J8.2 A (Unique) Peak into the Future: Prognosticating the Meteorological Landscape

Tuesday, 30 January 2024: 4:45 PM
Johnson AB (Hilton Baltimore Inner Harbor)
Cory J. Demko, CACI, Arlington, VA

Handout (5.3 MB)

In 1922, Lewis Fry Richardson integrated Vilhelm Bjerknes' primitive equations producing the world’s first numerically derived weather forecast by hand[1]. Over 100 years later, scientific, technological, and systematic advancements allowed humans the otherwise unthinkable power; acutely and accurately prognosticating Earths’ atmosphere, saving countless lives and alleviating catastrophic financial societal burdens. Today’s scientifically integrated Earth-system based observational and forecasting realm illustrates the continual and rapid growth (i.e., the scientific positive-feedback mechanism). Firmly within the Information Age grasp, humanity may experience a pivot with Artificial Intelligence (A.I.) emergence. How will the Earth Science domain continue illustrating the wider scientific community with responsible A.I. integration? Significant recent progress within machine learning and computer vision techniques illustrates the vast implications within not only the Earth Science community but evolving applications as well. How will quantum computing affect/alter numeric prediction? Currently, Large Language Models (LLM), for example, OpenAI’s Chat Generative Pre-trained Transformer (ChatGPT) significantly impacting business processes by, at times and when appropriate, continually automating previously rote tasks. Will the greater Natural Language Processing (NLP) discipline, capabilities, and applications impact the weather and Earth-System enterprise? An atmospheric scientist turned data scientist shares some thought-provoking ideas as this talk focuses upon the more “what if(s)” and perhaps coming full circle with Claude Shannon’s information theory, peering down the atomistic depths, and entropic growth as AI continues maturation.

[1]Peter Lynch, The origins of computer weather prediction and climate modeling, Journal of Computational Physics, Volume 227, Issue 7, 2008, Pages 3431-3444, ISSN 0021-9991,https://doi.org/10.1016/j.jcp.2007.02.034.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner