link to homepage

Institute for Advanced Simulation (IAS)

Navigation and service


Thu, 05 September 2019 Martin Schultz and Lukas Leufen attended a workshop on "Machine Learning in weather and climate modelling" at Corpus Christi college in Oxford. This workshop assembled more than 100 top-notch climate scientists and experts in HPC computer science and machine learning to present ongoing work and discuss the way forward. It became clear from the start that machine learning can likely play an important role in almost all stages of a weather and climate modelling workflow. Much discussed topics were the perceived need to impose physical constraints on the machine learning algorithms and quantify uncertainties. Martin Schultz's presentation on the IntelliAQ and DeepRain projects was well received and the positive response confirmed the research strategy followed by these projects.

Fri, 16 August 2019 At Osnabrück University Jonas Rebstadt sucessfully finished his studies with a master thesis titled "Deep Hyperresolution for Weather Forecasting". The goal is to develop a system that is able to increase the precision of rain forecast without exorbitant higher computational demand. The approach presented in this thesis is trying to increase the spatial resolution of a currently productively used forecast model developed from the Deutscher Wetterdienst (DWD) by training a neural network based on higher resolved radar images as target.

Thu, 01 August 2019 Severin Hußmann sucessfully finished his master studies at Humboldt-University of Berlin. His master thesis "Deep Learning for Future Frame Prediction of Weather Maps" focuses on applying data-driven deep learning methodologies to the field of weather forecasting, specifically air temperature over Europe. A future frame prediction model from the computer vision field is trained with the three input variables air temperature, surface pressure, and the 500 hPa geopotential in order to predict the air temperature itself. The experiments show that the model can make better hourly and even several-hour predictions than the persistence assumption.

Mon, 08 July 2019 Progress in data transfer: an essential aspect of the DeepRain project is the large amount of data used for training and evaluation of machine learning methods. A total of over 600 terabytes of data are currently being transferred from the German Weather Service to Forschungszentrum Jülich to be used on JSC supercomputers for deep learning. Today, the 100th terabyte was successfully transferred and integrated into the storage systems at JSC. This is an important milestone, as enough data is now available to carry out the first meaningful deep learning and analyses.

Wed, 22 May 2019 The Jülich Supercomputing Centre (JSC) has allocated two large data projects with a volume of several hundred terabytes for the DeepRain project. The first 30 TByte of meteorological model data have been successfully transferred from the German Weather Service to JSC and a prototype workflow for processing of these data has been established.

Fri, 12 April 2019 The DeepRain presentation by Martin Schultz at the European Geoscience Union conference in Vienna was well received. Machine learning attracts a lot of attention now in the research field of weather and climate. Fruitful discussions followed the talk, which may lead to future collaborations.

Fri, 08 March 2019 - The DeepRain team has just completed its second project meeting and the project partners return to their home institutions. The meeting was organized at the Institute for Cognitive Sciences at the University in Osnabrück and included a brief tutorial on Deep Learning, which is the expertise of Prof. Gordon Pipa's research group. Prof. Pipa's team presented their plans for neural network architectures that will be used to learn rainfall patterns from the ensemble model runs by the German Weather Service. The DeepRain partners continued their discussions about the Terabyte-scale data management and debated about validation methods and error characteristics, and how these may affect the performance of the neural networks. Careful analysis of errors and understanding the merits and limitations of deep learning in the context of weather data are key objectives of the DeepRain project.

Mon, 25 February 2019 - The DeepRain cooperation agreement has been finalized and signed by all project partners. This constitutes the formal basis for a fruitful collaboration among Forschungszentrum Jülich as coordinator, the German Weather Services (DWD), the Universities of Osnabrück and Bonn and Jacobs University in Bremen. "The DeepRain project adopts the principles of Open Science and Open Data. Therefore the collaboration agreement imposes as little constraints as possible, but some rules are necessary.", says Dr. Martin Schultz, who coordinates the project.

Wed, 20 February 2019 - The DeepRain project will be presented at the European Geophysical Union General Assembly in Vienna. Dr. Martin Schultz, the project coordinator, has been assigned to an oral presentation on Thursday, 11 Apr 2019, 16:15 h in the session on Machine learning for Earth System modelling.

Thu, 31 January 2019 – back from the Rasdaman training workshop on the campus of Jacobs University in Bremen, the DeepRain co-ordinator Dr. Martin Schultz sounds rather satisfied: “This workshop was very helpful to all ten participants. Not only did we learn a lot about the amazing technology behind Rasdaman and the thorough design concept, which not only follows but actually sets standards for geographic data processing, but it was also good to find a bit of time to actually work on samples from the actual data that will be used during the project. It was great to learn that we can obtain even more data from DWD than we had thought initially. At JSC we now have to get our heads together to organize the transfer and management of half a petabyte of weather model data. DeepRain clearly is one of the most exciting projects I have been working on in my career.”

Mon, 03 December 2018 - At the DACH Conference in Garmisch-Partenkirchen, Dr. Rita Glowienka-Hense from the University of Bonn will give the first presentation that is associated with the DeepRain project. The title of her talk on Fri, 22 Mar 2019 at 11:50 h will be "Partial correlation the natural correlation skill score".

Tue, 27 November 2018 – Yesterday, the partners of the DeepRain project met at Forschungszentrum Jülich for their Kickoff meeting. Two months into the project, the partners had already exchanged quite a few emails and discussed issues concerning the selection and management of data, the choice of deep learning methods, statistical approaches, and other issues. Some preliminary work has started and everyone was eager to get the project off the ground. The DeepRain team consists of scientists with rather different backgrounds who are all excited to work together in this challenging endeavor. It was difficult to keep the meeting on time, because so many interesting discussions were spun off.
Besides clarifying a number of formal project management issues, some agreements were reached how to approach the next steps. As common test case covering most aspects that need to be addressed during the project, we will focus on the Harz region during autumn. This choice was made, because the Harz represents a region with variable orography, and there are data available from earlier downscaling studies. Also, during autumn, rainfall is primarily driven by large-scale weather phenomena, which makes an easier start for the deep learning networks. Finally, the Harz is well covered by radar data and will allow an analysis of the effects from overlapping radar coverage. Firstly, we will work to establish the necessary data flows and reformatting, thereby making use of the Rasdaman database technology. In parallel, some first test neural networks are designed and decisions concerning the evaluation metrics and methods will be made. The project partners will meet again for a Rasdaman training workshop in Bremen in January 2019. The next project meeting will take place in Osnabrück in March 2019.