The Gridpoint Statistical Interpolation (GSI) software is used to convert the KPOL radar reflectivity into latent heating terms. The latent heating terms are applied to the Weather Research and Forecasting model during a pre-forecast. During the pre-forecast the latent heating terms for a series of radar observations are applied for the time period leading up to the approximate valid time of the radar observation. This approach is similar to that used by the High Resolution Rapid Refresh (HRRR) model over CONUS, but for this study over Kwajalein there is only a single radar and very limited non-radar observations. We experiment with different pre-forecast lengths and investigate the potential benefits of this approach in terms of the short term reflectivity forecast. Various simulations will be evaluated using metrics such as fractions skill scores and will be presented at the conference.