Temperature is continuously measured at weather stations throughout the country and at sea. For climate research, the KNMI uses measurements taken over decades. These measurements must be comparable with each other. Therefore, the KNMI is implementing improved homogenized temperature series. The new series provide a more reliable picture of how the climate is changing and how certain the results are.
The main findings have not changed. The climate in the Netherlands is getting warmer. Since the beginning of the twentieth century, the average temperature has risen by more than 2 degrees. As a result, it freezes less often and very hot summer days occur more frequently.
Ten years ago, the KNMI released a first version of homogenized series. Since then, continuous research has been conducted into possible further improvements. There was also a need for insight into the uncertainty of homogenization. This has led to new insights, also from outside the KNMI. This was the reason for this publication of a scientific report with a new version of the homogenized series.
What is homogenization?
Homogenizing measurement series means adjusting measurement data to make them better comparable over a longer period. Over decades, measurements may have been influenced because a weather station was moved or a new type of thermometer was used. This makes it harder to see how the weather and climate change, because other factors influence the figures.
This does not mean that the old measurements are wrong; they are just harder to compare with current measurements. Therefore, besides the original measurements, an adjusted, “homogenized” series of measurements is also made. Which series can be used depends on the purpose of the use. Therefore, both types of series remain available to everyone.
Homogenizing measurement series is complicated; therefore, the World Meteorological Organization (WMO) has established guidelines. The KNMI is basically cautious with homogenization. Adjustments are only made for known relocations or changes in instruments and after extensive research.
How does the KNMI make measurement data comparable?
In the Netherlands, temperature has been measured for over a hundred years near five locations: Den Helder, Groningen, Vlissingen, Maastricht, and De Bilt. These measurement series are very important for climate research because they cover a long period. Therefore, the KNMI chooses to homogenize only the measurement series of these locations.
After the publication of the first version of the homogenized series, further research was conducted. Based on the new insights obtained and thanks to feedback and comments from outside the KNMI, the methods were improved. Therefore, there is now a new version of the homogenized series.
To homogenize, measurements taken simultaneously in the old and new situation are usually used. This was possible at four of the five locations. In De Bilt, there were insufficient suitable simultaneous measurements available. Therefore, simultaneous measurements from two other locations were used there.
At four locations, the new homogenization also used measurements of wind, cloudiness, humidity, and (for coastal locations) sea water temperature. This was not the case in the first homogenization. The adjustments are thus better aligned with different weather conditions.
Adjustments were only made to older measurements, mainly in the first half of the twentieth century. This differs per location but ranges approximately from 1950 to 1970. All measurements after that remained the same and are therefore equal to the raw measurements.
Climate trends largely the same
The differences between the old and new homogenized series are generally small when looking at the main climate figures. Since 1901, the average annual temperature has risen by about 2 degrees. This applies to all locations and is the same in both versions.
Especially on very cold winter days and hot summer days, there are differences. This also affects the number of counted heat waves and cold waves. These counts are very sensitive to small changes. For example, if on only one day the maximum temperature is adjusted from 29.9 degrees to 30.0 degrees or vice versa, it can make a difference whether a heat wave is counted.
The KNMI often uses data from De Bilt in reports. In the raw measurements of this location, a total of 46 heat waves have been counted since 1901. In the first homogenization in 2016, there were 32. In the new homogenization in 2026, there are 39. In both cases, the conclusion is the same: there are many more heat waves nowadays. More than 40 percent of all heat waves in De Bilt since 1901 occurred in this century.
Data and source code public
The KNMI has published a scientific report about the new homogenization. The report was reviewed by experts inside and outside the KNMI. They also provided comments. These comments have been released, as well as the response from the KNMI authors. Finally, there is an overview of all adjustments made afterward in the report.
On the KNMI DataPlatform, raw measurements are available for all locations, along with both versions (version 1 from 2016 and version 2 from 2026) of the homogenized series. Processing in all systems may take a few more days; they will be accessible by early February at the latest. On KNMI Open Source GitLab, the source code is available that produced version 2 of the homogenization.