PhD Studentship: Machine Learning for Mapping Habitat Histories with the National Trust

Project Title: Machine Learning for Mapping Habitat Histories with the National Trust

Background

Hedgerow loss has been particularly marked since 1945. Even between 1984-1993 it is believed as much as 160,000km of hedgerow was lost. Hedgerows play a critical role in supporting a range of plant, animal and insect species, including pollinators.

They provide vital habitat connectivity and can help to prevent soil erosion. And they can also make a considerable contribution to carbon sequestration and storage through their biomass and the soils beneath them. As a result, the creation of hedgerows is recognised as an important component of plans to address the twin crises of climate change and the decline in nature, alongside sustainable food production.

Hedgerows are also heritage assets in their own right, telling stories of land ownership, division and management, social and political change, and linking to intangible heritage of craft skills, folklore and traditional ecological knowledge. They play an important role in defining the character of the landscape, and historically there were many different regionally distinctive hedging traditions and styles. As a result it is typically more beneficial to replace a lost hedgerow – both for landscape character and for the avoidance of archaeological remains.

Historic Ordnance Survey (OS) mapping is typically used as a source to identify former hedgerows. These may be identified as extant hedgerows, or as remnant fragmented hedgerows, isolated boundary trees, or linear arrangements of in-field trees. Linear earthwork banks (upon which hedgerows were typically planted) which may be apparent in remote sensing data such as lidar, can also be used to locate former hedgerows. While historic OS mapping and lidar data can play a critical role, therefore, in helping us to understand hedgerow loss – and thereby supporting opportunities for hedgerow re-establishment – it has not been possible to analyse these vast pools of imagery at a national scale – though Historic Landscape Characterisation (HLC) and similar exercises at county level have, in some cases, recorded historic boundary loss or even distinctive hedgerow types.

The Countryside Survey of 1978 provides the first firm data on hedgerow extent and distribution for the UK. This provides our current ‘baseline’ – but it is likely that this represents just a fraction of hedgerows present at the beginning of the 20th century. Just how hedgey was the British countryside? How hedgey could it be again?

Advert information