Convective-scale Regional Ensemble based Data Assimilation and Forecast System for the National Meteorological Service of Argentina

GARCIA SKABAR, YANINA; DILLON, MARIA EUGENIA; SAULO, CELESTE; SACCO, MAXIMILIANO; MATSUDO, CYNTHIA; CUTRARO, FEDERICO; GÓMEZ MAYOL, MAILÉN; RUIZ, JUAN; VIDAL, LUCIANO; RIGHETTI, SILVINA; ALVAREZ IMAZ, MILAGROS; DE ELÍA, RAMÓN; SALA, VERÓNICA; ETALA, PAULA; CAMPETELLA, CLAUDIA; LOYBER, PABLO; MININNI, PABLO; MIYOSHI, TAKEMASA; KALNAY, EUGENIA

Virtual

Workshop; 7th WMO Workshop on the Impact of Various Observing systems on NWP; 2020

Organización Meteorológica Mundial

Over the last few years, at the National Meteorological Service of Argentina (SMN) many efforts have been carried out to improve the quality of its Regional Numerical Weather Prediction (RNWP) system. A high resolution (4 km) Weather Research and Forecasting (WRF) model has been run operationally since 2016 for Southern South America, resulting in an improvement of the tools available for forecasters and users (García Skabar et al. 2018).Simultaneously, and being aware that both, data assimilation system and probabilistic forecasts are needful achievements for a developing country Meteorological Service, different experiments were accomplished using the Local Ensemble Transform Kalman Filter (LETKF). Positive impacts of assimilating conventional, satellite and radar data were documented over our region (eg. Dillon etal. 2016, Maldonado et al. 2019, Dillon et al. 2019).Motivated by these promising results and with the recent acquisition of a High Performance Computing (HPC) system at SMN, which increased 40 times the processing velocity in-house, the implementation of an operational high resolution LETKF-WRF for Southern South America is becoming possible. A successful joint venture between academia, weather service and a private company was needed to support the acquisition of this new HPC. With improved infrastructure, a convective permiting multi physics 40-member ensemble was designed to assimilate data fromthe following sources: surface stations, automatic stations, radar, geostationary and polar satellites, buoys and ships, radiosondes, aircrafts (AMDAR programme). A set of 40-member hourly analysis are generated along with 20-member 48-h forecasts every 6 hours. A ​ preliminary evaluation of this 4 km ​ Regional Ensemble based Data Assimilation and Forecast System will be shown. This represents a plausible strategy for countries that may not run global models, but still need to benefit from ingesting high resolution data into their operational systems. In a context where global models are increasing its resolution very fast, it is necessary to add value to the enhanced resolution, and this strategy proves to be adequate to fulfill this necessity.