Addressing uncertainty to improve urban tree management

Summary Description:

Eighty-percent of the UK population live in towns and cities. Urban trees provide numerous benefits to urban society, including air pollution removal, building energy conservation, urban climate regulation, and access to nature. Urban tree managers and government agencies are interested in assessing the magnitude and socially equitable distribution of urban tree benefit delivery, and building resilience under a changing climate. However, a national picture of urban forest cover, composition and quality does not exist. At the city scale, such information is occasionally available through local uptake of “i-Tree” tools.

This project aims to critically examine urban forest sampling protocols with a view to optimising i-Tree Eco surveying. By clarifying the surveying effort required and maximising output accuracy the project aims to increase the opportunity for cities to gain the inventory data required for evidence-based policy creation, and development of management strategies that maximise delivery of tree benefits to urban society.

Project Outcomes:

"Using statistically robust techniques, we found evidence to suggest that the efficacy of an i-Tree Eco survey plot design in estimating total tree populations, appears to be reliant on characteristics of the underlying areas when considering i-Tree Eco data, but not ProximiTREE or National Tree Map data. Unlike i-Tree Eco data, the ProximiTREE and National Tree Map datasets are fully observed, suggesting our results could be attributed to addit ional prediction variation introduced when simulating from i-Tree Eco data. Whilst addressed in the thesis, further work is required to establish the impact of the additional variation for i-Tree Eco simulations and whether the ProximiTREE and National Tree Map results are more reliable despite differing from i-Tree Eco in defining the term tree. Estimated populations errors have been presented and summarised for a variety of survey plot designs and tree density structures, with further images and tables provided in the appendices. Different tree density structures were assessed by simulating tree densities from existing data to cells overlaid on some area of interest. Simulated values were intended to provide tree densities that follow logically from the underlying area, but which may differ from tree densities observed in reality."

How have the results been used?

Phillip has presented and discussed his work in a variety of formats as he progressed through his PhD, including talks, posters and articles. Formally, he presented and discussed his work to statistical audiences through both a talk and a poster presentation at the Royal Statistical Society conferences in 2021 and 2022. He also presented his work through a talk at the Scottish Forestry Trust Bursary Students’ Seminar, an online talk for the Institute of Chartered Foresters, and presented a poster online at the Treescapes 2021 conference.

An article discussing Phillip's work was published in the Institute of Chartered Foresters members publication, TREES.