Show Notes
Host: Ryan Harris
Guest: Dr. Amy McGovern, Professor at University of Oklahoma Director, NSF AI Institute for Research on Trustworthy AI in Weather, Climate, and Coastal Oceanography & Senior Technical Staff Member
Description: A lot has changed in the AI-weather prediction space since the summer of 2023. Dr. Amy McGovern from the University of Oklahoma joins the Triple Point® again, a year since she last joined Jeff and I, to catch listeners up on the state of Artificial Intelligence in numerical weather prediction. We talk about the need to create policies and guardrails around AI that are broad and flexible enough that innovation can still happen alongside responsible AI principles and protection against nefarious uses. There are several areas where AI is going to continue accelerating weather prediction, but arguably none more so than in the way we assimilate data into our weather models. It's likely to be a space where, in a year, we will be looking back at how rapid a transition occurred. LISTEN HERE
References:
- White House Holds Workshop on Artificial Intelligence and Weather Prediction (NOAA)
- Transformational Artificial intelligence to Modernize the Economy against Extreme Weather Act or the TAME Extreme Weather Act (U.S. Senate)
- Congressman Lucas Introduces Bill to Establish NOAA as an Independent Agency (U.S. House of Representatives)
- Identifying and Categorizing Bias in AI/ML for Earth Sciences (Amy McGovern)
- OU Leads $6.6M Data Assimilation Consortium to Improve Weather Forecasts and to Develop Workforce (Univ. of Oklahoma)
- National Novel Writing Month (NanoWriMo)
- Approximating Observation Error Statistics Using Machine Learning Models (Amanda Burke)
- Machine Learning Meeting Data Assimilation (Peter Jan van Leeuwen, 2019)
- NSF’s Artificial Intelligence for Environmental Science (AI2ES)
- September 2024 NOAA Workshop on AI
- Ep 35 - Weather Outside the Box with Scott Mackaro (Triple Point®)
- Ep 26 - AI in Weather and Climate with Drs. Amy McGovern and John Williams (Part II) (Triple Point®)
- Ep 26 - AI in Weather and Climate with Drs. Amy McGovern and John Williams (Part I) (Triple Point®)
- Ep 27 - Weather Modification and Geoengineering (Triple Point®)
- 4DVAR = 4-dimensional variational data assimilation, defined as the system used by some modeling centers like ECMWF to combine a short-range forecast with observations to produce the best possible estimate of the current state of the atmosphere
Main
[01:11]
What has changed in the AI-weather space since Summer 2023?
[02:42]
Policy implications of AI in weather prediction
[06:44]
AI accelerating the world of data assimilation to advance weather prediction
[09:22]
Down the rabbit hole on geoengineering and weather modification
[16:13]
Out of the rabbit hole: handling bias and potential nefarious data spoofing in AI and weather
[22:26]
Academics like Amy on sabbatical and fulfilling that desire to innovate fast in the private sector
[29:32]
Amy's take on whether NOAA will still produce forecasts by 2034
[38:49]
Modified Lightning Round
[46:24]
Member discussion: