Differentiating Between Unutilised Gaps And Gaps For Access In Hedgerow Using Remote Sensing
Year: 2025
Type: report
Abstract: Introduction:Hedgerows are present throughout the majority of the United Kingdom, predominantly delineating boundaries, borders or limits of land (Baudry et al., 2000). They provide landscape-scale connectivity as well as increased biodiversity, pollination, and are a priority wildlife habitat in the UK (Wolton, 2009; Lawton, 2010; Graham et al., 2018; Norton et al., 2012; Montgomery et al., 2020). Hedgerows provide multiple ecosystem services such as increased agricultural production, biological control of pests and have the potential for flood risk reduction (Holden et al., 2019; Montgomery et al., 2020), yet are given little attention in policy when compared to widescale tree planting initiatives. Hedgerow gaps are found nationwide, primarily providing access between parcels of land, but can often develop due to over trimming, disease, and poor management. Gaps reduce connectivity of the landscape for migrating species, as well as reduce the amount of available habitat and food sources for wildlife (Graham et al., 2018; Staley et al., 2015). Understanding the distribution of hedgerow gaps at a high-resolution may provide much-needed information, which can be used to inform species connectivity models, flood risk analysis and potential locations for hedgerow planting and expansion. Differentiating between these gaps is relatively straightforward when performing ground-based field surveys, however, surveys of this nature are costly and time consuming. Utilising remote sensing datasets, including uncrewed aerial vehicles (UAVs), aerial imagery and satellite products, may enable semi-automatic differentiation between access gaps and unutilised gaps in hedgerows at the landscape scale. It is currently estimated that there are 3,983 ±597 km of hedgerows, 2,874 ±431 km of hedgerow gaps and 316 km of tall vegetation (vegetation over 3 m) along hedgerow boundaries within the East Riding of Yorkshire (Wolstenholme et al., 2022). Despite this, automatic identification of hedgerows and hedgerow gaps, the authors did not differentiate between access and unutilised gaps. Artificial intelligence and computer vision were utilised to automatically identify hedgerows from high resolution aerial imagery across the East Riding of Yorkshire Council, herein referred to as ERYC (Wolstenholme et al., 2022). Through leveraging existing datasets such as the Ordnance Survey MasterMap Topography layer (Ordnance Survey, 2017), which represents unique topographic features in Great Britain, hedgerow gaps were extracted from the hedgerow identification dataset. Three methods are tested for semi-automated differentiation of hedgerow gaps. The first (section 2) uses the existing hedgerow dataset produced by Wolstenholme et al. (2022) and assumes that most access gaps are located near the corners of fields, which landowners are most likely to use for access. The second (section 3) attempts to train a deep learning model, similar to how hedgerows were automatically identified by Wolstenholme et al. (2022), to identify and differentiate between the gaps. Finally, section 4 leverages multispectral products captured at both the field scale using a UAV and aerial/satellite datasets to assess if vegetation indices can differentiate between the gaps.
Authors: Wolstenholme, Joshua, Nolan, Joseph, Townsend, Elizabeth, Ahmed, Joshua, Parsons, Katie, Davidson, Giles, Thomas, Robert, Parsons, Daniel
Institution: University of Hull
Language: English
Subfield: Ecology
Field: Environmental Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7027718926
Cites:
Cited by:
Related to: