Kendzierski, Sebastian2018-08-062018-08-062017Badania Fizjograficzne, Seria A, Tom 68, 2017, s. 33-46.2081-6014http://hdl.handle.net/10593/23649The article presents the results of a study comparing the reliability of the GFS and WRF numerical models using different spatial resolutions. The study was carried out on the basis of a series of model and observational data for a period of 13 months from January 2015 to January 2016 with computational grid coordinates corresponding to the location of Poznań. Forecast quality was compared for basic meteorological parameters – air temperature, atmospheric pressure and wind speed. The quality of forecasts for the meteorological parameters included in the study decreased systematically in subsequent time steps in the case of both the WRF and the GFS models. It improved in the initial hours of simulation after a higher resolution was applied for the computational grid in the WRF model (1,75 km) as can be seen from the minimal Mean Error values and very high correlation. For forecast time t = 0–60 h, WRF generated better results than GFS, especially when it comes to the quality of forecasting the atmospheric pressure field.enginfo:eu-repo/semantics/openAccessforecast verificationGFSWRFair temperatureair pressurewind speedPoznańArtykułhttps://doi.org/10.14746/bfg.2017.8.3