Comparison of convective initiation and evolution in 3 km WRF simulations with and without the Kain-Fritsch scheme

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner
Thursday, 27 January 2011: 11:00 AM
Comparison of convective initiation and evolution in 3 km WRF simulations with and without the Kain-Fritsch scheme
615-617 (Washington State Convention Center)
Jeffrey D. Duda, Iowa State University, Ames, IA; and W. A. Gallus Jr. and M. Segal
Manuscript (308.9 kB)

A number of central U.S. warm season mesoscale convective systems were simulated using the WRF model at 3 km grid spacing both with and without the Kain-Fritsch (KF) convective parameterization scheme (CPS) to determine if the use of a CPS in simulations at such high resolution will improve a forecast over a simulation in which no CPS is used. Two different sets of modifications to the KF scheme code were implemented also in an effort to improve model skill: 1) The main input fields to the KF scheme - vertical motion, temperature, and water vapor mixing ratio - were averaged over 5 x 5 grid boxes before the KF scheme ran, and 2) the heat and moisture tendencies, as well as the convective precipitation amount, were averaged after the KF scheme ran. The effect of both modifications is to coarsen the model grid so that the grid spacing is closer to the range at which the scheme was meant to be used, i.e., around 25 km. The WRF Developmental Testbed Center's (DTC) Model Evaluation Tools (MET) product called MODE (the Method for Object-based Diagnostic Evaluation) was used to verify the skill of these simulations by comparing the QPF of the simulations to Stage IV multi-sensor precipitation data. Large-scale forcing terms such as low-level temperature advection, surface frontogenesis, upper level divergence, and differential vorticity advection were also computed in the vicinity (temporally and spatially) of the initiation of the convection in each simulation and will be compared to observations for the events and used to gauge model skill as a function of large-scale forcing.

Also, the sensitivity of the division of total precipitation between the microphysics and convective parameterization schemes to grid spacing was tested by running a large number of two-dimensional simulations at grid spacings ranging from 100 meters to 42 kilometers. Dramatic changes were found to occur in this division within a small range of relatively fine grid spacings.