Quantifying crop-climate risk across timescales to inform decision-making
Background
The impacts of climate change on food production have been addressed with a range of frameworks in response to an evolving series of societal questions. Two recent trends stand out. The increasing use of climate risk in framing assessments reflects the need for longer-term planning. The parallel increase in data-driven approaches is identifying adaptive responses that can be implemented now. The challenge now is to inform decisions by bridging these two timescales into an integrated understanding
This studentship builds on progress made at Leeds and in the Met Office. First, new process-based crop-climate models, using machine learning (ML) and satellite data, are being developed. Second, ML is being used to forecast weather days to seasons ahead, providing computationally cheap ensembles. Third, storyline approaches are being used to integrate understanding from crop-climate ensembles with analyses of large-scale weather patterns, nutrition security trade outcomes, and policy futures.
PhD Opportunity
The central concept of the studentship is to examine risk across timescales. At seasonal timescales, the recently-developed version of the GLAM crop model, using satellite data, will be used with ML weather forecasts to develop crop-climate ensembles of risks and responses to weather extremes. Across multi-year and multi-decadal timescales, storyline approaches will describe how patterns of correlated crop-climate risks linked to well-known patterns of climate variability, such as the El Niño-Southern Oscillation (ENSO), may change out to 2050 under different socioeconomic pathways
The studentship focusses on Kenya and the UK, which present diverse decision-making, technological and agro-climatic environments. We expect this diversity to provide insightful contrast, via a focus on multi-objective decisions, identifying both synergies and trade-offs.
Objectives:
1) Improve assessment of crop-climate risks by exploiting and developing data-driven ensemble approaches
2) Apply assessments of crop-climate risks in real-world decision-making exercises, through storyline approaches
3) Identify sources of uncertainty in crop-climate prediction from days to decades and explore how this determines the conditions under which actionable information can be produced.
The research builds on existing capacity at Leeds and at the Met Office. Kenya is the focus of multiple projects at Leeds. For example, we are working with an agro-advisory business to develop specific advice at the start of the rainy season for maize intercrop systems. The UK is the focus of the HiFi project and results show significant skill in predicting wheat yield months ahead of the harvest.
There is potential for both countries to be the focus of impactful new work by the student, e.g. for UK government monitoring of food security. The research aligns closely to wider Met Office research, e.g. ongoing work exploring climate risks to UK food security and the international food system.
Applicant Profile
Students with a strong background in physics, mathematics or statistics who want to apply these skills to food-related climate risks and the development of adaptive strategies.
Other information
https://www.gov.uk/government/publications/uk-food-security-index-2024/uk-food-security-index-2024
https://environment.leeds.ac.uk/see/staff/9387/dr-chetan-deva
https://environment.leeds.ac.uk/see/staff/1200/professor-andy-challinor
https://www.metoffice.gov.uk/research/people/edward-pope
Droutsas, I, Challinor, AJ , Deva, CR et al. (1 more author) (2022) Integration of machine learning into process-based modelling to improve simulation of complex crop responses. in silico Plants, 4 (2). diac017. ISSN 2517-5025
Watt-Meyer, O., Henn, B., McGibbon, J., Clark, S.K., Kwa, A., Perkins, W.A., Wu, E., Harris, L. and Bretherton, C.S., 2025. Ace2: Accurately learning subseasonal to decadal atmospheric variability and forced responses. npj Climate and Atmospheric Science, 8(1), pp.1-15.
https://iopscience.iop.org/article/10.1088/1748-9326/ac816d/meta
