Environmental & Climate Sciences Department Seminar

Global Climate Models (GCMs) are depicted as the holy grail of climate science. However, they are far from reliably simulating the current and future climate. One of the components that represents a source of errors and biases in these models is the convective parametrization scheme and the ad-hoc treatment of deep convection and entrainment. In this talk, I will focus on two tropical atmospheric events that GCMs struggle to simulate: Amazonian deep precipitating convection, and the Madden Julian Oscillations. I will present how idealized cloud-resolving models (CRMs) coupled with simulated and forced large-scale circulation can capture the essential dynamics of these events. In particular, when diagnosed from the CRMs, I will show that entrainment is merely a response to the convective regime and cannot have a constant rate as GCMs suggest. If time allows, I will also point out to another potential source of biases in the simulated mean climate stemming from the representation of numerical noise damping in the model dynamical core.