For many of us, inaccurate forecasts do little more than ruin a picnic or sully a recently washed car. We’re lucky. Weather is a sideshow to our daily lives. For others, though, advance knowledge of the weather directly affects their livelihoods, from farmers to construction workers to airline pilots. But there are times when weather affects us all, especially instances of severe weather like Hurricane Sandy. There, accurate forecasts can save lives, and those forecasts rely heavily on satellite data.
Unfortunately, in just a few years, part of our weather satellite system that is vital to forecasts will very likely go blind.
Berrien Moore, director of the National Weather Center, and many other experts who spoke with me cite Sandy as a prime example of how vital satellite data has become in weather forecasting and how terrible it would be if we lost any part of it. Typically, hurricanes that make their way up the East Coast end up veering back out to sea. But Sandy didn’t, and thanks to a European weather model that deeply incorporated satellite data, forecasters were able to predict its sudden turn west into the coast.
Though damage from the storm was still extensive, without satellite data the aftermath could have been far worse. In the months following Sandy, meteorologists reran the European model, but withheld information from weather satellites. It failed to predict Sandy’s left turn. “Satellites allow us a complete picture of the atmosphere,” says J. Marshall Shepherd, president of the American Meteorological Society, something weather balloons and Doppler radar cannot provide. “The satellite data is critical,” he adds.
“Sandy was just a warning,” Moore says. Without satellites, “that could have been a fundamental disaster.”
To appreciate how indispensable weather satellites are to forecasts today, we have to go back to the 1950s. Then, the science of weather prediction was just a couple of decades old, and incremental improvements had led to 24-hour forecasts. But the pace quickened when, in early 1959, the United States Navy launched Vanguard 2, the world’s first weather satellite. Though it didn’t return useable photographs, it paved the way for TIROS 1 a little over a year later. The squat cylinder beamed back images of cloud cover for a brief 78 days. Meteorologists couldn’t have been happier. Since then, our forecasts have only gotten longer and more reliable, so much so that many of us take them for granted.
“Much of that advancement is due to the assimilation of satellite data,” Shepherd says. “Take that away, are we going to go backwards in terms of our skill in weather forecasting? I think that’s a real possibility.”
Satellites Near and Far
Currently, the U.S. has 24 Earth-observing satellites in orbit. Their missions are widely varied, covering more than just weather. There are satellites that monitor tropical rainfall (key in our understanding and prediction of hurricanes), keep an eye on land-use change (important for urban development and habitat conservation), and observe the ice sheets that cover Greenland and Antarctica (an indicator of sea level rise). Each different program provides important data on our planet’s many systems.
The U.S. Earth-observing system consists of satellites in two basic orbits, geostationary and polar. Geostationary satellites sit high above the Earth at an altitude that allows their orbital period to match the rotation of our planet. One geostationary satellite can continuously observe one part of the Earth, sending updates as often as its mission requires. The GOES weather satellites, for example, beam back periodic updates which produce the stuttering images of cloud cover found on weather websites and TV news broadcasts. Such a high frequency of updates makes geostationary satellites valuable, but their distance from the Earth—about 22,300 miles—decreases their resolving power. They’re simply too far away to record in great detail.
That’s where polar orbiting satellites come in. They complement geostationary satellites by trading frequent updates for sharper images. Polar orbiting satellites sit much closer to Earth—generally about 500 miles up—and complete a trip around the globe every 100 minutes or so. Most Earth-observing satellites, weather and otherwise, fly in polar orbits. With the right sensor, they can image the entire planet twice a day, once on the day side, once on the night side.
Their role in meteorology is unquestionable. “The polar orbiting satellites give us the ability to do long term weather forecasts,” says Jane Lubchenco, outgoing administrator of the National Oceanic and Atmospheric Administration. “Currently, NOAA’s forecasts go out 7 to 10 days. If we don’t have a polar orbiting satellite, we would do 2 to 3 day forecasts. That’s a huge difference.”
Since the 1960 launch of TIROS, there has never been a moment when the U.S. hasn’t had a weather satellite in low Earth orbit. The latest is Suomi NPP, a next-generation satellite that was originally a testbed for advanced sensors and was only intended to last three years. Because of problems with its now-cancelled replacement, the National Polar Orbiting Environmental Satellite System, or NPOESS, it’s now being asked to serve for five years. Yet even that won’t be enough. It’s likely that sometime in 2016, for the first time in over 50 years, the U.S. won’t have a polar orbiting weather satellite.
