surface NMR. In the first section the most basic formulation is presented which reveals the complexity of the forward functional. For the quantitative description of the surface NMR signal the interaction of the spin system with the non-uniform, non-perpendicular and elliptically polarized secondary field is taken into account. Furthermore, for configurations with non-coincident transmitter and receiver loops, the vectorial relation of the spin magnetization to the these fields was considered. This allows a complete forward functional with suitable formulations for 1-D and 2-D conditions to be derived by integrating the data kernel of the forward functional to the respective dimensions. In the second section these data kernels are the basis for the inversion of surface NMR measurements to reconstruct models of sub-surface water content distribution. Besides a least-squares inversion of a model with a large number of layers and variable but constrained water content, which is most common in geophysical data inversion, a novel scheme with a small number of discrete layers whose boundaries are allowed to vary in depth is presented. Both schemes provide comparable models. Comparing these two approaches, the variable-boundary model gives a better quantification of the depths of layer boundaries and estimates of layer water content than does the model with fixed boundaries. But inversion with such a scheme can only provide useful sub-surface information if the variation of the water content in the sub-surface is sharp rather than gradational, and if an estimate of the number of geological units is known beforehand. Inversion of field data is in general ambiguous. This is particularly so in cases where there are considerable uncertainties in the measured data, which often occurs for the weak surface NMR signals in the presence of strong ambient background noise. A bootstrapping scheme applied to surface NMR inversion is introduced in Section 3. It provides a suitable tool to assess the ambiguity of the model of water content distribution and allows the assignment of confidence intervals for the shown example. In many NMR applications the relaxation time is the major source of information on properties of the object under investigation. Also in NMR applied to geo-materials the relaxation time can be a useful measure to estimate structural parameters, but determination of the sub-surface distribution of the relaxation constants is physically limited and technically challenging for surface NMR. In Section 4 the available techniques used for surface NMR are explained. It is demonstrated that in sedimentary environments the most easily accessible relaxation time T2* is rarely a valuable measure for host rock properties. In any case quantitative formulations for the derivation of T1 relaxation from surface NMR measurements are not yet available. The two inversion schemes and the bootstrapping techniques are applied to a real data example in Section 6. From both inversion schemes a consistent model of 1-D aquifer stratification is obtained. The comparison to borehole data from a nearby research drill-hole shows the capability and almost unique potential of the surface NMR technique in resolving discrete water-bearing zones. A model of similar spatial resolution of the water content distribution can only be interpreted by the combination of a series of borehole measurements, but can definitely not be obtained by any other surface geophysical technique. The development of surface NMR has undergone a rapid progress over the last two decades. Nowadays it has reached a mature phase in terms of available hardware, theoretical description of the forward functional and advanced inversion techniques. However, major topics for further research lie in (i) forward calculation of the loop magnetic fields in more complex environments such as varied topography or spatially complex resistivity distribution within the sub-surface; (ii) the quantitatively more precise formulation of the spin dynamics in the weak magnetic field of the Earth and the non-uniform loop fields and (iii) establishing reliable correlations of surface NMR determined relaxation times and hydro-geological parameters. A major drawback of applying surface NMR to groundwater studies is the presence of background noise. Typical signal amplitudes of surface NMR measurements are very weak and cannot be easily increased relative to the ambient noise level by technical means. Hence, surface NMR measurements are nowadays not feasible in many environments. Ongoing research by several groups worldwide, aimed at understanding, describing and recording surface NMR signals offers promise of further improvement. However, restriction to low noise environments will probably be the major issue and pose the greatest challenge to widespread acceptance of this promising technique in the near-future. © 2008 Elsevier B.V. All rights reserved.