A Nonhydrostatic Version of the Penn State–NCAR Mesoscale Model: Validation Tests and Simulation of an Atlantic Cyclone and Cold Front

Jimy Dudhia The Pennsylvania State University, University Park, Pennsylvania

Search for other papers by Jimy Dudhia in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

A nonhydrostatic extension to the Pennsylvania State University-NCAR Mesoscale Model is presented. This new version employs reference pressure as the basis for a terrain-following vertical coordinate and the fully compressible system of equations. In combination with the existing initialization techniques and physics of the current hydrostatic model, this provides a model capable of real-data simulations on any scale, limited only by data resolution and quality and by computer resources.

The model uses pressure perturbation and temperature as prognostic variables as well as a B-grid staggering in contrast to most current nonhydrostatic models. The compressible equations are solved with a split-time- step approach where sound waves are treated semi-implicitly on the shorter step. Numerical techniques and finite differencing are described.

Two-dimensional tests of flow over a bell-shaped hill on a range of scales were carded out with the hydrostatic and nonhydrostatic models to contrast the two and to verify the dynamics of the new version.

Several three-dimensional real-data simulations show the potential use of grid-nesting applications whereby the model is initialized from a coarser hydrostatic or nonhydrostatic model output by interpolation to a smaller grid area of typically between two and four times finer resolution. This approach is illustrated by a simulation of a cold front within a developing midlatitude cyclone, and a comparison of the front to observations of similar features.

The cold-frontal boundary was sharply defined at low levels and consisted of narrow linear updraft cores. At 2–4-km altitude this structure gave way to a more diffuse boundary with apparent mixing. Mechanisms are presented to explain these features in terms of inertial and shearing instability. Convection embedded in the frontal band formed a prefrontal line at later stages.

Finally, sensitivity studies showed that the frontal band owed its narrowness to the concentrating effect of latent heating. The frontal ascending branch was supplied by a strong easterly ageostrophic flow in the warm sector.

Abstract

A nonhydrostatic extension to the Pennsylvania State University-NCAR Mesoscale Model is presented. This new version employs reference pressure as the basis for a terrain-following vertical coordinate and the fully compressible system of equations. In combination with the existing initialization techniques and physics of the current hydrostatic model, this provides a model capable of real-data simulations on any scale, limited only by data resolution and quality and by computer resources.

The model uses pressure perturbation and temperature as prognostic variables as well as a B-grid staggering in contrast to most current nonhydrostatic models. The compressible equations are solved with a split-time- step approach where sound waves are treated semi-implicitly on the shorter step. Numerical techniques and finite differencing are described.

Two-dimensional tests of flow over a bell-shaped hill on a range of scales were carded out with the hydrostatic and nonhydrostatic models to contrast the two and to verify the dynamics of the new version.

Several three-dimensional real-data simulations show the potential use of grid-nesting applications whereby the model is initialized from a coarser hydrostatic or nonhydrostatic model output by interpolation to a smaller grid area of typically between two and four times finer resolution. This approach is illustrated by a simulation of a cold front within a developing midlatitude cyclone, and a comparison of the front to observations of similar features.

The cold-frontal boundary was sharply defined at low levels and consisted of narrow linear updraft cores. At 2–4-km altitude this structure gave way to a more diffuse boundary with apparent mixing. Mechanisms are presented to explain these features in terms of inertial and shearing instability. Convection embedded in the frontal band formed a prefrontal line at later stages.

Finally, sensitivity studies showed that the frontal band owed its narrowness to the concentrating effect of latent heating. The frontal ascending branch was supplied by a strong easterly ageostrophic flow in the warm sector.

Save