Clarification on wind direction code 99 in WDQMS

Clarification on wind direction code 99 in WDQMS

بواسطة - Sonia KEBAILI
عدد الردود: 3

Dear WDQMS experts,

I am reviewing data for a station in WDQMS. For a full month, we expect 744 values for the wind direction to have the station indicator appear in green. Out of these 744 observations, 261 have the wind direction coded as 99.

I understand that 99 represents a variable wind, not missing data. However, in WDQMS, the station indicator appears in orange, which seems to treat the 99 values as missing.

Could you please clarify how WDQMS interprets these 99 values? Should variable wind be considered as valid data for the station status, or is it automatically counted as missing?

Thank you very much for your guidance.

Best regards,


رداً على Sonia KEBAILI

Re: Clarification on wind direction code 99 in WDQMS

بواسطة - Cristina Prates
Dear Sonia,

Thank you very much for your pertinent question.
WDQMS results are derived from the data assimilation (DA) systems of various NWP models, and different modules may handle observations in different ways. The statistics provided in WDQMS are calculated for the zonal and meridional wind components (u and v), as these are the variables that are assimilated and provided by NWP Centres. Naturally, when wind direction is reported as variable, the u and v components cannot be computed by the DA.
That said, I would expect the ECMWF availability results not to be affected by this. When variable wind is reported, the departures for u and v cannot be calculated for that specific observation; however, the wind observation itself should not be considered missing. To investigate this further, could you please provide a concrete example where you have encountered this issue? For instance, a station ID that reported variable wind on a specific day.
With such an example, I will be able to examine the case in more detail.
Kind regards,
Cristina
رداً على Cristina Prates

Re: Clarification on wind direction code 99 in WDQMS

بواسطة - Sonia KEBAILI

Dear Cristina,


Thank you very much for your reply and for the clarification regarding the treatment of variable wind observations in WDQMS results.

With regard to the case in question, the station concerned is:

WIGOS-ID: 0-20000-0-60468

For the month of January, we identified 261 occurrences in which variable wind was reported with the value 99 for this station. At the same time, WDQMS (ECMWF) indicates that only 480 observations were received out of 744 expected observations.

Given this, we are wondering whether these 261 cases of variable wind may have contributed, wholly or partially, to the difference between the 744 expected observations and the 480 received observations shown in WDQMS. From a numerical point of view, the gap amounts to 264 observations, which appears very close to the 261 occurrences of variable wind recorded during the month.

For a quicker verification, you may also refer to yesterday’s data (23 March 2026) for the same station. On that day, WDQMS shows 15 observations received out of 24 expected, while the 9 missing observations correspond to cases of variable wind. The observations for that day are included in the attached file for your reference.


Thank you once again for your support and assistance.


Kind regards,


رداً على Sonia KEBAILI

Re: Clarification on wind direction code 99 in WDQMS

بواسطة - Cristina Prates

Dear Sonia, 

Once again, thank you for your valuable enquiry. I did investigate the station you mentioned, and indeed these wind observations are not included in the WDQMS statistics. I have actually included the text below in the WDQMS FAQ so that information is available in the WDQMS documentation.

Regards,

Cristina

WDQMS FAQs (https://confluence.ecmwf.int/display/WIGOSWT/FAQs)

How are “variable wind” measurements reported in SYNOP considered in WDQMS statistics?

In SYNOP reports, a variable wind (or light and variable conditions) is encoded with a wind direction of 99, indicating a valid observation rather than missing data. However, WDQMS statistics are based on the 10m zonal and meridional wind components from NWP DA systems. For 10m variable wind measurements, these components cannot be derived, so O–B departures cannot be calculated. As a result, such reports are excluded from WDQMS statistics. While meteorologically valid, variable 10m wind reports are not usable for DA‑based metrics in WDQMS. A high frequency of these reports can reduce effective data availability and may negatively affect the station availability scores for wind.