Martin Bergemann

- Building and Maintaining Real-Time Online Data Analysis Applications within the Distributed Control Software System - Karabo - which is developed by our group. - Developing Software to Support Scientists in Analysing Big Data (TB - PB). - Building Automated HPC Data Recording, Data Analysis and Data Visualisation Pipelines. - Maintenance and Development of Various Data Analysis and Recording Libraries.

• Applied supervised and unsupervised learning algorithms on various rainfall datasets to discover how extreme rainfall events are linked to the Monsoon over Australia. • Developed an object based radar image tracking algorithm which demonstrates that, contrary to general expectations, simulated rain- fall extremes are not better represented with increasing model reso- lution.

• Improved the forecast accuracy of rainfall near coasts in the tropics by implementing a simple cloud model to the UK-Met Offices’ numerical weather prediction system.

• Created high quality atmospheric multidimensional observationaldata sets by developing a pipeline for a 3D variational data assimilation and optimization software.

• Created the very first data set of coastline triggered rainfall by establishing an OpenCV based pattern recognition algorithm for mining satellite data. • Developed a simple stochastic Markov-Cain based model that is able to mimic coastal convection.