- Education and Science
How Accurate Are Weather Forecasts? Recent Weather Forecasting Improvements
While weather forecasters (also known as meteorologists) still make big blunders when predicting the weather, the reality is that weather forecasting has become a much more accurate science than it was just a generation or two ago. Are you wondering just how accurate are weather forecasts these days compared to the past? This article examines this interesting question.
Although people tend to fixate on the occasional weather forecasting mistakes that still occur, in actuality weather forecasting has made dramatic improvements in recent years. This marked improvement includes mainly short and long-term weather forecasts, as well as the first couple of days of what are considered medium-term weather forecasts (three to ten days out). There have also been big improvements in regards to major storm predictions, including winter storms and hurricanes. The one time-frame that weather forecasting has not improved all that much is medium-term weather projections between five and ten days. We look at why weather forecasting has advanced in recent years and how it can continue to improve.
How Do Meteorologists Make Weather Forecasts Like This One?
Short Term Weather Forecasts Are More Reliable Nowadays
Short-term weather forecasts are predictions that look out at what the weather will be within the upcoming two days. “Nowcasts” are forecasts up to twelve hours in advance. While forecasting errors are still occasionally made in the short-term range, they are much rarer than they were in previous decades (five day forecasts are as reliable today as three day forecasts were some two decades ago). There are a number of reasons why short-term weather forecasting has become a much more exact science in recent years, from better radar and satellite technologies to sophisticated computer models that use statistical methods to more accurately simulate future weather scenarios in the near-term over a regional area. Weather forecasting is also part art, as weather forecasters draw on their years of experience and research done on past storm forecasts to assess and make adjustments to what the satellites, radars, and computer models are indicating. Additional emphasis on learning from past storm behavior and forecasting errors to enhance the art of weather forecasting has also contributed to dramatic improvements in short-term weather forecasting.
The main reason why short-term weather forecasts are still occasionally wrong is due to small-scale weather features (known as mesoscale weather features) that are not accounted for by weather prediction models that focus on larger weather features. As computing power and short-term weather forecasting models become more powerful and higher resolution, they will be able to account for mesoscale weather features, reducing the margin of error in short-term weather forecasts.
Medium Range Weather Forecasting Is an Area In Need of Improvement
Forecasting weather in the medium range, from three to ten days, has seen a dramatic improvement in the early part of the range (days three and four), but much less improvement in the later part of the range (days five through ten). Weather forecasters can now predict the weather within four days with same amount of accuracy as was achieved looking out two days in the 1980s (Source: NOAA’s second-generation global medium-range ensemble reforecast data), which is a dramatic improvement. This improvement can be attributed to more computer processing power and better medium range weather forecasting models, as well as improvements in the art of weather forecasting, as historical precedents and past errors are taken into account by weather forecasters.
Days five through ten in the medium range of weather forecasting are still quite problematic for weather forecasters, due to the highly variable nature of the atmosphere, air masses, weather fronts, and storm systems. Even a slight and unexpected change in a weather feature can be amplified over time and dramatically change the weather five to ten days out. It is this time-frame that often gives weather forecasters a bad image with the public. When people look at the weather forecast for their picnic on Sunday that is six days away, and see sunny and warm weather predicted, then have their picnic cancelled by rain when Sunday arrives, they remember the forecasting error and think negatively of weather forecasters.
While medium-range weather forecasting models can provide predictions for the five through ten day time-frame, the models are rarely in agreement regarding what the weather will be five to ten days out, as they make different assumptions about how air masses, pressure systems, and storms will develop over the later part of the medium range. Weather forecasters have to make judgment calls regarding which model is the most reliable given the current macro weather pattern and seasonal factors to come up with a forecast for days five through seven. Some forecasting models perform better than others during particular macro weather patterns and at different times of the year, which is where the art of weather forecasting takes over. Sometimes, weather forecasters just take a blend of what the medium-range models are saying to formulate their five to ten day forecasts, which leaves a lot of room for variability and unexpected outcomes.
Long Range Weather Forecasting Has Also Improved
Long-range weather forecasts are more general indicators about what the climate will be like in a particular region. For example, weather forecasters cannot say that it will snow and be cold in a particular location on a particular day three months from now, but by using a much improved understanding of what drives seasonal climate and weather patterns, they can make fairly accurate predictions regarding whether a particular region will have a cold dry winter or a warm wet winter.
The reason why long-range weather forecasting has improved in recent years is because weather forecasters have a much better understanding of the macro climatological forces that affect long range weather patterns. For example, weather forecasters have learned in recent years that the El Niño–Southern Oscillation (ENSO), which is the periodic heating and cooling of the Pacific Ocean that is commonly known as El Niño (warming) and La Niña (cooling), has a dramatic impact on jet stream patterns throughout the world. The state of ENSO affects both storm tracks and the movement of air masses, which in turn affects the seasonal climate and weather patterns in a particular region. With an understanding of the current and predicted state of ENSO, weather forecasters can use this information and information collected about other similar macro climatological forces to develop relatively accurate long-range weather pattern forecasts.
