Vilhelm Bjerknes and his son Jacob began working through the initial problems of quantification. They came to the conclusion – and forced the meteorological establishment to come to it as well – that the existing models were completely unable to explain the observable life cycle of storms. In the 1920s they shook the establishment again by proving the importance of air-mass boundaries, which they called fronts, and by providing a conclusive mechanical explanation of the weather’s behavior which recognized storms as atmospheric waves.
Simple but functional numerical models might now be built, but a practical problem remained in a world before computers. The volume of calculations was so large that the process of forecasting was slower than the weather itself. To predict even the simplest storm only a single day ahead could take months of calculations. One British theoretician imagined a ‘weather factory’ in which 64,000 math workers, directed by the meteorological equivalent of a symphonic conductor, could barely keep up with the weather. And he vastly underestimated the problem. No wonder John von Neuman turned, in 1947, to weather forecasting as the perfect application for his new electronic computer. In April 1950, in Aberdeen, Maryland, under the guidance of von Neuman and meteorologist Jule Charney, the first successful numerical weather forecast was produced by a computer. Within a few years the computer results looked better than those of human forecasters.
That wasn’t saying much. Theoretical understanding of the weather had advanced in recent decades, but under the pressure of a daily schedule, the forecasters still had to rely on gut feeling – the old-fashioned experience of having seen some weather pattern before. They called it the ‘analog approach’ but might more honestly have called it educated guesswork. The results were poor for the obvious reason that no two storms are ever the same. A weather pattern today that looks like one last year will become something quite different by tomorrow.
Computers promised to solve those problems by bypassing the external features and working with the elemental equations of the weather process. The models were like electronic atmospheres within which mathematical storms could blossom. Meteorology suddenly knew no limits. With refinement of the models and closer weather observation, forecasters would be able to look a week, a month – why not an entire year into the future? The public would adore them. Until recently, faith in these principles was so great that you could almost overlook a persistent problem with the product: In practice, the computer forecasts still had to be judged and touched up with the old and unreliable methods. The models kept getting better, providing measurable improvements in the accuracy of short-range forecasts. But beyond one or two days, major inaccuracies crept in; and beyond five days, the forecasts proved nearly worthless. They still do.
The natural response was to blame the models or the sparseness of data. But already by the 1960s, an MIT meteorologist named Edward Lorenz had taken a different approach. He wondered whether the computers had so perfectly captured the functioning of the atmosphere that the forecasting errors were a manifestation of some unknown trait of the weather itself. Lorenz reduced a computer model to its essentials, then ran weather simulations from seemingly identical starting points. The results on the computer, as in the real weather, were wildly diverging patterns.
It was a fascinating observation. No wonder the new forecasts kept going wrong. Individuality appeared to be as fundamental to computer-generated storms as to those of the actual sky. Lorenz kept reducing the problem until he entered the realm of theoretical mathematics. There, he isolated the hidden and fundamental trait that he called chaos. You could say that pilots knew it all along: The weather is wide beyond the continents and wild beyond prediction. But Lorenz went farther than that. By separating chaos from its atmospheric effects, by giving it clear mathematical expression, he made one of the most important discoveries of our time.
And forecasters are bitter about it. The scientific recognition of chaos has only added to their personal gloom. Whatever hope they harbored of public adoration has now faded. They take it particularly badly that Lorenz was not awarded the Nobel Prize. Physicists and philosophers may glory in uncertainty, but meteorologists are scorned even in Scandinavia because by Monday they cannot be sure of Sunday’s weather. Chaos could let them off the hook, if anyone cared to let that happen, but of course no one does. The Weather Service’s recent lengthening of the forecast from five days to seven was the last cruel reflex of an old dream. The fact is that no improvement in the model and no amount of computer power can fix such a forecast.
So, quietly, there has been a shift. Gone is the talk of clairvoyance. The effort turns inward now to improved radars and automated observation posts designed not to extend the length of the forecast but to narrow its geographic scale. It may be possible, say the visionaries, to anticipate individual thunderstorms a few hours in advance. Chaos theory still permits that. The wild talk now imagines personal forecasts that will follow people across their daily navigational grids.
How tedious.
But rise above your disdain and remember Fitzroy. And pity the forecasters who are paid to predict the weekend’s weather. Lorenz has explained why they will get it wrong.
Richmond in the morning was cloudy. The Weather Service had an office at the airport, and so we walked in to sample the informed opinion there. One of us made a joke about the unexpected freezing rain in New York. The forecasters acted offended and refused to talk. We looked at their maps. The storm center had floated east from the Mississippi River to the hills of West Virginia. The warm front of the night before had buckled and jammed against New England’s stubborn winter. Just below it, the mid-Atlantic states remained locked in ice and would be for the rest of the day. However, a classic cold front now curved south from the storm center along the western slope of the Appalachians, out across Jackson, Mississippi, and into the Gulf of Mexico. The front was moving slowly and packing weather: Reported ceilings were low, and radar showed lines of building rainstorms that were especially active near Knoxville, Tennessee. We decided to try the storm at its roughest.
