2.1
Striving towards simulating all known atmospheres equally well: vertical coordinate and pressure-gradient force
Timothy E. Dowling, University of Louisville, Louisville, KY
This talk presents a review of two new algorithms in the Explicit Planetary Isentropic-Coordinate (EPIC) atmospheric model designed to accurately simulate both terrestrial and gas-giant atmospheres. The model now has a pure-sigma coordinate region underlying a hybrid sigma-theta region, which can be used as a planetary boundary layer for terrestrial planets or the deep atmosphere for gas-giant planets. In the sigma region, the potential temperature, theta, is treated as a prognostic variable and the pressure is a diagnostic variable of the vertical coordinate and the surface pressure, as is standard. The novelty of our approach is to do the complement in the hybrid region: we treat the pressure as a prognostic variable and the potential temperature as a diagnostic variable of the vertical coordinate and the pressure. The effect of heating enters exclusively through the vertical velocity, which is a diagnostic variable in this hydrostatic model, just as is the case in a pure-theta coordinate model. This contrasts with other hybrid-coordinate models for which potential temperature is treated as a prognostic variable, which we argue undermines the full potential of using hybrid isentropic coordinates.
Having added a sigma region to a model that once had a pure-theta coordinate, we now have all the problems that come with sigma, most notably the large errors from the horizontal pressure-gradient force (PGF) terms in steep topography. We have experienced these problems in Venus superrotation spinup simulations with topography. We argue that for hybrid-coordinate models, with or without topography, treating the PGF with a weak-formulation (finite volume) instead of a strong formulation (traditional finite difference) has advantages for both terrestrial and gas-giant atmospheres. We have developed a fully three-dimensional finite-volume version of the PGF that we fit to our existing C-grid, which is now the model's default, and explain how it works.
This research is supported by NASA's Planetary Atmospheres and Outer Planets Research Programs and NSF's Planetary Astronomy Program.
Session 2, Numerical modeling of planetary meteorology and climate dynamics
Wednesday, 20 January 2010, 4:00 PM-5:30 PM, B314
Next paper