Aktuelle, konkrete Themenvorschläge

Ecological Modelling Projects

• You will develop a mechanistic simulation model that includes horizontal and vertical soil water movement as well as explicit plant individuals with root dynamics. The project builds on previous modelling work in our group and extends it step by step toward an integrated system that can be used to test ecological hypotheses about pattern formation.
• This is a research-focused modelling project with a substantial programming component (C++ or similar; optional R interface). A basic understanding of soil properties and soild water dynamics is required.
• Primarily suited for MSc theses; advanced BSc projects are possible with adjusted scope.

Students join the ecological modelling group, participate in weekly meetings, and can use workspace in the lab.

Interested or curious? Contact Sebastian: shanss@uni-goettingen.de

## **Topic:** Development of a Modular Savanna Ecosystem Model Savanna ecosystems are shaped by tree–grass interactions, with vegetation dynamics primarily driven by rainfall. Understanding these processes requires simulation models that are both ecologically realistic and technically robust. In this project, you will develop a modular savanna simulation model focusing on vegetation–rainfall interactions. The aim is to build a clean, extensible C++ backend that can serve as a foundation for future ecological modelling projects. The thesis combines ecological modelling with substantial software development (C++, optional R interface). Emphasis is placed on modular architecture, code quality, testing, and documentation. The model will follow pattern-oriented modelling principles to achieve structural realism while remaining flexible for future research extensions. The project requires programming with C++. Experience with R is optional. Advanced tools such as Git, unit testing, and CI workflows can be learned during the project. The scope can be tailored for internship, BSc, or MSc level. *Students join the ecological modelling group, participate in weekly meetings, and can use workspace in the lab. Interested or curious? Contact Sebastian: shanss@uni-goettingen.de* ## **Topic:** Development of a Spatio-Temporal Rainfall Simulation R Package for Semi-Arid Ecosystems Rainfall in semi-arid ecosystems is highly variable in both space and time and strongly influences vegetation dynamics. This project reimplements and extends established stochastic rainfall models (temporal Markov-chain model and spatial SERGE model) as a reusable R package with a compiled backend (C++, for example). The resulting package will support standalone temporal and spatial simulation as well as combined spatio-temporal rainfall generation, providing tools that can be directly integrated into ecological modelling studies. The project includes a strong programming component (R + C++). It is suitable for internship, BSc, or MSc level. *Students join the ecological modelling group, participate in weekly meetings, and can use workspace in the lab. Interested or curious? Contact Sebastian: shanss@uni-goettingen.de* # :jigsaw: Research Software & Modelling Infrastructure Projects ## **Topic:** Extending NLRX for Generic Simulation Models in R Simulation experiments often require systematic exploration of large parameter spaces. NLRX is an R package designed for this purpose, but it is currently closely tied to NetLogo models. This project aims to generalise NLRX so it can support arbitrary simulation models callable from R. You will design and implement an abstraction layer that separates experiment design from model-specific execution. A toy ecological simulation model will serve as a proof-of-concept, and you will contribute to extending the package documentation. The result will strengthen reproducible and scalable simulation workflows in ecological modelling. Substantial programming in R is expected; advanced R programming and package development skills can be developed during the project. The scope can be tailored for internship, BSc, or MSc level. *Students join the ecological modelling group, participate in weekly meetings, and can use workspace in the lab. Interested or curious? Contact Sebastian: shanss@uni-goettingen.de* ## **Topic:** Performance Optimisation of NLRX Using the NetLogo Java API Large simulation experiments can involve thousands of model runs, making performance a critical factor. NLRX currently executes NetLogo simulations via file-based I/O, which limits efficiency. This project explores direct integration via the NetLogo Java API to significantly improve execution speed. You will prototype Java API integration for NLRX and benchmark performance gains across different simulation scenarios, contributing to faster and more scalable modelling workflows. A strong interest in programming is required. R knowledge is expected; Java experience is helpful but can be acquired during the project. The scope can be tailored for internship, BSc, or MSc level. *Students join the ecological modelling group, participate in weekly meetings, and can use workspace in the lab. Interested or curious? Contact Sebastian: shanss@uni-goettingen.de* ## **Topic:** Developing an Agent-Based Modelling Framework Beyond NetLogo NetLogo is widely used for teaching agent-based modelling because it is accessible, but it has limitations in scalability and software tooling. This project explores whether a modern workflow can combine NetLogo-like accessibility with improved scalability for research applications. A minimum goal is to explore existing approaches for ABMs in R and reimplement the classic wolf–sheep–grass model using your prototype framework — for example as an R package with a modular backend (preferably C++) and graphical output via tools such as Shiny or Quarto. The project connects teaching usability with research-grade modelling workflows. Substantial programming is expected. The scope can be tailored for internship, BSc, or MSc level. *Students join the ecological modelling group, participate in weekly meetings, and can use workspace in the lab. Interested or curious? Contact Sebastian: shanss@uni-goettingen.de*

Im Projekt "Management effects on wildfire risk and degradation of ecosystem services by eastern redcedar encroachment into the private sector of the Cross Timbers: A modeling and simulation approach" können Bachelor- und Masterarbeiten in den Bereichen

• Remote sensing
• Machine learning and neural networks using remote sensing data
• Agent-based modeling, point pattern analysis, and statistics

angefertigt werden. Mögliche Themen:

• Tree identification and segmentation
• Allometric equations
• Fire susceptibility of vegetation
• Vegetation dynamics

Kontakt: José Ortega (josemaria.ortegaballadares*uni-goettingen.de)
Details: PDF

Allgemeines zu Master- und Bachelorarbeitsthemen