Project at a Glance
Title: Improving the accuracy of mixing depth predictions from the mesoscale meteorological model MM5.
Principal Investigator / Author(s): Alapaty, Kirankumar V
Contractor: MCNC - North Carolina Supercomputing Center
Contract Number: 96-319
Research Program Area: Atmospheric Processes
Topic Areas: Modeling
This research work was initiated to determine modeling aspects that control the accuracy of the estimated mixed-layer depths over Central California, calculated by meteorological models. Most importantly, we focused on surface and large-scale atmospheric process representations and studied their effects in controlling the growth of the daytime boundary layer over this region.
First, using a 1-D boundary layer model, we studied and improved the representation of surface processes. We developed a technique to facilitate the assimilation of surface data into the model without damaging the modeled mixed-layer structures. We then demonstrated that the modeled boundary layer predictions are improved and are closer to the observations when using this technique. Next we developed and tested a formulation to improve the estimation of surface latent heat fluxes; it allows the vegetation component to be included explicitly in the land surface parameterization. We found that this formulation improved modeled surface fluxes and thus improved boundary layer predictions. These two new methods need further testing in a 3-D mesoscale model.
Second, using the Mesoscale Model, Version 5 (MM5) with a nested-grid configuration, we studied the interactions of large-scale processes with the growth of the boundary layer. We performed several sensitivity studies to understand the interactions of processes and to weed out insignificant aspects of model configurations. We found that increasing model vertical resolution from 32 to 62 layers did not lead to a significant skill increase in the model estimations. Inclusion of a large outer domain covering the East Pacific Ridge helped to yield better model solutions in the inner domains. Removal of analysis nudging in coarser domains in the lowest 1.5 km also improved the model solutions in the innermost domain in which no nudging was performed. Boundary layer initialization over the marine environment did not much improve the model solutions. Finally, application of a physically robust boundary layer scheme resulted in further improvements in the model predictions compared to a simple boundary layer scheme. Thus, we found that an enhanced model configuration did lead to improved mixed-layer predictions over Central California.
For questions regarding this research project, including available data and progress status, contact: Heather Choi at (916) 322-3893
Stay involved, sign up with ARB's Research Email Listserver