It's a dark and stormy night, 8.5km over the Midwest. Just after 10:30 PM, I'm standing aft of the cockpit of a NASA DC-8, while lightning flashes outside the cabin windows. A team of scientists recently took to the skies over Kansas and Nebraska to study how nighttime thunderstorms form over the Great Plains. The mission was called Plains Elevated Convection at Night, or PECAN. I was along for the ride.
Navigator Walter Klein points to a radar image on his tablet: a storm cell centred just over Kearney, Nebraska. On radar, the storm looks a little like a comet, with a core of red, orange, and yellow plunging toward the southwest, trailed by a tail of green curving off toward the northeast.
"That's a big thunderstorm," says Klein. "That will tear the wings off an aeroplane." He's grinning while he says it. Of course.
We've been chasing this storm, in an indirect, zig-zagging way, since about 8 PM, when we took off from a little municipal airport in Salina, Kansas. Deviating from the neat grid of our original flight plan, we've been flying back and forth in a series of east-to-west legs across the storm's path, which have brought us steadily closer to the towering black clouds full of constant, strobing lightning.
Instruments on board have been gathering data about the air flowing into the storm, because the moisture in that inflowing air triggers the storm's growth, keeps it going, and fuels its track across the plains.
That phenomenon is what scientists from around the country have come here to study, with a fleet of aircraft and an array of instruments. NASA is even using this mission to test some of those instruments for use in space. This 46-year-old plane is carrying tomorrow's weather satellites.
PECAN's goal is to open up what some meteorologists call the last frontier of weather prediction: big storms, like this one, that form at night over the Great Plains. They account for most of the region's rainfall, but we don't yet understand them very well.
A Meteorological Mystery
The science of weather prediction has made huge advances in the last few decades, but it's still hard to predict the nighttime storms of the Midwest. They work a little differently than other storm systems. Most thunderstorms form when warm, moist air from Earth's surface is forced to rise, either by a landform like a mountain, or by a "bump" from cooler air, like a cold front. Warm air is lighter than cool air, so once the warm air starts rising, it keeps going up, and it takes moisture with it. As the air rises, it starts to cool, and the moisture it contains starts to condense into clouds.
That process is pretty well understood, and but the nighttime storms that form over the Great Plains start differently. They're formed by something that happens at much higher altitudes.
"What often happens in the Midwest at night is that you have this, what's called a low-level jet, which is basically a stream of very fast-moving air that moves from the south to the north that carries a lot of this water vapour coming out of the south. And so where that low-level jet then interacts with another disturbance in the atmosphere, often means that's where the convection or that's where the storms will form," explained Richard Ferrare, a meteorologist from NASA's Langley Research Center and one of the principal investigators on the flight.
Sometimes, these nocturnal storms form in groups, which scientists call mesoscale convective systems. These groups of storms can last several hours and cover a few hundred miles of territory, and they're often very intense storms that dump large amounts of rain and cause deadly flash floods. Better prediction could help, but that requires a better understanding of nighttime convection in the skies over the Midwest.
In the Air and On the Ground
And that's why we're here. Tonight, there are a handful of scientists aboard the DC-8, running four instruments. That's actually a pretty small load for this aircraft, which carries 40 or 50 instruments and as many researchers on some of its missions.
With half the seats removed, leaving big empty patches of carpeted floor spaced between instrument racks and seats, it feels more like a lab than a plane -- until a moment of turbulence makes the floor shift under your feet, or until you step into the small bathroom in the back. The constant strobing of lightning in the dark outside the windows give the whole thing a mad science ambiance.
While DC-8 makes passes back and forth across the storm's path, it's not alone in the sky. Other aircraft, including a KingAir from the University of Wyoming, are much closer to the action. A P-3 Orion -- NOAA's famous "Hurricane Hunter" -- is flying through the swath of rain trailing the storm: the part that shows up as the green tail of the comet on Klein's radar display.
Keeping track of the storm on radar is pretty easy. Every instrument rack has a pair of monitors, and at any given moment, a handful of them show the radar image of the storm, overlaid with our baffling, twisty flight path. We're required to stay several miles away from the most intense areas of the storm, the yellow and red patches on radar, leaving the really rough flying to the KingAir and the Hurricane Hunter.
Peering out the window at the clouds and lightning, it's sort of nice to know we're not alone up here. 8500m below, even more researchers are busy on the ground. At six fixed sites around the region, researchers are launching weather balloons, or radiosondes, every two to three hours. Six mobile facilities are also launching a steady stream of radiosondes, and trucks with lidar, Doppler radar, and other instruments are also in position as the storm passes overhead.
