CHAPTER 11

It’s uncharted territory: You’re out there on your own, out in the middle of nowhere.

On May 1, the weather finally cleared on Key West. The plane carrying the lidar machine took off, refueled in Grand Cayman, and arrived in the Roatán airport at 2:00 p.m. Everyone rushed out to the airport to meet it, applauding and cheering when it finally touched down. Now our search for the lost city could begin.

The Skymaster is a twin-engine aircraft driven by what aviators call a push-pull configuration, with the two engines mounted in-line, one on the nose and the other at the rear of the fuselage. The plane’s most distinctive features are two struts or booms that extend behind the wings. Once a cheerful red and white, the paint job on this plane was full of patches and strips that had peeled off, and an ugly streak of oil ran down the fuselage from the forward engine. A big green lidar box almost filled the interior of the plane. This sleek, advanced, and costly piece of technology, so top secret that it had to be guarded by soldiers, was being schlepped around in a shabby flying tin can—or so it seemed to my inexpert eye.

After it landed, seven Honduran soldiers with M16s escorted the plane to a far corner of the airport, away from the public areas, where it could be kept secure. Nobody seemed to be paying attention anyway; the airport was small and the military was ubiquitous. The six soldiers, most barely older than teenagers, and the commanding lieutenant had been hanging around the airport, bored, for three days. They were excited at the plane’s arrival, and they marched around it, posing with their weapons while Elkins’s film crew shot footage.

The pilot, Chuck Gross, was a large, soft-spoken man from Georgia who addressed everyone as “sir.” He had recently returned from Iraq, where he had been flying classified lidar missions for the US military. He couldn’t disclose much, but I gathered that they involved, among other things, lidaring areas along patrol routes multiple times to detect tiny changes in topography. A new heap of trash or a fresh dirt pile suddenly appearing next to a route would often indicate the placement of an IED.

Gross mentioned he had a Cuban overflight number, which allowed him to fly through Cuban airspace. I asked him what would have happened if he’d had engine or weather trouble and had been forced to land in Cuba. After all, the plane carried classified military hardware, and relations with Cuba were at that time still in a deep freeze.

“First, I would have torched the plane on the runway.” This was, he explained, the standard protocol with airborne lidar. “In the desert, that’s what we would have done too: immediately destroy the equipment.” He added, “You should have seen the paperwork I had to do to get that Cessna out of the US.”

The technology of lidar was developed soon after the discovery of lasers in the early 1960s. Put simply, lidar works like radar, by bouncing a laser beam off something, capturing the reflection, and measuring the round-trip time, thereby determining the distance. Scientists quickly realized its potential as a mapping tool. Both the Apollo 15 and 17 missions carried a lidar machine on the orbiter, which mapped swaths of the moon’s surface. The Mars Global Surveyor, a satellite orbiting Mars, also carried a lidar machine, which bounced laser beams off the surface of Mars ten times per second. Over its ten-year mission, from 1996 to 2006, the Surveyor created a prodigiously accurate topographic map of the Martian surface, one of the supreme mapping projects of human history.

There are three types of lidar instruments: spaceborne, aerial, and terrestrial. On earth, aerial lidar has been used in agriculture, geology, mining, tracking glaciers and ice fields for global warming, urban planning, and surveying. It had numerous classified uses in the wars in Iraq and Afghanistan. Terrestrial lidar is currently being tested in self-driving vehicles and “intelligent” cruise control, which use lidar to map the ever-shifting environment around a car moving down a roadway, as well as to make detailed three-dimensional maps of rooms, tombs, sculptures, and buildings; it can re-create digitally, in incredibly fine detail, any three-dimensional object.

The target sites of T1, T2, and T3 would be mapped with this Cessna, the same one used over Caracol. As the plane is flown in a lawnmower pattern over the jungle, the lidar device fires 125,000 infrared laser pulses a second into the jungle canopy below and records the reflections. (The laser pulses are harmless and invisible.) The time elapsed gives the exact distance from the plane to each reflection point.

