Hourly, validated and automated in-situ ground-based meteorological observations in the Netherlands

These datasets are meant to provide meteorological observations in the Netherlands. Distinction is made between quickly available hourly and daily datasets which have only passed operational quality control, and datasets which have passed through manual validation (see figure under "Data aggregation and validation flow", below).

Using these datasets KNMI can calculate statistics to observe the climate in the Netherlands. Their use is not limited to climate science, and they are also intended for applications outside of meteorology or climatology.

Because of station relocations and changes in observation techniques, the data in these datasets are inhomogeneous. As a result, these series are not suitable for trend analysis. For climate change studies we refer to the homogenized daily temperature data of the five principal stations in the Netherlands or the updated monthly timeseries of the Central Netherlands Temperature. Trend analyses are presented in KNMI’23 Climate Scenario’s.

Timestamps in the file and filename are stored in UTC. For the Netherlands, local time is UTC+1 (CET) and during daylight saving times local time is UTC+2 (CEST).

The timestamp in the files indicates the end of the measuring interval preceding that timestamp. The filenames use an abbreviated timestamp indicating the total interval. See the table below for examples.

Individual variables may be measured and aggregated over a different time interval. This can be shorter or longer than the interval of the file. This is denoted in the description of the variable and the measurementType.duration field of the EDR metadata.

The unvalidated hourly and daily datasets are updated on a nightly basis with data from the preceding day. Validated hourly and daily datasets are updated during working days and can be delayed up to 10 working days. Validated data can be modified at any moment in the future if further issues are found with the data.

Monthly and yearly datasets are updated after all days in their respective period have been validated and can receive further updates up to months later.

It is highly recommended to use the MQTT Notification Service to be notified when the dataset receives new or updated files, to always get the latest file without unnecessary polling or redownloading. This is the fastest way to get new files as soon as they are available.

The datasets are available as NetCDF files. The files can be downloaded via the Open Data API. The datasets are also available as an Environmental Data Retrieval (EDR) collection. The EDR API makes it possible to query the dataset on specific variables in a spatio-temporal manner, without requiring the user to download full datasets.

The stations in these datasets are part of the WMO Integrated Global Observing System (WIGOS). The datasets contain the location, name, and the WIGOS identifier of each station.

A historical record of some of the station metadata is kept in the weather stations dataset on the Data Platform. More metadata and information about the stations can also be found in the WMO OSCAR database.

Some automatic weather stations consist of multiple sublocations where sensors are installed. This is especially the case at aerodromes with multiple sensors for aviation purposes. This dataset represents measurements from those sensors that are installed with minimal local effects, to ensure representative synoptic observations at each station.

Not all stations are present during the entire runtime of the dataset. Stations have been added and decommissioned since the start of this dataset. The start and end dates can also be found in the station metadata.

Observations from sensors located at Automatic Weather Stations are processed into time series of 10-minute intervals. From these 10-minute intervals the data is subsequently aggregated to hourly, daily, monthly and annual intervals. Aggregated values can only be calculated if data availability of the underlying datasets allow for it.

The methodology for aggregating depends on the variable and can vary. For example: Daily mean pressure is calculated as a mean of the 24 pressure values of that day (midnight to midnight UTC). The monthly maximum wind gust is calculated as the maximum of the daily maximum wind gust values of all days in that month, which in turn are the maximum wind gust values of the 24-hourly wind gust maxima (midnight to midnight UTC). The aggregation method for any variable can be found in the description of the parameter in the NetCDF, or in the collection metadata of the EDR.

The hourly data is thoroughly validated by KNMI experts. The quality of the data is assessed based on various criteria. Data is also compared with neighbouring stations and assessed based on the expertise of the validation team. If deviations are found, the data is corrected.  

Manual validation is carried out during office hours and may also take place in non-chronological order. In some cases, validation can take up to 30 working days. The exact availability of each dataset can be found in the table under “List of datasets”.  

Reasons for correction include:

Missing data

In the case of missing data points, 100% completeness is guaranteed for temperature, solar radiation and precipitation for all land stations. Other variables are supplemented as long as quality can be guaranteed. For example, in the event of a prolonged sensor malfunction, a variable cannot always be supplemented.  

Measurement errors

Measurement errors can arise due to faulty sensors, calibration problems or temporary malfunctions in the measuring equipment. These errors are detected by internal quality controls and by comparison with nearby measuring stations. Where possible, incorrect values are corrected or replaced.

Sensor drift

Sensor drift occurs when a sensor gradually starts to measure deviating values over time, for example due to ageing or wear and tear. This type of deviation is identified through long-term analyses and periodic checks, after which corrections are applied or the sensor is replaced.

Undesirable environmental influences

Environmental factors may temporarily differ, resulting in unreliable measurements. This is prevented as much as possible by the placement of the stations, but cannot always be avoided. This may concern maintenance, but animals or vehicles (e.g. helicopters at a drilling platform) can also influence measurements.