News

Looking Back at Workshop “Transformers for Environmental Science”

JSC, the Otto von Guericke University of Magdeburg, and the Technical University of Munich jointly organised a workshop on “Transformers for Environmental Science” on September 22 and 23, 2022. The workshop was co-sponsored by the ERC grant IntelliAQ and brought together about 40 participants in Magdeburg and up to 20 additional online participants, who discussed the potential of this new AI technology for environmental applications. The program included lectures on recent advances in transformer architectures and transfer learning as well as on prototype developments focusing largely on atmospheric research and remote sensing. Keynote presentations were given by Peter Düben (ECMWF), Pedram Hassanzadeh (Rice University, Houston), Duncan Watson-Parris (Oxford University), Lucas Beyer (Google Brain) and Jonathan Godwin (Google Deepmind). A poster session and panel discussion provided opportunities for an exchange of ideas. Large transformer models require huge amounts of data and constitute an attractive application for accelerated supercomputers such as JUWELS Booster. Within the atmorep compute time project, first steps towards training such a model for atmospheric research are being taken.

Paper on Ozone Forecast with Deep Learning published

Lukas H. Leufen, Felix Kleinert (both FZ Jülich and University of Bonn) and Martin G. Schultz (FZ Jülich) have published their latest research results of the study “Exploring decomposition of temporal patterns to facilitate learning of neural networks for ground-level daily maximum 8-hour average ozone prediction” in the Journal Environmental Data Science.  The study shows how the accuracy of deep neural networks for forecasting ground-level ozone can be improved by splitting long-term and short-term weather patterns. The article is available at https://www.doi.org/10.1017/eds.2022.9 .

Preprint “Representing chemical history in ozone time-series predictions – a model experiment study building on the MLAir (v1.5) deep learning framework” available

Felix Kleinert et al. submitted their article “Representing chemical history in ozone time-series predictions – a model experiment study building on the MLAir (v1.5) deep learning framework” to the Journal Geoscientific Model Development. It is now available as preprint for public discussion and review at https://gmd.copernicus.org/preprints/gmd-2022-122/ until July 6th 2022.

Paper on Explainable Machine Learning published

Scarlet Stadtler, Clara Betancourt (FZ Jülich) and Ribana Roscher (University of Bonn) published their study on “Explainable Machine Learning Reveals Capabilities, Redundancy, and Limitations of a Geospatial Air Quality Benchmark Dataset” in the Machine Learning and Knowledge Extraction Journal. In their study, they gained insights into the AQ-Bench dataset on air quality using explainable machine learning. The article is available at http://dx.doi.org/10.3390/make4010008.

Preprint “Global, high-resolution mapping of tropospheric ozone” available

The preprint of the article “Global, high-resolution mapping of tropospheric ozone – explainable machine learning and impact of uncertainties”  by Clara Betancourt et al. is now available at https://gmd.copernicus.org/preprints/gmd-2022-2/. It has been accepted for public discussion by the Journal Geoscientific Model Development.
The paper presents a data-driven ozone mapping workflow generating a transparent and reliable product. The global distribution of tropospheric ozone from sparse, irregularly placed measurement stations is mapped to a high-resolution regular grid using machine learning methods.

TOAR won second price of Open Data Impact Award 2021

The Stifterverband awarded the project “TOAR App, Tropospheric Ozone Assessment Report database” with the second price of the Open Data Impact Award 2021 (https://www.stifterverband.org/pressemitteilungen/2021_12_14_open_data_impact_award): “The award highlights the potential of Open Research Data for innovation and society.
The TOAR database is one of the largest collections of global near-surface ozone measurements. To further promote the replication of this data, Clara Betancourt, Jianing Sun and Sabine Schröder from Forschungszentrum Jülich are developing a smartphone app that will allow farmers to quantify ozone-related damage to their crops.”

IntelliAQ at Jülich’s Lecture Evening

Scarlet Stadtler will present developments in the IntelliAQ project at Jülich’s End-of-Year Lecture Evening on 18 Nov. 2021 (https://juelich-eveninglecture.de). The focus of the presentation will be on her interdisciplinary approach to develop AI techniques to find patterns within climate and air quality data.

TOAR-II FAIR Data

As an important milestone on the way to the second Tropospheric Ozone Assessment Report (TOAR), a new TOAR data infrastructure has been developed with a focus on FAIR data (https://www.go-fair.org/fair-principles/). First data has been uploaded to the TOAR-II database and more data will continuously be added. The REST API provides rich functionality for searching and accessing metadata and data, which are always delivered together. Read access is open to everyone under CC BY 4.0. The database and its metadata is well documented as are the related services (go to https://toar-data.fz-juelich.de for links to the services and the documentation). The TOAR data infrastructure will be assessed with respect to its FAIRness and sustainability through the Core Trust Seal (CTS, https://www.coretrustseal.org/). The CTS application was submitted on September 29, 2021.

Paper on “Context aware benchmarking and tuning of a TByte-scale air quality database and web service” published

The performance of one of the world’s largest databases of near-surface air quality measurements and its services have been benchmarked and tuned with good results. Specifically the on-demand processing of several air quality metrics directly from the database has been in focus. The work published online by Earth Science Informatics describes explorating and benchmarking in-database approaches for the statistical processing, which resulted in performance enhancements of up to 32%.