Probabilistic Predictions of Offshore Gas Hydrate and Submarine Permafrost Distribution Along the Alaskan North Slope

Abstract:

Methane gas hydrate is an ice-like solid found along most continental margins, including the circum-Arctic continental shelf. Gas hydrates form where suitable thermodynamic conditions exist and an adequate source of methane is present in excess of local gas solubility. Permafrost-associated gas hydrates occur within and below permafrost-bearing sediments on the shallow continental shelf and terrestrially on some coastal plains. Permafrost-associated gas hydrates formed when non-glaciated portions of the shelf experienced subaerial exposure during glacial episodes when sea level was much lower than today. Now submerged due to ocean regression, the relict submarine permafrost and associated gas hydrate deposits may release large quantities of methane gas as this shallow reservoir warms and degrades.

This presentation will detail a new workflow which integrates geospatial machine learning (GML) forecasts (via the U.S. Naval Research Laboratory’s Global Predictive Seabed Model) with a probabilistic numerical modeling framework (via Sandia National Laboratory’s Dakota and PFLOTRAN software), to forecast current and future thermodynamic conditions offshore along the Alaskan North Slope. The integrated modeling framework takes as input GML forecasts for basic marine sediment properties, including uncertainty estimates, predicted from observed seafloor data and sediment physics models. The continuous GML forecasts are in turn used as inputs to initiate 1-D ensemble simulations with PFLOTRAN, in which several thousands of realizations sample and propagate the input values along with their uncertainty estimates to provide probabilistic forecasts of the thermodynamic conditions on and below the seafloor. We will present initial probabilistic maps for the distribution of submarine permafrost, gas hydrate stability zone, and probability of encountering free methane gas offshore along the Alaskan North Slope. Model predictions will be compared against recent geophysical observations of submarine permafrost distribution (seismic and electromagnetic surveys).

SAND2020-7716 A