Delivering resilient power, road and rail networks by translating a tree failure risk model for multi-sector applications.

Lead Research Organisation: Lancaster University
Department Name: Lancaster Environment Centre


When storms cause trees to fall onto power lines, roads and railways this can pose serious threats to human life and disruption to electricity supplies and transport leading to large financial costs to both operators and users of these networks. There is growing evidence that climate change will lead to an increase in storminess in the UK so the problems associated with tree failure are likely to grow. This project aims to use a computerised system for predicting which trees are likely to fall onto powerlines, roads and railways during the types of storms that typically occur in the UK. This will allow the operators of these infrastructure networks to fell those trees that are most likely to fail and cause disruption. The project will make use of newly-developed techniques which employ airborne laser scanners to map trees and measure their key properties, which will improve our ability to estimate the susceptibility of trees to failure. So the outputs of the system will enable the operators of infrastructure networks to pro-actively manage trees in order to improve the resilience of their infrastructure to future storms.

The intensity and impacts of storms vary considerably over space and time so it is not possible to manage trees for all possible conditions. Therefore we will develop the tree failure prediction system so that it is able to use short-range weather forecasts (up to 5 days) which are the most reliable predictions of impending storm events. This will enable the system to predict which trees are likely to fail and cause disruptions to infrastructure networks during the forthcoming storm conditions. This information will help the network operators to draw up effective plans for responding to and recovering from storms, e.g. by organising field teams to be in the locations where greatest tree damage is likely to occur so they can remove fallen debris and repair the infrastructure. To be effective our tree failure prediction system will need to operate quickly and repeatedly so it can respond to regular updates in weather forecasts as storms develop. Also it needs to incorporate assessments of the large number of trees which surround the power, road and rail networks in the UK. Therefore, to achieve this, we will make use of the very powerful cloud-based computing technology that is now rapidly developing. The outputs of our system will be conveyed to users via an interactive web page which will support strategic decision-making and a mobile app that will support field teams.

Keywords: tree failure, storm, prediction, power supply, road, rail, decision-making, resilience.
The following organisations are stakeholders in the project and will form an advisory board to oversee our work: UK Power Networks, Scottish Power, Transport Scotland, Scottish Water, Bluesky International, Atkins Global.

Planned Impact

This project will have a direct benefit to infrastructure operators by ensuring that in the future they only manage trees that pose a significant risk to their networks. This will offer operators the advantage of being able to optimise the use of limited resources for tree management. This evidence-based pre-emptive tree management will reduce the probability of damage to infrastructure and therefore reduce the costs of repairs and compensation payments and disruption to customers/users. The project will also deliver a real-time forecasting system that will provide information on the location and timing of tree failures that are likely to occur during an imminent storm event. This will support the planning of effective response and recovery operations during storms in order to minimise the length of disruption to network functions and maximise the efficiency of spend on field operations.

The tree failure prediction model will also be of benefit to network operators when planning new or replacement planting of trees around infrastructure as it will enable simulation of future failure risk given different scenarios of species of tree used, spacing and orientation and the design of protective shelterbelts and profile planting. In this way the model will become an important element in the design of more resilient infrastructure networks in future.

A key impact of the model is that it will enable the selective management of individual trees that are at greatest risk of failure, as opposed to generic wide-scale tree and vegetation clearance programmes. This will generate important environmental benefits as it minimises disruption to habitat provision for biodiversity and other ecosystem services offered by trees such as their role in biogeochemical cycles, in particular carbon assimilation and their role in flood mitigation via their hydrological interactions. The selective management approach also minimises the unnecessary removal of trees which may be providing visual or noise screening of infrastructure for nearby residents.

Therefore, this project will have a series of positive impacts. Infrastructure networks will become more resilient to tree failure risks resulting in positive financial benefits for operators, reduced disruption to consumers and users, and fewer detrimental effects for business and the wider economy. These outcomes will be delivered at the same time as significant environmental benefits.


10 25 50
Description We have developed a robust, operational model of tree failure risk using improved representations of tree characteristics and context. We have developed a techniques for accurately delineating individual trees from LiDAR data, capturing the key biophysical attricbutes of each tree and using this data to drive the tree failure model. We have also developed the ability of the tree failure model to incorportae information on soil properties. These techniques have been applied to sections of the UKPN and Scottish Power electricity supply networks to simulate the impacts of tree failures on this infrastructure. The advantages of our approach has been demonstrated to a range of other stakeholders that have responsibility for managing different forms of infrastructure.
Exploitation Route The TREEFALL model will be valuable to the operators of infrastructure networs that are susceptible to damage and disruption caused by tree failures. The model allows stakeholders to identify locations at which to focus resources for tree management in order to maximise the robustness of their networks.
Sectors Energy,Environment,Transport

Description The TREEFALL model has been demonstrated to Scottish Power and UK Power Networks by applying it to parts of their electricity supply networks and we are developing a strategy for future work that will facilitate the operational use of the model. In 2018 and 2019 we have been in discussion with Network Rail and UK Power Networks about potential alternatives for funding future development and application of the model.
Sector Energy
Impact Types Economic

Title TREEFALL model 
Description A web-based model for predicting the risks of individual tree failures across landscapes. 
Type Of Material Computer model/algorithm 
Provided To Others? No  
Impact The model has been used to predict the risks to electricity supply networks due to tree failure. 
Description Collaboration with ADAS on TREEFALL 
Organisation ADAS
Country United Kingdom 
Sector Private 
PI Contribution We have worked with ADAS to develop the basis of the TREEFALL model.
Collaborator Contribution ADAS have collaborated on the project, under contract, to develop some aspects of the TREEFALL model and disseminate the finding of the project ot relevant stakeholders.
Impact The functioning model TREEFALL has resulted from this collaboration.
Start Year 2015
Description Dissemination of findings 
Form Of Engagement Activity Participation in an activity, workshop or similar
Part Of Official Scheme? No
Geographic Reach National
Primary Audience Industry/Business
Results and Impact Several presentations of the TREEFALL model to various stakeholders and research colleagues
Year(s) Of Engagement Activity 2015,2016,2017