A Machine learning assisted Cloud Population Model as a Parameterization of Cumulus Convection

Abstract:

A machine learning assisted cloud population model is coupled with the Advanced Research Weather Research and Forecasting (WRF) model to represent non-equilibrium fluctuations in cloud base mass flux associated with the life cycles and interactions among cumulus convection cells. In this cloud population model, the size distribution and associated cloud base mass flux of the convective cells are related to their previous state and the change in the convective area via a transition function. The convective area tendency in turn is assumed to depend on the convective mass flux tendency resolved by the course grid (25km) of the host model (WRF). The transition function is represented by a single hidden layer neural network trained by the evolution of convective cell size distributions in a 1 km grid-spacing WRF simulation run over the Australian Monsoon region. At every grid point of the host model, the cloud population model continuously predicts the cell size and cloud base mass flux distributions which are then fed to an entraining parcel model that calculates temperature, moisture and hydrometeor tendencies. These tendencies are averaged over the cells and provided to the host model. Several regional simulations are performed over tropical and mid-latitude domains to test this approach. The performance of the model in capturing key features such as the diurnal cycle of convection over land and ocean, meso-scale convective systems and the propagation of the Madden-Julian Oscillation are discussed.