Bachelor-/Master Thesis (f/m/x) - Machine learning for ionosphere modelling

The Institute for Solar-Terrestrial Physics focuses on the properties and interactions of the coupled ionosphere-thermosphere-magnetosphere (ITM) system and how it is driven by energy inputs from the sun and the underlying atmosphere. Within this field of research, the Institute for Solar-Terrestrial Physics at the Neustrelitz site in Mecklenburg-Western Pomerania focuses on research into space weather. The term space weather refers to the time-varying conditions on the sun and in the solar wind and their effect on the ITM system.

What To Expect

The Department for Solar-Terrestrial Coupling Processes investigates the variability of the ionosphere and thermosphere using radar, satellite, and optical data as well as physical models. This work serves to improve our understanding of the numerous complex electrodynamic coupling processes that take place between neutral gas, charged particles, and the Earth's magnetic field. The goal is to improve existing models and develop new ones. In our department, it is possible to work with numerous observational data and the latest models, supported by experts in the field. The aim of the thesis is to use machine learning to derive a new global time-dependent model for the total electron content of the ionosphere from a multi-model linear regression (MMLR) data set.

Your tasks

• screening of the MMLR data set and removal of noise and artifacts
• selection of a modeling approach from the spectrum of machine learning methods
• tuning of hyperparameters with grid search and Tensorboard
• validation of the results against the MMLR coefficients and against data on the total electron content
• implementation of the model as a Python library
• evaluation of the results in written form (possibly oral presentation)

Your profile

• currently enrolled in a university degree program (bachelor's/master's) in natural sciences (e.g., physics) or engineering (e.g., computer science, geoinformatics) or other degree programs relevant to the position
• good programming skills in Python
• Basic scientific knowledge will help you to quickly and confidently find your way around the sometimes complex information.
• ability to acquire new knowledge quickly and independently
• ability to work both independently and in a team
• a high level of motivation to work on your own initiative
• good knowledge of English

We look forward to getting to know you!
If you have any questions about this position (Vac-ID 3789) please contact:
Claudia Borries
Tel.: