The WIGOS Monitoring Centres that are currently providing near-real-time NWP monitoring reports are: DWD, ECMWF, JMA and NCEP for surface land stations and upper-air land stations. As it is based on the feedback from the assimilation systems, the monitoring information provided by the NWP Centres reflects the main characteristics - strengths and weaknesses - of each system. By combining the information across Centres the WDQMS monitoring tool exploits the strengths of each NWP assimilation system and provides a more robust statement regarding availability and quality than what would be possible from a single NWP centre.
5.1. Individual NWP Centre
Regarding the generation of the monitoring performance reports for both land surface stations and upper-air land stations, there are four NWP monitoring centres to choose from: DWD, ECMWF, NCEP and JMA. Note that the four 6-hourly reports provided by the participating WIGOS Monitoring Centres are based on the observations that were made available to the assimilation system at each Monitoring Centre, which is used as a proxy for the observations available on the GTS. The way the four different NWP data assimilation systems deal with observations varies, which explains some of the differences seen on the performance reports across Centres. Below are highlighted some of the procedures used in the different NWP centres that should be taken into consideration when comparing the results between NWP centres.
NCEP monitoring reports show only the data that pass a quality control step prior to DA, therefore part of the data available that is deemed to be of poor quality or duplicate is filtered out and not available to the DA. Another important aspect of its DA system is the assimilation of surface pressure observations provided by METAR reports - mainly over US - rather than from SYNOP reports when both reports are available in the same location, therefore some SYNOPs observations over this region may not show up in NCEP maps.
ECMWF, on the other hand, provides monitoring information of all observations available to the DA, even data that is deemed to be of poor quality and rejected and/or blacklisted by the system. The only case in which observations may not be available is when a serious coding issue prevents the acquisition system to extract any meaningful value. Also, when SYNOP and METAR reports are available from the same station, ECMWF gives preference to the SYNOP observation.
DWD provides monitoring information for all observations available to their DA system. However, in the case of surface pressure the model background value is not available if that observation is not used in the assimilation. Therefore, the surface pressure O-B departures cannot be calculated for that particular observation, and the O-B departures field is flagged (has missing value in the monitoring report) despite the fact that the observation itself has been made.
JMA provides O-B values for all observations available to their assimilation system even for the observations that are not used in the assimilation. It is worth mentioning that they tend to assimilate pressure observations every 3 hours rather than hourly.
5.2. Combined (All)
A summary combining all the performance reports from the different WIGOS Monitoring Centres is generated when option “All” is selected. This means that for each and every period, variable and monitoring category (availability and quality), there is a combined performance report available. The overarching rule for aggregating the information across NWP Centres is that the best data availability and quality result is used to generate a combined performance report. The combined availability performance report will be generated by taking the maximum availability result amongst all Centres (i.e. the information will come from the Centre with the highest total in the period). Regarding the quality, the result that will be retained is the one from the NWP model whose background field is closest to the observed value (i.e., the information will come from the Centre with the smallest O-B departures averaged over the period of interest).