Can we trust climate models?
- Catherine Louropoulou
- πριν από 5 ημέρες
- διαβάστηκε 2 λεπτά
Climate models are an important tool for understanding future changes in the Earth’s climate. Despite decades of scientific research, there are still many who question not only the accuracy of these models, but even the existence of climate change itself. Others may not doubt that this is happening yet doubt the ability of science to predict it — we often hear that “climate models are always wrong.”

However, the data show the opposite. For decades, climate models have consistently and impressively captured the actual trend of global warming.
A recent study titled “Evaluating the Performance of Past Climate Model Projections” (Hausfather et al., 2020), published by the American Geophysical Union, assessed the temperature projections of 17 different climate models, from the early 1970s through the late 2000s, comparing them with the actual temperature observations that followed. In general, the accuracy of these projections depends primarily on correctly representing the physics of the climate and on realistic assumptions about future CO₂ emissions. The results showed that when the researchers took into account the actual greenhouse gas emissions, the accuracy of the models increased significantly — 14 out of 17 matched the observed temperatures!
It seems, then, that we can indeed trust climate models — and, in fact, it’s essential that we do so. The planet is changing, and this change brings not only challenges for people, but also risks and opportunities for businesses.
Through the development of the RIBIA and RiskClima systems, E-On provides integrated solutions for assessing climate-related risks and opportunities linked to emissions and climate change. These tools support companies in their strategic planning for mitigating impacts and adapting to the future.
Article citation: Hausfather, Zeke & Drake, Henri & Abbott, Tristan & Schmidt, Gavin. (2020). Evaluating the Performance of Past Climate Model Projections. Geophysical Research Letters. 47. 10.1029/2019GL085378.