Below you'll find answers to the questions, we get asked the most about. We will continue updating these based on questions received.
Questions about the user account
In order to retrieve data from our services you need to subscribe to the service(s) in which you are interested. The api-key is linked to the data type (e.g. metObs) and the specific version of the API.
The reason for this is that DMI wants to be able to contact the users of the different services in order to inform users of planned service windows, breakdowns or new API versions.
In order to receive an API key you have to register as a user of DMI's API service. On the page User Creation you can read instructions on how to register as well as how to find your API key if you've already registered as a user.
If you forgot your password go to the page User Creation where you can find instructions on how to make a new password.
Your API key is valid as long as it is in use. If an API key has not been used for the past 12 months it will be deleted automatically. You can always contact DMI by using the contact form at https://www.dmi.dk/kontakt/frie-data/ to have your user information deleted by which your API key will also be deleted.
You can always contact DMI by using the contact form at https://www.dmi.dk/kontakt/frie-data/ and ask to be deleted as a user resulting in the deactivation of your API key and the deletion of the data registered about you. Out of consideration for traceability in relation to information security, deletion will not cover data registered in the logging of requests to DMI’s web services.
Questions about the webservices
You can access the API specification as well as Request & Response examples and information on data sources by choosing the appropriate data type in the menu on the left side of the screen.
You are able to access the current API documentation by choosing the appropriate data type in the menu on the left side of the screen and then choosing “Service documentation”.
DMI’s API is REST based and thus platform- and language agnostic.
It is not possible to include multiple parameters in a single query. If you want to query more than one parameter, you need to query each parameter separately or query all parameters and then filter out the parameters you don’t want afterwards.
If you want to query miltiple stations, you can use a bbox query to do so, if the stations of interest are within the same geographical area. Another option is to query each station separately or you can query all stations and filter out the stations you don’t want afterwards.
By registering as a user of the DMI API you will automatically receive a notification at the registered email address when a new version of the API is released.
There is a fair use limit to allow only 500 requests per 5 seconds. If more requests are received a HTTP 429 error is returned. We are continuously monitoring the use of the APIs and will adjust this limit to ensure that all users have fair access. We also recommend to use the bulk download functionality.
The API services are expected to have an availability around 99.9 %.
Downtime is not expected, but in case of planned downtime it will be announced under API status & contact.
You can report errors by using the contact form at https://www.dmi.dk/kontakt/frie-data/
If you have questions regarding data from third parties, please contact the authority responsible for the station or the data derived therefrom.
If you are experiencing that the API is out of service please check the API status page or contact The Agency for Governmental IT Services (SIT - Statens It).
Call SIT: +45 72 31 00 01
Questions about the data
Data retrieved from DMI may not be changed. DMI reserves the right to block or adjust the bandwidth for individual users if these strain DMI’s web services to a disproportionately wide extent.
You can see the update frequency by choosing the appropriate data type in the Confluence menu on the left side of the screen and then choosing “Data information” > “Parameters”.
Please note that lightning data does not have a specific update frequency, since the data is updated in real time.
If you want to download large quantities of historical data, we recommend that you use the bulk download service. The service lets you download .zip files, each containing historic data for a month. You are also able to download all historic data by selecting the file all.zip.
You can access the bulk download service by choosing the appropriate data type in the Confluence menu on the left side of the screen and then choosing “Service Documentation” > “Bulk Download Service”
The following bulk download services are available:
Data retrieved through the API is only available in JSON format, but DMI's open data can easiliy be imported directly into GIS. Please see our guide.
Data from our OGC API-features services (metObs, oceanObs, lightningdata and climateData) can be imported into excel. Please see our guide.
Observation data have not been subject to quality control nor processed in any way. Errors in measurements may occur. Errors are typically due to malfunction of instruments caused by wear and tear or exposure to weather and in some cases seawater and on rare occasions from vandalism. Wear and tear of the instruments are handled proactively by performing service checks at the stations on a regular basis and changing the instruments within the given time frame. Due to this, you have to be aware of the risk of faulty measurements when using DMI’s observation data.
DMI's forecast data will be made available during 2022
It is possible to download data from third parties through the DMI API if an agreement exists between the third party and DMI. The owner of the station is marked in the station list of the service.
If you have questions regarding data from third parties, please contact the authority responsible for the specific station or the data derived therefrom.
The data retrieved through DMI's Open Data API correlates to the data available to DMI. Sometimes you may experience that some data entries are missing. The root cause of this is typically a malfunctioning measurement instrument, which DMI will repair as soon as possible.
The coordinate datum used for the location of the stations is WGS84.
The 1x1 km grid is used as an intermediate basis to construct the spatial resolutions 10x10 km, 20x20 km, municipality data and country data. Under certain weather conditions the 1x1 km grid data can be quite imprecise, but by aggregating it to larger areas the uncertainty is reduced. Furthermore, 3rd party data, which DMI doesn’t have permission to redistribute, can be deducted directly from the 1x1 km grid data.
If you have questions regarding data, please contact the authority responsible for the specific station or the data derived therefrom.
You can see the owner of a station by selecting the appropriate data type in the menu on the left side of the screen and then choosing “data information” > “stations”
In the radar forecast on www.dmi.dk we use Gunnar Farnebäck's dense optical flow algorithm as implemented in OpenCV: https://docs.opencv.org/3.4/de/d9e/classcv_1_1FarnebackOpticalFlow.html
We also use a range of other methods (e.g. Lucas-Kanade optical flow) available in pySTEPS: https://pysteps.readthedocs.io
Radar data on www.dmi.dk is a special product that only shows the lowest elevation scan which is gathered every 5 minutes. This data is currently not a part of DMI’s open data service.
Depending on which tool or application used to visualize and process the GRIB file, an error in the geographical placement might occur due to a maximal decimal precision of three in GRIB files. Because of the high resolution of the model a precision of 4 or 5 decimals is needed in certain geographical areas.
In order to compensate you have to use the eastern and western reach together with the number of grids all of which are written in the header of the GRIB file, to calculate the grid distance in degrees as shown in the following equation:
A precision of at least five decimals is recommended when using the above equation.
Now use the calculated grid distance along with the western reach and the number of grids to calculate the accurate eastern reach by using the following equation:
- No labels