Second joint training course on Atmospheric Composition 2020

Course Introduction

From 16 – 20 November 2020, the European Space Agency (ESA), the European Organisation for the Exploitation of Meteorological Satellites (EUMETSAT), European Center for Medium-Range Weather Forecast (ECMWF) with the Copernicus Atmosphere Monitoring Service (CAMS) have organised the second joint training course on Atmospheric Composition with a focus on air quality and data interoperability.

The event brought together 30 participants from institutions and universities from 20 countries: Malta, Greece, UK, England, France, Switzerland, Poland, Italy, Norway, Portugal, Germany, Spain, Hungary, Netherlands, Austria, United Arab Emirates, Cyprus, France and Finland. The participants had a specialization level from graduate to PhD, with a majority of young researchers and some people working in local weather and environment agencies. Gender balance was ensured with 12 women and 18 men.

A focus was given to Copernicus datasets relevant for the thematic (Sentinel missions, Copernicus Atmosphere Monitoring Service) including a component from supporting missions and ground-based networks. The practicals made use of the tools developed from EUMETSAT and ESA support contracts to develop and work with Jupyter notebooks in training.

The training course was divided in two parts, each day had a morning and an afternoon session. The morning sessions were dedicated to lectures given by top-level scientists and experts from Europe and the afternoon sessions were focused on practicals. The last day of the event each participant was invited to present their own work from the training course.

This event followed the first joint course, organised in Nov 2019.

Specific Goals of the Course

• Present the state-of-the-art in air quality monitoring and modelling;
• Provide an overview of the different observations, tools and applications;
• Enhance the capacity on data access and analysis of existing and new satellite data products and monitoring services;
• Foster participants driven personal and team projects.

Course Platforms and Tools

The structure of the event, including lectures, streaming and practical sessions, required the deployment of several resources and tools. In coordination with ESA and ECMWF-CAMS, EUMETSAT led this activity with the support and contribution of all instructors.

• All course material was published on a dedicated page on the EUMETSAT training platform to provide:
  – A general introduction on the thematic (using the material of the EUMETSAT/ECMWF Atmospheric Composition MOOC);
  – Instructions for Data handling and access to the Hub;
  – Detailed program of the course with access information to the hosting platforms (Zoom) and virtual tools used to support the training (Slido for Q&A and Padlet for sharing of material and projects);
  – Overview of the main dataset
• A Padlet wall for participants to share their profiles and publish results and ideas
• A platform for participants to handle datasets
  A Jupyterlab was deployed with the selected Notebooks for the course. Specifically 16 Notebooks have been made available for the practical sessions. Participants had a specific directory to share development. The Notebooks also included specific contributions from ESA / S&T on the exercise with the HARP toolbox and workflows with the CAMS data provided by ECMWF.
• A Q&A tool
  A slido event was created to gather questions, keep track of answers and push the evaluation questionnaire. The use of polls has been proposed to lecturers, however, none was not used in the course.
• Internal communication channel
  A slack channel was made available to lecturers, instructors and course management to facilitate a “behind the curtains” coordination.
• Public Streaming
  Due to the large number of applications we considered important to offer the opportunity to join part of the course (lectures) to a wider public. For this reason, the morning sessions were streamed on youtube (EUMETSAT corporate channel).

Course Activities

The course comprised morning lectures (in a webinar format) and afternoon practical activities.

• Lectures were given by top-level scientists and experts from Europe. Each lecture lasted a maximum of 30 minutes, to avoid information overload, and the program followed a logical flow, “from general to particular”, starting with an overview (R. Munro, EUMETSAT; C. Zehner, ESA) and lessons on the basic principles of observations and models (V.-H. Peuch, ECMWF; A. Voulgarakis, Imperial College London –Univ. of Crete), moving towards several webinars on the current applications and science examples (C. Clerbaux, LATMOS-CNRS; A.M. Sundstroem, FMI; A. Colette, INERIS; H. Petetin, BSC;  H. Eskes, KNMI; A. Richter, Uni Bremen; A. Inness, ECMWF; J. Barre, ECMWF). The thematics covered mostly air quality and composition (both aerosol and chemical compounds) on satellite data and global/regional models, not neglecting the component of in-situ observations (P. Laj, Université-Grenoble-Alpes; L. Mona, CNR-IMAA;M. Van Roozendael, BIRA-IASB; A. Cede, LuftBlick). Please see the full programme below.
• The data discovery was conducted in sessions, involving the use of the JupyterHub. Four afternoon thematic sessions (two per day on Tuesday 17 Nov. and Wednesday 18 Nov) were proposed to participants to organise the work in groups. They could choose the preferred one on a “first come, first served basis”. Each session addressed practical issues, trouble-shooting, leading all participants to connect and run the proposed Notebooks. On Thursday afternoon (19 Nov) the participants worked in three groups starting to prepare for the project work.
• The project phase was built around groups. On the first day course organizers intervened in each group to discuss open questions on “basics” in modeling, atmospheric chemistry and remote sensing. On Thursday afternoon and Friday, the participants worked within the same groups on datasets to perform a small “personal project” whose extent, format and level of detail was kept free. Most participants made use of the JupyterHub provided. A first session was devoted to shaping the projects with a first draft to be published and a second one to the realization of it. An example of the analysis of one participant is reported in the link to padlet. Almost all participants (25 out of 29) were able to handle and analyse data and step in their project objectives, relating them to their needs and interests. Each group was tutored by a different instructor based on the specificity of participants to the group. The course organizers were present and rotating in the groups for further support.