Uncertainty in Science:
A case for robust energy solutions

Authors: Alexander Wanitschke, Berit Müller
With editorial assistance by Miriam Leich and Ludwig Hülk

As researchers in the renewable energy sector, we try to find solutions to real-life problems. How can we decarbonize energy and transport? Which technologies are preferable – which ones obstruct? What does transitional decision-making look like? Administrations on all levels turn to science, hoping to receive safe parameters to base their decisions on. But none of the parameters we deliver today are ever truly certain.

A good scientist accepts that no system or mechanism can ever be fully understood – uncertainties exist in all fields of research. Any results from a study or analysis, from recommendations for transition pathways to the number of charging points that are needed to supply electric cars in a specific region, are educated guesses based on historic evidence, simulations, and assumptions about future developments. These guesses can be quite accurate, when researchers know their trade, but they should never be viewed as universal truth.

This is, however, a mistake that policy makers commonly seem to make when planning for future needs. Complex problems that entail a lot of uncertainties are reduced to a handful of questions and the (scientific) answers to these questions is often wrongfully used as justification for decisions that may heavily influence the fate of future generations. But who can blame them – the world has become incredibly complex and it seems now very difficult for decisions makers to find certainty in the solutions to the problems they are in charge of.

Why is it now more difficult to design robust policies than a few decades ago? One answer lies in a general increase of interconnectedness. A good example for this can be found in energy industry itself: The different sectors, such as power, heat, and transport, used to be almost completely independent of each other, which made it easier to predict the consequences of change within the sector. Today, through sector coupling, a necessary step for the deep penetration of renewable energy, all sectors of energy are increasingly interlinked and influence each other, which increases the overall uncertainty regarding predictions on energy.

A second reason for today’s increased uncertainty lies in impending, game changing technological innovations. Digitization, for instance, bears the potential to change many of so far well-understood problems. New digital mechanisms like service automation are not simply added to existing mechanisms, but change or terminate them. Autonomous driving is a good example. Full automation can alter a vehicle’s function so profoundly that it changes its concept entirely. While this kind of conceptual revolution is nothing new to us (as we have gone from horse wagon to combustion car), we can only know that we don’t know many of the consequences of such a change.

A third reason for uncertainty lies in an increased awareness and recognition of society’s diverse values. Today’s pluralism demands that a large range of perspectives and values be taken into political and – more importantly – scientific account. Applied to modern energy systems, this means that the aim has shifted from the simple task of supplying energy cheaply and perhaps reliably to now also supplying it at the very least sustainably, decentralized, transparently, and socially fair. This is not a negative development, however, it adds immense complexity to the energy problem and increases the pressure on decision-makers and scientists in this field.

So what are we to do about it? As researchers, we need to detach ourselves from the role of “master consultants” for society as a whole. Political decisions can’t be based solely on scientific foreknowledge and prophetic scenarios. We should rethink the way we do energy research and pay more attention to the uncertainties which inevitably frame our results, whether we want it or not. In light of uncertainty, the more popular term for robustness, resilience, is not a blank buzz word, but has actually replaced optimality as the guiding principle of designing energy solutions.
Energy transition is tedious, so we must think of better ways to communicate our research results and their reliability for future purposes. Political decisions must be informed by robust expert advice and supported through public dialogue – with science and everyone else. We cannot easily overcome the increasing complexity of things, but if we stop ignoring the obvious uncertainties, we take a step into the right direction. If we find out what we don’t know, we will gain even better understanding of the energy transition problem, increase public trust in our work, and finally stop the kind of path-dependent decision-making that previous generations of scientists have already burdened us with.
Uncertainty classification, multi-objective optimization and exploratory modeling are just a few of many powerful approaches in this direction.