Workshop 'Predictability in Natural and Social Sciences'


KüNO Online Workshop 'Predictability in Natural and Social Sciences'

3 May, 2022

9 am to 1 pm


The first KüNO Online Workshop ‘Know how meets know why - Collaboration of natural and social sciences in German coastal research’ that was held on 6 Sep, 2021 with about 60 participants, provided the opportunity to discuss a wide range of topics arising in interdisciplinary cooperation. It addressed most common problems and misunderstandings in collaboration and showed possible solutions. The workshop concluded that a deeper exchange on specific topics would be desirable. One specific topic that was identified was the one of predictability in different natural and social science disciplines which this KüNO workshop now addresses in detail. 

Topic of this 2nd workshop:

Intuitively speaking, predictability is the antonym of uncertainty. If predictability is high, uncertainty is low and vice versa. However, each scientific disciplines has its own understanding what is considered to be predictable. Specifically, when natural and social scientists talk about predictability, they usually don't mean the same thing. While in many quantitative natural and social sciences it is common practice to explicitly characterize uncertainties for any measured quantity, the predictability is treated more implicitly in more qualitative oriented social science disciplines. But even if the uncertainty in measure quantities is low, predictability via numerical models can be be limited, because even a small degree of uncertainty can lead to unpredictable outcomes in complex systems.

In all cases, an adequate characterization of what is predictable and what is not is the basis for meaningful policy recommendations. Hence, a basic understanding of different disciplinary viewpoints on predictability and the time scales involved is essential for interdisciplinary collaboration. For this reason, the goal of this workshop is to clarify how predictability is defined in different research fields and what this implies for the communication of interdisciplinary research results.


After discussing general challenges of interdisciplinary collaboration in a first workshop last year, the exchange on the topic of 'Predictability' was now deepened here. The speakers from different marine research areas gave insights into their views on the topic and triggered lively discussions among the more than 30 participants. Gary Polhill from the James Hutton Institute in the UK gave a keynote on predictability in complex systems. In breakout groups, discussions were held on how to communicate uncertainties to stakeholders and the public. Finally, it was concluded that researchers should not only communicate their findings in scientific journals, but also engage in adapted communication with stakeholders and society. To avoid confusing statistical uncertainties with results that are not certain, confidence intervals should be emphasized more. A summary can be downloaded here: SUMMARY (pdf).


Introduction to predictability
Jochen Hinkel (Global Climate Forum)
Predictability of coastal morphology (pdf available here)
Christian Winter (Kiel University)
Predictability of species populations and ecosystem dynamics (pdf available here)
Kai Wirtz (Hereon)
Predictability of economic behaviour (pdf available here)
Andrea Wunsch (Kiel Institute for World Economy)
Predictability of institutional change
Achim Schlüter (Leibniz Centre for Tropical Marine Research)
Keynote: Predictability in complex adaptive systems (pdf available here)
Gary Polhill (The James Hutton Institute)
Coffee break
Discussion in breakout groups
  • In which disciplines/fields can scientific progress increase predictability?
  • How can limits of predictability be communicated to policy and decision makers?
  • Limits of predictability and public discourse (e.g. with regards to Corona)


Kai Wirtz (Hereon), Christian Winter (Kiel University), Jochen Hinkel (Global Climate Forum)

Wrap up
Jochen Hinkel (Global Climate Forum)
End of the workshop




Gary Polhill presented his views on predictability in complex systems.