Research and analysis

Methodology Note: Mapping the government estate

Published 10 June 2026

1. Purpose

This summary note outlines the methodology used by Defra to identify the spatial extent of the Government land holdings. This work was undertaken to support the Government Estate Nature Plan, but is designed to be useful for further applications.

2. Policy context

To produce the Government Estate Nature Plan, it is essential to know where the government estate is. Existing datasets (such as those used in the State of the Estate report) defining the extent of government land holdings are insufficient when applied for environmental policy analysis purposes. This is because the State of the Estate does not include information on the shape of individual land parcels in the government estate. HM Land Registry records spatial data for UK land titles, and so was identified as a potential solution to produce the required map of the government estate. The use of HM Land Registry enables data extraction for all government departments using a single method and from a single data source which we can reasonably expect to be comprehensive.

3. Findings

Our analysis indicates that the total land area of the government estate is 562,606 hectares, based on the principal landholding departments in government.

It should be noted that the State of the Estate: Annual Data Publication 2024-25 (UK Government, 2025 indicates the total area of the government estate in England to be 546,700 hectares. The difference between these figures is due to methodological differences in how the spatial data has been sourced and compiled. The State of the Estate figure remains the official estimate for statistical purposes.

4. Methodology

4.1 Extracting data from HM Land Registry

HM Land Registry records spatial data for UK land titles. For this work we created a data model which identified titles in HM Land Registry relating to government departments and their associated agencies and bodies (hereafter referred to collectively as ‘organisations’).

Of the 24 ministerial departments, 13 were included in our analysis. These were selected based on the highest expected land areas. The remaining departments were not known to have any significant land holdings. These 13 departments were:

• Department for Culture, Media and Sport
• Department for Education
• Department for Energy Security and Net Zero
• Department for Environment, Food and Rural Affairs
• Department for Transport
• Department for Work and Pensions
• Department of Health and Social Care
• Foreign, Commonwealth and Development Office
• HM Treasury
• Home Office
• Ministry of Defence
• Ministry of Housing, Communities and Local Government
• Ministry of Justice

The associated organisations for each department were identified using the government’s organisations webpage. Additional research was carried out to identify historic organisation names using the most authoritative information available; the availability of this information varied by organisation.

The Levenshtein’s distance technique was used to identify titles where there have been spelling errors during the entry of organisation names into the HM Land Registry data.

The identified list of titles was then used to extract the relevant land parcels for each organisation from the National Polygon Dataset. This resulted in a spatial dataset which included extent information for all titles in the government estate. Overlapping and adjoining land parcels belonging to the same organisation were then combined to create a single non-overlapping map of each organisation’s land holdings.

4.2 Ministry of Defence estate

Due to the sensitive nature of Ministry of Defence land, HM Land Registry data was not used to identify the department’s estate. Instead, a dataset was supplied by the Ministry of Defence of their land holdings for use in the Government Estate Nature Plan.

4.3 Validation

Steps were taken to validate the government estate dataset to a high standard. Comparisons to the figures from the State of the Estate dataset found largely similar results across departments. Differences were within the expected realms and can be explained by the differing methodologies used. Where possible departments and large land holding organisations were contacted to validate the dataset produced with respect to their estate. Where appropriate, changes were made to the methodology to reflect feedback from the validation. Changes were not made where feedback related to changes reflected in a more recent version of the HM Land Registry data than the August 2025 dataset used for this analysis, or where feedback reflected inaccuracies in HM Land Registry data which should be corrected by the department or organisation with HM Land Registry.

The validation process is ongoing, and the dataset may be subject to further revision in response to feedback.

4.4 Processing the government estate dataset for use in the Government Estate Nature Plan

The government estate dataset was clipped to an England boundary using ‘Countries (December 2024) Boundaries UK BFC’. This removed any sections of the estate below the mean high water mark or outside of England. The boundary layer has a simplified geometry resulting in estates near or overlapping the mean high water mark being clipped.

The government estate dataset was then dissolved by department into a single estate polygon using the ArcGIS Pro Dissolve tool. This was necessary to remove overlapping land parcels that are listed under different owners.

The analysis to calculate overlap area between government estate polygons and environmental datasets was performed in ArcGIS Pro 3.4.0.

5. Data assumptions

The HM Land Registry states the following data quality issues regarding the National Polygon Dataset:

Overlaps, gaps, slivers and underlaps all exist in the dataset. Many are genuine and are acceptable but there are some that exist which should not.

Some inconsistencies in the dataset include:

small gaps and overlaps

technical overlapping polygons - These are polygons where small overlaps exist due to mapping inconsistencies. They do not describe a real conflict of registered titles.

actual overlaps – These are where two or more registered titles include the same land or property and will be resolved upon application.

In addition to the above, we have found topological errors in the data: for example, polygons with only 2 vertices.

The above issues from the input dataset persist in our output dataset of government land holdings; although where possible we have worked to mitigate them.

6. Limitations and caveats

There are some limitations and caveats to be taken into consideration when using the government estate data:

• There are known overlaps in the data where land ownership is listed with HM Land Registry under multiple departments. For the calculated area figure these overlaps have been removed.
• There are known overlaps between government and non-government land holdings. Some, but not all of these are known to be legitimate overlaps where multiple proprietors exist. These overlaps were not resolved as part of this work. Our intention is to make departments and their associated organisations aware of these overlaps; to enable them to be resolved on a case-by-case basis where required.
• Not all of the government estate is registered with HM Land Registry; HM Land Registry is working with government departments towards complete registration of public sector land.
• Although designed to handle spelling errors, the methodology used to identify HM Land Registry titles may still result in errors of omission or commission. For instance, a title may be rejected as belonging to a department because the spelling was too different or the historical organisational name was not captured. Or a title may be incorrectly identified as belonging to a department if an unrelated company name is very similar to organisation name.
• HM Land registry records land held under freehold or on a long leasehold over 7 years. The produced government estate dataset therefore also only considers holdings of these types. Depending on the start date of the leaseholds, some leases may end in less than 7 years’ time. Information on lease length is not included in the HM land registry data and therefore this has not been accounted for.
• This analysis has not considered restrictions placed on titles; some titles may have restrictions which commit them to an immutable purpose or prevent their use for certain purposes.
• This analysis has not considered where government estate land is leased to third parties.
• This analysis has not considered land parcels which are owned by a third party but in practice managed by a government department or organisation. This analysis has not considered land parcels which are owned by a third party but to which a government department or organisation has access rights.
• Land ownership is accurate to August 2025. Because some of the government’s land is held on a temporary basis figures may change year on year, for example, Homes England hold land temporarily prior to freehold transfer to homeowners upon construction completion. Land acquisitions and disposals since August 2025 may impact accuracy of the data.
• Clipping the estates to the Countries (December 2024) Boundaries UK BFC artificially reduced the footprint of parcels along the mean water mark or English border. This is due to the countries layer polygon having a simplified geometry.

7. Note published

10/06/2026