Paper No. 190-1
Presentation Time: 10:05 AM
SOLVING THE PUZZLES OF FLUID PRESSURE ANOMALIES IN SHALLOW CLAY ROCK FORMATIONS (Invited Presentation)
Investigations related to siting nuclear waste repositories are changing understanding of subsurface environments. A notable example is the finding that anomalous pore fluid pressures appear to be far more common in shallow (< ~ 1 km deep) clay rock formations than previously thought. As used here, pressure anomalies (PAs) are high or low values of fluid potential (hydraulic head) that appear to be isolated in the subsurface and imply a net outward or inward flow of pore fluid. While PAs are well known in relatively deep sedimentary sequences, their presence in shallow clay rocks in stable onshore settings is surprising; pressure-altering processes that can explain deep PAs, including sedimentary loading, vigorous tectonism, diagenesis, and thermal generation of oil and gas are not occurring in these environments. It nevertheless appears that shallow PAs, like deep ones, are hydrodynamic responses to forcing processes that perturb pore fluid pressure faster than readjustment caused by fluid flow. Removal of overburden by erosional downwasting and ice sheet retreat causes rock to mechanically dilate, which can decrease pore fluid pressure and may explain why most shallow PAs discovered so far are fluid potential lows. Interpreted as hydrodynamic responses, shallow PAs contain information about two aspects of the subsurface that are challenging to characterize, namely permeability at scales too large to test and forcing that is too slow to observe directly. For example, PAs have offered useful constraints on permeability values over dimensions of kilometers that are relevant to repository siting. There are also instances where the cause of PAs is not understood, perhaps signaling an as yet unidentified forcing mechanism. However, interpretations must be tempered by the possibility that some pressure data are affected by unresolved systematic errors or that hydrodynamic forcing is an inadequate conceptual model. Ongoing analyses and data from new investigations may allow addressing these uncertainties.