Air traffic controllers are terrific people dedicated to helping pilots complete their flights smoothly and safely. That is a true statement until the smooth part, or maybe even the safe part, comes into conflict with the only absolute requirement in ATC, which is to separate airplanes under its control from one another by required minimum distance.
I have run up against this conflict between what pilots need and what ATC requires a number of times, and I haven't come up with a way to resolve the situation without one side giving in. And it is always the pilot who has to yield, because the force of law requires controllers to issue clearances that maintain minimum separation between airplanes under positive control. The same laws require pilots to fly the clearance issued. The only alternative is for the pilot to declare an emergency, ignore the clearance and deal with the consequences later.
My most recent encounter with the absolute requirements of controllers to keep their jobs by guaranteeing separation, and my requirement to keep the wings attached to my Baron, occurred while flying over the Indiana and Ohio border region early this summer. I was flying home from Wichita, Kansas, to New York with my usual planned fuel stop at Port Columbus International Airport in Ohio. The forecast for almost the entire route included the chance of thunderstorms.
Forecasting the development of thunderstorms can be either pretty easy or something that is only a little more than guesswork. The easy forecast is when there is a strong atmospheric feature such as a cold front moving into very warm and moist air. The warm, moist air is the fuel for a thunderstorm, and the destabilizing effect of the front is the trigger to start the storms building.
The tough forecast is when the air is warm and moist but there is no defined front or pressure change to destabilize the air and launch storm growth. The best the forecasters can do is tell us that the elements for thunderstorms are present, but they can't know where and when a storm may actually form.
I write down a host of data about every flight, including air temperature at my cruise altitude, on a form that I made years ago. It was a pleasant 27 degrees C at Wichita, and level at 7,000 feet I noted the air temp was 17 C. The air was perfectly smooth with only a thin layer of clouds far above me and at least 10 miles of visibility in haze. As you might recall from ground school, stable air cools gradually with an increase in altitude, while unstable air cools very quickly as you climb. The most stable air is an inversion in which the air temp aloft is warmer than it is on the ground. Cold air simply refuses to rise into warmer air, so with an inversion, colder air is trapped at the surface.
The most unstable air is, of course, a thunderstorm. The storm forms because the warm air near the surface lifts into the cooler air above. Moisture in the rising air cools, and this adds fuel to keep the convection process going. The freezing level is a cap to weak convection, but once the rising air bursts through the freezing level, it is off to the races and a cloud can grow into a full-blown thunderstorm.
As I droned across Missouri and Illinois, the weather changed little during a smooth ride in hazy skies. But as I crossed over into Indiana, I noticed scattered small, puffy clouds forming below me. This was not a good sign. Before long the puffy clouds were building through my cruise altitude at 7,000 feet. I looked at the air temperature, and it was 10 C. I had left the stable air behind. Why the air cooled by 7 C in a short distance was impossible for me to know, as it apparently was for the forecasters or they would have said to look for storms over Indiana but not Illinois, but they didn't.
The XM Weather display on the Garmin G600 was showing the forecast freezing level at 14,000 feet. Before long it was clear that some building cumulus had punched through that level and the moisture inside the clouds had formed into droplets big enough to be detected by my weather radar. This all happened in about 15 minutes of flying that covered about 50 miles of distance.
As is typical in this type of weather situation, only a small fraction of the clouds was forming into thunderstorms. The other clouds were topping out at 10,000 to 12,000 feet, and there was space to fly between most of them while only flying through the ones where I could clearly see a low top. The ride was bumpy even in the small clouds, with lots of up and down drafts and the always-surprising splat of rain that the radar didn't show.
By the time Indianapolis controllers handed me off to Dayton approach, serious weaving was required to stay out of the big buildups. Dayton, Ohio's airspace is not particularly busy, but the controller quickly had his hands full because every pilot on the frequency was calling and insisting on a turn this way or that to avoid a buildup. The jets descending into the area or departing all needed turns off the standard routes, and there were four or five light airplanes in the sector all needing deviations right or left, and each time I missed one cloud I needed another turn to dodge the cloud behind it.
Before long the controller was going down the chute because pilots could not tell him how far they needed to fly on that heading, or which way they wanted to turn next. I'm sure that on his radar scope more than a pair of airplanes were pointed at each other at various times, and he had no way of knowing how far they would continue on and if they would become too close, violating separation standards.
I became the first problem the controller just simply couldn't ignore. My clearance was to fly directly from the Brickyard VOR to Port Columbus Airport, but that is not a line that would show on his radar scope like an airway or other standard procedure. Plus, I had already asked for and received clearance to deviate left and right several times, so I'm sure he lost track of which way I intended to go. A regional jet descending into the tangle was also deviating off the arrival path and was now headed down through my altitude. The controller's only option was to order me to turn right immediately to 140 degrees, and I could hear a fair amount of stress in his voice. But one of the biggest storms in the immediate area was just to my right, and that's why I was turning left. A turn to 140 degrees would have put me right into that thunderstorm.
