This presentation discusses the definition of randomness, the sources of randomness in the physical system (the Universe) as well as in the formal mathematical system. I discuss how randomness, through nonlinearity and chaos, the second law of thermodynamics, the quantum mechanical character of small scales, and stochasticity, is an intrinsic property of nature. I then move to our mathematical system and show that even in this formal system we cannot do away with randomness, and that the randomness in the physical world is consistent with the origins of randomness suggested from the study of mathematical systems. Many examples are presented ranging from pure mathematical processes, to natural processes, to social processes, and to life in general, which clearly demonstrate how the combination of rules and randomness produces and explains the world we live in. Finally, a possible explanation is provided as to why rules and randomness cannot exist by themselves but instead, they have to coexist.

How to cite: Tsonis, A.: On the Origins of Randomness, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-1503, https://doi.org/10.5194/egusphere-egu26-1503, 2026.

EGU26-1503

Download Close

Climate models, in the form of Earth System Models (ESMs), are central to both climate science and climate-informed decision-making. Technically, they are complex, nonautonomous, and chaotic mathematical and computational representations of the Earth’s interacting “spheres”, including the atmosphere, oceans, cryosphere and biosphere. In practice, they are widely used to produce conditional projections of future climate under prescribed forcing and emissions scenarios - and remain the only tools capable of doing so within a physically consistent framework. However, despite their indispensable role, the interpretation of ESM outputs is fundamentally constrained by multiple sources of uncertainty, normally grouped as internal climate variability, uncertainty in future forcing scenarios, and uncertainty arising from model formulation.

While internal variability and scenario uncertainty have traditionally received most attention, attempts to partition uncertainty (e.g. [1]) have shown that it is actually the latter, model uncertainty, which is responsible for most uncertainty in climate projections. But despite its relevance, model uncertainty is frequently treated only implicitly, commonly subsumed under the broad label of “structural uncertainty”, with limited clarity regarding its definition or impact. This raises several unresolved questions of direct relevance to both modelling and decision-making efforts: what constitutes structural uncertainty in contemporary ESMs, how does it propagate through ensembles and projections, and how does it affect downstream socio-economic impact assessments and climate risk analyses? 

In this presentation, I will discuss these questions by reviewing recent advances in the characterisation and interpretation of model uncertainty (including results from my collaborators and myself, e.g. [2-4]) and examining their implications for the use of climate projections in scientific inference and decision-making. I conclude by identifying key conceptual and methodological gaps that must be addressed to improve confidence in climate information under persistent structural uncertainty.

With many and wholehearted thanks to all my collaborators, mentors, friends, colleagues, students and funders who made the journey to this award possible.

References:

[1] Lehner, F., Deser, C., Maher, N., Marotzke, J., Fischer, E. M., Brunner, L., Knutti, R., and Hawkins, E. (2020): Partitioning climate projection uncertainty with multiple large ensembles and CMIP5/6. Earth System Dynamics, 11, 491–508. https://doi.org/10.5194/esd-11-491-2020

[2] de Melo Viríssimo, F., Stainforth, D. A., and Bröcker, J. (2024): The evolution of a non-autonomous chaotic system under non-periodic forcing: A climate change example. Chaos, 34, 013136. https://doi.org/10.1063/5.0180870

[3] Martin, A. P., Bahamondes Dominguez, A., Baker, C. A., Baumas, C. M. J., Bisson, K. M., Cavan, E., Freilich, M., Galbraith, E., Galí, M., Henson, S., Kvale, K. F., Lemmen, C., Luo, J. Y., McMonagle, H., de Melo Viríssimo, F., Möller, K. O., Richon, C., Suresh, I., Wilson, J. D., Woodstock, M. S., and Yool, A. (2024): When to add a new process to a model – and when not: A marine biogeochemical perspective. Ecological Modelling, 498, 110870. https://doi.org/10.1016/j.ecolmodel.2024.110870

[4] de Melo Viríssimo, F., and Stainforth, D. A. (2025): Micro- and Macroparametric Uncertainty in Climate Change Prediction: A Large Ensemble Perspective. Bulletin of the American Meteorological Society, 106, E1319–E1341. https://doi.org/10.1175/BAMS-D-24-0064.1

How to cite: de Melo Viríssimo, F.: Is it possible to have confidence in climate information with structurally uncertain models?, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-21589, https://doi.org/10.5194/egusphere-egu26-21589, 2026.

EGU26-21589

Download Close