Breakdown: Why forecast accuracy decreases beyond 7 days

Breakdown: Why forecast beyond 7 days are less accurate

MEMPHIS, Tenn. (WMC) - The day to day forecast can be difficult but forecasting beyond seven days is extremely tough. When you forecast beyond seven days and try to get into specific things like rainfall totals or an exact temperature for a certain hour, it’s close to impossible.

There are so many limitations beyond seven days and according to experts those limits on long range forecast aren’t likely to change in the near future. Experts point out that even if meteorologist were able to obtain all of the data needed along with the perfect understanding of all the complexities of forecasting, even then sound forecast could only extend out to about 14 or 15 days. The reason the 2019 study found this to be the case is due to the constant changing and chaotic nature of the atmosphere.

The American Meteorological Society concurs. Their current statement on the limits of prediction, has been around since 2015. The AMS states that “presently, forecasts of daily or specific weather conditions do not exhibit useful skill beyond eight days, meaning that their accuracy is low.”

Although the AMS strongly discourages against issuing specific forecasts beyond eight days, there are other forecasting agencies that forecast up to 30 and as far as 90 days in advance. Although these agencies take a lot of flack about these detailed long-ranged forecast, they continue to predict and put out forecast 30 to 90 days out. Many meteorologists feel that has led some to mistrust in meteorologists.

These long-range forecasting approaches from some weather agencies have sparked criticism from many meteorologists for being specific and not communicating uncertainty or low confidence. Owner and operator of Thermodynamic Solutions and Consulting Meteorologist Beth Carpenter, feels these long range specific forecasts, are “not feasible and should not be trusted.”

Studies reveal that a seven-day forecast can accurately predict the weather approximately 80 percent of the time, while a five-day can accurately predict the weather 90 percent of the time. However, the percentage declines with a 10-day or longer forecast and is only accurate half the time.

Meteorologists forecast models provide predictions on various different atmospheric elements like temperature, wind and pressure but the models have biases and can’t take every element into account. It’s impossible to collect data from the future, models use estimates and assumptions to predict future weather. The atmosphere is constantly changing, which make these estimates from the models less reliable the farther you go into the future.

Meteorologist continue to develop ways to generalize long-range outlooks which uses probabilities and likelihoods.

You may have noticed this type of wording in forecasts, for example with winter or spring outlook. In these forecast, they are non specific and use phrases like “above-average chances of a warm winter” or “above-average hurricane activity is likely.” These forecasts differ in that they don’t mention exact temperatures or precipitation amounts so far in advance.

Probabilistic forecasts are to try to predict the likelihood of something happening. These type of forecasts have improved a lot due to a better understanding of the ocean and the atmosphere and a rise in computing power.

Copyright 2019 WMC. All rights reserved.