| dc.description.abstract | Ice sheet dynamics play a vital role in the Earth’s climate and a key question is if melting of the ice sheets contributes to sea level rise. The dynamics of ice sheets are controlled by rheological behavior at the grain-scale. Rheological behavior can be described by a flow law of the form:
ἑ = Aσ^n/d^p exp(−Q/RT)
in which n and p denote respectively the stress and grain size exponent. The rheology of ice at high stress (>1 MPa) is attributed to dislocation creep, with n=3 and p=0. At low stress (<0.1 MPa), there is evidence for a decrease in stress-exponent which has been attributed to grain boundary sliding
(GBS) with n=1.8. It is assumed that the contribution of grain boundary processes, such as diffusion creep, which would be characterized by n=1, are insignificant in ice. However, ice creep experiments show a sharp increase in strain-rates of a factor of 5-10 close to the melting temperature (Tm). Optical reflectivity measurements and thermodynamic calculations indicate that above T∼262 K liquid films are present on the surfaces and grain boundaries of ice. The thickness of these films increase in ice close to Tm, with a thickness of 2-5 nm at T=272 K. A fluid film of this size has nearly the same viscosity as bulk water. The question addressed in this thesis is if flow by grain boundary melt transport can contribute to the deformation of ice close to Tm and can account for the observed increase in strain-rates. This process is expected to be controlled by the slowest of three steps: (1) melting, (2) transport through the grain boundary and (3) freezing. Hence, it can be envisaged as an analogue of pressure solution creep in
rocks. In this thesis, pressure solution theory is modified to construct a grain-scale model for melt-assisted creep in polycrystalline ice. This model is modified to explain the behavior of a wire cutting through ice,since it is expected that this process is controlled by the same kinetics. An attempt is made to perform wire regalation experiments to verify these models and to test if experiments show the same stress and wire diameter/grain-size exponent. This theory predicts that melt-assisted creep in polycrystalline ice is controlled by the transport-step, as melting and freezing rates are 3 orders of magnitude faster. The mechanism can be described by n=1 and p=3. At low stress (≤0.1 MPa) and small grain size (1 mm),
the polycrystalline model predicts a transition to melt-assisted creep at T∼272.6 K. A stress-exponent of n=1 is observed in some creep experiments on polycrystalline ice close to Tm. Experimental data on the wire regalation velocity confirms the dependency on wire diameter, but there is no conclusive creep data close to the melting temperature to contradict or confirm the grain-size exponent in polycrystalline ice. Melt-assisted creep may play a role in the rheology at the base of ice sheets, where the temperature is close to Tm. However, there is substantial ice core data that show an increase in grain size close the base of the ice sheets. This suggest that grain growth by migration recrystallization is occurring, which does not favor grain size sensitive (GSS) creep. In some layers of the ice sheet, the presence of impurities inhibits grain growth and melt-assisted creep may be the dominant deformation mechanism. Conventional flow laws for ice rheology do not consider this effect and may underestimate strain-rates in these layers. However, in these layers a strong CPO is observed and it remains questionable whether this can be reconciled with flow by grain boundary melt transport as dominant deformation mechanism. | |
