Central and northern Argentina are one of the most favorable regions for the occurrence of high impact meteorological events. These events can generate intense precipitation, large hail and/or extreme winds, causing enormous damages to the population. Therefore, it is essential to make progress in improving very short-term weather forecasts (0-6 hours) of this type of events.

In this work, a very-short term forecast model based on extrapolation techniques of the reflectivity field is developed and evaluated. This method assumes that precipitating systems can be described from their movement, estimated from the latest available information and consists in two steps. Firstly, it estimates a vector field describing the displacement of the reflectivity field. Secondly, this motion field is used to extrapolate the reflectivity field displacement using a semi-lagrangian advective method. In this work we compare different variants in the algorithm that estimates the displacement of the reflectivity field and its impact on 2-hr reflectivity field forecasts.

Experiments were performed for a case that represents the evolution of a convective system over central Argentina. We use synthetic radar data generated from a numerical simulation with high resolution that explicitly resolved convection dynamics.