The lidar beam does not actually penetrate foliage. It does not “see through” anything in fact: The beam will bounce off every tiny leaf or twig. But even in the heaviest jungle cover, there are small holes in the canopy that allow a laser pulse to reach the ground and reflect back. If you lie down in the jungle and look up, you will always see flecks of sky here and there; the vast number of laser pulses allow lidar to find and exploit those little openings.

The resulting data is what lidar engineers call a “point cloud.” These are billions of points showing the location of every reflection, arranged in 3-D space. The mapping engineer uses software to eliminate the points from leaves and branches, leaving only bounce backs from the ground. Further data crunching turns those ground points into a hill-shade picture of the terrain—revealing any archaeological features that might be present.

The resolution of the lidar image is only as good as how well you keep track of the position of the plane flying through space. This is the greatest technological challenge: In order to achieve high resolution, you need to track the plane’s position in three dimensions during every second of flight to within an inch. A standard GPS unit using satellite links can only locate the plane within about ten feet, useless for archaeological mapping. The resolution can be refined to about a foot by placing fixed GPS units on the ground underneath where the plane will be flying. But an airplane in flight is being bounced around by turbulence, subjected to roll, pitch, and yaw, which not even the finest GPS unit can track.

To solve this problem, the lidar machine contains within it a sealed instrument that looks like a coffee can. It contains a highly classified military device called an inertial measurement unit, or IMU. This is the same technology used in cruise missiles, allowing the missile to know where it is in space at all times as it heads toward its target. Because of the IMU, the lidar machine is listed as classified military hardware, which cannot leave the country without a special permit, and even then only under highly controlled conditions. (This is another reason why there was a long lag-time in the use of lidar at Third World archaeological sites; for years the government prevented the IMU from being used outside the country in civilian applications.)

Aerial lidar can achieve a resolution of about an inch, if there is no vegetative cover. But in the jungle, the canopy causes the resolution to drop precipitously, due to many fewer pulses reaching the ground. (The fewer the pulses, the lower the resolution.) The Belizean rainforest around Caracol, where the Chases had used it in 2010, is thick. But it doesn’t come anywhere near the density of Mosquitia.

The first lidar flight over T1 took off the next day, May 2, 2012, at 7:30 a.m., with Chuck Gross at the controls and Juan Carlos Fernández acting as navigator and running the lidar machine. We all went to the airport to see the plane off, watching it rise into the Caribbean skies and wink into the blue across the Gulf of Honduras, heading for the mainland. It would take three days to map the twenty square miles of T1. If all went well, we would know in four days if T1 held anything of interest. After that, the plane would shift to T2 and T3.

The plane returned from its first mission in late afternoon. By nine in the evening Sartori confirmed that the data was clean and good; the lidar machine was operating flawlessly and they were getting enough ground points through the forest canopy to map the underlying terrain. While he had no images yet, he saw no technical reason why we wouldn’t get detailed terrain maps.

After the second day of flying, on May 3, Juan Carlos came back with intriguing news. He had seen something in T1 that didn’t look natural and had tried to photograph it through the windows of the Skymaster. We gathered in his bungalow to look at the photos on his laptop.

It was my first glimpse of the valley. The photos, taken with a shaky telephoto through scratched Plexiglas, were not clear; but they showed two squarish white objects that looked like the tops of carved limestone pillars, opening into an area of low vegetation that was square in shape. The feature was on a brushy floodplain in the upper end of the valley. Everyone crowded around the laptop, squinting, pointing, and talking excitedly, trying to make sense of the pixilated images that were so tantalizingly ambiguous—they could be pillars, but then again they could be trash dropped from a plane or even the tops of two dead tree stumps.

I pleaded to accompany the third and final flight over T1, despite the logistical issues it posed. There was no room in the plane, but after some discussion, Chuck Gross agreed that he might be able to clear out a tiny space for me to crouch in. He warned me it would be mighty uncomfortable over six to seven hours of flying time.

On May 4, we arrived at the airport as the sun was just rising above the curve of the ocean, the plane throwing an Edward Hopper shadow across the tarmac. The soldiers guarding the plane greeted us sleepily. Now that I was about to be a passenger I looked at the plane more attentively, and I did not like what I saw.

“What’s with that oil streak?” I asked Chuck.

“Don’t worry about that,” he said. “I’m topping it off every day. In one flight it won’t lose enough to make a difference.”

