Long-term verification of HIRLAM and ECMWF forecasts over Southern Europe: History and perspectives of Numerical Weather Prediction at AEMET

The evolution of Numerical Weather Prediction (NWP) activities at the Spanish State Meteorological Agency (AEMET) is reviewed not only for historical reasons but mainly to demonstrate the value added by Limited Area Models to deterministic and probabilistic short range forecasting. This paper highlights the decisive role of international cooperation in the general progress achieved by the NWP area of research. In fact, most of the activities here described have taken place in collaboration with the HIRLAM Consortium. The main milestones of HIRLAM model development are analyzed. Long-term performance assessments of HIRLAM operational runs illustrate the positive trend of forecast quality brought by the different scientific developments. Besides, the evolution of the European Centre for Medium-Range Weather Forecasts (ECMWF) operational forecast skill in South-western Europe is presented. HIRLAM complements the global model forecasts in the short range and adds value especially for some surface parameters. The HARMONIE system is the result of the collaboration between HIRLAM and ALADIN Consortia addressing km scale operational forecasting. Some promising results using this new system are shown and the main ongoing work to improve HARMONIE is outlined. Innovative developments on Regional Ensemble Prediction Systems (SREPS) conducted over the last decade at AEMET – within HIRLAM and in collaboration with several NWP centres – are complementing the probabilistic ECMWF forecasts for the short range. Finally, the main achievements and future prospects of SREPS in the mesoscale are presented.

► We use milestones of NWP activities at AEMET to show de value of operational LAM.
► International collaboration has played a crucial role for NWP progress in Europe.
► HIRLAM has always provided more accurate surface parameters than ECMWF.
► Parameters having a high social impact are very much improved by km-scale HARMONIE.
► SREPS and GLAMEPS provide uncertainty assessment that complements ECMWF in the short-range.