ZMT Seminar Series: Beyond Modelling

Image of sky with storm clouds

 

BEYOND MODELLING invites talks with a mathematical or computer simulation modelling perspective and is open to a wide audience with an interest in modelling and its application to real-world phenomena.

With this seminar series we however strive to reach out beyond their immediate field to also address researchers working in observational or field work. Interdisciplinary and disciplinary modelling approaches are equally welcome.

We also welcome pure and abstract models, e.g., mathematical or computer models of intrinsic “beauty,” including highly simplified toy models. We naturally encourage a clear, or at least indirect, interpretation towards a real-world phenomenon. Topical fields of particular interest are the basics of the tropical climate and related coastal processes, such as ecological and societal dynamics and impacts.

 

The talks take place in person at the ZMT Seminar Room in Fahrenheitstr. 8


INVITED SPEAKERS (chronological order, latest talk first)

 

JANUARY 2023

January 18, 2023 | 11.00 - 12:00

Title: Diurnal warm layers and atmospheric convection in kilometer-scale coupled simulations

Radomyra Shevchenko

Speaker: Radomyra Shevchenko, Max-Planck-Institut für Meteorologie, Hamburg

Abstract:

The phenomenon of sea surface temperature (SST) anomalies created by oceanic diurnal warm layers has been extensively studied for the last decades, but the assessment of its importance for atmospheric convection has come within reach only very recently, thanks to the development of kilometre-scale simulations. We use the output of a global coupled simulation with a 5km horizontal grid spacing and near-surface ocean layers of order O(0.5m) to explicitly resolve both atmospheric convection and diurnal warm layers. As expected, the simulations produce daily SST fluctuations of up to several degrees. The increase of SST during the day causes an abrupt afternoon increase of atmospheric moisture due to enhanced latent heat flux. This increase is followed by an increase in cloud cover and cloud liquid water content. However, although the daily SST amplitude is exaggerated in comparison to reanalysis, the impact on cloud cover and cloud liquid water content only lasts for 5-6 hours. Moreover, the global daily average of these quantities is not influenced by their increase. All in all, we conclude that the global short-timescale impact of diurnal warm layers is negligible.


January 16, 2023 | 12.00 - 13:00

Title:Deep Learning the Colour Patterns of Coral Reef Fishes

Tsun Fung Yau

Speaker: Tsun Fung Yau | University of Freiburg

Abstract:
Animal colour patterns serve a variety of ecological functions related to e.g. signalling, camouflage, mimicry or defence against predation. Studies have long used colour patterns as a model system for understanding evolution since they provide exceptional access to phenotypic diversity and coral reef fishes display a rich diversity of colour patterns and are phylogenetically diverse, making them a unique group of targets for studying colour patterns. Nevertheless, the phenotypic diversity in reef fishes is not explainable by their genetic diversity and the evolution of colour patterns and their significance for diversification are still poorly understood. Colour pattern is a phenotypic trait that involves colour, geometry, morphology and vision systems. With the well-established techniques for encoding genotypes, we are now limited by our ability to encode this complex phenotypic trait objectively and quantitatively. In the commonly used colour pattern quantification methods, colour and geometry are not simultaneously addressed and colour pattern diversity is confounded with body shape diversity. We present here a workflow that disentangles colour pattern from body shape using morphometric methods and automatically extracts colour patterns across a broad taxonomic range. The workflow consists of two parts, (1) machine-assisted landmarking using community-sourced fish images and (2) colour pattern quantification using a deep perceptual metric. We show that using deep-learning-based techniques, colour and pattern can be concurrently assessed, allowing new possibilities in research that will enhance our understanding of the complex evolution of colour patterns.

DECEMBER 2022

December 13, 2022 | 14:00 - 15:00

Title: Mechanisms driving symbiont shuffling in corals

agostino..merico

Speaker: Prof. Dr. Agostino Merico, Head, WG Systems Ecology, ZMT

Abstract:
The success of scleractinian corals in oligotrophic waters of the tropics is the result of an endosymbiotic association with unicellular photoautotrophs called zooxanthellae (the symbiont). By providing photosynthetic products, the symbionts satisfy most of the corals’ metabolic needs. Increasing sea temperature causes the breakdown of this association, the expulsion of the symbionts and, in severe cases, the death of corals. Observational evidence suggests that the shift in the relative abundance of resident symbiont populations within the coral host (symbiont shuffling) enables corals to transcend their thermal tolerance limits because some symbionts are more thermally tolerant than others. The mechanisms of symbiont shuffling, however, are poorly understood. I will present here a new trait-based, acclimation dynamics model with which we show that classic competition theory can explain symbiont shuffling when the competitive abilities of different symbionts are driven by their thermal tolerances. We also show that rapid symbiont shuffling can occur (1) with the presence of a positive feedback, according to which some of the symbiotic benefits received by corals are re-allocated to symbiont growth, thus yielding even higher benefits, or (2) by keeping a background population size of non-dominant symbionts, which enables them to become abundant once conditions become favourable. Our results narrow down the mechanisms that could help reconcile the different patterns observed in symbiont population dynamics and provide new model theories that can be tested with laboratory experiments.



