Reanalysis for S2S AI Workshop - Objectives

The workshop will address the emerging and urgent need for high-resolution datasets that are produced using the extensive observations of the earth system and the new generation of faster, higher spatial resolution models for performing the reanalysis. Higher resolution reanalysis that can account for measurement and model uncertainties are necessary to develop a new generation of AI weather forecast models at S2S time scales to support energy, agriculture and other critical economic sectors. AI training datasets for these time scales have differing requirements than needed to train models for shorter lead times. The workshop will deliberate on these requirements, including optimal spatial and time resolution of datasets, observational data needed, and state-of-the-art reanalysis models and methods. The workshop will also provide a timeline for completing the critical technical challenges and begin a roadmap guiding the cooperation of NOAA and DOE to producing a reanalysis that meets the needs of an AI training dataset for all AI S2S model development. The workshop will lay the groundwork as just a first step in a cooperative partnership between NOAA and DOE leading to the development and implementation of the reanalysis as an S2S AI training dataset. To this end, the workshop will produce a white paper laying out the objectives and timelines for next steps for socializing with NOAA and DOE leadership.

Objectives:

• Introduce NOAA and DOE/ANL technical teams
• Deliberate on the technical specifications (such as optimal spatial/temporal resolution and model suitability) required to generate high-resolution reanalysis datasets, specifically for S2S (subseasonal-to-seasonal) forecasting.
• Evaluate existing hurdles in observational data availability, data assimilation (DA) techniques, and computational capacity that currently limit the development of datasets necessary for training AI foundation models.
• Identify high-priority project opportunities that align NOAA and DOE/ANL priorities, and determine the structural framework for future collaborative working groups or project annexes.