All told, there are over 100 instruments on the ground and in the air, from 27 universities and three federal agencies. "Each instrument is capable of giving us an additional piece of the puzzle, so there's not one instrument that can give us the complete description of the atmosphere," says meteorologist Tammy Weckwerth of the National Center for Atmospheric Research, one of PECAN's lead scientists. "There's not just one agency that could put together a successful PECAN project, so we really needed the collaboration."
PECAN is a huge, complex operation, but like all good science, it began with an unanswered question. "It started with just a handful of scientists meeting in the hallway, basically, to talk about, why is there such a nighttime precipitation max in the central US," says Weckworth, "and of course, we have some hypotheses and theories, and but we don't really know."
Tomorrow's Weather Satellites
By about 1:40 in the morning, the comet shaped storm is losing some of its structure; it's shaped more like an anglerfish now, still diving southwest, and patches of showers have cropped up all around it, showing up as pale blue blobs on the radar. Those blue blobs are good news for RainCube, an instrument package about the size of a cereal box that measures rainfall. It's one of the future satellite payloads being tested on PECAN, and NASA plans to put it into orbit in the next few years.
RainCube is a six-unit cubesat -- large by cubesat standards, but still much smaller, and much cheaper to launch, than today's satellites. That's the real advantage, according to Jet Propulsion Laboratory engineer Brad Ortloff, because it means NASA could launch several RainCube satellites at a time. There are already satellites in orbit that measure rainfall in basically the same way as RainCube, but those larger satellites usually only get one pass over a storm system before the storm dissipates, so it's hard to track the lifetime of a storm system in much detail. A constellation of cubesats could pass over the storm, one after the other, and gather data about how the storm lives and dies. One day, that kind of data could help improve the models that predict storms.
"With the DC-8, we can show that we can operate from an airborne platform," Ortloff tells me. Operating in space comes with its own set of challenges, of course, but airborne testing is a good intermediate step between ground tests and orbit.
I've caught him just after a busy few minutes of data collection as the plane passed over a lake. On the monitor in front of him, a yellow line marks where the signal bounces off the ground below us -- it's flat, because this is Kansas, after all. That line gets much thicker and brighter over the lake, a reading that the researchers use to calibrate RainCube. In an image of a rain cell, the ground is a bold yellow line against the blue background, while the rain cell is an amorphous shape, solid at the centre but wispy toward the edges.
Sometime in the next couple of years, a single RainCube satellite will blast off into near Earth orbit as a technology demonstrator. NASA hasn't set a timeframe yet for launching the full constellation of operational cubesats, but the agency has shown increasing interest in cubesats in recent years, mostly because they're smaller and less expensive than traditional full-sized satellites, and the form factor is pretty versatile.
"JPL and NASA are getting very excited about cubesats," Ortloff says. In fact, on earlier flights, PECAN tested another future cubesat called Microwave Atmospheric Sounder on a Cubesat, or MASC, a millimetre-wave radiometer that measures temperature and water vapour profiles in the atmosphere.
Laser Beams in the Sky
One of the challenges of flying a mission like this is that you have to make sure not to hit other planes with the big laser beam. As the DC-8 flies along, an instrument called LASE fires short laser pulses at the atmosphere above and below the plane, and the light that reflects back gives the scientists a look at the sky. Sadly, the lasers are invisible, and there's no "pew pew" sound effect, but the beam could still blind pilots of other aircraft if it happens to hit them in the eyes.
"Air Traffic Control tries to keep aeroplanes out of here, and our pilots are always looking out for aeroplanes," Ferrare tells me, adding, "It's an extremely, extremely, extremely small risk." LASE fires straight up and straight down, and it's pretty hard to look straight up from an aircraft cockpit and even harder to look straight down (glass-bottomed aeroplanes haven't really caught on yet). Just to be safe, if there is an aircraft directly above or below the DC-8, the team shutters the laser until it's clear, and the flight crew manoeuvres carefully. "When we turn, we try to make sure that the turns are shallow, ten degrees or less, so that the laser's not going off in some direction," says Ferrare.
LASE fires two laser pulses at a time: one that's absorbed by water vapour, and one that's not. When the light gets scattered back from water vapour and aerosol particles in the atmosphere, scientists use the ratio of those two wavelengths to get a profile of how much water vapour is in the air. "That's important, because water vapour is the major driver for these severe storms," says Amin Nehrir, a research scientist at NASA's Langley Research Center.
