Mini-hackathon S&L

by Chiel van Heerwaarden, Wageningen University

A first glimpse of variability over land

In preparation for the upcoming Madrid hackathon, the Storms and Land theme met in Wageningen from 6 to 8 April 2023. With multiple positions that have recently started, our goals were to get to know each other better, and to shape up a document that will be a first overview of the ability of the storm-resolving models to capture the climate over land.

Our first focus will be on variability over land in the broadest sense. With the unprecedented global resolution of storm-resolving models we see for the first time mesoscale phenomena, such as cloud systems, mountain winds, sea breezes, or heterogeneity-driven circulations appearing at such a detail level that we can directly compare their statistics directly to field observations. Our hypothesis is that this will largely influence (and hopefully ultimately improve) resolved variability in thermodynamic variables, wind, precipitation, and derived variables such as river runoff or solar power production.

With the expertise and research focus of the different groups, we have an interesting collection of modes of variability that the Storms and Land theme will study in detail in the coming months. The ETH Zurich group is analysing temperature and drought, Wageningen University is studying solar and thermal radiation variability and river runoff, MPI-M and Wageningen together are investigating whether this generation of models is able to pick up soil-moisture precipitation coupling well. Also, the University of Lissabon is evaluating the simulated surface temperatures, while the University of Bern is analysing precipitation in the Alps and the representation of atmospheric blockings in the storm-resolving simulations.

What makes nextGEMS challenging and also exciting for that reason, is that the step from research ideas to actual output in the form of graphs is for most of us far more complex than ever before. This is due to the very large amount of data produced by the storm-resolving models and the different ways of storing those. Postprocessing model data requires advanced engineering skills, for which we are crucially dependent on the (fortunately excellent) support teams from MPI-M and ECMWF.  Nonetheless, after three days of hacking, we were able to bring our first plots to the screen and are warmed up for the Madrid hackathon, where we will hopefully find the first answers to our research questions.