Forecasts for severe weather, including hurricanes, excessive rainfall and storm surges, are expected to improve thanks to several recently announced upgrades.
In June, the National Oceanic and Atmospheric Administration announced that its flagship weather model, the Global Forecast System, was getting a “significant upgrade” to include a new dynamical core called the Finite Volume Cubed-Sphere — the first upgrade to the core in more than 40 years. At the same time, the National Weather Service office in San Juan, Puerto Rico, plans to expand coastal storm surge watches to include Puerto Rico and the U.S. Virgin Islands, while displaying excessive rainfall outlooks online.
NOAA also has upgraded the resolution of radar systems aboard hurricane-hunting aircraft so they can collect more data when they fly into a storm or a tropical depression in the western half of the Atlantic. “NOAA is driving toward a community-based development program for future weather and climate modeling to deliver the very best forecasts by leveraging new investments in research and working with the weather enterprise,” Neil Jacobs, Ph.D., acting NOAA administrator, stated in a press release.
The most noteworthy improvements to the GFS include the Geostationary Operational Environmental Satellite system, which is a joint effort of NOAA and the National Aeronautics and Space Administration. The GOES-16 satellite has an advanced baseline imager, which is a sensor that views the Earth with 16 spectral bands (compared to five on the previous generation of GOES). The bands include two visible channels, four near-infrared channels and 10 infrared channels.
Forecast models use each of the channels to indicate elements on the Earth’s surface or in the atmosphere. With the new satellite, the information coming into the models is not only more extensive, but also is redrawn faster. It provides three times more spectral information, four times the spatial resolution and more than five times faster temporal coverage than the previous system.
“We’ve found it useful to monitor situations and look for surface situations that have formed,” says Michael Brennan, branch chief of the hurricane specialist unit for the National Hurricane Center. “We have pictures of the clouds every five minutes. That allows us to animate the images. In previous satellites, we could only get images every 30 minutes, so going from 30 to five is a pretty big deal. We can actively monitor the structure of these systems so there is more data going into the analysis.”
In the hurricane hunting aircraft, NOAA has upgraded the on-board radar systems’ resolution so they can collect more data over the same area. The propeller-driven planes fly directly into a storm or tropical depression to collect as much real-time data as possible. They are limited to storms in the western half of the Atlantic.
Meteorologists say the Finite Volume Cubed-Sphere, FV3, will boost the Global Forecast System’s accuracy, but they add that the U.S. system is still less accurate than the European system. “My guess is it will still lag behind the European center’s medium-range model, but we are well on our way to catching up,” says Dr. Richard Rood, a professor who teaches climate change courses at the University of Michigan. “The European center can run models on high resolution with more bells and whistles than the U.S.”
The European model’s superiority showed when superstorm Sandy hit New York and New Jersey in 2012. “The European model predicted the westward turn of Sandy two days earlier, while the American model didn’t predict it until a couple of days later,” Rood says. “Historically, you can find more examples where the European center model does a better job than the U.S model, but Sandy was the archetype that triggered a lot of the current discussion.”
Dr. Jeff Masters, a meteorologist and co-founder of Weather Underground, which is part of IBM, says the European model’s superiority is based on the amount of information that goes into it. “They do such a brilliant job of initialization,” he says. “The Europeans benefit from having so many diverse people come in to contribute.”
A Closer Look
While the European and GFS models get all the attention, storms are forecasted using multiple models that gather as much information as possible. Regional models also add data as storms near landfall.
Globally speaking, Rood says, hurricane trackers should gain better forecasting with the technology upgrades. “A whole portfolio is used,” he says. “We know the intensity forecasts, which were quite bad in the old forecast system, are significantly improved.”
Ken Graham, director of the National Hurricane Center, says his teams are indeed working on improving storm-intensity forecasts, taking into account more water-temperature and wind measurements. “We’ve gotten better at tracking the hurricanes, but we still need more science to understand the intensity,” he says. “When we measure water temperature, that’s just for the surface. But what about below the surface? What is the difference between the surface and the hurricane stirring up cold water from below? That will require more study.”
Rood continues: “It’s not just the oceans but the winds aloft. We’ve seen hurricanes go over the warmest parts of the Gulf of Mexico and weaken. We need to understand more about the winds above.”
Rood says that in the old version of the Global Forecast System, readings for the sea level pressure of a storm’s eye were “too intense.” They will be more accurate in the new system, which upgrades the way cloud physics are calculated, as well.
Other data can come from taking a statistical approach to weather patterns from prior seasons, such as El Niño events in the Pacific, water temperatures in the Atlantic and vertical wind shear. “They don’t work out perfectly, but it’s purely statistical,” says Phil Klotzbach, a research scientist at Colorado State University. “We know what the conditions are now, and we look back at previous years and see what kind of storm activity we had.”
