We investigate the ability of viscous-plastic (VP) sea-ice models with an elliptical yield curve and normal flow rule to reproduce the shear and divergence distributions derived from the RADARSAT Geophysical Processor System (RGPS). In particular, we reformulate the VP elliptical rheology to allow independent changes in the ice compressive, shear and isotropic tensile strength parameters (P*; S*; and T*, respectively) in order to study the sensitivity of the deformation distributions to changes in the ice mechanical strength parameters. Our 10 km VP simulation with standard ice mechanical strength parameters P* = 527.5 kN m(-2), S* = 56.9 kN m(-2), and T* = 50 kN m(-2) (ellipse aspect ratio of e=2) does not reproduce the large shear and divergence deformations observed in the RGPS deformation fields, and specifically lacks well-defined, active linear kinematic features (LKFs). Probability density functions (PDFs) for the shear and divergence of are nonetheless not Gaussian. Reducing the ice compressive strength (with constant S* and T*) or increasing the ice shear strength (with constant P* and T*) both results in shear and divergence PDFs in better agreement with RGPS distributions. The isotropic tensile strength of sea ice does not significantly affect the shear and divergence distributions. When considering additional metrics such as the ice drift error, mean ice thickness fields, and spatial scaling of the total deformations, our results suggest that reducing the ice compressive strength P* (while keeping S* constant, i.e. reducing the ellipse aspect ratio) is a better solution than increasing the shear strength to improve simulations of the Arctic seaice cover with the VP elliptical rheology.
