skip to primary navigationskip to content

Coomes Group: Getting epidemiological: applications of hyperspectral remote sensing to detecting, predicting and managing the spread of plant diseases

NERC / GFC / OTHER

Supervisor: David Coomes (Plant Sciences
Co-Supervisors: Susan Wiser (Landcare Research, New Zealand); John Dymond, (Landcare Research, New Zealand); John Carr (Plant Sciences)

Importance of the area of research concerned:

Introduced diseases are highly destructive in forests, eliminating tree species that were once common and radically altering ecosystem processes. Tracking the spread is valuable for forest management particularly if patches of resistance can be identified remotely, but traditional approaches are time consuming and costly. Airborne hyperspectral imaging is an emergent technique for tracking diseases. Hyperspectral imagers measure light energy in hundreds of narrow wavebands, making it possible to identify changes in the physico-chemical properties of canopies infected with a disease. This project will develop methods to monitor the spread of ash dieback (Hymenoscyphus fraxineus) in Europe and/or Myrtle rust (Austropuccinia psidii) in New Zealand. Ash dieback is predicted to cost the UK economy £15 billion, while myrtle rust threatens the multi-million-dollar manuka honey industry.

Project summary:

New hyperspectral imaging approaches will be developed to map the distribution of ash and ash dieback in UK forests, and Myrtaceous plant species + associated disease in New Zealand forests. We will use a spectranomic approach, developed by Asner and Martina (2009) to predict the physico-chemical properties of leaves from their spectral information, allowing us to map species and disease status remotely.

What will the student do?

Working within a team of New Zealand researchers, the student will conduct fieldwork at selected sites in the North Island. Firstly, a field spectrometer will be used to record the leaf spectra of hundreds of species of New Zealand plant, and these leaves will be analysed for 20 physico-chemical traits. Chemometric modelling will then be used to predict trait values from leaf spectra (Asner and Martin 2009). Secondly, an airborne spectrometer will be used to map forests at contrasting sites in northern New Zealand and by identifying individual tree crowns on the ground, a predictive model will be constructed to generate maps of species and disease status (see Vaughn et al 2018). Together, these two approaches should allow Myrtaceous species to be mapped and their disease status assessed. These data will be used to parameterise a SAR epidemic model, which would be used to predict the spread of myrtle rust within New Zealand. A similar approach will be applied in the UK, where we have already developed approaches for species and disease detection within single woods.

References:

  • Asner G.P. & Martin RE (2009) Airborne spectranomics: mapping canopy chemical and taxonomic diversity in tropical forests. Frontiers in Ecology and the Environment https://esajournals.onlinelibrary.wiley.com/doi/abs/10.1890/070152
  • Vaughn et al. (2018) An approach for high-resolution mapping of Hawaiian Metrosideros forest mortality using laser-guided imaging spectroscopy. Remote Sensing 2018, 10(4), 502; https://doi.org/10.3390/rs10040502
  • Biosecurity New Zealand website: Myrtle rust.https://www.mpi.govt.nz/protection-andresponse/responding/alerts/myrtle-rust/

Applying: To the Cambridge NERC C-CLEAR DTP programme: https://nercdtp.esc.cam.ac.uk/

 

Filed under: , , , ,

The Department has carried out a comprehensive COVID-19 risk assessment process and has opened to allow research work to take place. To ensure the safety of our staff, a range of measures to reduce building occupancy and allow strict social distancing have been introduced, including increased cleaning and hygiene regimes. We are currently not accepting visitors so please continue to contact us by email until further notice.