Looking Ahead Cautiously
Whether it’s the European or U.S. model, researchers analyze much of the same data. That data includes the El Niño Southern Oscillation, which measures water temperature near Chile; the temperatures in the Western Atlantic; and air pressure between the Azores Islands and Iceland, as well as other factors. It doesn’t take much of a difference in those numbers to have a major impact on conditions.
In the Atlantic this season, temperatures are about a half-degree Fahrenheit above last year’s and approximately 1.5 degrees below the average. Hurricanes need warm water, so the cooler temperatures are a good sign.
“Think about the difference between 32 degrees and 33,” Masters says, referring to water freezing or not freezing. Michael built strength so quickly last year because the Gulf of Mexico was 2 to 4 degrees above normal for the time of year.
Given the conditions now, Masters supports the early predictions for below-average activity in the Atlantic this year. “I want to see cold ocean temperatures, a lot of dry air from Africa and strong upper-level winds that create a shearing force that rips apart hurricanes,” he says. “So far this June, we have two out of those three conditions.”
The fact that this year has been the wettest January through May in history, along with an El Niño pattern in the eastern Pacific, should also have a dampening effect on hurricanes. Conversely, one area that could be hit hard this year is Hawaii because the water temperature there is hovering around 80 degrees. Masters says that 1 more degree is all that’s needed to help fuel a hurricane.
Looking at recent storms, Rood says that the new FV3 was running alongside the older version of the GFS last year when Florence stalled over North Carolina. “The FV3 predicted the stall better,” Rood says, referring to the storm slowing after it made landfall and dumping torrential rains that caused flooding throughout the Southeast. “The U.S. forecasts were quite good on the total amount of precipitation,” he says.
Not everyone agrees with the June forecast for a slow Atlantic hurricane season. “The May and June 2019 indexes will be the most negative we’ve seen since the late 1950s,” says Dr. Hal Needham, founder of Marine Weather and Climate, which provides data-driven flood risk analysis. “I expect storms to track farther west toward the U.S. coastline, rather than stay out over the Atlantic.” That means more storms reaching the southern part of the United States. Needham isn’t a proponent of climate change, but he acknowledges that the slowing of upper level winds is a “signature of climate change.”
Context and Climate
The technology and systems upgrades to improve forecasting come at a time when scientists are discussing whether the personality of storms is changing. Some researchers say storms are getting wetter and producing epic rainfall totals, but others point to more powerful storms becoming more prevalent.
A reduction in the temperature disparity between the poles and the tropics is often seen as a reason that storms stall. When the difference in temperature is greater — colder at the North Pole and warmer at the equator — the upper level winds that keep a hurricane moving are stronger. That creates classic wind- focused storms, such as Michael. When the Arctic is warmer, upper atmosphere winds weaken, and storms slow or stall. This is what happened with Florence and Harvey.
Fed by the warm waters of the Gulf of Mexico, Harvey stalled and flooded the greater Houston area, dumping 51.88 inches of water on Mont Belvieu, a record amount from a single storm in the continental United States, according to the National Weather Service. Experts also say the atmosphere contains more water vapor than it is used to, another indicator of climate change.
The North Carolina Institute for Climate Studies looks at historical data to determine what could be happening in the future. “The science is pretty clear, and we have evidence from a number of sources that climate change is definitely happening — and it’s human caused,” says Dr. Carl Schreck, a research scholar with the institute. “It’s not just the rising thermometers. It’s not just looking at the glaciers. We have a number of variables called climate change indicators that all point in the same direction.”
With Michael making landfall in the Florida panhandle last year at 160 mph and Irma slamming the Keys two years ago, it may appear that increasingly powerful storms are becoming more prevalent. However, only four Category 5 hurricanes have made landfall in the United States, according to NOAA: Michael, Andrew in 1992, Camille in 1969 and the Labor Day Hurricane of 1935.
Katrina hit New Orleans as a Category 2 storm, and Florence was downgraded to a 2 or a 1, which has raised questions about the effectiveness of the Saffir-Simpson scale, which looks primarily at wind speeds to assign a category rating. Calls are now being made for other scales to project a storm’s strength based on other factors.
“The rating system should stay as it is because it has been around a long time and people react to it,” Masters says. “I think we’re going to have to do something like that for rainfall, as well.”
Brennan says the idea is to provide as much information as possible on the various hazards that a storm is creating. “The hazards are what people really care about,” Brennan says, “so we try to have more hazard-specific information.”
More specific forecasts of that nature might help people understand that, for instance, rainfall or storm-surge risks may be rising even as wind-speed risks are decreasing. As the technology being used to forecast storms improves, more-specific forecasts become a possibility.
“The storms are going to come,” Klotzbach says. “We just try to be as accurate as we can be.”
This article originally appeared in the August 2019 issue.