Climate Modeling

Model development and evaluation, simulation setup and execution, application of machine learning methods to climate science

Climate models represent the climate system’s components and mechanisms numerically. They rest upon physical principles and approximations for hardly resolvable process chains (“parametrizations“). As such, they provide a laboratory for investigating climate at all times during Earth’s past and future.

We employ models of varying complexity to analyze the climate system on various scales in time and space. Our model development focuses on including new processes or modeling processes in more detail, especially for simulating land use and isotopes. These developments help us improve the quality of simulations since they implement processes that are crucial for future projections (land use, carbon dioxide removal) and because they provide a direct way of comparing simulations of climate’s past to past climate data. We carry out experiments simulating conditions at varying times in Earth’s history and future as well as idealized experiments aimed at furthering the understanding of the Earth System.

The variety of topics that can be explored with models is immense. As such, they contribute to all other research areas in our group.

Model-specific research questions

  • How does the simulated climate depend on the complexity of models? What is the necessary and sufficient complexity to simulate various processes in the climate system?

  • How sensitive is the climate system to different external forcings and feedbacks?

  • How does simulated variability depend on the mean climate state?

  • How can we improve climate models to better agree with past climate data and to sufficiently simulate the future climate trajectory in the coupled Earth-Anthroposphere system.


On-going work, methods & modeling tools

  • Climate models of varying complexity from energy balance models (e.g. TransEBM [1]), models of intermediate complexity (e.g. Plasim [2]) to Earth System Models (e.g. MPI-ESM)

  • Isotope-enabled models [3]

  • Agent-based modeling of land use decision-making

  • Land surface modeling

  • Hydroclimate simulations, i.e. monsoon systems

  • Transient simulations of the Last Deglaciation

  • Impact modeling

  • Modeling of sustainable technologies [4]

Thesis projects & collaboration

Thesis project ideas that could be explored are

  • improving (future) climate assessments based on model intercomparison,

  • sea ice dynamics [2],

  • effect of resolution on modeled climate,

  • effects of climate variations on attribution studies,

  • modeling impacts & damages of climate change,

  • modeling of climate extreme,

  • simulating water isotopes [3],

but many more options are possible. If you are interested in modeling projects, contact us as described here. Having some experience with programming, statistics, data analysis, and/or climate physics is helpful but not necessary.

We are looking forward to learning about your ideas!

You might also want to check out our team members’ work on modeling impacts and the carbon cycle, understanding climate's variability across time scales and atmospheric dynamics. 


[1] Ziegler, E., & Rehfeld, K. (2021). TransEBM v. 1.0: Description, tuning, and validation of a transient model of the Earth’s energy balance in two dimensions. Geoscientific Model Development, 14(5), 2843–2866.

[2] Adam, M., Andres, H. J. & Rehfeld, K. The role of dynamic sea ice in a simplified general circulation model used for palaeoclimate studies. in Book of Extended Abstracts of the 6th ECCOMAS Young Investigators Conference 386–395 (2021).

[3] Bühler, J. C., Roesch, C., Kirschner, M., Sime, L., Holloway, M. D., & Rehfeld, K. (2021). Comparison of the oxygen isotope signatures in speleothem records and iHadCM3 model simulations for the last millennium. Climate of the Past, 17(3), 985–1004.

[4] May, M. M. & Rehfeld, K. Negative Emissions as the New Frontier of Photoelectrochemical CO2 Reduction.Advanced Energy Materials 2103801 (2022) doi:10.1002/aenm.202103801