Denys Grombacher

Studies of parameter correlations in surface NMR using the Markov chain Monte Carlo method

Surface nuclear magnetic resonance is a technique capable of providing insight into subsurface naquifer properties. To produce estimates of aquifer properties (such as the spatial distribution of nwater content and parameters controlling the duration of the nuclear magnetic resonance signal), an ninversion is required. Essential to the reliable interpretation of the estimated subsurface models is nan understanding of the uncertainty and correlation between the parameters in the estimated models. nTo quantify parameter uncertainty and correlation in the surface nuclear magnetic resonance inversion, na Markov chain Monte Carlo approach is demonstrated. Markov chain Monte Carlo approaches nhave been previously employed to invert surface nuclear magnetic resonance data, but the primary nfocus has been on quantifying parameter uncertainty. The focus of this paper is to further ninvestigate whether the parameters in the estimated models exhibit correlation with one another; nequally important to building a reliable interpretation of the subsurface is an understanding of the nparameter uncertainty. The utility of the Markov chain Monte Carlo approach is demonstrated nthrough the investigation of three questions. The first question investigates whether the parameters ndescribing the water content and thickness of a layer exhibit a strong correlation. This question nstems from applying concepts known to electromagnetic surveys (that the layer thickness and layer nresistivity parameters are strongly correlated) to the surface nuclear magnetic resonance inversion. nA water content–layer thickness correlation in surface nuclear magnetic resonance would not have nlarge effects for quantifying total water content but would affect the ability to identify layer boundaries. nThe second question examines whether the parameter controlling the duration of the nuclear nmagnetic resonance signal exhibits a correlation with the water content and layer thickness parameters. nThe resolution of surface nuclear magnetic resonance typically does not consider the duration nof the signal and focuses primarily on the distribution of current amplitudes that form the suite of ntransmit pulses. It is common to treat regions with short-duration signal with greater uncertainty, nbut it is important to understand whether the signal duration controls resolution for medium to long nduration signals as well. The third question explores if the parameter uncertainty produced by the nMarkov chain Monte Carlo approach is consistent with that produced by an alternative approach nbased upon the posterior covariance matrix (for the linearised inversion). The ability of the Markov nchain Monte Carlo approach to more thoroughly explore the model space provides a means to nimprove the reliability of surface nuclear magnetic resonance aquifer characterisations by quantifying nparameter uncertainty and correlation.

Comparison of stabiliser functions for surface NMR inversions

Surface nuclear magnetic resonance is a geophysical technique providing non-invasive aquifer ncharacterization. Two approaches are commonly used to invert surface nuclear magnetic resonance ndata: (1) inversions involving many depth layers of fixed thickness and (2) few-layer inversions nwithout predetermined layer thicknesses. The advantage of the many-layer approach is that it nrequires little a priori knowledge. However, the many-layer inversion is extremely ill-posed and nregularisation must be used to produce a reliable result. For optimal performance, the selected nregularisation scheme must reflect all available a priori information. The standard regularisation nscheme for many-layer surface nuclear magnetic resonance inversions employs an L2 smoothness nstabiliser, which results in subsurface models with smoothly varying parameters. Such a stabiliser nstruggles to reproduce sharp contrasts in subsurface properties, like those present in a layered subsurface n(a common near-surface hydrogeological environment). To investigate if alternative stabilisers ncan be used to improve the performance of the many-layer inversion in layered environments, nthe performance of the standard smoothness stabiliser is compared against two alternative stabilisers: n(1) a stabiliser employing the L1-norm and (2) a minimum gradient support stabiliser. Synthetic nresults are presented to compare the performance of the many-layer inversion for different stabiliser nfunctions. The minimum gradient support stabiliser is observed to improve the performance of nthe many-layer inversion for a layered subsurface, being able to reproduce both smooth and sharp nvertical variations of the model parameters. Implementation of the alternative stabilisers into existing nsurface nuclear magnetic resonance inversion software is straightforward and requires little nmodification to existing codes.

