In WDQMS the quality check of observations is NWP based. In the case of upper-air observations, the NWP data assimilation systems provide the background fields (very short range forecasts) of temperature, humidity and wind that are used as independent estimates against which observations are compared, i.e. the so call O-B departures. The statistics based on O-B departures for each of the variables, in this case root mean square error (rmse), is then compared against observational requirements of global NWP defined in OSCAR/Requirements (https://www.wmo-sat.info/oscar/applicationareas/view/1). The atmospheric column is split into two layers and rmse computed for these two layers: “Trop”, is for the first level up to 100hPa inclusive, and “Stra”, is for the 100hPa level up to the top level reported. The 6-hourly and daily maps display the highest value of rmse of the two layers “Trop” and “Stra”.