Racing the Weather

Cross-country strategies from an air race

January 1, 2006

The weather that dominated the week of the 2005 Air Race Classic was an echo of the whole long summer: A stubborn high-pressure system — a necklace of H's — choked the midsection of the country. Good flying weather? Well, essentially yes, but not the prettiest, or the most fun. Hot, hazy days blanketed visibility down to marginal levels at times, though the skies remained persistently clear above.

And the tailwinds? The Holy Grail of the air race pilot? Elusive, unorganized, hiding at altitudes between the 3,000-, 6,000-, and 9,000-foot levels on the winds aloft forecast.

We had to find them.

How to race

The overall strategy to win a multileg cross-country air race such as the Air Race Classic — an all-women's race with its roots in the All-Woman Transcontinental Air Race — is to optimize your groundspeed to best your handicap by the greatest margin. To do this, you need to pick a departure time when the winds aloft are most advantageous, and once aloft, pick an altitude at which the winds are optimum — without hunting around a bunch to find them.

You also want to fly the straightest course possible, and you must stay VFR. If a team is found to have entered instrument meteorological conditions or files an IFR flight plan, it is disqualified. We also were required to file VFR flight plans for each leg — it's disqualifying to fly without one. This means picking a time of day (or day of the race) to fly a given leg when you won't likely have to vector yourself around storms or IFR weather — even a benign layer of clouds can prove problematic if it requires you to alter course or change altitude. You need to fly the legs in sequence, and what may be a great day for a leg two hops ahead of you may harbor poor conditions for the leg you need to fly to get to that leg farther along the course.

Both aspects of the strategy hold great lessons for pilots launching on VFR cross-country flights.

I served as the weather officer for an experienced race team, Gretchen Jahn (pilot) and Ruby Sheldon (copilot), to gain exposure to the race environment — and to help the team make weather decisions that would allow it to place higher in the final standings. The 2005 course began in West Lafayette, Indiana, and traced a polygon through Winona, Minnesota; Beatrice, Nebraska; Bartlesville, Oklahoma; Shreveport, Louisiana; Walnut Ridge, Arkansas; Tullahoma, Tennessee; Athens, Ohio; and Frankfort, Indiana, to end back at West Lafayette.

Since the race took us on a circular tour of the Midwest and south-central United States over several days, we dealt with terrain, the Midwestern plains, and haze off the Gulf of Mexico and Mississippi River and lingering around the Ozark and Ouachita mountains in Arkansas and the Cumberland Plateau in Tennessee. A couple of large-scale weather patterns affected the flight. We used forecasts to determine where to spend the night, and on-board satellite weather (using a portable system) to determine the best altitude to fly, check on advancing weather, and formulate a plan for the next leg.

Modern tools

While sensing devices that collect current wind and pressure data (and the weather balloons to carry those devices aloft) have been used in aviation since its early days, the computer-generated models we use today to forecast winds and weather were born in the late 1940s, and grew exponentially in the 1960s and 1970s as computer processing power increased. Early cross-country racers in the 1920s and 1930s could only call ahead to the next stop for conditions, and rely on experience and what they found along the way for determining the best time and altitude to fly.

In 2005, we still started our weather-data gathering with a call to flight service — but we weren't walking up to the counter of a flight service station like pilots did nearly 40 years ago, in the race's early days. Back then, before the consolidation of FSSs across the country, it was far more likely that the race would start at an airport with its own FSS — or one nearby would send a specialist (which we know as a "briefer") to discuss the weather synopsis and forecast with the racers. It was deemed more efficient than having 50 or 60 additional calls to FSS on race morning. But with staffing problems at FSSs nationwide particularly challenging in 2005, the Terre Haute FSS in Indiana couldn't spare a specialist for the race, and, in advance of the race, we discussed the weather with a professor from Purdue University's aviation department instead — in addition to our required calls to FSS, of course.

During our preliminary discussion the afternoon before the race began, we looked at a cold front projected to move through the north-central states — and impose itself on the first and second legs of the course in northern Illinois, southern Minnesota, northern Iowa, and eastern Nebraska. But by the time 7 a.m. on June 21 rolled around, the convective activity associated with the front had disappeared, and only clear skies lay ahead. We were going — everyone was required to start in order with the group once the first go/no-go decision had been made — now the only question was: How high to fly?

