Project Description

Emerging infectious diseases of trees such as ash dieback are treated as a national emergency in the United Kingdom and their financial impact is estimated at £15 bn. Both the international scientific community and the media have underlined that under the current framework, eradicating recent outbreaks is a laborious task and new diagnostic tools must be developed. Real-time sensors for detecting and characterising early decay in trees can play a vital role on monitoring and maintaining potential outbreaks. To that extent, the current project aims at developing a set of non-destructive geophysical tools fine-tuned for large-scale forestry applications and woodland management.
Tree decays and compartmentalisation of decays are robust diagnostic criteria for EIDs. Due to the importance of decays to the overall health status of trees, numerous drilling approaches have been suggested for assessing the internal structure of trunks and detecting hidden cavities. Although drilling methods are accurate and reliable, nonetheless, destructive techniques can cause irreversible damage to the tree, making it more vulnerable to fungi, pathogens and pests. In addition, drilling methods are constrained to a single point and fail to provide a coherent and comprehensive image of the internal structure of the trunk.
Near surface geophysics (NSG) define a set of techniques aimed at evaluating the internal structure of the subsurface in a non-intrusive manner. NSG techniques like electrical resistivity tomography (ERT) and ground penetrating radar (GPR), have been extensively used in a diverse set of applications ranging from glaciology and hydrogeology to civil/environmental engineering and landmine detection. The project will focus on developing and advancing already existed state of the art processing NSG techniques tuned for forestry and arboriculture applications. This is an inter-disciplinary project that will bridge the gap between geophysics and forestry/arboriculture. Extensive fieldwork and laboratory measurements are expected, combined with an equal amount of theoretical development, machine learning, inversion and numerical modelling.
Research Objectives
  • Develop novel processing frameworks for detecting decay in tree trunks using NSG, and in particular ground penetrating radar (GPR) and electrical resistivity tomography (ERT). Complement the data with advanced numerical modelling and laboratory experiments.
  • Use advanced NSG for mapping tree roots offorest and urban trees.
  • Explore the applicability of machine learning and artificial intelligence for automatic monitoring of trees using NSG.
Methodology
  • Data acquisition via dedicated fieldwork in Cruickshank botanic garden and in parks and woodlands in Aberdeen(CASE partner Aberdeen City Council).
  • Complement the data with advanced numerical modelling and laboratory experiments.
  • Python-based tools such as gprMax and BERT will be used for the numerical simulations.
  • Different protocols, methodologies and measurement configurations will be tested, e.g. full 3D electric resistivity tomography of tree trunks.
  • Building machine learning pipelines using Python (Tensorflow, Pytorch etc.) for processing and
  • analysing GPR and ERT data for forestry and arboriculture applications.
Requirements
The project requires an enthusiastic student who is keen to work with an inter-disciplinary group tackling global challenges via creative and innovative approaches. The work has a strong numerical and computational aspect, so a background in applied-mathematics, signal processing, machine learning, and geophysics are essential. Moreover, the student should have a strong background in Python and/or R,Matlab. The project will be based in Aberdeen with lots of experimental work and fieldwork in nearby areas.
Essential & desirable candidate skills
Essential: Bachelor and Master in Geophysics/Geosciences or any technical discipline with a strong computational element.
Desirable: Experience with Python/Matlab. Strong background in inversion, numerical modelling and machine learning. Prior experience on acquiring and processing geophysical data.

Supervisors

Iraklis Giannakis

Primary Supervisor:

Profile: Iraklis Giannakis
Email: iraklis.giannakis@abdn.ac.uk
Institution: University of Aberdeen
Department/School: School of Geosciences

Alastair Ruffell

Secondary Supervisor:

Profile: Alastair Ruffell
Email: a.ruffell@qub.ac.uk
Institution: Queen's University, Belfast
Department/School: School of Biological Sciences

Michelle Pinard

Additional Supervisor:

Profile: Michelle Pinard
Email: m.a.pinard@abdn.ac.uk
Institution: University of Aberdeen
Department/School: School of Biological Sciences

Additional Supervisor:

Dr Antonios Giannopoulos
University of Aberdeen, School of Engineering
Lab website: https://www.eng.ed.ac.uk/about/people/dr-antonis-giannopoulos
Online Profile: https://scholar.google.com/citations?user=0GKjYwQAAAAJ&hl=el
Email address: a.giannopoulos@ed.ac.uk

References

1.A. Broome, D. Ray, R. Mitchell and R. Harmer, ”Responding to ash dieback (Hymenoscyphus fraxineus) in the UK: woodland composition and replacement tree species,” Forestry, An International Journal of Forest Research, vol. 92, pp. 108–119, 2019.

