Tropical cyclone activity in nested regional and global grid-refined simulations

The capacity of two different grid refinement methods-two-way limited area nesting and variable-mesh refinement-to capture Northwest Pacific Tropical Cyclone (TC) activity is compared in a suite of single-year continuous simulations. Simulations are conducted with and without regional grid refinement from approximately 100-20 km grid spacing over the Northwest Pacific. The capacity to capture smooth transitions between the two resolutions varies by grid refinement method. Nesting shows adverse influence of the nest boundary, with the boundary evident in seasonal average cloud patterns and precipitation, and contortions of the seasonal mean mid-latitude jet. Variable-mesh, on the other hand, reduces many of these effects and produced smoother cloud patterns and mid-latitude jet structure. Both refinement methods lead to increased TC frequency in the region of refinement compared to simulations without grid refinement, although nesting adversely affects TC tracks through the contorted mid-latitude jet. The variable-mesh approach leads to enhanced TC activity over the Southern Indian and Southwest Pacific basins, compared to a uniform mesh simulation. Nesting, on the other hand, does not appear to influence basins outside the region of grid refinement. This study provides evidence that variable mesh may bring benefits to seasonal TC simulation over traditional nesting, and demonstrates capacity of variable mesh refinement for regional climate simulation.