Wind profiling

Prior to the briefings, I looked at several additional sources of information to help divine the answer to this question. Starting with the winds aloft forecasts for stations along the route, I then turned my attention to the NOAA (National Oceanic and Atmospheric Administration) profiler network — a system of 35 unmanned Doppler radar sites located in 18 states in the central United States and in Alaska. These sites provide hourly wind data at altitudes from the surface up into the flight levels. Links to data from the sites are found on NOAA's profiler Web site.

The profiler station at Blue River, Wisconsin, sat to the northeast of the last half of our course on the first leg to Winona, so I downloaded and studied the profile to determine a couple of things.

First, what were the winds doing presently? And second, what were the trends — were the winds increasing or decreasing, or shifting direction? And last, but perhaps most important, how well did they match the winds aloft forecast for the present period? This information would help me determine how valid that forecast was, and if the weather was doing something different than the National Weather Service had predicted. Hey, everyone knows that the weather doesn't always behave according to the models....

Profiler stations circled most of our route, in locations such as Slater, Iowa; Lamont, Oklahoma; De Queen, Arkansas; Winnfield, Louisiana; Bloomfield, Missouri; and Wolcott, Indiana. After I picked the ones closest to the route (and in the direction where the weather was expected to come from), I noted the actual winds aloft and on the surface at these points.

I then assessed the winds aloft forecasts for predictions of change. Of course, that's the way it always is — the wind changes positively on one leg and negatively on another. That's where the strategy comes in — when to move, and how long to wait? We could gamble and wait, with a promise that conditions might improve, or we could move on and take what we had, with the chance that those following behind us would luck out with better winds.

Another tool in our quiver was a tablet computer (a Sony U750 packaged as a NavAir EFB; see " Pilot Products," June 2005 Pilot) with moving-map GPS and weather datalink capability from AirGator and WxWorx. Sitting in the back of the airplane, I would be able to download from WxWorx's satellite-based system updated weather and winds aloft. This would prove a valuable tool — except during the very few times when I was unable to get good satellite reception because of our direction of flight — the satellites are best viewed when the antenna can "see" the southern sky through an aircraft window. The data available included maps of winds at various altitudes and on the surface, up-dated hourly, which I used more often than any of the other products, given the lack of hazardous weather on our route.

We had no choice about when to fly the first leg, and we faced headwinds aloft as a result of our required departure that morning. The answer to the altitude question? Stay low.

Our leg time told the tale: It was our slowest for the race and gave us an average headwind component of about 8 knots.

Decision time

The first go/no-go decision we made on our own was whether to press on from Winona to Beatrice right away, or to pause for better weather or winds.

During the previous day's weather discussion, Tom Carney, head of the department of aviation technology at Purdue, had briefed us on how to determine the location of a front through noting various METARs in the region through which the front is expected to move. Generally speaking, you look at the temperature-dew point spread (for changes in the air-mass composition), the barometric pressure (which typically changes as the front approaches and passes), and the wind direction and speed. For example, Carney stated that typically winds are out of the south and relatively weak ahead of a cold front, and they shift to the northwest and build as the front passes.

The cold front had marched on, and it was a relatively weak one. Therefore, the only hazardous weather hinting at us was the normal afternoon convective activity found in the Great Plains in the spring, summer, and early fall. A small clump of storms was already building in the Dakotas and western Nebraska. With a launch by 11 a.m., we would get to Beatrice by 1 p.m., and be able to take off again before the heat of the day hit — and brought more storms with it. Like all VFR pilots, we had to balance decreased performance on the leg ahead with other considerations.

But we would wait for better conditions if the winds weren't in our favor. The winds looked good, so we took a chance and pressed on. Soon we were riding a 17-knot tailwind component at 6,500 feet — but it didn't last. We posted about a 2-knot gain overall for the leg. Too much hunting around for the right altitude probably cost us. The lesson: Sometimes picking a reasonable altitude for winds and sticking with it give you more in return than continuing to cast around for better — and losing speed to climbing in the process.

We didn't spend much time at Beatrice, either. A host of storms threatened to the west, and we wanted to get south of them before they crossed our path. Besides, the winds out of the south were supposed to build overnight.