After climbing from Richmond through warm cloud layers, we flew past Raleigh to the point over coastal North Carolina where we could turn west and head directly for the front. The route from there would take us over the heights of the Smoky Mountains, past Asheville, North Carolina, and down into Knoxville, where within the hour the airport had reported a 500-foot overcast and 2 miles visibility in heavy rain.
Westbound through the cloud tops at 8,000 feet, we crawled toward the mountains against sixty-mile-per-hour head winds. A dark wall marked the front ahead. As we approached, we made out bulbous and hooked cloud shapes indicating power and turbulence. Onboard the airplane we had a device, known as a stormscope, that plots the direction and distance to lightning strikes. It showed the first ones now, ahead about fifty miles. Lightning means thunderstorms. We strapped down hard into our seats.
The pilot’s claim to have known always about chaos is of course a sort of vernacular conceit. Edward Lorenz discovered and could describe a core element in the very functioning of history. Pilots discover the immediate sky and can describe only their current confusion. That may explain the sensitivity of the Richmond forecasters, who, working at an airport office, must have been tired of listening to pilots’ ignorance and presumption.
For example, this Carolina front into which we now prepared to fly was for meteorologists on the ground nothing dramatic or difficult to understand. A wedge of cold dry air was driving under warm moist air and forcing it up to altitudes where it cooled and condensed into cloud droplets, which collected into rain. Within the clouds, the condensation released the moist air’s store of latent heat in a molecular process opposite to that of evaporative cooling. The released heat caused the air to rise higher and condense faster, which in turn released more heat. Accelerated in places by uplift from the mountains, the chain reaction raged along the entire length of the front.
When later Paul Kocin studied my weather map, he said, ‘Yup,
a cold front.’ He glanced at me with the disappointment I had come to expect. ‘But you know, it really wasn’t anything unusual.’
He was right. It was like flying into a slow, sustained explosion. The rain pounded at the windshield and tore paint from the wing’s leading edges. Turbulence slammed the airplane from above and below, rocked it onto its side, stretched us against the seatbelts, and at times shook the instrument panel so violently that we had trouble focusing on the instruments. But that makes it sound worse than it was. You can fly an airplane like you ride a horse, refusing to be intimidated. It is one of the inside tricks of storm flying – based on the knowledge that airplanes are the most weather-worthy of vehicles, strong and capable beyond the imagination even of their pilots.
Passengers are not easily reassured because in the alien world of flight they lack a useful sense of scale. But here is some advice for nervous airline riders: The only reason to grip your seat is to keep the seatbelt from bruising your thighs. It may help in reverse to know that the trickiest turbulence is not rough at all but is the layered shift in the wind known by pilots as shear, which happens close to the ground and causes a notoriously seamless sink. There are solutions for that, too.
It was just as well that we had no passengers for Knoxville. Our ride kept getting rougher. The pilots with me were apparently unafraid. The man at the controls fist-fought the airplane with determination. I watched him carefully, alert for fatigue or any slackening of control. We measured the head winds in places now at eighty miles an hour. The clouds were swollen with rain and so dark that we switched on the cockpit lights. The weather at Asheville, one of our escapes if the weather became unflyable, had dropped nearly to the limits of the instrument approach.
We considered diverting while we still could but got a report that Knoxville, ahead, was holding steady. We continued westward. As we crossed the highest mountains, the stormscope showed lightning strikes off the right wing and ahead to the left. We heard the crackling on the radio. Lightning is an electrical reckoning. When it occurs between oppositely charged clouds, or between the clouds and the ground, airplanes will not get in the way. But airplanes do get hit by lightning – or a mild form of the same thing. As they fly through heavy rain or snow, they build up a static charge which normally bleeds off through metal wicks attached to the trailing edges. But if the wicks can’t keep up, the charge builds until, with a flash and a bang, a stroke of lightning takes care of it. The stroke will usually not ignite the fuel, but it may damage a circuit or – as occurred to me during that first winter of my cargo flying – burn or blow small holes in the airplane. Airplane repairs are expensive. So we grew concerned when, with lightning around, the rain turned to sleet and our radios started hissing with static and began to fade. It was a sign that the wicks could no longer keep up.
The static charge grew so large that eventually it knocked out our ability to talk to air traffic control. This was less of a problem than it might seem, since there were no other airplanes out there and the controllers could see us on radar and knew our intentions. The static charge continued to build. We grew more concerned when next we lost the navigational radios as well, one by one, until for a while we banged through the crashing rain and sleet by compass and clock alone.
It would be ridiculous to say that we were not then afraid. The sky had gathered around us in a malicious display of its power. But the control of fear is a necessary part of the inner work of flight, and one of the reasons no doubt that each of us was there. We still had an escape route open to us – a 180-degree turn and a retreat downwind and down-weather, on compass alone if need be, into the warm coastal air of the Carolinas. We talked it over. We decided to keep going.
