Estimates of effective stress beneath a modern West Antarctic ice stream from till preconsolidation and void ratio

Preconsolidation stress recorded in subglacial sediments provides important information about subglacial effective stresses. It is commonly used to reconstruct past effective stresses from sediments left after ice retreat. In this article, we use properties of sub‐ice‐stream till samples to estimate effective stresses beneath a modern West Antarctic ice stream. Two previous estimates of sub‐ice‐stream effective stress were derived for the Upstream B (UpB) area of Ice Stream B from shear wave velocities (50 ± 40 kPa, Blankenship et al 1987) and borehole water level measurements (63 ± 24 kPa, Engelhardt & Kamb 1997). However, geotechnical tests performed on samples of the UpB till have shown that if subjected to effective stress of 50–63 kPa this till would have significantly lower porosity (?0.32–0.35) and higher strength (?‐22–28 kPa) than it apparently has in situ (?0.4 and ?2kPa). We derive new estimates of sub‐ice‐stream effective stress using: (1) Casagrande's construction applied to the results of six confined uniaxial tests, and (2) a combination of void‐ratio data for 51 till samples and 3 experimentally constrained equations describing compressibility of the UpB till under normal consolidation, overconsolidation and in the critical state. Casagrande's method yields an upper bound on effective stress of 25 kPa for four till samples and values of 13, and 4.4kPa for two other samples. The void‐ratio approach gives 11.7 ± 2.6 (normal consolidation), 18.3 ± 4.4 (overconsolidation) and 2.0 ± 0.8 kPa (critical state). These new, lower estimates of effective stress are consistent with the low till strength that has been independently measured and inferred from recent theoretical ice‐stream models. Our interpretation of data on till void ratio in terms of sub‐ice‐stream effective stress means that we can qualitatively evaluate the nature of the vertical distribution of this stress in the UpB till layer. We infer that in the sampled top 3 m of till the effective‐stress distribution is non‐hydrostatic, probably close to lithostatic. The results may be useful in future modeling of ice‐stream behavior and may aid efforts to delineate paleo‐ice streams based on their geologic record.