Atom feeds use the Atom Syndication Format – an XML language used to share web content, for example for news pages that share news. These feeds can also be used to subscribe to the latest news or to a certain type of news (e.g. sport, finance, etc.). A feed reader informs you when new articles are available. Feed readers are implemented in several web browsers (including Internet Explorer and Firefox). If you search for data via SMHI’s Atom feed and receive an XML file that is difficult to read, this could be because your web browser does not have a feed reader. This can be solved by finding and installing a feed reader plugin.
SMHI uses feeds to logically split up data and make it readable and searchable for both humans and machines. Different hierarchical structures are used for the datasets.
Feeds for oceanographic observations
These feeds are divided by parameter. The main feed page is a list of all the oceanographic parameters made available by SMHI. When you have selected a parameter you then choose from a list of stations and vessels (ferryboxes) that measure your chosen parameter. Finally you select the time interval in which you are interested. Depending on data availability there can be up to three files; one with archive data from when measurements first started at the station up until midnight the previous day, one file with data for the last hour, and one file with data from the last 24 hours. Hourly and daily files are generated each hour wherever data is available.
The results are presented as CSV (Comma-Separated Value) files, which can be read with any text editor or a spreadsheet program such as Microsoft Excel. The files often contain a large amount of data and since Excel is limited to about 1 million rows there is a risk of the data volume being truncated.
Feeds for oceanographic and meteorological model and analysis data
Oceanographic and meteorological model data (analysis stage) as well as meteorological analysis data are divided into time steps. The main feed page is a list of all the years for which data is available. Years are broken down into months and then days, and, depending on the model, data is available either every hour or every sixth hour.
The output file is in GRIB (GRIdded Binary) format and includes data for several parameters and for a specific geographic area. The geographic area is described in the GRIB file.
The GRIB format is not that simple to use. You need an understanding of the format, a special API to read the data, a code table for the parameter name mapping and codes for coordinate conversion.