A recent study reveals the shortcomings of computer simulations used to formulate energy policy, highlighting in particular its unreliable assumptions and lack of transparency.The study evaluates a key model from the 1980s, highlighting the need for more effective simulation methods such as "sensitivity audits" to improve accuracy and decision-making. Dr. Samuele LoPiano stressed that these limitations must be acknowledged in order to strengthen democratic debate and decision-making on energy policy.

The widespread adoption of nuclear power, predicted by computer simulations more than four decades ago, stands in stark contrast to current reliance on fossil fuels. A recent study showed that this difference highlights the need to strengthen these simulations.

To assess the effectiveness of current energy policies, researchers revisited an influential model from the 1980s. The model predicts a significant increase in nuclear power use. Energy policy determines the production and use of energy, with profound consequences for employment, spending, climate and security. These policies are developed through simulations or mathematical models that predict elements such as electricity demand and technology costs. However, these predictions sometimes veer wildly off target.

The team found unreliable assumptions in simulations that inform energy policy and called for greater transparency about their limitations, according to findings recently published in the journal Risk Analysis. To address this issue, they propose new ways to test simulation results and be forthright about their uncertainties. This includes methods such as "sensitivity audits," which evaluate model assumptions. The goal is to improve modeling and open up decision-making.

Lead researcher Dr Samuele LoPiano, from the University of Reading, said: "Energy policy affects everyone, so it is worrying if decisions rely on just a few models without questioning their limitations. By questioning assumptions and exploring what we don't know, we can make better decisions. We must accept that no model can perfectly predict the future. But by confronting the limitations of models, the democratic debate on energy policy will be improved."

Reference: "Interpreting the Modeling Process of Energy Policy Development" by Samuele LoPiano, Máté János Lőrincz, Arnald Puy, Steve Pye, Andrea Saltelli, Stefán Thor Smith, and Jeroen van der Sluijs, November 14, 2023, Risk Analysis.

DOI:10.1111/risa.14248

Compiled source: ScitechDaily