2.A. Santini, L. Ghelardini, C. De Pace, […] and J. Stenlid, ”Biogeographical patterns and determinants of invasion by forest pathogens in Europe,” New Phytologist, vol. 197, pp.238-250, 2012.

3.A. M. Ellison, M. S. Bank, B. D. Clinton, […] and J. R. Webster, ”Loss of foundation species: consequences for the structure and dynamics of forested ecosystem,” The Ecological Society of America, vol. 3, pp. 479-486, 2005.

Expected Training Provision

The PhD candidate will spend at least one month per year with the Aberdeen City Council (CASE partner) working along foresters on forestry management and tree inspection. The supervisory team combines a strong background and expertise in signal processing, machine learning, NSG (UoA, QUB, UoE) and forestry (UoA). Our team of investigators brings together early-to mid-career researchers with senior, experienced staff of internationally recognised standing. Academic support and further training will be provided to the PhD student throughout the duration of the project. The PhD candidate will attend conferences related to both NSG and forestry, enhancing her/his knowledge on the topic and expanding her/his network. The student will have the option to undertake the modules GL5059 (Near Surface and Environmental Geophysics), GL5060 (Inversion Theory) and GL5709 (Machine Learning in Geophysics). Some of these modules are taught partially by the PI Dr Giannakis for the MSc in Geophysics at UoA, and will aim at equipping the PhD student with all the necessary skills to pursue a career in near surface geophysics, computational geosciences and machine learning; rapidly growing sectors with plenty of career opportunities.

Impact

Ash dieback is a prominent Emerging Infectious Disease (EID)that has invaded the UK in 2012 and it has spread majorly in central England. Less than 5% of the ash trees are immune to this disease and it is predicted that most of the ash trees in the UK are going to be affected and die in the next twenty years. Acute Oak Decline (AOD)is a particularly aggressive EID that can lead to tree mortality within a period of 3-5 years. AOD has been introduced to the UK in 2006 and since then has rapidly spread mostly in the central part of England. Ash dieback and AOD have already infected thousands of trees in the UK, nonetheless their impact is dwarfed in comparison to the millions of trees affected by Xylella-Fastidiosa in Italy. Xylella-Fastidiosa is a vector-transmitted, slow progressing bacterium that has spread in Italy with devastating effects to the overall population of olive trees. Italy has declared state of emergency since 2015 and is now under European quarantine control. The international scientific community and the media have underlined that under the current framework, eradication of EIDs is a laborious task and new forestry approaches should be developed for monitoring and diagnosing EIDs. The outputs of this project will change the paradigm in forestry monitoring by introducing advanced 3D geophysical methods coupled with state of the art processing schemes and machine learning. This will have a direct impact to the forestry community, the stakeholders from the private sector and government bodies associated with woodland management and heritage preservation. The Natural Environment Policy and Climate and Environment Policy Service of Aberdeen City Council is a CASE partner of this project, applying and testing our newly developed tools in the field for managing urban trees, forests and woodlands within their jurisdiction. Most importantly, our project will answer the ongoing call for more advanced monitoring tools for detecting, containing and overall managing current and future outbreaks of EIDs; a timely problem with an exponential growth, primarily affected by modern lifestyle and climate change.

Proposed Timetable

0 -6 months: Literature review.

6 -12 months: Data acquisition (fieldwork, experimental, synthetic).

12 -24 months: Adjust and tune already existed processing schemes for forestry and arboriculture.

24 -36 months: Develop novel processing and machine learning schemes.

36 -42 months: Writing thesis.

Conference attendance:

July 2024 5-days : IGARSS 2024 to be held in Athens, Greece.

March 2025 5-days : EGU 2025 to be held in Vienna, Austria.

2025 –2026 1-day : Attend one of the four annual LOTA meetings held in London, UK.

November 2026 2-days : National Tree Officers Conference 2025 to be held in London, UK.

Every year the student will be spending at least one month at Aberdeen City Council(CASE partner)working along with foresters on managing urban and forest trees.

QUADRAT Themes

  • biodiversity
  • earth-systems
  • environmental-management

Partners

A CASE partnership has been confirmed with Aberdeen City Council.

View All Projects