VALIDATION EXPERIMENTS FOR A NESTED GRID-POINT REGIONAL FORECAST MODEL

Some experiments are carried out with a nesting strategy that is free of numerical problems near the lateral boundaries of the driven regional forecast model. As a first step, the high‐resolution, nested grid‐point model is compared with a high‐resolution, global spectral model in order to show that the differences are small over the region of interest. Subsequently, some intercomparison experiments are made with two versions of the nested model. These additional tests are used to determine the most efficient nesting configuration that will give the desired level of accuracy. It is found that the proposed configuration is able to produce very accurate integrations over a small region. © 1990 Taylor & Francis Group, LLC.