The response of a viscous-plastic dynamic-thermodynamic sea-ice model to a constant wind forcing is tested in an idealised setting. Bjornsson et al. (2001) have shown that the granular model of Tremblay and Mysak (1997) gives good results in such a set-up compared to a polynya flux model. Here it is shown that these results can be duplicated using the computationally more efficient elliptic yield curve approach by Hibler (1979) and modified Coulombic yield curve by Hibler and Schulson (2000). Some care is, however, required regarding the parametrisations of the ellipse. In addition it is shown that the new ice thickness formulation of Mellor and Kantha (1989) does not allow for proper polynya formation in the bay. In contrast the new ice thickness formulation of Hibler (1979) is found to give good results. Finally we propose a parametrisation of the ice demarcation thickness (<i>h</i><sub>0</sub>) in Hibler's formulation, based on wind speed and ice thickness.