CTP-SAI-083
Breaking Breeding Bottlenecks: Combining remote sensing and machine learning to rapidly phenotype field crop yield potential
CTP-SAI-083
Breaking Breeding Bottlenecks: Combining remote sensing and machine learning to rapidly phenotype field crop yield potential
Dr Kanthu Joseph Mhango (HAU), Dr Edwin Harris (HAU), Prof Jim Monaghan (HAU), Dr Edwin van der Vossen (Solynta)
BACKGROUND
This PhD project will develop high-throughput, automated phenotyping tools for rapid evaluation of field crops, using the potato as a model crop. Current capacity for rapidly genotyping crops is substantial, but phenotyping crop responses in the field is now the limiting stage. The project will integrate the output from reflectance sensors on unmanned aerial vehicles (UAV) with radiative transfer models (RTMs) to estimate values of critical traits like leaf area index (LAI) and chlorophyll intensity. These traits will be used to develop machine learning models for benchmarking yield potential and simulating the impacts of agroclimatic change on yield. The resulting high throughput methods will impact breeding and agronomy programs by enabling predictions of yield-potential in large breeding populations. The PhD research will therefore appeal to a candidate with a plant science background but with high interest in programming and applied remote sensing.
Spectral reflectance data collected through UAVs can build a complete picture of light absorption in a canopy. Mechanistic models, such as radiative transfer models (RTMs), are ideal for inverting spectral signatures into LAI and Chlorophyll intensity with sound mechanistic principles but remain largely untested in potatoes. With reliable estimates of light absorption, plant breeding programs can be equipped with powerful selection traits for improved photosynthesis and mechanistic crop growth models based on irradiance. This project will leverage advances in sensor technology and data science to calibrate RTMs for potatoes, extract light absorption parameters and generate methods for dynamic prediction of Radiation Use Efficiency (RUE). The RUE models will be used in a knowledge-guided machine learning framework to enhance the accuracy of process-based crop growth models. The student will therefore optimize mechanistic crop growth models, applying crop physiology concepts to provide decision support tools for plant breeding and precision agriculture.
OBJECTIVES AND APPROACHES
1. Optimization of LAI, Chlorophyll intensity and light absorption estimation from UAVs using Radiative Transfer Models
The student will conduct cutting-edge research to integrate physics-based radiative transfer models with spectral data captured by UAVs for optimizing the prediction of physiological state (e.g. leaf area index) and photosynthetic efficiency (e.g. chlorophyll intensity) in potatoes. This objective will ultimately integrate data on the spectral properties of plants with remote sensing, machine learning and constraint-based optimization to generate knowledge-guided predictive models for crop growth.
2. Comparative analysis and development of crop growth models based on light absorption
The project will develop and test mechanistic crop models that will predict yield development by integrating spatial and temporal determinants of light interception, dry matter production and allometry. Together with studies on the temporal evolution of dry matter content, the project will contribute significantly to addressing key knowledge gaps in potato crop growth modelling.
3. Development of a knowledge-guided predictive engine for crop growth and yield
The student will develop a knowledge-guided predictive engine of crop growth and yield development that takes full advantage of remote sensing data sources. It will have a significant data science component that will investigate and quantitatively assess analytical frameworks for remote sensing data assimilation, quality control, imputation and constraint-setting for optimization algorithms with minimal oversight requirements. Additionally, state-space models, machine learning and traditional calibration curves will be compared for forward extrapolation of crop growth processes with non-linear variables.
PRIMARY LOCATION OF THIS PHD
This PhD will be based and registered at Harper Adams University. The data science will be based at the Centre for Agricultural Data Science (harper-adams.ac.uk). Glasshouse and field work will be based at the Centre for Crop and Environmental Science (harper-adams.ac.uk) and Solynta
TRAINING
This project will give the student substantial experience in applied remote sensing for high throughput phenotyping with spectral data, plant physiology and mechanistic crop growth modelling, advanced machine learning, design of automated data collection and management in large-scale field and controlled-environment experiments and collaborating with a commercial sponsor. The student will embed part of their field experimentation within Solynta’s research program, giving the student the opportunity to engage with a cutting-edge potato breeding company with advanced phenotyping capabilities.
The CTP – SAI (https://www.ctp-sai.org) is a groundbreaking partnership between leading businesses, charities and research providers offering outstanding training for the agri-food sector. Students will have access to training opportunities through their University to complement their scientific development. This will be augmented by training in key bioscience skills to enhance employability and research capability through the CTP-SAI.
There will be additional training to enhance employability and research capability. All CTP-SAI students will receive Graduate Training in Leadership and Management as well as personal development skills training from MDS (www.mds-ltd.co.uk).
INDUSTRIAL PLACEMENT
Placements are a key feature of CTP and UKRI-BBSRC expects all doctoral candidates on a CTP programme to undertake a placement. Placements can be in the form of research placements (3-18 months duration) or used more flexibly for experiential learning of professional skills for business and/or entrepreneurship. All placements are developed in collaboration between the partners with input from the doctoral candidate.
APPLICATION AND ELIGIBILITY
Contact Dr. Kanthu Joseph Mhango (jmhango@harper-adams.ac.uk) for an informal discussion on the research content of this PhD.
This studentship will begin in October 2025. The successful candidate should have (or expect to have) an Honours Degree (or equivalent) with a minimum of 2.1 in Plant Science, Applied Statistics, or other related science subjects. Students with an appropriate Masters degree are particularly encouraged to apply.
We welcome UK, EU, and international applicants. Candidates whose first language is not English must provide evidence that their English language is sufficient to meet the specific demands of their study. Candidates should check the requirements for each host organization they are applying to, but IELTS 6.5 (with no component below 6.0) or equivalent is usually the minimum standard.
This studentship is for four years and is fully funded in line with UKRI-BBSRC standard rates. These were for 2024/25, an annual maintenance stipend of £19,237, fee support of £4,786, a research training support grant of £5,000 and conference and UK fieldwork expenses of £300.
To be classed as a home student, candidates must meet the following criteria and the associated residency requirements:
• Be a UK National or,
• Have settled status or,
• Have pre-settled status or,
• Have indefinite leave to remain or enter
• Be an Irish National
If a candidate does not meet the criteria above, they would be classed as an international student and must demonstrate the ability to meet the supplement in fees required for an international student.
Anyone interested should complete the online application form before the deadline of 5th January 2025. Interviews will be held during January and February 2025.
Please contact recruitment-ctp-sai@niab.com for further application details.
