WCRP’s global km-scale climate modeling hackathon: The nextGEMS legacy goes global!

Last week, from May 12th to May 16th, the World Climate Research Programme (WCRP) brought together more than 600 people to participate in the first global climate modelling hackathon, building upon the experiences gained throughout the preceding nextGEMS hackathons. Other initiatives, such as WarmWorld, EERIE, Destination Earth, and ESMO supported this endeavour.

During the hackathon, teams from Europe, Southeast Asia, Australia, and North and South America gathered at 10 different locations, the so-called nodes, to collaborate, exchange, and advance research in kilometer-scale climate modeling.

One of the nodes was hosted by the Max Planck Institute for Meteorology (MPI-M) in Hamburg. Here, over 120 participants were welcomed by Bjorn Stevens, director of the institute’s Climate Physics Department, who stressed the importance of exchanging and collaborating between different generations of climate scientists.

In this spirit, the participants spent the week collaborating with local “Science Teams” and gaining insights into a variety of research topics, using data from the newest runs of the ICON and IFS–FESOM models. The group “Energetics of Tropical Rainbelts”, led by Hans Segura, for example, worked on extracting the spectrum of convective coupled equatorial waves and Paulina Czarnecki’s team examined the high cloud top temperatures in storm resolving models to compare them between different models. A list of all of the topics investigated at the Hamburg node can be found here.

Previous to the hackathon, the scientists were encouraged to participate in a plotting challenge to get acquainted with the new model outputs and available data. This type of challenge was developed throughout previous nextGEMS hackathons to motivate participants’ creativity and generate interesting plots about phenomena of their choice. Below you can find a collection of the submissions. Can you guess, which plot won the challenge this time around?

During the hackathon, the groups were supported by a dedicated Data Support Team, made up of scientific programmers, data scientists, and model developers from the German Climate Computing Centre (DKRZ), the MPI-M, ECMWF, and AWI. This technical support group was also a feature that proved to be very helpful during previous nextGEMS hackathons.

Internationally, the different nodes stayed connected through virtual meetings, joint chat groups, and so-called cross-cutting activities. The latter focused on technical skills or data sources, useful for addressing a range of scientific questions and acting as a connector between scientists working on different topics.

However, our participants did more than hack and research together. They also had the opportunity to network and attend various side events throughout the week. For instance, Prof. Caroline Muller from the Institute of Science and Technology in Austria (ISTA) gave an inspiring keynote on the characteristics of deep convective storm systems, as well as the environmental and internal processes driving their growth. Some of the most attractive activities organized at the hackathon were the visits to the wind tunnel of the Hamburg University and DKRZ’s supercomputer Levante.

On the final day, each group presented their work, highlighting their results, lessons learned, and future research activities. Henning Franke’s group, consisting of Theresa Mieslinger, Hartmuth Borth, Chao Li, Quan Liu, Frederik Vieira Fischer, and Lianet Hernández Pardo, tested the hypothesis that mesoscale convective systems (MCSs) over tropical oceans are often triggered by mesoscale surface wind convergence. They looked at the evolution of surface temperature, precipitation, and how the wind converged and diverged at low levels — before and right after marine storms formed in the models. Their work showed how different models captured phenomena that impact storm formation. An animation of the mean precipitation flux and wind convergence up to 10 hours before the storm was detected can be found here. Going forward, the group wants to work on improving their code, extend the analysis to all simulations and to other trackers, and explore and deepen the connections of their work with other groups.

Concluding the event with a full-circle moment, Bjorn Stevens led a lively discussion around the concept of hackathons for the climate modelling community. Specifically, noting how such a format can and should be used beyond nextGEMS to boost intense networking and facilitate cooperation for better km-scale models while providing tools that support the users’ interaction with these models.