How do the ups control the downs within deep convective clouds?

In convective storms, the air that goes down is almost as important as the air that goes up. These convective downdrafts influence the termination of updrafts, near surface air properties, surface winds. However, predicting downdrafts under specific atmospheric conditions is difficult because their properties depend on several interrelated processes. In a new study led by Dr. Julia Windmiller, scientists from the Max Planck Institute for Meteorology and the Australian Research Council Centre of Excellence for Climate Extremes used the new capabilities of global storm resolving models to investigate what controls the properties of downdrafts over tropical oceans globally. They show that, primarily due to the preceding updraft and associated precipitation, up to 75% of the mass flux and 37% of the velocity variations of downdrafts can be predicted.

Evaporation of condensate from a cloud can cause air to become negatively buoyant and sink. This downward movement of air is called a downdraft. By injecting air from higher levels into the atmospheric boundary layer, downdrafts can modify the atmospheric conditions in the boundary layer. Downdrafts can also end the life cycle of a cloud and result in strong, sometimes destructive, wind gusts at the surface. The basic driving forces for downdrafts are well understood. For example, we know that evaporation of rain and the associated cooling of air is usually critical in causing the air to sink. Even though the basic driving forces are known, many interrelated processes contribute simultaneously to the strength of the downdraft, making it difficult to predict the strength of a downdraft under specific conditions. This prediction is, however, required to strengthen our conceptual understanding of convective systems.

Given the central role of downdrafts, Windmiller et al. improve our understanding of how best to predict downdraft properties as a function of updraft and environmental properties. Previous studies have investigated this question using simplified theoretical models or high-resolution but small-domain atmospheric simulations. This study addresses this question using an atmospheric simulation whose model domain spans the globe but still explicitly resolves rain-forming clouds. Focusing on convection over tropical oceans, the authors study 30 000 rain-forming clouds, and their associated downdrafts, under very different, naturally varying environmental conditions. The resulting average time evolution of the precipitation rate at the surface as well as the profiles of rain water, updraft and downdraft mass flux are shown in Fig. 1.

Machine learning techniques and traditional statistical methods agree on the result that the strength of the downdraft can be well predicted if the strength of the updraft that caused the downdraft or, even better, if the amount of rain that an updraft produced is known. Surprisingly, the authors find that downdrafts can be predicted only slightly better if other environmental conditions of the air surrounding the downdraft are known, such as the temperature and/or humidity profiles. The small importance of environmental humidity is surprising, as the evaporation rate is higher under drier conditions, which should thus increase the strength of the downdraft. One possible explanation for this result is that the mean humidity on larger scales is not a good indicator of the relative humidity of the air within which the condensate evaporates. Another explanation is that, also considering the impact of environmental humidity on the starting properties of convective downdraft, a number of competing effects exist which work in opposite direction for different convective events and tend to balance in general.

Original publication

Windmiller, J., Bao, J., Sherwood, S., Schanzer, T. & Fuchs, D. (2023). Predicting convective downdrafts from updrafts and environmental conditions in a global storm resolving simulation. Journal of Advances in Modeling Earth Systems, 15: e2022MS003048. doi:10.1029/2022MS003048.


Dr. Julia Windmiller
Max-Planck-Institut für Meteorologie
julia.windmiller@we dont want