NOVEMBER 2022

November 17, 2022 | 15:00 - 16:00

Title: Impacts of convective treatment on tropical rainfall in ICON-NWP simulations from simple to complex configuration

hyunju jung scaled

Speaker: Hyunju Jung | Karlsruhe Institute of Technology, Germany

Abstract:
Tropical rainfall prediction is a challenge in current numerical weather prediction (NWP) models. To improve rainfall prediction in the tropics, it is essential to understand convective nature of the system and interactions of convection with large-scale features such as convectively coupled to equatorial waves which have potentials to improve rainfall predictions in the tropics as they control a substantial portion of rainfall variability in the tropics. Yet, NWP models often struggle to realistically produce convectively coupled equatorial waves. Resolved convection is thought to be key for the improvement by allowing for multi-scale interactions between convection and large-scale circulations. This presentation will discuss about mean tropical rainfall and its variability by assessing their representations in the ICON-NWP experiments between explicit versus parameterized convection. First, a set of ICON-NWP aquachannel simulations will be introduced, convective treatment of which comprises different combinations of deep and shallow convection, and we show time-mean behaviors of the aquachannel simulations, focusing on mean rainfall. To understand different behaviors among the simulations, we will present a novel diagnostic tool for a physically consistent comparison between simulations with different representations of convection. After demonstrating the time-mean behaviors of the simulations, equatorial waves are identified by using complementary wave identification methods. Finally, we will move onto realistic global ICON-NWP simulations and their ability of producing equatorial waves. These realistic simulations demonstrate that large-scale equatorial waves are fairly robust between explicit and parameterized convection and across various horizontal resolutions. This talk highlights our understanding of how different convective treatment alters mean rainfall and convectively coupled equatorial waves, and ultimately provides implications for forecast skills in the tropics.


November 15, 2022 | 14:00 - 15:00

Title: Latitudinal variation in heterotrophic nitrogen fixation associated with sinking marine particles (part II)

subhendu.chakraborty

Speaker: Dr. Subhendu Chakraborty, Senior Scientist, WG Systems Ecology, ZMT

Abstract:
Nitrogen (N2) fixation by heterotrophic bacteria associated with sinking marine particles is expected to have an important contribution to marine N cycle, but a mechanistic understanding of its regulation and significance is still unknown. We develop a mathematical model for unicellular heterotrophic bacteria growing on sinking marine particles and can fix N2 under suitable environmental conditions. We find that the interactive effects of polysaccharide and polypeptide concentrations in particles, sinking speed of particles, water temperature, and surrounding O2 and  concentrations determine the N2 fixation rate inside particles. Our model suggests that anaerobic processes, including heterotrophic N2 fixation, can take place in anoxic microenvironments inside sinking particles even in fully oxygenated marine waters. In contrary to photosynthetic N2 fixation that is more prevalent in low latitude waters, the spatial range of heterotrophic N2 fixation associated with sinking particles is more widespread. The modelled rates are similar to bulk rates measured in the aphotic ocean, and our study consequently suggests that particle-associated heterotrophic N2 fixation contributes significantly to oceanic N2 fixat.


November 10, 2022  | 10:30 - 11:30

Title: Optimal Trading Algorithms: A Machine Learning Approach

 Sandra Marion Kam Tsemo

Speaker: Sandra Marion Kam Tsemo, African Institute for Mathematical Sciences, Senegal

Abstract:
In this seminar, I will present a method that addresses the problem of finding optimal selling strategies for filling a large execution order. If the execution of a large order is not done efficiently, the resulting bad filling prices may have serious consequences on the market equilibrium. I will present an efficient selling strategy based on a trading algorithm that maximises the volume-weighted average price. The method involves Machine Learning tools such as Reinforcement Learning, Q-Learning, and Stochastic Optimal Control techniques. I will present numerical examples showing that the algorithm significantly improves trading performance and regularly meets or even exceeds market benchmarks. Although applied here to stock markets, Reinforcement Learning and Stochastic Optimisation can be used for addressing a variety of problems, including pure learning problems, dynamic resource allocation problems, or sequential decision problem during an epidemic.


November 1, 2022 | 14:00 - 15:00

Title: Latitudinal variation in heterotrophic nitrogen fixation associated with sinking marine particles

subhendu.chakraborty

Speaker: Dr. Subhendu Chakraborty, Senior Scientist, WG Systems Ecology, ZMT

Abstract:
Nitrogen (N2) fixation by heterotrophic bacteria associated with sinking marine particles is expected to have an important contribution to marine N cycle, but a mechanistic understanding of its regulation and significance is still unknown. We develop a mathematical model for unicellular heterotrophic bacteria growing on sinking marine particles and can fix N2 under suitable environmental conditions. We find that the interactive effects of polysaccharide and polypeptide concentrations in particles, sinking speed of particles, water temperature, and surrounding O2 and  concentrations determine the N2 fixation rate inside particles. Our model suggests that anaerobic processes, including heterotrophic N2 fixation, can take place in anoxic microenvironments inside sinking particles even in fully oxygenated marine waters. In contrary to photosynthetic N2 fixation that is more prevalent in low latitude waters, the spatial range of heterotrophic N2 fixation associated with sinking particles is more widespread. The modelled rates are similar to bulk rates measured in the aphotic ocean, and our study consequently suggests that particle-associated heterotrophic N2 fixation contributes significantly to oceanic N2 fixation.


OCTOBER 2022 

October 27, 2022 | 14:00 - 15:00

Title: The atmosphere as a driver for diversity

Miguel Garrido Zornoza

Speaker: Miguel Garrido Zornoza, Niels Bohr Institute, Copenhagen University

Abstract:
Spatial structure has been shown to play an important role in the ecology of different modes of interaction, such as predation or competition. However, most studies focus on micrometer-sized structures such as the impact of bacterial colony formation upon predation by (bacterio)phages. We here present a numerical investigation of the ecological impact of the atmosphere on these modes from a population dynamics approach combined with a simple lattice model of the atmosphere. In our model, advection, molecular diffusion and aerosolization and deposition fluxes are combined with system-specific interactions (that from a bacteria-phage predator system) to suggest their possible role in patterning formation and steady-state population densities. Preliminary results point towards a discrepancy from the mean-field approximation, and to advection as the driver of spatial structure in population density.