Unrealistic Expectations Cause False Narratives About Weather Forecasts

I have been thinking about this and finally decided to put something on”virtual” paper. For some of you, this may feel like an opinion editorial. For others, it may sound like a slight rant. As a scientist, professor, and former president of the American Meteorological Society, I have come to realize that unrealistic expectations cause false narratives about the accuracy of weather forecasts. Here’s why I say that.

This thought crystalized for me while taking a question from a listener on a podcast yesterday. The question was, “Why is it so difficult to track hurricanes?” I was initially baffled by the question, but then I realized that I was thinking about it from my perspective and not the person asking the question. Hurricane track forecasting is an area of substantial progress in modern weather prediction. The graphic below shows the reduction in average track errors in the Atlantic Basin from 1970 to 2020. Clearly, there are dramatic improvements in the 1 to 5 day range. Today, the average error at 1-day is less than 50 nautical miles. In the early 1970s, it was 2 to 3 times that amount. Today, a 3-day forecast is better than a 1-day forecast in 1970.

Another example is the narrowing of the “cone of uncertainty.” As University of Miami hurricane expert Brian McNoldy writes in his blog (a must-read by the way), “The size of the cone is fixed for every forecast of every storm during an entire hurricane season, but the size slowly evolves from year to year. If the storm is moving quickly, the cone will appear more elongated and if the storm is moving slowly, the cone will appear more compact… but it’s the exact same cone.” Jake Reyna tweeted a graphic from McNoldy (below) illustrating that the cone has narrowed. Guess what that means? We have gotten better at forecasting track. Intensity forecasting has lagged but was ironically pretty solid for Hurricane Ian.

I truly believe some people believe that we have absolute forecast skills or capabilities to tell them the exact track a hurricane will take. We cannot and never will be able to do that. That’s what Jake Reyna means with “limit of predictability.” It is also why forecasters issue information with a measure of uncertainty (the cone). While I firmly believe that new risk communication tools may be needed going forward, for now, it is important for people to understand what the cone conveys. During Hurricane Ian, I saw people evacuate from one part of the cone to another part of the cone. The cone suggests there is 67% chance that the center of the storm will be anywhere in the cone so do not focus exclusively on the center line.

Unrealistic expectations are also seen with precipitation forecasts. I have always found it odd that people perceive “20% chance of rain” as meaning there is “0% chance of rain.” People ask questions like, “Is it going to rain over my dog’s water bowl in the left corner of their back yard at 12:37 pm?” Ok, I am being hyperbolic, but I think you get my point. Weather radar and some modern apps can help extrapolate such information, but guess what? There will always be uncertainty so probabilistic information is delivered. You see it with snow forecasts too. Meteorologists might call for 3 to 6 inches of snow (uncertainty). If 3 inches falls, some will say the forecast was incorrect because they “wishcasted” for the higher amount. While that sounds bizarre, I see it all of the time.

Other unrealistic expectations are related to limits of predictability in time. I cannot tell you how many times this question has presented itself to me – “I am having a ________in 2 months and it is outside, it is going to rain?” Unfortunately, the only answer that is credible is to look at the climatological possibilities for the date in question. Weather forecasting capabilities have limits of about 10 to 14 days. A Pennsylvania State University press release noted, “Unpredictability in how weather develops means that even with perfect models and understanding of initial conditions, there is a limit to how far in advance accurate forecasts are possible….” In a study by University scientists, they confirmed the longstanding hypothesis by Edward Lorenz. The Massachusetts Institute of Technology meteorologist and mathematician gave us the chaos theory and posited that there is an inherent limit of predictability. Armed with this information, you should be skeptical of certain Twitter or Facebook posts seeking clicks, shares, or likes.

Human biases also shape perspectives on weather forecasts. Forecasts are right more often than wrong. However, people tend to remember the wrong forecasts, especially if it impacted them in some way. As I wrote years ago in Forbes, “A field goal kicker could make every single kick during football season, but what if he misses the “big one” in the championship bowl game? He may be ridiculed or criticized, but is he a bad kicker? Probably not, but he did miss a kick with great impact. Weather forecast outcomes are very similar.” Around the time, marketing professional Sravanthi Meka tweeted, “I work in marketing and customer service. 90% of client interaction post service is a negative experience. People remember negative experiences more.”

Weather forecasts are quite good, and they are certainly better than the expert predictions for last week’s University of Georgia vs University of Tennessee football game (Go Dawgs!). However, it is important to temper expectations on what weather forecasts can deliver. Additionally, in this era of cute weather icons and Apps, try to avoid being “App-notized.” The Weather Apps can tell you some things but possibly not what you need to know in evolving weather situations.