All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 39 13 1
PDF Downloads 16 7 0

Impact of a Newtonian Assimilation and Physical Initialization an the Initialization and Prediction by a Tropical Mesoscale Model

View More View Less
  • 1 Malaysian Meteorological Service, Jalan Sultan, Selangor, Malaysia
Restricted access

Abstract

This study illustrates the capability of Newtonian nudging and physical initialization in improving the initialized state and forecasts in the Florida State University high-resolution regional tropical mesoscale model. In particular it is shown that this form of initialization leads to major improvement in the precipitation forecasts. The precipitation spinup, precipitation phase shift, and overactive convection over the tropical region that are inherent characteristics of prediction models initialized through analyses produced by an intermittent data assimilation scheme are minimized by the initialization.

It is shown that the coupling of Newtonian nudging of the nondivergent component of the wind and surface pressure with physical initialization of cumulus parameterization, surface latent heat flux over the rain areas, and outgoing longwave radiation enables the model to build the required temperature, moisture, and velocity divergence distribution to produce the desired precipitation. Of these, the initialization of cumulus parameterization and surface latent heat flux show the most impact. The cumulus initialization as illustrated was done through a reanalysis of the humidity fields over the cloud depth via a “reverse Kuo” algorithm. The algorithm constrained the Kuo cumulus parameterization scheme of the model to precipitate according to prescribed rates. The initialization of surface latent heat flux was done through a reanalysis of the humidity field at the lowest level of the model through a “reverse similarity” algorithm that constrained the surface latent heat flux of the model to be similar to prescribed values.

Abstract

This study illustrates the capability of Newtonian nudging and physical initialization in improving the initialized state and forecasts in the Florida State University high-resolution regional tropical mesoscale model. In particular it is shown that this form of initialization leads to major improvement in the precipitation forecasts. The precipitation spinup, precipitation phase shift, and overactive convection over the tropical region that are inherent characteristics of prediction models initialized through analyses produced by an intermittent data assimilation scheme are minimized by the initialization.

It is shown that the coupling of Newtonian nudging of the nondivergent component of the wind and surface pressure with physical initialization of cumulus parameterization, surface latent heat flux over the rain areas, and outgoing longwave radiation enables the model to build the required temperature, moisture, and velocity divergence distribution to produce the desired precipitation. Of these, the initialization of cumulus parameterization and surface latent heat flux show the most impact. The cumulus initialization as illustrated was done through a reanalysis of the humidity fields over the cloud depth via a “reverse Kuo” algorithm. The algorithm constrained the Kuo cumulus parameterization scheme of the model to precipitate according to prescribed rates. The initialization of surface latent heat flux was done through a reanalysis of the humidity field at the lowest level of the model through a “reverse similarity” algorithm that constrained the surface latent heat flux of the model to be similar to prescribed values.

Save