I tuned one radio to an inactive frequency and rhythmically keyed the transmitter, hoping to spark a discharge to help the static wicks dissipate the airplane’s electrical shield. I don’t know whether the trick worked, but we avoided any damaging strikes. Past the mountains the weather eased, the static wicks did their job, and the navigational radios sprang back crisply. Soon afterward we made contact with a Knoxville controller, who mentioned laconically that we seemed to have come through ‘some pretty good cells.’
We had not seen the ground now for several hours. Angling for the approach into Knoxville, we descended rapidly through continuing rain and cloud. Ten miles north of the airport we hooked onto the instrument approach and began to ride its electronic beams like a downsloping rail to the runway. Five hundred feet off the ground, the clouds hinted at the green fields. Seconds later we caught the motion of trees sliding by below. The runway’s pulsing approach lights materialized ahead, floating in the mist. We emerged from the clouds, crossed the runway threshold, and touched down on the glistening runway.
That afternoon we continued down the front in rain and ice, and spent the night in Montgomery, Alabama. By morning, the storm center had crossed Nantucket and was heading into the North Atlantic, where within a day, deprived of its sustaining temperature differences, it would quietly collapse into the Icelandic Low, the birthplace of European weather. The end of its weakened cold front curved through central Florida. We flew to Orlando and made an approach through gentle clouds that now seemed like old friends. Then we headed west and along the Gulf Coast finally flew into the clear skies behind the system. In mid-flight we radioed for a weather briefing. Already a powerful new system was bringing blizzards to the Great Lakes. The news encouraged us not to regret the old storm’s passing. We turned north, in search again of our landscapes of solitude.
5
Slam and Jam
At this point in our starting evolutionary development, hardly more than a century of human flight, the sky has become crowded with our egocentric species, each of us on self-important missions, and wanting to go first. Nowhere is this more true than in the New York area, where the runways of Newark’s International Airport, for instance, now rank among the most heavily used in the world. Night after day, in good weather and in bad, the airplanes bear down on them. Their traffic is relentless. Drivers on the adjacent New Jersey Turnpike can count on the distraction: the procession of lights inbound to the runway, the graceful touchdowns, the taxiway parades, the miraculous, banked, nose-high departures. The equipment out there is complex, capable, even exotic, but it is the sheer quantity of it that commands our attention. The big orange radar that stands beside the turnpike can never stop turning.
The radar sweeps the sky beyond the eye, keeping watch on the intertwined arrivals and departures from New York’s three major airports. LaGuardia and Kennedy each handle a third of a million flights annually, and Newark, which used to be called Sleepy Hollow and is still thought of as a lesser airport, is in reality even busier than the others, accounting for another half-million flights a year. Because jets fly fast and turn wide, these three airports, which once stood distinctly apart, now lie atop one another. Adding to the tangle, each of the smaller airports – White Plains, Teterboro, and Islip, to name just three – produces its own heavy flows of arrivals and departures. And just overhead pass the en route flights cruising to and from Boston, Philadelphia, and Washington, D.C. The result is the most congested air space in the world, a chunk of sky through which much of American air traffic daily flies, a special place where the usually reliable ‘big sky’ theory of collision avoidance simply does not apply.
It was because of the congestion that I chose Newark to flesh out my impression that, despite what the public has been led to believe, no immediate danger lurks within the system of air traffic control. This may come as a surprise. If there is one thing that nearly everyone can agree on, it is that air traffic control is critical to the safety of flight. Decades of moviemaking and superficial reporting have contributed to the idea that controllers ‘guide’ airplanes, that the task allows no room for error or inattention, that controllers must have superhuman reflexes and cool nerves, that only split-second timing and fast computers keep disaster at bay, that passengers’ lives hang in the balan
ce because of old and unreliable equipment – and that the work of air traffic controllers as a consequence is impossibly burdensome. These images jibe so neatly with people’s sense of helplessness in flight that they have acquired the force of an accepted reality and have become the necessary starting point for any conversation about air traffic control.
But that reality is a myth. Controllers do not puncture it because it gives them leverage with the public and because they themselves have come to believe in it. To be sure, the potential for collisions exists, all the more so in the high-pressure environment of the big-city sky. Concern for safety is the bottom line of all aviation – in the control room as well as in the cockpit. But even in a place like New York, the controllers’ real concern is with a set of work rules which operate narrowly atop the nearly absolute safety provided already by pilots and aircraft designers. Mistakes by controllers have led to accidents, but only as one link in a chain of failures. Air traffic control’s main function is to provide for the efficient flow of traffic and to allow for the best possible use of limited runway space – in other words, not to keep people alive but to keep them moving.
Like jugglers, controllers are practiced at handling constellations of flying objects. There is an important difference, though. When jugglers get distracted, their constellations tumble to the ground, but when controllers make mistakes, or lose their radars or radios, their airplanes continue to fly. Even if these jugglers were to stop suddenly and walk away, the elements of their constellations would on their own eventually slow down, take in the situation more or less calmly, and by following a variety of well-accepted procedures discover places where they could land softly. Imagine juggling on a low-gravity planet using smart balls that knew how to navigate and to talk to one another and could find ways not to collide.
Aloft Page 11