Fighting fatigue

Early on, my call to climb high ended up costing us a knot or two — we descended from 8,000 back to 5,500 along the way — but we did all right on the leg from Beatrice to Bartlesville, arriving there around 3 p.m. At this point, we were all fighting the afternoon circadian lull. After 839 nautical miles, we also were fighting plain-old fatigue. We had a hotel room ready in case we decided to stay, and at first we put our names on the list to remain overnight — this is how the race keeps tabs on where participants are.

As we took a break, it was time to look at the weather again. While the overall weather picture didn't change much as we considered the leg from Bartlesville to Shreveport, a couple of other factors played into our next decision. First, the visibility as we progressed farther south was deteriorating — and we noted that METARs from points south the previous morning showed visibilities hovering around the IFR/MVFR mark. And second, the winds for the leg beyond, from Shreveport to Walnut Ridge, looked like they would be better the next morning (June 22) than during the next afternoon.

It was a big guess — but after discussing it, we decided to head for Louisiana. Another bonus to making the flight that evening? We would point away from the sun into the haze instead of into the sun the next morning.

Although we didn't plan to fly that high (5,500 feet), because of fatigue we decided to sip some oxygen from the system on board the airplane — another tool to consider on cross-country flights. Fifteen minutes of the good stuff left our pilot well prepared for the flyby and landing at Shreveport. We were the only ones to go so far that day — but would it pay off?

We woke to visibility right at three miles with clear skies — and waited past the start time to gain a little better perspective. Once en route visibilities came up to five miles, we launched and stayed low. An initial altitude of 1,500 feet agl gave us a 10-knot tailwind, but on the way up we'd seen a 15-knot tailwind, so we slid back down to 1,000 feet agl to avail ourselves of the extra push. This would be our second-best leg, and we were pleased with the decision to stay in Louisiana.

At this altitude, weather was less on our minds than the obstructions and towns we'd pass over. Luckily, we were flying over rural areas, and the combination of a good obstruction database on our GPS and a quick eye on the sectional chart kept us clear of cell towers and such. Still, when you're clicking along at 188 knots indicated, the world goes by fast — so we were glad for our three sets of eyes focused outside the cockpit. It may not make sense for a pilot on a standard VFR cross-country flight to fly so low for a few knots — but if headwinds force you into a corner, it's nice to know that this trick can extend your range.

The wind slacked off to near nothing by the time we hit Walnut Ridge, so we waited until after lunch to see if afternoon heating would bring any change to the story. We could find none, so we called it a one-leg day and saved Tullahoma for the next morning.

Nothing, and more nothing

The good news: The high-pressure system still dominated, meaning little or no convective weather. The bad news: The high-pressure system still dominated, meaning little or no wind. But a front threatened to push into Indiana by the end of the last day of the race (June 24), so we determined we would finish on day three. Had we waited, and had the front come through, we would have faced strong headwinds on the leg from Athens to Frankfort. Testimony to how little the wind mattered on the day we did fly these three legs? With course headings of 90 degrees on the first leg to Tullahoma, 27 degrees on the second leg to Athens, and 295 degrees on the last leg to Frankfort, we posted groundspeeds of 190.9, 189.7, and 189.8 knots, respectively, flying at about 1,000 to 1,500 feet agl. It just wasn't worth it to climb any higher — there were simply no winds to be found in any direction.

And the front that was forecast to plague day four? It too disappeared — probably banished by the gods of air racing. We came in a respectable tenth overall (out of 39 race teams), separated from the top spot by a little more than 3 knots groundspeed.

That's OK. There will be other races — and many other cross-country flights — in which to use the wind and weather to our advantage.


E-mail the author at julie.boatman@aopa.org.


Links to additional information about cross-country weather strategies may be found on AOPA Online.


'Fly' Real-World Scenarios in 'WeatherWise' Online Course

Low ceilings and restricted visibilities are aviation's most deadly killers. But fortunately, with a little knowledge, you can minimize the risks these conditions pose. And you can do it without leaving the ground.

By using special Microsoft Flight Simulator scenarios, you can "fly" many of the conditions explained in the AOPA Air Safety Foundation's new online course, "WeatherWise: Ceiling and Visibility."

The scenarios are available at the end of the free course. Or you can download them directly from AOPA Online if you have Microsoft Flight Simulator 2004 version 9.1 or above.