Tornado Warnings Have Improved Considerably
While weather forecaster may never be able to accurately predict exactly when and where tornadoes will strike on any given day, due to the extremely random nature of severe thunderstorms and tornadoes, their ability to warn the public of an imminent tornado touchdown or when a tornado has reached the ground has improved dramatically in recent years. Back in the 1980s, the average warning time for a tornado was four minutes. It is now thirteen minutes, with far fewer false warnings, which is crucial when providing people enough time to find cover from an oncoming tornado.
The additional tornado warning time and better warning capabilities are due to a sophisticated weather radar technology, called Doppler radar that was implemented during the 1990s and improved since its initial installation. Doppler radar allows weather forecasters look inside of thunderstorms to see wind patterns that are indicative of a tornado forming or a tornado on the ground. This additional insight allows weather forecasters to issue tornado warnings several minutes faster than was previously possible and reduce the number of false warnings.
Forecasting Hurricanes and Major Storms Has Had a Marked Improvement
Another area of meteorology that has seen a dramatic improvement in recent years is forecasting of hurricanes and major winter storms. Predicting where a hurricane is headed and where it will make landfall has improved dramatically since the 1980s. In 1985, a hurricane's path was forecast by the National Hurricane Center in a 366 mile-wide cone. As of 2008, a hurricane's path is predicted within a 111 mile-wide cone. This improvement in hurricane path predictions represent an impressive 67 percent increase in prediction accuracy. "This expected improvement will continue during the next decade or so as we fly aircraft inside the inner core of the storm to learn more information," stated Douglas Young, a meteorologist at the National Oceanic and Atmospheric Administration (NOAA) in Washington, D.C.
Only two generations ago, it was fairly routine for the Northeastern United States to experience a surprise blizzard or two during a winter season. Weather satellites and forecasting models were just not powerful enough to accurately predict the explosive dynamics of major winter storms that emerge from the Gulf of Mexico or develop rapidly off the East Coast of the United States. Over the past twenty years, both weather forecasting models and the art of weather forecasting have improved to such a degree that major winter storms are usually predicted with high level of accuracy. This new era of winter storm weather forecasting became obvious to the meteorological community in 1993, when the so-called “Storm of the Century”, which formed in the Gulf of Mexico, went up the East Coast of the United States, and dumped snow on an unusually large area, was accurately predicted by forecast models and weather forecasters days before it formed. In regards to current large storm forecasting, Young stated, "Our forecasts, on average for large storms, were about 90 percent accurate".
The Improving Trend of Hurricane Path Predictions
How Modern Weather Forecasting Technology Helped Forecasters Predict Hurricane Sandy Days Before Landfall
Hurricane Sandy, which affected the Caribbean and Northeastern United States in late October 2012, was predicted to make landfall on the East Coast of the United States six days in advance by the European Centre for Medium-Range Weather Forecasts (ECMWF) weather forecasting model. Although many other weather forecasting models disagreed initially with the ECMWF forecast and many weather forecasters were skeptical, the National Hurricane Center shifted the forecasted track for Sandy to an East Coast landfall four days before landfall. By then, the ECMWF model was predicting a New Jersey landfall, which turned out to be correct. The advanced warning of Sandy’s landfall in New Jersey undoubtedly saved countless lives and reduced property damage in the Northeastern United States, as people got out of harm’s way and prepared their properties for the storm.
An analysis of the Hurricane Sandy forecasts by NOAA came to some troubling conclusions about how the ECMWF weather forecasting model was able to predict Hurricane Sandy’s landfall along the East Coast of the United States accurately six days in advance. NOAA’s findings included a post-storm analysis by the European Center for Medium-Range Weather Forecasts, which ran the ECMWF weather forecasting model without the satellite data from the United States polar-orbiting satellites. The analysis found the ECMWF model sent Sandy out to sea when the satellite data from the United States polar-orbiting satellites was removed from the data set. The reason why these findings are troubling is that due to management problems and cutbacks in NOAA’s weather satellite program, the replacement of key polar-orbiting satellites may not occur before the current ones fail, leaving weather forecasting models and weather forecasters without the important data the satellites provide, which could leave the United States venerable to unexpected hurricane landfalls in future years.
A special thanks to Gary Arnold, fellow weather enthusiast, for helping with editing and adding to the content of this Hub.
How Accurate Is Weather Forecasting
How Accurate is Weather Forecasting?
Other Weather Articles
- How a Major Category 3 Hurricane Would Affect New Yo...
It is not a matter of if, but when, a major Category 3 hurricane will make a direct landfall in New York City and cause major devastation to the mid-latitude city. The City of New York and its surrounding areas would be wise to have contingency plans
© 2013 John Coviello