@misc{81582,
  author = {Thomas Haiden and Martin Janousek and Frédéric Vitart and Maliko Tanguy and Fernando Prates and Matthieu Chevalier},
  title = {Evaluation of ECMWF forecasts},
  abstract = {This report provides a summary of ECMWF’s forecast performance, covering medium, sub-seasonal and seasonal forecast ranges. Headline scores have been adopted by ECMWF in collaboration with its member states to monitor the evolution of various aspects of forecast skill. The report gives updates on these scores, as well as supplementary scores to help provide a more complete assessment of forecast skill. An important recent change has been the increase in resolution of the ensemble forecast. With the implementation of model cycle 48r1 in June 2023 it now matches the high-resolution run, and this brings clear improvements in forecast skill. The primary focus of this summary is the medium range, specifically the forecast performance for upper-air variables. It is shown that in this respect ECMWF continues to have an overall lead among centres. For surface variables, other centres have partially taken the lead, especially in the short range, but significant improvements of ECMWF forecasts due to model cycle 48r1 can be seen there as well, such as a reduction of large 2-m temperature and 10-m wind speed errors in the ensemble forecast. In the sub-seasonal forecast range, the increase of the frequency of forecasts from bi-weekly to daily and the increase of the number of ensemble members from 50 to 100 has enhanced the effective value of the forecast. On the seasonal timescale, the change of Pacific SSTs from near-neutral to El Niño conditions in 2023 and return towards neutral conditions in 2024 was well predicted. For the first time this report also includes scores from machine-learning forecasts (AIFS), higher-resolution forecasts (DestinE continuous Extremes Digital Twin), and hydrological forecasts from the Copernicus Emergency Management Service (CEMS), in addition to atmospheric composition forecasts from the Copernicus Atmosphere Monitoring Service (CAMS).},
  year = {2024},
  journal = {ECMWF Technical Memoranda},
  number = {918},
  month = {09/2024},
  publisher = {ECMWF},
  url = { },
  doi = {10.21957/52f2f31351},
}