I told the controller I couldn't turn right, and he repeated the 140 heading. He had no choice and I had no choice. I simply told him, "45FM is canceling IFR." I have done this several times over the years, and controllers are always stunned, even disbelieving, when I do it. He asked me to say again, and I did. He still didn't want to believe and asked again, so I said, "45FM is canceling IFR and I am out of here." Not exactly textbook language, but it seemed to work.
I had a clear path between the cells, flat terrain below, and I descended to 2,500 feet with plenty of cloud clearance and continued on VFR to Columbus, dodging the rain shafts that were now falling out of the bottoms of the thunderstorms. I could see on the moving map that I was clear of the Dayton Class C airspace. I stayed tuned to the Dayton frequency and listened to the near chaos as the pilot of another Baron canceled IFR and descended below the cloud bases.
The key for me is that I had kept an escape route by being in a position to cancel IFR if there was no visual route around the cells. The Dayton controller didn't have an escape route and eventually needed at least one airplane to fly into a storm that eyeballs and radar said would be dangerous if not deadly. At that instant, his near-term problem was between me and the RJ, and he picked me for a route through the storm.
The FAA has long recognized this problem in which thunderstorms become obstructions as hard as mountain ridges, at least as far as pilots are concerned. When storms form in the really busy airspace, or sometimes even when they are forecast, the FAA's flow control system just keeps airplanes on the ground. The FAA knows that if the same situation of rapidly building thunderstorms would occur in the busy airspace over New York or Chicago or Dallas or other congested areas, the conflicts I experienced near Dayton would be unmanageable, so ground stops and gate holds are issued.
If the FAA's NextGen ATC system delivers as promised and IFR airplanes are packed in more closely, what will happen when pilots simply have to turn left or right to avoid a thunderstorm? In the FAA's dream, there will be another airplane not far away, and any turn off course will create a conflict. Part of the NextGen plan is that each airplane will broadcast its intent, along with its position, altitude and velocity. The intended flight path will come from the navigation system. In my case that day, a NextGen world would have seen that my intent was to fly directly from Brickyard to Columbus. And we could have also all seen that the RJ intended to descend and land at Dayton. "Intended flight path" is interesting and useful information on a day without storms, but totally useless and even misleading information when thunderstorms are popping.
In the busiest airspace, the FAA uses procedural separation to backstop controllers and their radar vectors. That's why in congested airspace we fly the same seemingly arbitrary routes and altitudes in and out. These routes keep IFR airplanes safely apart without constant intervention by controllers and allow more densely packed traffic than if airplanes flew random routes with only last-minute controller vectors to keep them apart. But procedural separation through standard routes goes out the window when a thunderstorm squats on the route.
The superior navigation accuracy of the proposed NextGen system will be great for flying around fixed obstacles such as high terrain, or even along an established route close to other traffic. But thunderstorms blow up this whole concept because we can't know exactly where and when they will form, and where they will move. All we know for certain is that airplanes can't now or in the future safely fly through thunderstorms.
Though I have no global solutions to the thunderstorm problem, my advice to light-airplane pilots is to always leave an escape route. When the clouds start to build, don't ask for higher altitude because you can't climb high enough to top a thunderstorm. When the spacing between the buildups closes and you can't see a clear path around the cells ahead, you need to take action such as descending below the bases or even turning around. Thunderstorms can form a box canyon that is nearly as unyielding as the granite canyons in the West.
Finally, you must know that weather radar can't detect the existence of very significant turbulence until a buildup is well on its way to becoming a full-blown thunderstorm. An onboard weather radar is going to give you more warning than will the Nexrad image sent down from satellite, which is at least five minutes old, but neither one can tell you that a cloud contains serious turbulence until it's too late. A rapidly building cumulus cloud that can deliver nothing more than a few very sharp jolts for a jet with its high wing loading can produce a wild, maybe even uncontrollable, ride for a light airplane.
Bottom line: Don't rely on the controllers to keep you out of trouble. Yes, they have very good weather radar, and, yes, they do their best to pass on weather reports and ride reports from other pilots. And, yes, they can almost always approve deviations around clouds or radar targets you see. But, when the sky fills with building thunderstorms and airplanes, controllers will at some point have to fall back to their ultimate responsibility and turn airplanes away from each other, even if that means one of them has to turn into a thunderstorm. It's their job, but it's our job to always be in a position to legally turn down that clearance.