Master Defense by Reyk Börner

Title: Modeling diurnal sea surface warming in the tropical ocean -- An interactive boundary condition for idealized atmospheric simulations

Abstract: One of the key challenges in climate science today is to understand how clouds organize. An accurate description of cloud processes is important for predicting extreme weather events, calculating the global energy balance, and reducing uncertainties in climate models. Over the tropical ocean, observations evidence that atmospheric dynamics are coupled to the ocean. This air-sea interaction is regulated by the variability of sea surface temperature (SST), which can oscillate with amplitudes of 3°C and more between day and night. In contrast, idealized numerical studies of the tropical atmosphere often assume a constant, uniform SST. Recent simulations with imposed SST variations – such as a diurnal cycle – show different patterns of cloud clustering compared to the constant SST case. These findings suggest the need for more realistic representations of ocean-atmosphere interaction. Consequently, several studies have represented the upper ocean as a responsive slab with a fixed heat capacity. However, we argue that such slab models may be insufficient because, by ignoring wind-induced water mixing, they neglect a crucial wind effect on sea surface warming.

Here we present an idealized, one-dimensional model of heat transfer in the upper few meters of the ocean, forced by insolation and atmospheric conditions. Unlike slab oceans, our model includes important processes such as wind-driven turbulent diffusion, near-surface heat trapping, and skin cooling, while retaining conceptual simplicity. Using a combination of ship measurements and buoy data, we estimate the model parameters by Bayesian inference and show that our model reproduces key features of observed diurnal warming. Aiming to be versatile, computationally affordable and easy to code, the model is suited as an interactive surface boundary condition for high-resolution cloud-resolving simulations. Thus, our work could support a next step towards more realism in idealized simulations of the tropical atmosphere, seeking to help build the bridge between conceptual modeling and real-world process understanding.

 Supervisors: Jan O. Haerter, Romain Fiévet, and Peter Ditlevsen