Modelling our landscape from space using fine resolution data to achieve a more sustainable future
MAX RAYNER *1, HEIKO BALZTER 1,2 AND LAURENCE JONES 3
1 Centre for Landscape and Climate Research (School of Geography, Geology and the Environment), University of Leicester, Leicester, LE1 7RH, UK; mjr59@leicester.ac.uk
2 National Centre for Earth Observation, University of Leicester, University Road, Leicester, LE1 7RH, UK
3 UK Centre for Ecology and Hydrology, Environment Centre Wales, Deiniol Road, Bangor, Gwynedd LL57 2UW, UK
Earth’s landscapes are in a critical state and governments, conservation organisations and landowners face vital decisions regarding our future relationship with the environment. Many countries are pursuing a “natural capital approach” to environmental management where landowners are financially rewarded for maintaining healthy environments. A key part of the natural capital approach is modelling the processes across our landscapes, such as flood protection and habitat quality, that provide benefits to society and the environment. These models are often based on land cover maps produced from satellite imagery.
Here we present a new, highly detailed land cover map (S2LC_2018) produced from images captured by the Sentinel-2 satellite. We compare the composition of the landscape presented in S2LC_2018 with two less detailed land cover maps. Focusing on two important habitats within the UK, we use spatial statistics to compare the abundance and distribution of woodland and grassland across the landscape.
Woodland cover more than doubled from 4.1% to 9.1% between the most and least detailed map. This came at the expense of ‘cultivated land’, the dominant land cover of the region, which decreased by 21%. The mean size of woodland and grassland areas was smaller, and they were more spread out across the landscape in the more detailed map.
The level of detail in land cover maps can affect models of the environment by altering the abundance of different habitats and their distribution across the landscape. We conclude that new fine-resolution maps are the best option to accurately model natural capital.