Can we predict the weather? New tools for an old problem

Today’s weather forecast is based on a simulation of established laws of physics and mathematics. Despite important progress in computing power, observations, and mathematics, predicting the weather reliably for more than a few days has always been a major challenge. In the late 1950s and 1960s, Lewis Fry Richardson, and later, Edward Lorenz, set up the basis for the understanding of this challenge, using observations, phenomenological arguments, and low-order models. Here, we first summarize their findings, then discuss some of the recent advances in this domain, made possible through the advent of new numerical and theoretical tools.