Abstract/Summary

Climate change triggers frequent adverse weather, increasing flood risks and impacting crop production amid growing food demands. My project aims to develop a method for integrating such complex systems to aid decision-making and demonstrate it, for example, focusing on the impact of tillage on flooding and farm production. This includes exploring the influence of tillage on surface runoff and crop yield by integrating evidence using diverse data types. By comparing tillage preference methods, my research aims to identify approaches that reduce flooding, maintain soil structure, support sustainable crop production and promote climate action measures. The findings have implications for promoting responsible agricultural practices. I chose the Bayesian modelling approach for its versatility in representing complex systems and flexibility in handling diverse data types. This research utilised Bayesian modelling, preferred over other techniques (Frequentist, Deterministic, Dynamic models etc.) because it reflects interrelated predictability and allows conditional dependencies between probabilities. On the one hand, the Bayesian belief network (BBN) model can function as a meta-model for integrating and analysing outputs from multiple levels within a modelling framework. And on the other hand, BBN can serve as a decision support tool for enabling decision makers to make informed and valid decisions. By employing the BBN model, my research aimed to capture uncertainties among key variables across multiple domains, enhance the adaptability of Bayesian modelling to various data types and its ability to quantify variable responses quantitatively and qualitatively. My research introduces a unique way of reviewing the literature to construct a BBN model which can be used by researchers and scientists from academia and R&D departments in the industry. This research contributes to developing decision-support tools for complex interactions such as sustainable farm production and is usable for the local farmers & growers by providing probabilistic inferences to guide decision-making in complex environments.