Host map of Ash trees in Great Britain

This is an Ash (Fraxinus excelsior) host map for Great Britain at the 1 km x 1 km scale. The units are in terms of ha or ash per 1km x 1km grid square. There are 100 ha in 1km x 1 km, so this can be thought of as a percent.
It was created generating a 20m x 20m land use layer based on the UKCEH Land Cover Map for 2019, deriving areas of broadleaved tree cover for each cell of the land use layer and estimating the percentage of ash within broadleaf, which were assigned dependent on land use and region (see more details about the methods in data preparation and processing activities). The percentage of ash and area of broadleaf trees were combined to give an area of ash for each land use in cell and this was aggregated (summed) across all land use types to give a 1km by 1km ash area map.

The distribution map were produced in four steps:

1. Create a 20 m x 20 m land use layer

Land use categories were based on the UKCEH Land Cover Map for 2019 at 20m x 20m resolution (Morton et al., 2020, available at https://doi.org/10.5285/643EB5A9-9707-4FBB-AE76-E8E53271D1A0). Individual cells were recategorized if their cell centres intersected the Forestry Commission sub-compartment databases for England or Scotland, and NRW database for Wales (available from Forestry Commission geodata, http://www.forestry.gov.uk/pdf/nigreatbritain.pdf/$FILE/nigreatbrita...). Based on species composition, cells were assigned to FC ownership as: “FC ash”; “FC mixed broadleaf”; and “FC other species”. Unallocated cells that intersected the NFI woodland cover map (NFI, 2020) were assigned as: “NFI woodland” (> 0.5 ha). The remaining cells were assigned based on the CEH landcover categories: woodland was combined into “Small woods” (< 0.5ha); Urban and suburban squares were grouped together as “Urban”; and all other cells were assigned as “Openland”.

2. Derive areas of broadleaved tree cover for each cell of the land use layer

Tree cover density and dominant leaf type layers (Herrmann et al., 2017, available at: https://land.copernicus.eu/pan-european/high-resolution-layers/fores...) covering Great Britain were projected onto the OSGB grid. Tree cover density was filtered to include only broadleaf dominant 20m x 20m cells and then resampled to give average values for each cell that aligned to the land cover map. Percentage cover scores were combined with hectares of woodland to give hectares of broadleaf cover for each land use in each cell. Finally, hectares of ash were calculated by multiplying the area of broadleaf by the average proportion of ash for the respective land use and region.

3. percentage of ash within broadleaf was assigned dependent on land use and region

Additional broadleaf cover was included using an estimate of small woody features (SWF) (Langanke, 2019, available at https://land.copernicus.eu/pan-european/high-resolution-layers/small...). This was projected and resampled to a grid matching the land use layer. To avoid double counting all squares with tree cover density values greater than zero were removed from the SWF map. SWF were divided into urban and non-urban land uses before they were combined with the values in Table 1 (see Table_1_Summary_proportion_ash_by_land_use_and_region_to_generate_host_map.csv as part of Data File). An additional step was used to fill squares with missing SWF data in the source data layer. This was achieved using universal kriging with elevation for urban and non-urban cells independently. All analysis was conducted in R 4.2.0 using sf (Pebesma, 2018), terra (Hijmans, 2022), gstat (Pebesma, 2004).

4. Aggregation to 1km x 1km scale

The percentage of ash and area of broadleaf trees were combined to give an area of ash for each land use in cell and this was aggregated (summed) across all land use types to give a 1km by 1km ash area map.