Jamie Bantock

Tree and forest climate interactions are fundamental to a sustainable future and societal wellbeing. Trees are at the heart of the current political discourse, and the UK government is preparing to launch a strategy to accelerate tree planting and improve the management of existing trees and woodlands. The world is becoming increasingly urbanised, and urban trees are well recognised for their environmental, health and wellbeing benefits. Trees across the rural and urban landscape are going to play a central role as we move towards a net-zero emissions economy. The ability to measure, monitor and map the health and status of the UK’s trees is therefore essential to the UK’s future treescape and urban green infrastructure. Currently, tree identification can be achieved with high spatial resolution panchromatic imagery, but we need to go beyond this. We need to be able to not only map species within mixed assemblages, but also characterise the health and size of the trees. Furthermore, we need to be able to measure the particular configurations of urban environments and the small-scale but widespread plantings that are likely to feature prominently in planting programmes. Importantly, we need to be able to monitor changes over time – to quantify carbon sequestration; assess vulnerability and detect the onset of climatic stress or disease outbreak to facilitate early intervention; and to measure the success and monitor compliance of tree planting programmes. This requires going to the spatial scale of tree crowns but capturing the spectral information that will provide the information for classification and characterisation. Advances in both sensors and process understanding is closing the gap between canopy reflectance properties and its functional meaning, opening the possibility of detailed studies at the scale of individual trees and their responses to global change from space. Species identification and mapping has been demonstrated from airborne hyperspectral sensors, enabling species mapping across forests and urban areas. In this project we will push the limits of leading CMOS TDI sensors and optimise system configuration to develop a new platform for the classification, characterisation and monitoring of trees across urban and rural landscapes. Band selection will provide information on both plant health and will feed into classification algorithms developed from extensive ground truthing data. In this pathfinder phase of the project, our objectives are to (could include): 1. Refine band selection based on key vegetation characterisation indices and sensor constraints (from starting point of bands a,b,..f) 2. Develop classification algorithms from airborne hyperspectral data and extensive ground-truthing data sets. 3. Collect new airborne data specifically for one target market: railways 4. Produce the system requirements document for TreeView 5. Produce the mission requirements document for TreeView 6. Perform market analysis of end users across local and national government, commercial and research sectors