Research about dynamic compensation of convective mass flux

The schematic vertical cross sections for HYMACS are depicted as follows: (a) the scale of a convective updraft and a mass-compensating environment compared with the size of grid columns; (b) how subgrid-scale mass convergence or divergence is treated in a grid column; and (c) how grid-scale dynamics respond to the subgrid-scale mass source and sink. (From Ong et al. 2017)


In atmospheric convection, convective drafts bring air mass up or down, and a wider environment brings air mass back in compensation. All these processes occur within a ~100-km-wide model grid cell, so the models cannot resolve convection and instead parameterize the bulk effects of both convective drafts and the environment. However, with increasing computer resources, models can run with ~10-km-wide grid cells, wider than convective drafts but narrower than the environment. Is there any consequences of imposing environmental compensating mass flux within a ~10-km-wide grid cell?


Instead of imposing local compensation of convective mass flux, a hybrid mass flux cumulus scheme (HYMACS) dynamically resolves mass compensation, making the dynamic response to a prescribed mass lifting to change less with changes in grid-cell width.

Broader Impact

My master thesis points out a problem and a remedy in representing convection in weather and climate models when computer resources allow finer model grids than in tradition, which may lead to model development that improves prediction, which improves our lives.

Project Status

I have switched to other projects while other people keep going, e.g., see Cited by


Ong, H., Wu, C. M., & Kuo, H. C. (2017). Effects of artificial local compensation of convective mass flux in the cumulus parameterization. J. Adv. Model. Earth Syst., 9(4), 1811-1827.