Observation

Overview

The Met Office MIDAS Open archive provides quality-controlled UK land surface station observations at daily and hourly resolution, spanning eight datasets (daily rain, temperature and weather; hourly rain, weather, mean wind, radiation and soil temperature). Data reaches back to the 19th century for some stations.

MIDAS Open is not the (paid, forecast-oriented) Met Office Weather DataHub. It is the historical archive, published as annual releases: each release covers data up to the end of the previous complete year, so the most recent ~6–12 months are not yet available. The provider always reads the latest release.

Access requires a free CEDA account. The archive is browsable anonymously, but downloading files needs a bearer token, which the provider mints automatically from your CEDA credentials. Register at services.ceda.ac.uk and configure the credentials via WD_AUTH__CEDA=<username>:<password> (environment variable) or Settings(auth={"ceda": "<username>:<password>"}) (Python).

Station metadata comes from each dataset’s station-metadata.csv catalogue. Observations are one BADC-CSV file per station and year. A station may transmit several report types for the same period (for example an overnight and a daytime 12-hour reading alongside a 24-hour one); these are collapsed to a single value per calendar day (maximum for max-type parameters such as temperature_air_max_2m, minimum for min-type ones). For the daily temperature extremes this recovers the true daily extreme regardless of which report types a station transmits; the other daily quantities (sunshine, snow depth, precipitation) are reported once per day, so the aggregation is effectively a de-duplication there. Hourly datasets carry one reading per hour.

Not every station measures every parameter — a station/parameter combination MIDAS doesn’t have simply contributes no rows.

The quality column carries MIDAS’s raw five-digit MESQL quality-control flag verbatim; each digit encodes a different aspect of the observation’s quality (see the MIDAS QC flag documentation), so a larger number is not “worse”.

License

MIDAS Open is © Crown Copyright, Met Office, published under the UK Open Government Licence v3.0. See the dataset catalogue record for further information and usage conditions.