RMI operates a network of 17 automatic weather stations in Belgium. These weather stations report meteorological paramaters such as air pressure, temperature, relative humidity, precipitation (quantity,duration), wind (speed, gust, direction), sunshine duration, shortwave solar radiation and infrared radiation every 10 minutes.

The air quality measurements are interpolated on a 4x4 km² grid using Corine Land Cover.

The ceilometer CL51 employs a pulsed diode laser LIDAR technology, where short, powerful laser pulses are sent out in a vertical or near-vertical direction. The reflection of light (backscatter) caused by clouds, precipitation or other obscuration is analysed and used to determine the cloud base height, the cloud layer height and the amount of clouds (in octas) in different layers.

Since August 2019, users of the RMI smartphone app are able to send an observation of the meteorological conditions at a certain place and a certain time. The observations provide information about the weather conditions and potentially severe weather to the other users and to RMI. The collection of citizen weather reports is a valuable complement to the information obtained with the classical instruments like stations, radar and satellite. The data can be exploited for nowcasting, warnings and model verification, and eventually in assimilation. A general introduction of the data and their characteristics can be found in Reyniers et al. (2023). A basic quality control is implemented on the received observations via a plausibility check. This plausibility check determines whether an observation is plausible, suspicious or false, by comparing it to the INCA-BE nowcasting system using a simple thresholding scheme. INCA-BE is RMI's operational nowcasting system described in Reyniers et al. (2021). There is no strict spatial extent since there is no restriction at the input side: users can send observations from all over the globe. The bulk of the observations are received from within Belgium. Note that the plausibility check is not available for reports from outside Belgium.

Total lightning observations, i.e., cloud-to-ground (CG) strokes and intra/inter-cloud (IC) pulses, in Belgium. BELLS data is given in Universal Time! (local time winter = UT +1; local time summer = UT + 2). Each record contains following fields: 1. Network type, 0 = LF, 1 = VHF-TOA, 2 = VHF-ITF, 3 = VLF 2. Year, 1970 to 2032. 3. Month, with January as 1 and December as 12. 4. Day of the month, 1 to 31. 5. Hour, 0 to 23. 6. Minute, 0 to 59. 7. Second, 0 to 60. 8. Nanosecond, 0 to 999999999. 9. Latitude of the calculated location in decimal degrees, to 4 decimal places, -90.0 to 90.0. 10. Longitude of the calculated location in decimal degrees, to 4 decimal places, -180.0 to 180.0. 11. Altitude in meters, -4950 to +60535. This field will be 0 if the altitude is not provided in the lightning data. 12. The altitude uncertainty in meters. This field will be 0 if the altitude uncertainty is not provided in the lightning data. 13. Estimated peak current in kiloamps, -9999 to 9999. 14. VHF Range normalized power, -9999.0 to 9999.0 15. Multiplicity for flash data (1 to 99) or 0 for strokes. 16. Cloud pulse count for the lightning record. This field will be -1 if the lightning did not contain the cloud pulse count attribute. 17. Number of sensors participating in the solution, 2 to 99. 18. Degrees of freedom when optimizing location, 0 to 99. 19. The error ellipse angle as a clockwise bearing from 0 degrees north, 0 to 180.0 degrees 20. The error ellipse semi-major axis length in kilometers, 0 to 50.0km. 21. The error ellipse semi-minor axis length in kilometers, 0 to 50.0km. 22. Chi-squared value from location optimization, 0 to 999.99 23. Rise time of the waveform in microseconds, 0 to 99.9 24. Peak-to-zero time of the waveform in microseconds, 0 to 999.9 25. Maximum rate-of-rise of the waveform in kA/usec (will be a negative rate if discharge is negative), -999.9 to 999.9 26. Cloud indicator, 1 if Cloud-to-cloud discharge, 0 for Cloud-to-ground 27. Angle indicator, 1 if sensor angle data used to compute position, 0 otherwise 28. Signal indicator, 1 if sensor signal data used to compute position, 0 otherwise 29. Timing indicator, 1 if sensor timing data used to compute position, 0 otherwise 30. Flash ID, a unique integer value identifying the flash grouping a stroke belongs to or 0 if not set. This product is not publically available yet.

Hail products are derived from the observed vertical profiles of radar reflectivity and the NWP vertical profiles of temperature. Three types of products are generated. poh : probability of hail of any size (larger than 0.5 cm diameter)expressed in %. posh : probability of severe hail(larger than 2cm)expressed in %. mesh : maximum expected size of hailexpressed in mm of hailstone diameter. All products are generated every 5 minutes. This product is not publically available yet.

Bird density profiles are derived from weather radar volume data in real time, by the vol2bird algorithm as described in Dokter et al. (2011, 2019). The vol2bird algorithm exploits the radar reflectivity characteristics of different scatterers in the atmosphere, in order to distinguish biological from non-biological radar echoes. Once biological scatterers are isolated in the volume files, the reflectivity of these scatterers is converted in an estimate of the bird density per vertical layer of 200m, using a mean cross section of 11 cm2. The vbird profiles are provided for the following radars, with the radar owner in parentheses: Jabbeke (RMI), Wideumont (RMI), Helchteren (VMM), Zaventem (Skeyes), Herwijnen (KNMI), Den Helder (KNMI), Neuheilenbach (DWD), Essen (DWD), Abbeville (Météo-France) and Avesnois (Météo-France). References: - Dokter A.M., Liechti F., Stark H., Delobbe L., Tabary P., Holleman I., Bird migration flight altitudes studied by a network of operational weather radars, J. R. Soc. Interface, 8, 30–43, 2011, DOI 10.1098/rsif.2010.0116 - Dokter A.M., Desmet P., Spaaks J.H., van Hoey S., Veen L., Verlinden L., Nilsson C., Haase G., Leijnse H., Farnsworth A., Bouten W., Shamoun-Baranes J., bioRad: biological analysis and visualization of weather radar data, Ecography, 42, 852-860, 2019, DOI 10.1111/ecog.04028

The composite is produced every 5 minutes using the reflectivit data from the Wideumont and Jabbeke weather radars. This product is not publically available yet.

Weather Model 'Alaro' A numerical forecast model is a computer program that simulates the evolution of the atmosphere. Available parameters : Temperature, Max. temperature, Min. temperature, Dewpoint temperature, Wet-bulb potential temperature, Cloud cover (low, medium, high, tot.), U and V wind component, U, V and vertical velocity, Stationary boundary layer (SBL) Gust, Total precipitation, Large scale rain, Convective rain, Large scale snow, Convective snow, Relative Humidity, Specific humidity, Geopotential, Mean sea level pressure, Orography, freezing level / 0°C isotherm, Convective available potential energy (CAPE)

UV Index derived from spectral measurements with a Brewer UV spectrophotometer. This product is not publically available yet.