Considerations on domain location according to the jump of resolution between the driving data and the nested regional climate model within the Big-Brother experiment.

Abstract:

Many studies have shown the importance of choosing the domain size adequately for dynamical downscaling with nested regional climate models. It is well known that domain should not be too large to avoid large departure from the driving data, and not be too small to provide sufficient distance from the lateral inflow to allow a full development of the small-scale features resolved by the increase resolution. Although practitioners of dynamical downscaling are well aware that the jump of resolution between the driving data and the nested regional climate model impacts the simulated climate, the issue has never been properly study. Larger is the jump of resolution, larger is the distance from the lateral inflow to fully develop the small-scale features permitted by the increase resolution.

Our investigation compares direct nesting to achieve a grid mesh of 0.15^o from driving data at 3.6°, 1.8^o, 0.45° and 0.15° using the perfect-prognostic approach of the Big-Brother protocol. The results show that the small-scale transient-eddy component struggles to be fully developed with reduced resolution of the driving data. Overall, this study suggests that domain location (i.e. domain of interest or subsequent nested domains) must be chosen carefully according to the jump of resolution to allow the optimal development of small-scale features allowed by the increase resolution of the nested model.