CROME Crop Maps – Terra Sulis Woodlands

The Rural Payments Agency (RPA) retrospectively publishes crop maps of England each year. Four years of data are now available: 2016, 2017, 2018 and 2019. These maps are derived from Sentinel-1 (radar) and Sentinel-2 (optical) satellite imagery and use a standardised methodology and land cover classification and are comparable one year to the next. Unusually they are presented as a tessellation of vector hexagons of which there are some 32 million hexagons covering England. Each hexagon is between 60 – 80m is size. In this novel format the data are split into Ceremonial Counties which are individually manageable, but for all of England the full data set become almost unusable. This is compounded by its temporal nature with a new full data set being added each year. For this reason the data have been converted to raster format based on the 50 m grid defined by the Ordnance Survey Terrain 50 data set.

The CROME maps use a standard set of land cover classes, although the number of classes has expanded somewhat since 2016. The classes fall into the broad categories Cerial Crops, Leguminous Crops, Energy Crop, Grassland, Non-Agricultural Land, Water, Trees and Unknown or Mixed Vegetation. The individual arable crops, of which there are 70, are not of direct interest to the project. The classes of interest are as follows:

Broad Category	Specific Class

Arable Crops	70 cereal and leguminous crops
Grassland	FA01 Fallow Land HE02 Heathland and Braken PG01 Grass
Non-Agricultural Land	NA01 Non or sparsely vegetated land
Water	WA01 Water
Trees	TC01 Perennial Crops and Isolated Trees NU01 Nursery Crops WO12 Trees and Scrub, short Woody plants, hedgerows
Unknown or Mixed Vegetation	AC00 Unknown or mixed vegetation

For the years 2016 and 2017 the data from RPA are organised as four regional datasets, whereas for the years 2018 and 2019 it is organised into 40 counties. Each dataset has 32,000,000 hexagons, presenting an onerous data management task. To make the data processing manageable the vector hexagons were rasterised into grids based on the 50 m grid cells of the Ordnance Survey Terrain 50 data set. The coordinate space for this is given in the table below. The raster versions are more manageable and amenable to data processing in a raster GIS system.