You’ve no doubt been hearing in the news about Hurricane Helene over the past few weeks. Helene made landfall in the Big Bend region of Florida on September 26 as a Category 4 hurricane with sustained winds of 140 mph. But as so often is the case with tropical systems, the story is not only about the winds but also the water. Usually that means storm surge, and while that was a major issue with Helene there was some water-related damage with this storm that is a bit out of the ordinary with tropical systems—tremendous (in fact, historical) flooding damage in the mountains of western North Carolina, eastern Tennessee, and southwestern Virginia. Recovery operations are still going on, and those communities will be impacted for years if not decades to come.
While it by no means compares with the impact to the lives of those directly affected by Helene, the air dispersion modeling community has also been affected in terms of availability of meteorological data for modeling.
Hourly surface data typically used for creating meteorological input files for use in AERMOD are archived at the National Oceanic and Atmospheric Administration’s National Centers for Environmental Information (NCEI) in Asheville, NC. NCEI has one of the most significant archives on the planet, with data back as far as the 18th century and continuing to archive more than 200 terabytes of data each month.
Unfortunately, Asheville is one of the areas most devastated by flooding from Helene, which has in turn impacted NCEI. NCEI has confirmed that while its archived data are safe its internet service provider has limited functionality, and it isn’t clear when it’ll be back up and fully operational. So, in the meantime access to meteorological data for modeling is down (as of this writing, when I went to the page where I look for data I get a 404 Error).
So, what are your options at this time? Well, the Guideline on Air Quality Models (Section 8.4.2(e) of 40 CFR 51, Appendix W) specifies that if you’re using National Weather Service data (i.e., what is available from NCEI) you need five years of representative data; “EPA’s Meteorological Monitoring Guidance for Regulatory Applications” goes on to specify that EPA prefers five consecutive years from the most recent, readily-available data. So, if you’ve already got some data for an area but don’t have up through 2023, it may be possible to lean on the “readily-available” specification and argue that using data that aren’t the most recent is good enough in this case.
Another option could be to generate prognostic data for the area you’re modeling—although the creation of those data may be affected by NCEI’s data access issues as well.
Of course, if you do have some regulatory modeling you need to do in the near future (internal modeling is different) you’ll very likely want to talk with your State agency about how to handle this. While I would like to think that most of them would be pretty understanding in this situation, if your analysis is controversial you should be aware that opposition groups could cite the “most recent” language from the guidance to call the conclusions of your analysis into question.