Student projects

Bachelor and Master students

Michael Debertshäuser (Predictability of Föhn events)

Daniel Köhler (Tropopause inversion layer and humidity)

Peter Neuroth (Hydrodynamic instabilities)

Moritz Rehage (Cloud detection)

Anton Seyfert (Error propagation)

Tassilo Wagner (Deep Learning Physical Simulations)

 


Available student projects

Several students projects are available in BINARY, it is anticipated that the projects will be embedded into the main work of the overall project.

If you are interested in a Bachelor or Master project (Atmospheric Physics or Computer Sciences), please have a look on the list of projects and contact directly the responsible PI. If there is no suitable project, you could also contact us directly via binary@uni-mainz.de. It is also often possible to shift themes dedicated to Master projects to Bachelor projects and vice versa.

updated on 13 October 2021


List of projects:

Classification of cloud structures

Clouds consist of myriads of small particles. However, on larger scales pattern are formed via the interaction of different processes in an emergent way. It is not clear how this structure formation process really works, but obviously the environmental conditions (e.g. winds, temperature, humidity, aerosols) have a strong impact on the formation of different pattern as can be seen in observations. In this project, cloud structures as seen from satellite should be classified. Using machine learning algorithms, a large data set of satellite images (e.g. Meteosat) will be investigated in order to identify different classes of cloud pattern automatically. The results should be investigated statistically in order to obtain a first estimate about the frequency of occurrence of distinct cloud pattern. In a second step, the identified cloud structures should be combined with meteorological data sets (reanalysis data from ECMWF) in order to provide information about preferential conditions for different cloud pattern.

Contact: Prof. Peter Spichtinger

Target group: MSc Computer Sciences/Atmospheric Physics

 

Determination of radiative forcing for ice clouds using machine learning

Clouds crucially influence the energy budget of the Earth-Atmosphere system. They partly reflect or scatter incoming solar radiation back to space, thus less energy as compared to clear sky scenarios is transferred into the system (cooling, albedo effect). On the other hand, infrared radiation as emitted from Earth’s surface is partly absorbed by clouds and re-emitted, leading to warming of the system (greenhouse effect). For ice clouds, both effects are of comparable size. Thus, the resulting net effect depends on some parameters characterizing the environmental conditions. In this project, radiative transfer calculations using a very simple and idealized setup should be carried out for an ensemble of several parameter settings. From the results a new but simple model should be developed using machine learning techniques to describe the net radiative effect of ice clouds as a function of parameters. This new model then can be used for first estimations of the radiative effect of ice clouds for similar environmental settings.

Contact: Prof. Peter Spichtinger

Target group: MSc Computer Sciences/Atmospheric Physics

 

Detection of convective clouds from geostationary satellite data

(Erkennung von konvektiven Wolken anhand geostationärer Satellitendaten)

Based on the MSG cloud mask or other properties (Cloud top temperature, Cloud optical thickness), the task would be to identify convective cloud systems with various criteria, e.g. cloud cluster size, time evolution of the cloud field (dissipation time), based on a manual detection scheme.
Additionally, the convective clouds / cloud clusters should be detected with an e.g. gradient based detection algorithm. Once a detection algorithm is functional, the organisation of convection with the help of organisation indices should be described and analysed.
The combination of these approaches could build the foundation for an improved machine learning algorithm.

Contact: Prof. Holger Tost

Target group: BSc Atmospheric Physics/Computer Sciences

 

Organisation of convective clouds based on superparameterised model output

(Analyse des Organisationsgrades konvektiver Wolken basierend auf Modelloutput)

Model simulations applying a convective superparameterisation can provide detailed global information about small scale features of convection. The project aims at analysing the degree of organisation of convection with typical organisation indices across the CRM cells within one GCM grid cell. This should be performed for various regions and seasons. A key point of the analysis is whether the degree of organisation is consistent across the GCM gridcells.

Contact: Prof. Holger Tost

Target group: BSc Atmospheric Physics/Computer Sciences

 

Simulations of organised convection with a cloud resolving model system

This project will develop a scenario with a cloud resolving model (CM1 or SAM) in which the organisation of convection can be simulated. With the help of sensitivity simulations, conditions under which organisation occurs will be determined, such as surface temperatures, evaporation flux strength, details of cloud microphysics, etc.

Contact: Prof. Holger Tost

Target group: MSc Atmospheric Physics/Computer Sciences

 

Online diagnostics of convective organisation using a superparameterised GCM

Within this project the candidate is expected to develop online-diagnostics for convective organisation of convective events simulated with a superparameterised GCM. For this purpose, these diagnostics have to be implemented into the superparameterisation and simulations analysing the degree of convective organisation shall be performed. The degree of organisation should be compared to observations from the satellite platforms.

Contact: Prof. Holger Tost

Target group: MSc Atmospheric Physics/Computer Sciences

 

Comparison of high resolution forecast and reanalysis data from ECMWF

Within BINARY ERA5 reanalysis data will be compared with high resolution measurement data. The focus will be on the identification of regions close to the tropopause which show signs of turbulence. However, the ERA5 data is available on a three dimensional spatial grid which has a lower grid point density in the horizontal than the ECMWF forecast. The differences in each horizontal direction are on the order of a factor three. Given these differences it can be assumed that turbulence might be resolved differently in the two ECMWF products. However, albeit its lower resolution the ERA5 data is preferred, due to the long coherent data set. This work should identify similarities and differences between common measures of turbulence in ERA5 and the forecasts.