The resulting view looks like a curtain hanging from the lower stratosphere down the ground, running along the plane's flight path. "This is really one of the only instruments in the world that has that capability," says Nehrir, but LASE is over 20 years old. Originally designed in 1994 for NASA's ER-2 high altitude jet, it was -- like RainCube and MASC -- meant to be a technology demonstrator for a future satellite technology, but that never came to pass. Instead, NASA retrofitted LASE to fly on the DC-8, mostly by adding the upper panel, allowing the laser to fire upward as well as downward, which, says Nehrir, "allows us to look deep into the upper troposphere and lower stratosphere where small changes in water vapour can really affect our climate."
But NASA hasn't abandoned the idea of one day putting LASE in space. According to Ferrare, "Amin [Nehrir] himself is designing, hopefully, a successor to this instrument that will be much smaller and fly not only in the DC-8 but also in other aircraft, and hopefully as a precursor to something that will go in space."
Flying the Stormy Skies
Even in the early hours of the morning, the lightning show outside is intense. We're flying level with the storm, and it looks much bigger from up here, at relatively close range, than it would from the ground. Everything outside the windows looks black, until a flash of lightning illuminates a mountain range of clouds for a split second, just long enough to give an impression of vastness and power. And then it happens again, every second or two, flickering through the storm.
In the forward half of the plane, they have cut the lights, making it easier to watch the show. I've been glued to the windows since take-off, bouncing from one side of the plane to the other as we turned, pressing my face against the glass like a small child in order to cut out some of the glare from the cabin lights. Even after six hours in the air, it's mesmerising. Even PECAN's crew members and scientists pause to glance out the windows on their way to and from the restroom or the small galley in back, which offers a microwave, an esky and a coffee pot.
Coffee is essential, since the flight won't land until about 4:00 AM. Tonight, the crew ditched the original flight plan right after take-off, because the storms they expected had fizzled, and this new one looked promising. That's part of the challenge of studying storms that meteorologists don't really know how to predict yet -- it's hard to make sure everyone's going to be in the right place at the time to get the data.
And that's why it's so useful to chase these storms in a plane. "We did some ground tests where we just kind of went out and tried to find some rain, but unfortunately, by the time we got everything set up, the rain we were hoping to capture had dissipated," says Ortloff. "So the great thing about the DC-8 is it will take us to wherever the rain is."
Of course, making that happen for the scientists keeps the flight crew on their toes. Two mission directors sit aft of the cockpit, in swivel chairs in front of a huge pair of panels, each covered in at least a hundred switches, knobs, and lights. They're the interface between the flight crew up front and the scientists in the back.
Each instrument's team talks with Ferrare about where to fly next, and he relays those requests to the mission directors, who in turn talk with the pilots and navigators. The navigators have plotted out 38 course changes on tonight's flight -- and for one of them, this was scheduled as a training flight. He's getting plenty of practice.
It's up to the navigators and pilots to talk to Air Traffic Control and get clearance for course changes. Sometimes that's tricky earlier in the flight, but by the early morning hours, there's little traffic to worry about, according to mission director Matthew Berry.
"And the areas that we want to work are where everybody else is trying to avoid," he says.
Just the Beginning
At 4:00 AM, when the DC-8 touches down in Salina after a long night of science, everyone spills out onto the tarmac and into the little terminal to debrief. Lightning flickers on the horizon; PECAN may be finished for the night, but the storm isn't.
And the PECAN team will be doing it all over again in a few hours. Forecasters will start working on the coming night's weather forecast around 8:00 AM, and at 11:30 AM, the scientists will gather to hear the forecast and plan the night's mission.
Of course, the fieldwork is just the beginning. "It's basically an intense six-week period this year, but the analysis of the data are going to take several years," says Ferrare. For the first year, the 70 scientists involved in PECAN will basically have dibs on the data, but after that, it will be open to the public. The researchers will share the data with weather modelers, who will compare their models to what the storm actually did, and then see how differently their models behave when they plug in the new data from PECAN.
"It helps them try to figure out if there's something missing that they're not representing, either a measurement that would be required, or maybe their model isn't representing the measurements correctly, and they need to fix something in the model or have a better representation of some process in the model," explains Ferrare.
That work, ultimately, may lead to better predictions of flash floods, hail, dangerous winds -- and even the familiar bane of the Midwest, tornadoes. "First we need to collect the data, improve our understanding of these nighttime thunderstorms, and that will lead to improving the skill of the numerical models and forecasts will improve after that," says Weckwerth.