Implicit learning of convective organization explains precipitation stochasticity.

Accurate prediction of precipitation intensity is crucial for both human and natural systems, especially in a warming climate more prone to extreme precipitation. Yet, climate models fail to accurately predict precipitation intensity, particularly extremes. One missing piece of information in traditional climate model parameterizations is subgrid-scale cloud structure and organization, which affects precipitation intensity and stochasticity at coarse resolution. Here, using global storm-resolving simulations and machine learning, we show that, by implicitly learning subgrid organization, we can accurately predict precipitation variability and stochasticity with a low-dimensional set of latent variables. Using a neural network to parameterize coarse-grained precipitation, we find that the overall behavior of precipitation is reasonably predictable using large-scale quantities only; however, the neural network cannot predict the variability of precipitation (R2 ∼ 0.45) and underestimates precipitation extremes. The performance is significantly improved when the network is informed by our organization metric, correctly predicting precipitation extremes and spatial variability (R2 ∼ 0.9). The organization metric is implicitly learned by training the algorithm on a high-resolution precipitable water field, encoding the degree of subgrid organization. The organization metric shows large hysteresis, emphasizing the role of memory created by subgrid-scale structures. We demonstrate that this organization metric can be predicted as a simple memory process from information available at the previous time steps. These findings stress the role of organization and memory in accurate prediction of precipitation intensity and extremes and the necessity of parameterizing subgrid-scale convective organization in climate models to better project future changes of water cycle and extremes.

Significant Amplification of Instantaneous Extreme Precipitation With Convective Self-Aggregation.

This work explores the effect of convective self-aggregation on extreme rainfall intensities through an analysis at several stages of the cloud lifecycle. In addition to increases in 3-hourly extremes consistent with previous studies, we find that instantaneous rainrates increase significantly (+30%). We mainly focus on instantaneous extremes and, using a recent framework, relate their increase to increased precipitation efficiency: the local increase in relative humidity drives larger accretion efficiency and lower re-evaporation. An in-depth analysis based on an adapted scaling for precipitation extremes reveals that the dynamic contribution decreases (-25%) while the thermodynamic is slightly enhanced (+5%) with convective self-aggregation, leading to lower condensation rates. When the atmosphere is more organized into a moist convecting region and a dry convection-free region, deep convective updrafts are surrounded by a warmer environment which reduces convective instability and thus the dynamic contribution. The moister boundary-layer explains the positive thermodynamic contribution. The microphysic contribution is increased by +50% with aggregation. The latter is partly due to reduced evaporation of rain falling through a moister near-cloud environment, but also to the associated larger accretion efficiency. Thus, a potential change in convective organization regimes in a warming climate could lead to an evolution of tropical precipitation extremes significantly different than that expected from thermodynamical considerations. The relevance of self-aggregation to the real tropics is still debated. Improved fundamental understanding of self-aggregation, its sensitivity to warming and connection to precipitation extremes, is hence crucial to achieve accurate rainfall projections in a warming climate.