Insights from the first cross-training between EOSC-Life and EOSC-Nordic
EOSC-Nordic does not only happen on a governance and strategic level. We, the project participants, work with scientists trying to make or utilise what is to be the European Open Science Cloud. This cannot be achieved without training.
For the climate use cases and demonstrators in Work Package 5 (WP5), the role of EOSC-Nordic is to support cross-border collaboration and in particular the usage of High-Performance Computing (HPC) infrastructure across countries. As part of this work, WP5 has joined forces with another EOSC project, EOSC-Life.
Collaboration with EOSC-Life
EOSC-Life brings together the 13 Life Science ‘ESFRI’ research infrastructures to create an open, digital and collaborative space for biological and medical research.
There are many technologies in common between the EOSC-Nordic project and the EOSC-Life, and all the tools from the EOSC-Life toolchain have been adopted by EOSC-Nordic for the Climate Science demonstrator.
The Climate Science demonstrator project concerns itself with ecosystems only found at high latitudes (read more here). These ecosystems are poorly represented in climate models, and to reduce uncertainties and improve climate predictions, field ecologists, environmental scientists and climate modellers need to work together while having a different level of technical expertise and scientific background.
Galaxy as a common platform
As a common ground, we chose the web-based platform Galaxy. Galaxy is an open platform for accessible, reproducible, and transparent computational research. Galaxy is also listed as a service in the EOSC marketplace and a key partner in EOSC-Life. Thanks to WP5 and the climate science demonstrator, we set the ground for a collaboration between EOSC-Nordic and EOSC-Life with Galaxy interlinking them, which, ultimately, provided us with a sustainable solution for our project.
Therefore, in the 1st EOSC-Life Training Open Call, the European Galaxy team was awarded two grants to develop climate science e-learning material and training opportunities for its user communities.
So, on the 26th and 27th of October 2020, we trained 20 field ecologists, environmental scientists, climate modellers and biologists interested in improving climate models by better representing terrestrial ecosystems. They learned how to compose and execute repeatable and reproducible modelling workflows with ‘FATES’ (Functionally Assembled Terrestrial Ecosystem Simulator) for improving climate models.
Unsurprisingly in these Covid-19 times, the workshop became an online workshop. While we had the experience of organizing workshops online, it is still a bit challenging when it comes to engaging with new/potential users of new services. To provide a better training user experience and to better facilitate exchanges during the workshop, we had not only three instructors, but also two helpers.
We used the available Training Infrastructure as a Service. We not only get sufficient computing and storage resources from this, but instructors can also monitor the progress of their students during the hands-on exercises.
Welcoming participants and taking the time to explain the logistics is key for a successful workshop. We used Zoom and a shared document through hackMD to interact with our learners, especially during the hands-on work.
Finally, we had our first lecture given by Rosie Fisher, one of the main developer of FATES: FATES is a numerical terrestrial ecosystem model for use in climate models that simulates and predicts growth, death, and regeneration of plants and subsequent tree size distributions. Her lecture was very enlightening to understand the need for such a model and how modelling with FATES can be used for reducing climate prediction uncertainties. All the software is fully open and available on an online GitHub repository and Rosie insisted on the importance of collaboration.
We then gave a short introduction on Galaxy which can be followed here.
And we went through the CLM-FATES Galaxy tool Hands-on in a work-along session and then allowed everyone to work at their own pace, using the hackMD to ask questions. We also had a meet-up Q&A at the end of the day to answer questions. This was a great opportunity to get feedback on the training material and to understand what we should emphasize in our next training. Using workflows to compose and automate tasks were quite new for many scientists and they all asked to get more training. This is a success for us, as we strongly believe using workflow management systems can speed-up scientific research and facilitate cross-border exchanges.
The scientists attending the workshop learned to produce a Galaxy workflow, extract it from their Galaxy history and publish it to the workflowhub.eu where each workflow is assigned a unique SEEK ID and is properly versioned. In a near future, the WorkflowHub will also be connected to Zenodo to get a Digital Object Identifier (DOI). This approach significantly increases the FAIRness of FATES simulations or any workflows developed within Galaxy.
During the second day, we focused on teaching CLM-FATES in Galaxy Climate JupyterLab. The Galaxy interactive Climate JupyterLab is a co-design platform and the goal with FATES is to facilitate model development. This platform is meant to be used for:
- preparing new input datasets
- testing new versions of FATES, including code changes
- developing new notebooks for analysing and showing the model results
- creating interactive publications (notebooks)
- teaching purposes
The main advantage over the Galaxy CLM-FATES tool is that you can run any versions of FATES (recompile it whenever necessary), including developments that are not released yet.
This training was a fantastic opportunity for EOSC-Nordic to foster and advance the take-up of the European Open Science Cloud by Nordic scientists. The results of the survey we did after the training clearly show that our event was successful with more than 90% of our learners saying that the event met their expectations. We got a lot of feedback for improving the tools for running FATES on Galaxy and many requests from both the biodiversity and climate communities for introducing new tools in the Galaxy platform.
This is a step in the right direction and a step towards Open Science becoming a reality, not only at a policy level.