ONE Planet DTP - Predicting farm profitability under future climate and market uncertainty. (OP20272)

Project Description: The PhD project will develop a farmer behaviour predictive model to investigate the implications that 1) climate change events (e.g. changes to rainfall and temperature patterns and extremes), 2) market events (e.g. changes in prices and its volatility) and 3) agricultural policies have on farmer’s decisions and how these decisions may affect farm profitability of farms in England and Wales. The PhD project will provide a tool to evaluate how the level of uncertainty faced, the role of farmer and farm characteristics may affect farmers making the right decisions in terms of farm profitability. Understanding farmers’ behaviour under uncertain weather, market and policy events is key for analysing how farmer’s decisions affect farm profitability. Usually, research on farm profitability tends to focus on the economic elements associated to it such as inputs and output quantities, prices and policy support levels disregarding farmer’s expectations on weather events. Likewise, research on the effect of climate change on agriculture production disregards farmer’s behaviour under uncertainty. Integrating all these aspects will be fundamental parts of this research. This PhD project combines the use of available information from historical farm business survey data and level of policy support received with weather information from the Met Office and the use of primary data collected from surveys to farmers to help validating the model. A combination of pattern recognition using machine learning and spatial econometric modelling will be used to identify, simulate weather, market conditions, and predict future farmer decisions on agricultural production and the associated farm profitability across farms across England and Wales. Key specific skills gained as part of the training for this project are learning about predictive modelling, machine learning and econometric methods as well as learning R/Python programming languages.