PhD Defense by Irene L. Kruse

Title: Chasing the Storms: A Simulation and Observation-Based Exploration of Mesoscale Convective Systems and Cold Pools, from the Midlatitudes to the Tropics

Abstract: Convective storms. We have most probably all experienced them. In Northern Europe, sporadically in the summer season. In the tropics, several times a week in the rainy season. The most organized storms highly contribute to extreme rainfall, both in Europe and in the tropics. The frequency of these very storms is increasing with a warmer climate, making the mechanistic understanding of the upscale growth of a storm, and the predictability of its onset essential today.

In this dissertation we explore the effect of a large diurnal cycle in surface temperatures, typical of tropical land, on organizing small convective clouds into large storms - mesoscale convective systems (MCSs) - and how these MCSs can in turn organize the surrounding moisture field into a breeding ground for even more organized storms, that can persist in less favorable environments, like over the ocean, hinting at a path to cyclone formation. We then focus on cold pools (CPs), a defining feature of precipitating convective clouds. CPs are both a key ingredient in the organization of these into larger systems, and an easy-to-interpret measure of storms from the ground. After learning key properties of real-world CPs by analyzing measurements over 10 years from a weather tower in Northern Europe, we turn to the practicality of portable automatic weather stations in regions of the world with less weather monitoring capacity. We set these up in two locations in Senegal, and develop an AI-based nowcasting tool with long short-term memory neural networks (LSTMs), trained to predict the advent of CPs in the city of Dakar.

With the combination of observation and simulations, we thus dive into the core of MCSs and their associated CPs, and turn our findings into an application useful to society in improving convective weather forecasting.