Uncertainty Discussion: What is? - What I learned from four years of spending to understand this topic.

Uncertainties inevitably arise when dealing with complexity. Over the past four years working on this topic, I've realised how intricate and layered it is. At different points in history, this topic was swayed away by other topic areas—risk management and, recently, future adaptation pathways—which leaves the fact that “uncertainty discussion” as a sole topic is always tough and has always been associated with different topic areas.

 The experts in this area add another challenge to uncertainty studies, encompassing diverse perspectives, expertise, and use of this topic. Many people are working on this topic, but their approaches and interpretations differ significantly. For instance, probability theory is often seen as a common language when discussing uncertainty, but its meaning and uses vary. A statistician’s uncertainty concept differs from a climate modeller to a seismic hazard modeller, even though they use similar statistical frameworks. This difference in interpretation adds complexity and makes it challenging to communicate uncertainty.

I encountered this firsthand during an interview during my PhD when a senior expert asked me, “Whose definition of uncertainty are you using?”  I didn’t fully understand why I was asked that question then, but it stayed with me. As a student, it was daunting. I couldn’t grasp the question's significance in my second year of a PhD. However, over the next two years, I delved deeply into the subject, exploring different types of models—statistical models, climate models, AI, and machine learning models. While I didn’t necessarily use all these tools, I needed to understand their nuances to fully comprehend the perspectives of the experts I was engaging with and the latent meaning of the question asked. Eventually, I wrote a paper reflecting on this question, which was part of my PhD thesis. 

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4888698

 

This question was so important that it triggered different avenues for me to explore. It helps me understand uncertainty's various interpretations and frameworks, which are crucial to bridging gaps between disciplines and advancing meaningful research. The challenge with uncertainty is that individuals often dedicate their entire careers to studying a specific research discipline- seismic hazard, flood hazard, landslide hazard, or tsunami hazard. They are constantly challenged by uncertainty in their scientific models. However, because of the current system of deductive scientific research approach, despite their will, they can’t focus much on ‘uncertainty as a sole focus”. They also need to deliver the output of their scientific model despite a significant uncertainty component. So, they tend to focus more on reducing uncertainties using available and widely acknowledged techniques like sensitivity analysis or Monte Carlo simulation. Where the responsibility of scientists in terms of uncertainty stops. However, natural scientists in all these domains are constantly bothered by the fact that “they are not doing enough about uncertainties, and most of the time, they can’t do anything because of the complexity involved in this topic”.