HadUK-Grid Methods
How the HadUK-Grid climate datasets have been produced.
The gridded data sets are based on the archive of UK weather observations held at the Met Office. In addition, the Rainfall Rescue project added monthly rainfall data for the pre-1961 period.
Observations from the UK’s land-based surface observations network are used to inform a process which includes regression and interpolation to generate values on a 1km x 1km grid, taking into account factors such as latitude and longitude, altitude and terrain shape, coastal influence, and urban land use.
Gridded data of this type alleviates the impact of the routine opening and closing of individual weather stations.
A full description of the HadUK-Grid dataset can be found in the science paper HadUK-Grid – A new dataset of gridded climate observations. The methods used to generate the monthly and annual grids are described in more detail in a paper published in the International Journal of Climatology, vol. 25 (2005), pages 1,041-1,054, which can be viewed here: Generation of monthly gridded data sets for a range of climatic variables over the UK . The methods used to generate the daily grids are described in more detail in the report 'The generation of the daily gridded data sets of temperature and rainfall for the UK' (PDF, 2.27 MB).
Grids are also provided for a number of 30-year climatological averaging periods (1961-1990, 1981-2010 and 1991-2020). Grids for the 20-year 1981-2000 baseline climate average have been provided for use alongside other products from the UKCP18 project . These averages are simply calculated by averaging or summing the monthly or annual gridded data sets for each variable.
To help users combine the baseline data sets with the UK Climate Projections, values have also been generated for the 5, 12, 25, and 60 km projection resolutions and for countries, administrative regions and river basins.
Each coarser resolution grid box value is an average of the 1 × 1 km grid cell values that fall within it. Averages have been calculated for each month, season and the year as a whole. Each regional value is also an average of the 1 × 1 km grid cell values that fall within the specified geographical region.