How the physics timestep controls tropical cyclone frequency in the high resolution Community Earth System Model

Abstract:

High resolution Earth system models (ESMs) have shown improved capability to simulate tropical cyclones (TCs) over the past decade. While previous work has noted the potential sensitivity of modeled TC climatology to dynamical cores, physical parameterizations, and surface model components, less attention has been paid to more subtle design aspects such as physics-dynamics coupling strategy and model timestep choice. With multi-model ESM comparisons becoming more popular, it is important to understand mechanisms that contribute to model spread in order to better constrain uncertainty in future projections.

In this presentation, we assess TC climatologies from three historical 25km simulations of the Community Earth System Model (CESM). All three runs use identical codebases; the sole difference between the configurations is the timestep of the subgrid parameterization suite (i.e., the number of dynamics timesteps per physics timestep). Simulations with a physics timestep of 450s produce approximately twice as many TCs per year as with 1800s, with almost identical differences seen in integrated metrics such as TC days and accumulated cyclone energy. We show that mean climate fields associated with TC genesis are largely insensitive to physics timestep, implying short-term variability leads to this response. Analyses of both grid point extrema and the spatial scales associated with cyclonic relative vorticity in the tropics show that a competition to remove column instability arises, with longer (shorter) timesteps being dominated by the deep convective (large-scale microphysics) schemes. The "winner" of this competition determines the spatial scales of short-term vorticity aggregation and subsequent TC genesis frequency. We complete additional sensitivity experiments to confirm this behavior by producing a "timestep-like" response in TC climatology while holding the physics timestep fixed. This work highlights the impact of physics-dynamics coupling on extreme weather events simulated by high resolution ESMs.