Uncertainty and Predictability of Rainfall for Agricultural Planning in Kenya

Background

Rainfed agriculture is a vital source of food and income for Kenyan farmers. Agricultural yields are highly impacted by variations in rainfall on timescales from days to decades. Future projected changes in rainfall over East Africa are uncertain, with the range of projections spanning increases and decreases in rainfall under future climate change. Understanding projected changes in agriculturally-relevant rainfall can directly inform adaptation strategies, while providing warnings and advisories to agricultural stakeholders can help planning and decision making on shorter timescales.

In this PhD you will:
• conduct high-impact physical science research underpinning the development of decision-relevant agricultural advisories for Kenya,
• explore the meteorological drivers and predictability of rainfall across Kenya on a range of timescales, and uncertainty in future projections,
• exploring the potential for introducing probabilistic outputs in agricultural advisories.

PhD Opportunity

This PhD will conduct high-impact physical science research that will support agricultural decision making and future planning in Kenya. A key part of the project is integrating probabilistic outputs and uncertainty into agricultural advisories for a changing climate.

Firstly, you will design metrics describing rainfall characteristics relevant to agriculturalists, tailored for use in probabilistic and risk assessment frameworks. Engagement with relevant stakeholders and literature will allow you to identify quantities to investigate (e.g. onset).

The project will take a seamless approach, spanning timescales from sub-seasonal to multi-decadal. Meteorological drivers of the target quantities across timescales will be explored, to understand the important modes of variability and potential predictability across timescales. Different modes of variability are expected to dominate at different timescales. Understanding these drivers will indicate on the timescales on which there is predictability.

Multi-decadal timeseries and future climate projections (including large ensembles) will be used to explore recent and future changes and variability in these metrics, to assess evolving risk to agriculture in Kenya, and to explore and quantify uncertainty in future projections. This will bring changing risk context to advisories, which are currently focused on short-term risk. Understanding the key drivers of recent variability will be key for exploring (and potentially reducing) the uncertainty in future projections.

There is scope to explore machine learning approaches.

The PhD supervisors collaborate with Kenyan organisations to improve agricultural advisories; through these connections you may engage with stakeholders, to further explore how uncertainty in rainfall forecasts and projections impacts decision making, and how probabilistic advisory information can be tailored (e.g. using storyline approaches) to support future planning in the face of uncertainty.

Applicant Profile

The project will provide exciting opportunities for students with a strong background in maths, physics, statistics, computer science or meteorology.

Other information

Caroline Wainwright: https://environment.leeds.ac.uk/see/staff/12868/dr-caroline-wainwright
Andy Challinor: https://environment.leeds.ac.uk/see/staff/1200/professor-andy-challinor
Chetan Deva: https://environment.leeds.ac.uk/see/staff/9387/dr-chetan-deva
Ongoing project: https://environment.leeds.ac.uk/directories0/dir-record/research-projects/1986/ispark-innovation-in-sustainability-policy-adaptation-and-resilience-in-kenya