The goal of the The Big Data and Earth Sciences: Grand Challenges Workshop is to bring thought leaders in Big Data and Earth Sciences together for a three day, intensive workshop to discuss what is needed to advance our understanding and predictability of the Earth systems and to highlight key technological advances and methods that are readily available (or will be soon) to assist this advancement. With the ever growing quantity and quality of hyper-dimensional earth science data (satellite and ground based observations and cutting edge Numerical Weather Prediction (NWP) models), the advancements in machine learning (e.g. supervised, unsupervised and semi-supervised learning techniques), and the progress made in the application of Graphical Processing Units (GPUs) and GPU clusters, we now have an unprecedented opportunity and challenge to engage these computational advances to improve our understanding of the complex nature of the interactions between various earth science events, their variables and their impacts on society (flooding, drought, agriculture, etc.).
At SC21 last week, Smarr delivered a racing reprise of this (ongoing) HPC tour de force in his talk, The Rise of Supernetwork Data Intensive Computing. Presented here are a sampling of his comments (lightly edited) and quite a few of his slides.