Palaeoclimate and Climate Dynamics - Research

Research Overview

We investigate synoptic scale atmospheric dynamics at different times in Earth's history and model Earth surface responses to climate change for the past, present and future. More specifically, we employ climate models to simulate climates that are in equilibrium with reconstructed palaeoenvironmental and orbital forcings at different times in the Late Cenozoic. This global climate modelling is complemented by the application of dynamic statistical models and machine learning techniques to capture the links between synoptic scale atmospheric phenomena, such as the Antarctic Oscillation, and meso- to microscale responses of atmospheric variables, mountain glaciers, erosional processes and topography. Our research allows us to work on a wide range of problems - from constraining the climate and surface uplift history of the Alps to predicting glacier and local atmospheric responses to contemporary climate change.

Research Topics


Reconstructing and understanding the different climates in Earth’s history allows scientists to put geological records, palaeofauna and -flora, and early human migration in a climatic context, which in turn allows us to understand underlying mechanisms our observations. Climate models, or General Circulation Models (GCM‘s), simulate global climate based on our physical understanding of atmospheric processes. They are primarily used to investigate atmospheric dynamics and contemporary climate change, but have also been applied to improve our understanding of past climates and Earth system dynamics. We use GCM’s in tandem with proxy-based palaeoenvironmental reconstructions to simulate Late Cenozoic climate.

Climate - Tectonics Interactions


The atmosphere is linked to tectonics in several ways. Mountain building affects climate by blocking air flow and changing temperature and pressure gradients. In turn, climate affects uplift through climate-driven erosional processes. Furthermore, atmospheric science and an investigation of past climates can benefit the geological community by reconstructing and quantifying the climatic context for geological processes. For example, oxygen isotopes are used in palaeoaltimetry to estimate the height of mountains at specific times in geological history. However, differences in climate can produce similar magnitudes of change of isotopic ratios (Fig. 1). Quantifying the climatically induced isotopic changes therefore allows for isotopic signal separation, which ultimately leads to more robust palaeoelevation estimates. This is what the research project APE (Apline Palaeoclimate/-elevation Experiment)  is currently attempting.

Climate - Earth Surface Interactions

Climate interacts with the Earth’s surface in several ways: It directly shapes the surface by river incision and hill slope processes (dependent on precipitation amount and characteristics), frost cracking (dependent on temperature and water availability) and aeolian (wind induced) erosion; it indirectly shapes the Earth’s surface by modifying vegetation cover and type, which in turn affects surface erosion. Detecting patterns in erosion-relevant climate change through geological time helps provide climatic context for erosion rates calculated from geological archives; machine learning techniques applied to palaeoclimate simulations can help identify and explain these patterns.


Climate - Glacier Interactions

The research of climate-ice interaction focuses on mountain glaciers in Chile and Norway. Mountain glaciers not only shape the landscape via glacial erosion, but also serve as a water reservoir for local communities, and as a renewable energy source. A better understanding of climate-glacier interactions therefore benefits different scientific communities, as well as policy makers. In Norway, for example, melt water flow contributes ca. 15% to the hydroelectric power generation, which makes up more than 95% of the total power production in the country.

Empirical Statistical Downscaling

Empirical Statistical Downscaling (ESD) is an umbrella term for methods that statistically relate local expressions of the climate system, such as temperature or precipitation recorded at weather stations, to larger-scale atmospheric conditions, such as the Antarctic Oscillation (AAO) and El Niño/Southern Oscillation. For example, coarse climate model output with location-dependent systematic errors in rainfall may be fitted to local-scale observational data of rainfall. The resulting statistical model may then use these large-scale conditions as predictors for local-scale rainfall. This method is well established in the climatological and meteorological community. It is (computationally) less expensive than dynamical downscaling based on regional climate models, and takes in-situ geographical realities, such as local topography, into consideration without the need to explicitly parameterise these.

While dynamical downscaling over South America has significantly advanced in the past two decades, the ESD community in South America is incipient and the potential for ESD in the region. We explore this potential with the application of classic [e.g. Fig.1; Mutz et al., 2021] and innovative approaches to ESD.

Research Methods

Numerical Modelling

We use a wide range of numerical and analytical models to simulate climate and understand various mechanisms in the climate system. These range from wind trajectory models to full-fledged Atmospheric General Circulation Models (AGCMs).

Statistics and Machine Learning

While models based on our understanding of atmospheric physics give us insights into the workings of the atmosphere, statistical models and machine learning techniques let us objectively detect and describe patterns in our data and allow us to model processes empirically to complement our numerical modelling. The empirical tools we use range from classic cross-validated multiple regression procedures, as used in weather forecast and statistical downscaling, to the application of efficient clustering techniques and artificial neural networks.

High Performance Computing

The sophisticated atmospheric models and machine learning algorithms require exceptional computational facilities. We use a high-performance Linux cluster computer that is owned by the Earth System Dynamics research group and maintained by the computing centre of the University of Tübingen. Additionally, we have access to the flagship supercomputer Mistral of the German climate computing centre (Deutsches Klimarechenzentrum, DKRZ).

Research Projects

REAL – Reconstructing Eastward Propagation of Surface Uplift in the Alps

The goal of the REAL project is to refine the uplift history of the Alps by integrating stable isotope palaeoaltimetry and palaeoclimate modelling in a novel and interdisciplinary approach to palaeoaltimetry. REAL is part of the DFG (Deutsche Forschungsgemeinschaft) priority programme MB-4D ("Mountain Building Processes in 4D") and aims to refine the uplift history of the European Alps by investigating the question of asynchronous surface uplift.

MB-4D (external link)

EarthShape – Earth Surface Shaping by Biota

EarthShape is a priority research programme (Schwerpunktprogramm, SPP) funded by the German Research Foundation (DFG). The German-Chilean research initiative explores how biologic processes (in addition to climate and tectonics) shape the Earth surface and modulate the impact of climate change on the Earth surface.
EarthShape (external link)

Q-TiP – Tipping points of lake systems in the arid zone of Central Asia

Q-TiP is a project funded by the German Ministry for Education and Research (BMBF) and investigates the climatic controls on water resources in Central Asia with regard to the development of arid environments and their relevance for scenarios of ongoing and potential future climate change.
Q-TiP (external link)

APE – Alpine Paleoelevation/-climate Experiment

APE – Alpine Paleoelevation/-climate Experiment (ongoing, funded)

APE is part of the DFG (Deutsche Forschungsgemeinschaft) priority programme MB-4D („Mountain Building Processes in 4D“) and aims to link Earth surface processes to lithospheric dynamics with the aid of palaeoclimate simulations and stable isotope altimetry.
MB-4D (external link)