It’s no secret that people here in the South have a fascination with snow. It is not very common, relatively speaking, so any mention of the word prompts posts, shares and a full range of emotions. In the midst of Sunday football, I have watched closely as meteorologist colleagues try to be a credible vaccine for a viral post about snow in the southeastern U.S. over 10 days from now. What’s the fallout from such posts?
A post on a popular Georgia pop culture site included one weather map indicating the possibility of a foot or more of snow in parts of the South, including the Atlanta area, by the second week of January. It has gone viral. The site is awesome, and one that many follow, so it is not surprising how widespread the post has been. Friends have been messaging me all day asking if it is true. Such maps and information are certainly available to anybody, which creates a challenge for the meteorology community. I witness all types of platforms posting long-range snow forecasts. Colleagues like James Spann in Alabama and Brad Panovich in North Carolina routinely push back on what we call “snow pornography.” It plays on the snow lover’s desire, and the social media culture rooted in clicks, views, or shares.
When I saw the model information being shared, I cringed. To the site’s credit, it mentioned that such information could change. However, the context was missing, and it was probably better not to post it at all. Atlanta News First meteorologist Ella Dorsey commented on Instagram, “This is just one run of the Euro model. The next run showed 0.2 inches of snow for the same period.” She went on to explain that a weather pattern shift to colder temperature is expected, but information this far out is not reliable for precipitation forecasts.
Meteorologist Tyler Mauldin also said in the same thread, “A little early to be posting! Forecasting snow, it’s hard enough three days out, much less 10 days.” As I recently wrote, “These ‘pick your favorite mode scenarios’ often disappear in the next model run…. credible meteorologists should not convey information based on one model run.”
As weather professionals, we know the limitations of long-range or single model runs. However, many people do not and will post or share information without proper context. So what Dr. Shepherd! I hear you. It may all seem innocent, but there are serious implications of “snow porn.” When those forecasts do not verify, people question the credibility of the weather community.
When these “fantasy storms” never develop, it amplifies false narratives that weather forecasts are always wrong. As I have written many times, forecasts are right most of the time, but people have biases or shortcomings in how they consume the information. They may not understand percent chance of rain, limitations of weather Apps, or other metrics of uncertainty. There is also a tendency to latch on to the rare bad forecast and overlook the more numerous correct ones. Most people will remember a kicker’s field goal miss in the Super Bowl, and forgot that he made every kick all season.
It also “clouds” the boundary of expertise. Meteorologists are trained experts who understand the limitations, tendencies, and biases in weather model information. My colleagues and I are aware of fantasy snowstorms and hurricanes that dissipate in the next model run or when things are ten or more days out. A 2019 study out of Pennsylvania State University has describe the barriers in numerical model weather predictability.
Part of the challenge is that many people do not really understand how forecasts are made. The atmosphere is a fluid within a rotating frame of reference. There is a whole lot of physics, calculus, and complex computational processing going on y’all. Those calculations are further affected by slight differences in the initial condition of the model, data quality or scale, and changes in the weather scenario between runs. To mitigate against these challenges, we use numerous models or ensembles (multiple manifestations) to reduce uncertainty. We do not anchor forecasts in one single outcome or scenario.
Another challenge is that once bad, viral forecasts are in the public space, confirmation bias sets in. People have seen or heard what they want to happen and will often resist credible explanations. We saw this during Hurricane Helene when colleagues were being threatened or attacked for explaining that we cannot control hurricanes. Confirmation bias and “wishcasting” are powerful motivators if you want to believe that a snow scenario is possible. I will defer to my psychology colleagues for deeper explanations. However, trust me, it is as real as the irrational overload on French toast ingredients purchased at the very mention of snow in the South.
I encourage you to stop, take inventory of the source, and process the information carefully before pushing the “share” button or running to the grocery storm. Having said all of that, conditions are trending towards a colder pattern for the South by the middle of January. That means snowfall could be in future cards, but it is certainly too early to confirm base on a model you are seeing in late December.
