Second white paper on community guidelines on the use, value and applicability of emerging technologies in climate and weather applications - Deliverable D2.7

This second white paper provides an overview of trending and emerging technologies related to climate and weather applications, focusing on novel extreme-scale systems, programming models for heterogeneous computing and machine learning (ML). It updates the findings highlighted in the first white paper by providing a summary of relevant efforts available in literature, as well as some early results from these fields.

Tracking of state-of-the-art solutions and technologies has been carried out during the ESiWACE2 project to identify recent and relevant work focusing on trends in High Performance Computing (HPC), use of hardware accelerators and machine learning for weather and climate modelling. The document summarises the main results from this tracking activity providing pointers and references to interesting articles.

In order to track some of the earliest work in the field, a second workshop on Emerging Technologies for Weather and Climate Modelling was held on the 7th of October as a virtual event. The workshop organised in the frame of the ESiWACE2 project aimed to bring together HPC experts from academia and industry and climate/weather scientists from the EU and USA to discuss about some of the latest developments, research opportunities and efforts. The workshop agenda was organised around three topical sessions, for a total of 14 talks, about: European Exascale hardware (Session 1), Programming models and hardware interplay (Session
2) and Machine Learning (Session 3). More than 130 people from 22 countries world-wide registered to the workshop with Germany, UK, Italy, France and USA being the most represented countries.