The trouble started on May 5, 1994 when, with little fanfare, President Bill Clinton signed one of the 80-some directives he would issue during his time in office. The Soviet Union had collapsed two and a half years earlier, and the federal government was looking to trim budgets inflated by the Cold War. Clinton’s order directed the Department of Defense and the National Oceanic and Atmospheric Administration to merge their weather satellite systems.
Administration officials hoped the merger would save hundreds of millions of dollars. While the arranged marriage between NOAA, NASA, and the Air Force may have sounded good on paper, the way it was carried out “was flawed from day one,” says Tom Young, former director of NASA’s Goddard Space Flight Center and chairman of an independent review of NPOESS.
By 2003, the Government Accountability Office had grown concerned over NPOESS’s progress, or rather lack thereof. They issued a report, something that would happen more or less annually for the next several years. David Powner, director of IT management issues at the GAO, authored those reports. “It was very common for us to go up and tell Congress a delay was forthcoming and that there was an overrun somewhere,” he says. “Very common.”
Most of the technical delays can be traced to development of the Visible Infrared Imaging Radiometer Suite, the satellite’s most important sensor and an incredibly complex piece of equipment. It was to be built by the Santa Barbara Research Center, which like many defense contractors, was many layers removed from the top. A subcontractor to subcontractors. That sort of isolation doesn’t always foster efficiency or accountability.
Making matters worse, the NPOESS program manager was an Air Force officer with no experience working on space projects, according to a later report commissioned by NOAA and confirmed by a former NASA official. Those circumstances, combined with technical delays wrought by the complexity of the device, meant VIIRS wouldn’t be completed until December 2009.
Each VIIRS delay caused a domino effect. Because the many sensors on NPOESS needed to be integrated into one coherent system, a single missing piece set back the entire project. Other sensors were troublesome, and some were eventually cut from the manifest. But none were as problematic, or essential, as VIIRS.
The Golden Age
Meanwhile, other Earth-observing satellite systems were firing on all cylinders. Between 1999 and 2004, nine satellites were successfully launched. Though many were only expected to last three to five years, all but two remain in orbit today. And that doesn’t count earlier satellites that are still beaming back data.
We are in the golden age of Earth observation. But all golden ages eventually come to an end.
One of the darlings of the Earth-observing community is Landsat. Since its first satellite was launched in 1972, the program has never had a gap in coverage. Some of its success is due to luck. Landsat 5 was only decommissioned earlier this year after nearly 29 years in operation, giving it the record for the longest running Earth-observing satellite. But the rest is because of commitment and good planning. Landsat 8 entered orbit earlier this year, and there are plans on the table for 9 and 10. Thanks to Landsat, we know exactly how the Earth has changed, in two-week increments, for the last 40 years.
Other successes abound. MODIS sensors aboard the twin Terra and Aqua satellites continue to deliver daily updates of the Earth’s surface years beyond their expectations. EO-1, launched in 2000, was only anticipated to last for 12 months. Today, it’s still recording over 200 different wavelengths of light reflected off the Earth’s surface. And NASA would have been pleased with GRACE if it had lasted the five years it was intended. Instead, it’s turning 11 this year and still helping scientists track the movements of water, ice, and magma to better understand ocean currents, plate tectonics, and more.
However, the future of our Earth-observing capabilities is anything but robust, even when you don’t consider the looming gap in weather satellite coverage. Of the 24 satellites in orbit today, it’s expected that just six will be working by the end of the decade. Two recent launches, the Orbital Carbon Observatory and Glory, failed to reach orbit due to rocket malfunctions. Only six satellites are scheduled to be launched between 2014-2020, not enough to keep up with the rate of attrition. We’re living on borrowed time.
We’ve known this for a while now. In 2007, the National Research Council issued a report on the overall status of the U.S. Earth observation system. What they found wasn’t promising. “The extraordinary U.S. foundation of global observations is at great risk,” they wrote.
Today, more than five years later, the situation hasn’t improved. Of the 15 satellite missions reviewed for that report, “I believe two of those are actually on track,” says Dennis Lettenmaier, a hydrogeologist at the University of Washington and member of the NRC committee. Budget shortfalls have jeopardized nearly every program. “Notionally, at least, there was enough money to do all those things, so it wasn’t supposed to be about there not being enough money,” he says. That changed when the economy soured. When NASA started running short on funds, it went looking for programs to cut. Satellites that were many years away from launch got the ax. “NASA basically just dropped them all,” Lettenmaier says.
