Predicting the Weather: Not As Easy As It Looks


A few weeks ago, I noted that insurers in the US reported more losses from severe storm events in the first quarter of 2017 than at any other time in the last 20 years. If there’s a flip side to this news, perhaps it’s that at least one source is predicting fewer storms than average in the Atlantic this year during hurricane season. But another source—NOAA—is predicting more storm activity than usual.

These predictions, even as they somewhat contradict each other, have gotten amazingly precise. Specifically, a report from the Tropical Meteorology Project at Colorado State University is calling for “11 named storms, four hurricanes, and two major hurricanes…. This is slightly below the 30-year average of 12 named storms, six hurricanes, and two major hurricanes. A major hurricane is one that is Category 3 or stronger.”

Of course, the report’s authors note, it takes only one poorly placed storm to cause major damage, and there’s no telling yet how many of the storms that do develop will make landfall. They point out that in 1992 only six named storms occurred—which sounds like a good thing until we remember that one of them was Hurricane Andrew, a Category 5 storm that is still the most destructive in US history, killing 65 people and causing more than $26 billion in damage. Historically, the authors say, there is no strong correlation between the number of storms and how many reach land. The CSU prediction “is based on a combination of 29 years of statistical predictors, combined with analog seasons exhibiting similar features of sea-level pressure and sea-surface temperatures in the Atlantic and eastern Pacific Oceans.”

Then, however, we have a different outlook from the National Oceanic and Atmospheric Administration, which operates the National Hurricane Center: “For the upcoming Atlantic hurricane season…forecasters predict a 45% chance of an above-normal season, a 35% chance of a near-normal season, and only a 20% chance of a below-normal season.” NOAA’s forecast says we have a 70% chance of 11 to 17 named storms; five to nine of those might become hurricanes, and two to four major hurricanes—that is, Category 3, 4, or 5. NOAA’s predictions are based, in part, on a weak El Niño.

An article coming up in the September issue of Stormwater will look at the complex process of forecasting hurricanes—the models and historical data forecasters use, the real-time data-gathering process as a storm is developing, and how emergency responders, stormwater managers, and other make use of the information available—for example, whether and precisely when to issue evacuation orders. In the meantime, hurricane season begins in June; you can find not only forecasts, but also real-time information on the National Hurricane Center website. SW_bug_web

  • John Folks.

    It should be noted that the over the last 20 years our population has increased and that people are living in places at densities that had not existed 5,10, 20 years ago. Also, the value of dwellings has increased significantly in the last 20 years, so the cost of repairing damage also has gone up. So your your assumption that the increased storm damage, based on costs, ( the insurance companies looking only at their costs) is due to increased storms is not conclusive, and in fact a stretch to justify your premise.

  • John Folks.

    It also must be noted that both UC’s and NOAA’s prediction of hurricanes over the last five years have been way off, both groups predicted numbers were to high. Until models are developed that can predict tomorrow’s weather with statistically valid accuracy, one must always take weather models’ predictions as just best guesses.


Leave a Reply

Enter Your Log In Credentials