The Effects of a New Boundary Forcing Approach on Model's Near-Surface Variables

Physical parameterization schemes are among the main sources of uncertainty in numerical weather/climate predictions. Numerical prediction models can be used to assess regional climate information through dynamical downscaling of general circulation model solutions, increasing prediction/projection uncertainty. In order to reduce the uncertainty in downscaled global fields, a new boundary forcing that combines scale-selective bias correction with precipitation assimilation was proposed. In this study, the effects of this new approach on long-term simulations of a regional spectral model were examined and revealed that near-surface temperature and winds were improved, in comparison with a control solution in which only a scale-selective bias correction was applied. Similar to the spectral nudging, scale-selective bias correction prevents internal states generated by the regional model that are inconsistent with the large-scale solution. Satellite-based estimates were used in the precipitation assimilation procedure to improve the regional model's cumulus-convection scheme.