Atmospheric Physics Building,Department of Physics, University of Oxford, Parks Road, Oxford, OX1 3PU
Dr Adrian Tompkins, International Centre for Theoretical Physics (ICTP)
Andrea Simpson (andrea.simpson@physics.ox.ac.uk)
Abstract
Over 30 years ago, it was discovered that deep convection in idealized simulations of radiative-convective equilibrium (RCE) using convection permitting models could undergo a spontaneous transition from a random state to a highly organised state in which convection is clustered in one part of the domain. In the intervening period, we have understood how this clustering can affect the tropical energy budget and how diabatic forcing, in particular cloud-long wave radiative forcing, can lead to clustering. However, the details and onset characteristic of clustering remain highly model-specific. In this talk, we will show how the characteristics of the bistable random/clustered states can be reproduced in a simple stochastic toy model, and use this to derive a dimensionless parameter that predicts whether a given model will cluster. We will then move to examining aspects of organisation in tropical observations, showing how both clustered and random type situations can be found on the mesoscale in the warm pool region, and how these states relate to tropical wave dynamics. We will argue that understanding convection organisation on the mesoscale is important for understanding basin-wide tropical energy budgets.