A Multichannel, Low Noise Surface NMR Receiver System with Wireless Connections to Signal and Reference Coils

Summary Surface NMR holds great promise as a tool in groundwater measurements due to its unique direct sensitivity to water, but the method currently suffers from a number of drawbacks which limits its widespread applicability. Among these drawbacks are a low signal to noise ratio which limits the use of the method in many places of interest and a low production rate which makes the method costly in field campaigns. Hence there is a need for research further advancing the technology. In this paper we report on the development of a new multichannel, low noise surface NMR receiver system with wireless connections to reference coils. The receiver system works as a completely independent add-on to existing transmitter systems and consists of a number of independently operated data acquisition boxes connected with WiFi and synchronized by GPS. The internal electronic noise level of the system is 1.2 nV/sqrt(Hz). The timing jitter between data acquired in different boxes is less than 100 ns.

Removal of Co-frequency Powerline Harmonics from Multichannel Surface NMR Data

Powerline harmonics is often the dominant noise source in surface NMR data and can completely overwhelm the NMR signal. Several methods have been successfully implemented for removal of powerline harmonics including notch filtering, multichannel filtering and model-based subtraction. However, the performance of these methods can be problematic when one of the powerline harmonic components has a frequency close to or coincident with the Larmor frequency, referred to as a co-frequency harmonic. Removal of the co-frequency harmonic can distort the NMR signal causing erroneous estimates of water content and relaxation rates. To solve this problem we propose an extended method of multichannel model-based subtraction of powerline harmonics. In this method, the co-frequency component is modeled in the primary coil and a reference coil on noise-only data recorded immediately prior to the NMR excitation. The phase and amplitude relationships between the components measured in the two coils are calculated. The relationships are used to predict the co-frequency harmonic component parameters in the primary coil during the NMR decay based on NMR signal free, synchronously recorded reference coil data. The mathematical framework is developed and we give examples of the efficiency of the method based on synthetic NMR signals embedded in noise-only data.

Potential for Square Wave Transmitters in Surface NMR

An alternative transit strategy for the surface nuclear magnetic resonance technique is presented. The current waveform traditionally used for excitation in surface NMR is described by a sinusoid oscillating at the Larmor frequency for a finite duration (20-40 ms). However, in principle any current waveform whose spectrum contains non-zero energy at the Larmor frequency may be used to perturb the subsurface magnetization. We investigate the utility of square waves and triangle waves for surface NMR excitation. We aim to exploit: 1) the potential to adapt existing time domain electromagnetic (TEM) transmitters for application in surface NMR, and 2) the potential to allow for modulation of the current amplitude during a single pulse independent of the instantaneous transmit frequency. Although a square wave transmitter represents a less efficient use of power compared to a sinusoid, it offers the potential for a single transmitter to be used to conduct both the surface NMR and TEM experiments. Additionally, the use of square waves allows pulse width modulation to be exploited to provide independent control of the current amplitude throughout the pulse improving flexibility in the design of current waveforms feasible in surface NMR.

Near-surface geophysics for informed water-management decisions in the Anangu Pitjantjatjara Yankunytjatjara (APY) lands of South Australia

The Anangu Pitjantjatjara Yankunytjatjara (APY) Lands of South Australia is an arid environment and the population relies largely on groundwater resources for potable water and agricultural needs. Historically, locating productive wells in the region has been hit-and-miss and even if a water source was found, the quality may be unreliable. In this project, we seek to improve the water security in the APY lands by demonstrating that surface Nuclear Magnetic Resonance (NMR) and Time-Domain Electromagnetic (TEM) geophysical measurements are able to map local aquifers and quantify ground water resources, thereby optimizing site selection for potential future wells. Surface NMR is directly sensitive to water and TEM measurements detecting the electrical conductivity structure and able to image the subsurface over large areas - all entirely non-invasively and with minimal risk of disturbing sites of importance to the local Aboriginals.