The other problem lies in the distinction between research and operational satellites. Typically, NASA will launch a satellite as a research platform and distribute the data to scientists. Eventually all satellites stop working, but the data doesn’t necessarily stop being useful. Someone then has to launch a replacement.
“NASA says, that’s not research anymore, that’s operations. Somebody else needs to do it,” Lettenmaier says. “The problem is, who’s the somebody else? Well, the somebody else they usually look to is NOAA, but NOAA never seems to have a budget for these things.”
That was partially because of NPOESS. Its numerous cost overruns and missed deadlines were straining NOAA’s budget. The program was turning into an albatross.
In 2006, Congress forced NPOESS to restructure, but it wasn’t enough. “It was really too late,” says Young, the former Goddard director. “They had done too many things that had the program in jeopardy by that time.”
NOAA and NASA soon became adamant that the Air Force leave the program, and eventually that’s what happened. In early 2010, the Obama administration annulled the three-way marriage. NPOESS was scrapped and reborn as the Joint Polar Satellite System, which was was run by NOAA and NASA alone. The Air Force is back at the drawing board developing a new, independent platform.
To date, JPSS has launched one satellite, Suomi NPP, the sensor testbed mentioned above. It’s a carryover from NPOESS and the defunct program’s only tangible result. It’s also our only weather satellite in low Earth orbit. The longest anyone reasonably expects it to last is until early 2016, but it may stop functioning as early as next year—it was originally designed to operate for three years, not five. All satellites have a life expectancy, a standard to which they are engineered. After that time, probability suggests they’ll succumb to the harsh conditions in space: extreme heat and cold, constant radiation, solar flares, space debris, and so on. Some, such as Landsat 5, operate for decades beyond expectations. But that doesn’t guarantee every one will.
Of course, everyone hopes Suomi NPP will be one of the lucky ones. The first true JPSS satellite won’t be operational until late 2017.
So what lessons can we learn from NPOESS?
For one, the program was based on the assumption that civilian and defense weather satellite systems could be merged. They couldn’t, at least not in a way that didn’t create unwanted compromises. For example, the military prefers observations in the morning, civilian meteorologists in the afternoon; the Air Force doesn’t need certain types of information, such as temperature and humidity, data that is critical to civilian forecasts; and civilian meteorologists don’t need to see clouds at night, but the military does.
Furthermore, the three organizations pooled their resources, but not necessarily their expertise. The interagency office that ran the program was “on an island by itself,” Young says, and missing on that island were NASA’s decades of experience putting satellites in orbit. As a result, there wasn’t anyone who could help the project when it started to founder. Worse, there wasn’t anyone who could realize there were problems in the first place.
Finally, these sorts of projects are never easy. “Space programs are very difficult,” Young says. “To make these things really work, you’ve got to have some strong, crisp leadership, and maybe we’ve gotten away from that in some of our space activities.” It’s also the tenor of our time, he adds. The space program doesn’t have the same popular or budgetary support it once did, making it difficult to lead with the boldness that once defined our space program.
A Perilous Future
With NPOESS dead and JPSS years from readiness, NOAA is scrambling. “We are looking at a lot of different options and doing everything possible to not have further delays,” Lubchenco says of JPSS. There’s also a study underway looking for ways to mitigate the gap. “We don’t have the report back from that yet, but we’re pulling out the stops to try to figure out how to deal with this while keeping our fingers crossed that maybe, maybe these satellites will last longer than the models suggest they will.”
In response to one of Powner’s GAO reports, NOAA has created a website where the public can submit mitigation strategies for consideration. Dozens have been submitted so far, but it’s anyone’s guess as to which are being seriously considered. No one in the agency has hinted that they’re close to a viable solution.
“We are in a precarious position. We don’t know how long NPP is going to last,” says Moore, the National Weather Center director. “I honestly believe we have become complacent about our infrastructure,” he adds. “I understand we’ve got huge budget issues, but these are fundamental things we ought to be doing. Weather waits for no man.”
It’s probably too late to completely eliminate the gap. We simply can’t build satellites fast enough, at least not with NASA’s or NOAA’s current budgets. If we’re lucky, one of NOAA’s mitigation strategies will pan out. “We’re usually good at solving problems once we decide they’re a problem,” Young says.
Hopefully, we’ve decided this is a problem, because it could become a very serious one in the near future. “We are not a weather ready nation,” Moore says. “And Sandy showed just how close we came. Suppose we had lost a satellite?”
We may find out yet.