Cloud processes act on a whole range of scales, from submillimeter cloud droplet formation to large 1000 km cloud systems. Their evolution influences other processes over this scale range, all the way to planet scale. It is, for instance a large player, and at the same time, a large unknown in climate change.
One of the least understood cloud processes is drizzle. Drizzle formation modulates cloud properties and evolution, and affects the water cycle of the Earth. Since drizzle formation involves cloud droplets of all sizes, it requires extensive computational time. Hence, we often use simplified methods called parameterizations in weather and climate prediction models to obtain a bulk estimate of how fast and how many cloud droplets collide with each other or collide with bigger drops to form drizzle. However, many models continue to have inadequate representation of drizzle formation, calling for the need to improve these parameterizations.
In a coordinated effort with several researchers, lead by Christine Chiu, we have developed three 3 new parameterizations for autoconversion, the process by which two cloud droplets collide to form one rain (drizzle) drop, see this open access paper. To this end we combined aircraft measurements and results from a detailed bin-microphysics model that models the collision processes in clouds. Firstly, we have build an extremely accurate machine learned model for autoconversion and accretion process parameterization. We also used the extensive data set to derive new classic parameterizations where the autoconversion rate was assumed to have a power=law dependence on several cloud properties. Both methods found a strong dependence of the autoconversion process on the drizzle number concentration, a new finding that is missing in existing parameterizations. To understand this better, we developed new analytical expressions for autoconversion and accretion based on the stochastic collection equation.
This finding is unexpected, because the autoconversion process represents the coalescence between cloud droplets and is causally only related to cloud properties. However, drizzle number concentration does contain information on the width and evolution of the drop size distribution and hence indirectly on the autoconversion rate. In the past parameterizations have been proposed that include rain water concentration, but rain water is directly affected by the accretion process, while the rain drop concentration is not, and hence is mostly affected by autoconversion.