Failure of Anisotropic Sea Ice Erland M. Schulson Ice Research Laboratory Thayer School of Engineering Dartmouth College Hanover,NH ONR Workshop on "PIPS 3.0" Monterey,CA 7-9 July 1998 Satellite images of the ice cover on the Arctic Ocean often show patterns of intersecting leads of 20-40o acute angle 2a (Marko & Thomson 1977, JGR 82,979; Erlingsson 1991, Annals Glac. 15,73) . Similar patterns of 2a=38-45o are seen in the laboratory in specimens of columnar-grained sea ice deformed within the brittle regime under across-column biaxial compression (Iliescu and Schulson 1998, unpublished). The "minileads" are symmetrically oriented w.r.t. the direction of the more compressive stress. Satellite images occassionally show wing-like leads (e.g., Schulson and Hibler 1991, J.Glac.,37,319; SHEBA 1997/8). Wing cracks are also seen in the laboratory (Cannon et al. 1990, Acta.metall. mater.,38,1955; Schulson 1990, ibid. 1963; Schulson 1997, J.Phys. Chem.B 101, 6254). These observations suggest that similar failure mechanisms operate in the field and the lab, even though the two scales differ by about six orders of magnitude. If so, then the formation of leads may be viewed essentially as a process of brittle compressive failure, preceeded perhaps by plastic flow. Non-plastic deformation should then be included in "PIPS 3.0". Observations in the lab show that compressive "minileads" are shear faults that develop rapidly near the peak on the compressive stress-strain curve. The faults are localized bands of damage which consist of wing cracks and splay cracks (Schulson et al. 1998, submitted to JGR). The splay cracks (previously termed "feather cracks" by Smith and Schulson 1993, Acta metall. mater. 41,153) are secondary tensile cracks (wings are primary cracks) that originate from one side of sliding inclined cracks. Splay cracks are important because they create closely spaced sets of slender fixed-free microcolumns which bend under compression. It is suggested that the shear fault is triggered when the near-surface microcolumns break under frictional sliding of their free ends, like the breaking of teeth in a comb under a sliding thumb. Modeling of the terminal failure stress based upon this mechanism is in reasonable agreement with observation, and suggests that to a first approximation the terminal strength scales directly with fracture toughness Kc and inversely with the product (1-m)(sh)0.5 where m is the ice/ice coefficient of sliding friction, s is the microcolumn slenderness ratio and h is the column length. Confinement raises the failure stress. Assuming that a similar process operates in the field, it is suggested that: (i) intersecting leads form under low-confinement biaxial compression within the regime of brittle (vs ductile) behavior; (ii) leads are oriented symmetrically w.r.t. the maximum principal stress and define conjugate planes; (iii) leads are localized bands of damage and form when a deformation-induced instability develops within the material of the ice sheet; (iv) the instability is triggered by the breaking of fixed-free columns under frictional shear end loading; (v) the terminal failure stress of an ice sheet (i.e., the stress acting when the lead propagates all the way through the sheet) should be several orders of magnitude lower than in the lab owing to the liklihood that the size of the field cracks that constitute pre-fault damage scale with the size of the ice sheet; and (vi) a frictional sliding or Coulombic failure criterion is probably more appropriate than one based upon plastic flow to describe lead formation. With the objective of improving the physical content of "PIPS 3.0", it is propsed to study the formation and properties of meter-sized "minileads", using IRL's new in-pond loading frame situated at CRREL. Questions to be addressed include: Does the failure mode change from splitting to shear faulting (i.e. leading) to spalling as the ratio of minor to major stress increases? Changes of this kind have been observed in the laboratory (Schulson and Nickolayev 1995,JGR 100, 22383) and are predicted from recent numerical modeling of lead formation (Hibler and Schulson 1998, submitted to JGR) Within the low-confinement regime, does the inclination of the minileads w.r.t. the direction of higher compressive stress increase as the ratio of the minor/major stress increases, in accord with the Hibler-Schulson (1998) modeling? To what degree, if any, does additional anisotropy introduced by crystallographic c-axis alignment affect lead orientation? What is the failure stress of floating blocks of ice containing refrozen minileads and how does the strength vary with lead length,width and orientation and with the thickness of the new ice within the lead? In other words, to what extent do refrozen leads change inelastic behavior from isotropic (in the loading plane ) to anisotropic?