Predictability experiments with a baroclinic model

Abstract

Abstract A series of model predictability experiments has been performed using the baroclinic primitive equation spectral model developed by Staniforth and Daley (1977). These experiments were designed to show the difference in predictability decay rates for geostrophic and ageostrophic initial perturbations. The difference in predictability decay rates has implications for the optimal specification of the initial state as discussed in Daley (1980). The results showed that the predictability decay rates from initial ageostrophic perturbations were smaller than those from geostrophic perturbations for large vertical scale. The predictability decay rates of the two types of initial perturbations tended to be similar for vertical scales smaller than 100 m. Results from a model containing a moist convective adjustment parameterization indicated a larger predictability decay rate from initial ageostrophic perturbations.