Ignorance of Model Uncertainty and its Effects on Ethics and Society Using the Example of Geosciences

Numerical modeling plays an important role in geosciences when it comes to the prediction of future states of the Earth. The predictive capabilities of numerical models building on machine learning or computational modeling become increasingly relevant in political decision making. However, models are fraught with uncertainties, which currently are rarely made transparent. If we ignore maximally realistic uncertainty quantification in numerical modeling as value-adding knowledge about the unknown, we are likely to exacerbate ethical problems related to geoscientific modeling. In this chapter, we review the role of models in geosciences and types of model uncertainty. We discuss the communication of model uncertainties from societal and ethical perspectives. Currently, users of models, e.g., political and economic decision makers, do not appreciate predictive models that carry more knowledge about uncertainties. Rather, they often perceive them as more uncertain. This makes it easy for the geoscientific community to ignore the need to quantify uncertainties. We argue that it is crucial to develop improved and standardized uncertainty quantification approaches and communicate their outcomes when modeling and predicting complex systems such as the Earth.