# More accurate climate predictions

Predicting the future climate is a grand mathematical challenge. Our climate is a chaotic dynamical system in which minor disturbances can have a major impact in the longterm. Researcher Keith Myerscough of CWI has developed new simulation methods to improve the accuracy of such long term predictions. He will defend his thesis on Monday, 24 August, at Utrecht University.

The chaotic behaviour means climate experts usually use a probabilistic approach, calculating the probability of certain scenarios by running their mathematical models several times with varying input. The climate is a balance of a huge amount of small-scale events such as air and sea currents, clouds, turbulence and pressure areas. It is impossible to correctly capture all these events on the microscale. This forces scientists to make modeling choices that inevitably affect the statistical properties of simuations.

Myerscough's work addresses this issue by improving the statistical accuracy of the mathematical models. He demonstrates the possibility of integrating models from both molecular dynamics and fluid physics to simulate atmosphere dynamics. This method is statistically more realistic than current models. Used in climate models, they will result in more accurate predictions.

This research is funded through the Free Competition programme of the Netherlands Organisation for Scientific Research (NWO).

*Image: Simulation of vortices in circulation, CWI*