The Anadarko Hot Ice 1 well was cored as part of a project to study the occurrence of gas hydrate on the North Slope of Alaska. The observations and measurements made at the drill site along with the subsequent core analysis are described in five individual reports published in this Memoir. This report deals with the nuclear magnetic resonance (NMR) measurements made on the recovered core from the Hot Ice 1 gas-hydrate research well.

Samples from sands recovered during phase I of the coring of the Hot Ice 1 well came from the permafrost zone. The NMR measurements were conducted on frozen core samples, samples in which the ice was allowed to completely melt, and samples that were cleaned, dried, and resaturated with brine. Not all measurements were done on all samples. For the low-field-strength system used to measure the NMR response, only hydrogen nuclei in unfrozen fluid contribute to the signal.

Four core-derived samples were measured at a temperature of 5C (23F). They had NMR porosities of 7.1% (37% core-derived He porosity), 9.4% (40% core-derived He porosity), 9.8% (42.4% core-derived He porosity), and 11.6% (31.6% core-derived He porosity). The NMR spectra for all the frozen samples were very similar. The NMR porosities represent the percentage of the sample volume containing unfrozen brine. The He porosities were measured on cleaned and dried core plugs. Four models for the way ice could form when the unfrozen sample is frozen where investigated. For each model, the frozen NMR spectra were computed from the unfrozen spectra. None of the models provided a satisfactory match to the measured frozen spectra.

An anomalous result of the study was a significant difference between NMR spectra measured on melted samples and the measurements done after cleaning and resaturation with brine. The differences are essentially a shift to slower relaxation times for the resaturated samples without any change in the shape of the spectrum. The simplest explanation for this is a reduction in surface relaxivity produced by cleaning and resaturation.

For most sedimentary sandstones, the NMR geometrical mean decay time combined with porosity can be used to predict permeability. For the phase I recovered sandstones, the prediction is worse than commonly seen. Also, a commonly used default formula provides a very poor fit to the data and underestimates the permeability on average by a factor of 25. This suggests a much larger surface relaxivity than commonly seen.

The phase I recovered sands, when water saturated, exhibit five different spectrum types. Most of the samples fall into the first two identified types, which differ only in the presence of a small percentage of faster relaxing small pores in type 2 and their absence in type 1.

The unconsolidated sand samples recovered in phase II were recovered from unfrozen sediments. The NMR was done only on cleaned and dried samples. They had very similar spectra classes as the phase I samples.

Unlike the phase I samples, an excellent permeability estimate can be obtained from the NMR measurements. It has the same form as the standard estimator, but the multiplicative constant is about 12 times the value used in the standard estimator. This is very likely caused by a much larger than normal surface relaxivity or large internal magnetic gradients.