Data Machine learning

Understanding and mitigating the impacts of greenhouse gas (GHG) emissions and the drivers of nature loss are critical priorities for governments, corporations, and financial institutions. Intensive agricultural practices playa key role contributing to GHG emissions and nature-related risks. In England, approximately 105,200 farm holdings cover 8.7 million hectares of utilised agricultural land. Estimating GHG emissions at the farm level presents challenges due to the lack of open-source farm-level data. To address this, we employed natural language processing (NLP) techniques for entity matching and unsupervised learning, mapping farm names to polygons with missing owner or farm name information. Using this approach, we have produced an open-source farm-level dataset that contains information on 117,116 farming entities, including addresses, land areas, crops, production output, and coordinates. This resource can potentially support financial institutions and corporations in estimating farm-level GHG emissions, facilitating the development of targeted decarbonisation strategies for England’s agricultural sector.

Dataset 

RPA_fields_with_owner_name.csv 

Columns: 

fid: Unique identifier. 

OWNER/OPERATOR: Corresponding owner of an agricultural holding according to LR or operator if also under Countryside Stewardship. 

Address: Address of the farm. 

District: Companies House registration number of the entity. 

Region: Region of the farm. 

Company Registration No. (1): Companies House registration number of the entity. 

Proprietor (1) Address (2): Address of the registered owner from LR records. 

Proprietor Name (2): A second name appearing on the LR records. 

Company Registration No. (2):  Companies House registration number of the entity. 

Proprietor (2) Address (1): A second Address appearing on the LR records. 

SBI: Single Business Identifier. 

Area(ha): Farmland area associated with an owner-operator entity or farm, derived from LR or Countryside Stewardship sources, or assigned using a Voronoi model. 

      2. Farm_SBI_mapping_with_Farm_Area_overlap.xlsx 

Columns: 

SBI: Single Business Identifier. 

Farm ID: Unique identifier assigned to farm names extracted from online directory sources. 

Farm Name: Name ascribed to the farm through Voronoi mapping (LR non-ownership fields only). 

Farm Area: Area of voronoi-designated farm entity (hectares) 

SBI Area:  Entity area according to SBI allocation 

SBI-Farm overlap (%): Percentage of SBI Area overlap with Farm Area 

Aggregated overlap (ha): Area overlap of all the farms intersecting within SBI polygons 

Farm-SBI Area Overlap (ha): Area overlap between SBI and Farm polygons 

     3. OWNER_SBI_mapping_with_Area_overlap.xlsx 

Columns: 

SBI: Single Business Identifier. 

Owner ID: Unique identifier for this database assigned to owners or associated operators from Land Registry (LR) and Countryside Stewardship scheme sources. 

Owner Name: Corresponding owner of an agricultural holding according to LR or operator if also under Countryside Stewardship. 

Owner Area: Area of owner-designated farm entity (hectares) 

SBI Area: Entity area according to SBI allocation 

Owner-SBI overlap (%): Percentage of SBI Area overlap with Owner Polygonal Area 

Aggregated overlap: Area overlap of all the ownership polygons intersecting within SBI polygons 

Owner-SBI Area Overlap (ha): Area overlap between SBI and Owner-designated polygons 

     4. country_side_owner_data_without_geo.csv 

Columns: 

fid: Unique identifier. 

ORG_NAME Organisation Name: Entity that is participating in the countryside stewardship scheme.  

Area(ha): Farmland area associated with an owner-operator entity or farm, derived from LR or Countryside Stewardship sources, or assigned using a Voronoi model. 

Crop: Crops assigned to the polygons (under stewardship) through CROME. 

     5. RPA_with_owner_with_geometry.gpkg 

 This dataset contains polygonal geometry data representing the intersection of RPA (SBI) polygons with owner-defined polygons. The file, stored in GeoPackage (.gpkg) format, provides spatial information on areas where land ownership boundaries overlap with designated RPA polygons.  

     6. country_side_owner_data_with_geometry.gpkg.zip  

This dataset contains polygonal geometry data representing the intersection of Countryside Stewardship polygons with owner-defined polygons. The file, stored in GeoPackage (.gpkg) format, provides spatial information on areas where ownership polygonal boundaries overlap with designated countryside polygons. 

      7. final_RPA_with_owner.gpkg.zip 

      This dataset is the same as “RPA_with_owner_with_geometry” but without geometry.