Question: Do ERA5 and ECMWF forecast show differences in regions of turbulence?

Contact: Dr. Daniel Kunkel

Target group: BSc/MSc Atmospheric Physics

Data: ECMWF ERA5 and forecasts

 

Interpolation between modeled and observed quantities along flight tracks

An important task while comparing model and measurement data is to combine the information from the coarsely gridded model data to the high resolution (but limited) measurement data. A common way is to interpolate the model data in space and time onto the flight track. However, often differences arise simply to the fact that the model data is available on different grids, in particular varying in the vertical. Moreover, various measured parameters can be used for the interpolation, e.g., pressure, potential temperature or geometric height. This project should compare various interpolation methods, in particular in order to correctly interpret high resolution data in turbulent situations.

Question: Which differences arise from different interpolation methods between model and measurement data?

Contact: Dr. Daniel Kunkel

Target group: BSc/MSc Atmospheric Physics/Computer Sciences

Data: various ERA5 data products; airborne measurement data

 

Signal-to-noise ratio in highly resolved in situ measurement data

Noise is a common issue of high resolution measurement data. In fact, noise can become larger than atmospheric variability of target quantitiues and measurements become unreliable. Methods applying optical spectroscopy often suffer from statistical noise but the measurement signal can also be affected by optical interferences and fringes of varying frequencies. Under certain circumstances such fringes can be identified and eliminated to enhance the signal-to-noise ratio. This allows then to have more reliable data for example in the vicinity of turbulence. This work will therefore use the raw signal of a spectrometer which is initially affected by potential noise. Spectral filter, wavelet analysis, or principle component analysis are potential tools to detect and eliminate or account for the noise.

Question: How can the signal-to-noise ratio be increased in optically derived high resolution measurement data?

Contact: Dr. Daniel Kunkel

Target group: BSc/MSc Atmospheric Physics, Mathematics, Physics

Data: spectral airborne measurement data

 

Optimized method to calculate mixing ratios from spectral measurement data

Commonly, a raw signal from an optical spectrometer needs to be converted to mixing ratios of trace species for further analysis steps. On the basis of theoretical considerations final quantities are derived using iterative approximations or fits towards a comparable empirical signal. However, there are not only several different fitting methods but also differences in the numerical implementation leading to a number of degrees of freedom. This work should lead to a consistent, accessible fit routine which allows to study the impact from various input parameters on the final mixing ratios.

Question: How does the fitting method affect the final mixing ratio?

Contact: Dr. Daniel Kunkel

Target group: MSc Atmospheric Physics/Computer Sciences/Physics

Data: spectral airborne measurement data

 

Determination of tropopause and boundary layer heights in radiosoundings and ECMWF reanalysis data

The heights of the tropopause and planetary boundary layer can be determined by using various parameters and methods. Furthermore, these heights often differ between model and observational data products. Nevertheless, the tropopause and the boundary layer height take on a crucial role when it comes to separating different flow regimes in the atmosphere. In particular, close to both of these interfaces turbulence is expected to occur frequently. This project aims to identify the heights of the tropopause and boundary layer from observations using radio soundings as well as from reanalysis data.

Questions: How do tropopause and boundary layer heights differ between reanalysis and observational data?

Contact: Dr. Daniel Kunkel

Target group: BSc/MSc Atmospheric Physics

Data: ERA5, Radiosoundings DWD

 

Distribution of ozone in the vicinity of the tropopause

Ozone is one of the most frequently observed trace gases and can be used as a marker for the tropopause, i.e. the ozonopause. Radio soundings of ozone have a high vertical resolution, however, with the caveat of being sparsely available in time and space. In turn, airborne measurements of ozone have complementary characteristics, being available over larger larger areas in the horizontal but with lower coverage in the vertical at most locations. Reanalysis data also often include information about ozone. This project aims to study how the different information from observations and reanalysis data differ from each other. In particular, whether differences are larger under certain atmospheric conditions, e.g., close to the jetstream or the tropopause, where mixing affects the ozone distribution.

Question: How can observations and reanalysis data of ozone be used to identify regions which are subject to mixing?

Contact: Dr. Daniel Kunkel

Target group: BSc/MSc Atmospheric Physics

Data: ERA5, ozone soundings (WMO, GAW), airborne measurements

 

Climatology of PV anomalies in ERA5

PV anomalies represent a specific dynamic feature which fosters mixing of different air masses. Processes associated with the modification of PV anomalies may be related to high wind shear as well as diabatic modifications. Ultimately, these processes determine the lifetime of the PV anomalies. This project aims to study the temporal and spatial distribution of PV anomalies in the ERA5 data set. For this metrics of varying complexity can be applied to identify the PV anomalies and their consequences for mixing.

Question: What is the spatial and temporal distribution of PV anomalies in the region of the tropopause in ERA5?

Contact: Dr. Daniel Kunkel

Target group: BSc/MSc Atmospheric Physics/Computer Sciences

Data: ERA5