NMR
Surface nuclear magnetic resonance (surface NMR)
The surface NMR (nuclear magnetic resonance) technique has a significant advantage over other hydrogeophysical methods: the water molecules, especially the protons within them, are directly detected through their properties at the atomic level. All other geophysical methods only allow indirect water detection through measurements of the physical properties of the subsurface, such as electrical conductivity or permittivity, which depend on the water content but also on other properties of the rock. This increases the ambiguity of the results and thus the risk of misinterpretation. In a hydrogeophysical NMR measurement, in contrast, only the nuclear spins of the protons in the water molecules are excited.
NMR has been used as a geophysical borehole technique for oil exploration since the 1980s (e.g., Dunn et al., 2002). At the same time, the development of a non-invasive NMR method for groundwater exploration, the so-called surface NMR technique, began. Since its basic applicability was first demonstrated in Russia (Semenov et al., 1989), ongoing worldwide research has begun to continuously improve and expand the surface NMR technique (e.g., Legchenko et al., 2004; Dlugosch et al., 2011; Behroozmand et al., 2014). The simplest, one-dimensional application of this technique is also called magnetic resonance sounding (MRS). Similar to the TEM technique, a cable coil is deployed on the surface, serving as both a transmitter and a receiver. In transmit mode, a pulsed alternating current with the Larmor frequency of the protons is applied. The resulting electromagnetic (EM) field in the subsurface excites the proton spins and deflects them from their equilibrium position in the Earth's magnetic field. After the excitation is switched off, the spins return to their equilibrium position, inducing a voltage that can be measured at the receiver. The measurement signal corresponds to an exponentially decaying relaxation curve. Its initial amplitude is proportional to the number of excited protons and thus to the total water content in the sensitive volume, while the decay behavior, represented by the relaxation time, is linked to the average pore size. By varying the current strength, i.e., the excitation pulse, different exploration depths are achieved. In 2D and 3D applications of the surface NMR method, the transmitter and receiver are separated to increase the resolution (Hertrich et al., 2009; Dlugosch et al., 2014).
The maximum possible exploration depth roughly corresponds to the diameter of the coil used (usually 10 to 150 m), but also depends on the electrical conductivity of the subsurface. To perform the inversion of the NMR data as accurately as possible, additional electrical conductivity measurements (TEM or geoelectrics) are necessary. Since the excitation frequency (Larmor frequency) depends on the strength of the Earth's magnetic field, magnetic measurements also accompany every surface NMR campaign. In addition to the actual NMR measurement coil, additional reference coils are often installed to characterize the EM ambient noise and thus improve data processing (Dalgaard et al., 2012; Müller-Petke and Costabel, 2014).
The measured data are fitted using a model during inversion (Müller-Petke et al., 2016) (Fig. 4b). The model describes the distribution of water content and relaxation time as a function of depth (Fig. 4c). The total water content measurable with surface NMR (Fig. 4d) provides a useful estimate of the effective porosity (Legchenko et al., 2004). Capillary-bound water has very short relaxation times and is not detected due to the relatively long instrument dead times (=time required to switch from transmit to receive mode). Therefore, as shown in the data example, the silty layers are represented by underestimated water content, while the effective porosities of the sand aquifers are very accurately estimated at 25 to 35%. Using relaxation times as a measure of the mean pore size, the distribution of hydraulic conductivity at depth can be calculated by appropriate calibration (Legchenko et al., 2004; Mohnke and Yaramanci, 2008).
Numerous studies have demonstrated the suitability of surface NMR for various hydrogeological questions, such as the characterization of aquifers in loose sediments (e.g., Mohnke and Yaramanci, 2008; Günther and Müller-Petke, 2012; Knight et al., 2012) and of fractured and karst aquifers (e.g., Vouillamoz et al., 2005, 2014; Girard et al., 2007). Furthermore, the ongoing technical and methodological development of the method is expanding its range of applications. The reduction of instrument-related dead time, for example, now allows measurements in the vadose zone (Walsh et al., 2014; Costabel and Günther, 2014). Newly developed pulse sequences improve resolution and enable measurements in magnetically heterogeneous environments (Legchenko et al., 2010; Walbrecker et al., 2011; Grunewald et al., 2014; Grombacher et al., 2016). Strategies to accelerate measurement progress will significantly increase the practical applicability of surface NMR in the future (Costabel et al., 2016; Grunewald et al., 2016), so that continued increasing acceptance within hydrogeology can be expected.