As I crawled on board, my dismay deepened. The interior of the Cessna, once a rich velvetized fabric in burgundy, was now worn, greasy, and faded; much of the inside appeared to be held together with duct tape. It smelled of Eau de Old Car. Parts of the plane had been sealed with acrylic caulk, now peeling out in strings. As I tried to maneuver around the giant lidar box into the micro-space provided, I bumped my elbow into a panel, which fell off.

“No worries, that always happens,” said Gross, reseating it with a blow from his fist.

I marveled that a plane as unsafe and decrepit as this one looked would be used to carry a million-dollar scientific instrument. Chuck firmly disagreed. “No, sir,” he said. “This plane’s a perfect platform for the job.” He assured me the 337 Skymaster was a “classic,” and a “great little aircraft.” Unlike a King Air or a Piper Navajo, he said, this craft was ideal, with a fuel efficiency that would allow us to spend “six hours on station.” Even though it was forty years old, it was “totally dependable.”

“What if we go down?”

“Wow,” said Chuck, “what a question! First thing, I’d look for a clearing to set it down in. It’s uncharted territory: You’re out there on your own, out in the middle of nowhere, no two-way coms.” He shook his head—unthinkable.

Despite my worry, I had a lot of confidence in Chuck because I had learned of his feats of flying; at the age of eighteen, he had soloed across the Atlantic, one of the youngest pilots to do so. I hoped the aircraft’s deficiencies were mainly cosmetic. I told myself a world-class pilot like Chuck would never fly a plane that wasn’t safe.

I jammed myself behind the lidar box: no seat, my knees in my mouth. Juan Carlos was right in front of me. He was concerned about how I would fare; I sensed he was worried I might get airsick and vomit down the back of his neck. He asked if I’d had anything to eat or drink that morning. I said no. He casually mentioned how grueling it was out there, flying low and slow over the jungle for six hours straight, banking steeply turn after turn, tossed around by thermals, sometimes dodging vultures. The A/C on the plane was broken, he said; we would be sealed in a metal tube flying in full sun. The plane had no bathroom. If you had to go, you went in your pants. I tried to assure him I would be an exemplary passenger.

Elkins gave me a GoPro video camera and a still camera with a telephoto lens and asked me to take more pictures of the mysterious white pillars and anything else of interest I spied down below.

Chuck Gross climbed into the pilot’s seat and began running down the checklist, while Juan Carlos jacked his laptop into the lidar box. He showed me the flight plan he had programmed on his computer screen, dozens of parallel lines crisscrossing the valley, designed to maximize coverage while minimizing flight time. In addition to being a lidar engineer, Juan Carlos was also a licensed pilot, enabling him to work seamlessly with Chuck.

We took off from Roatán and were soon winging over the glittering Bay of Honduras, the mainland looming up ahead. It was a gorgeous day, the sky dotted with fluffy white cumulus. Far ahead, where the blue mountains of Mosquitia rose up, we could see the cloud cover was sparse and high. As we flew inland, the settlements along the coast gave way to scattered hamlets and agricultural fields alongside slow brown rivers. The land mounted into forested foothills, where hundreds of ragged patches of clear-cutting came into view. Plumes of smoke rose from the jungle in every direction.

The logging holes eventually disappeared and we were flying about four thousand feet over unbroken, precipitous forest. Chuck maneuvered his way through the mountains as we approached T1. An hour out of Roatán, Juan Carlos pointed out the rim of the valley in the distance, a wall of green mountains with a sharp notch in them. Chuck eased the plane to a lower altitude and we cleared the rim at a thousand feet, which gave a tremendous view of the landscape. As the land dropped away beyond the rim, I was struck by the valley’s picturesque topography, the ring of mountains embracing a gentle, rolling landscape divided by two rivers. It really did look like a tropical Shangri-la.

The plane leveled off at an altitude of about 2,500 feet above ground, and Juan Carlos booted up the lidar machine, picking up where they had left off the day before. As the lidar bombarded the canopy with laser pulses, Chuck steered the Cessna in parallel lines across the valley, each four to six miles long, in a pattern that, on the computer screen, looked like a gigantic weaving. The plane was buffeted by thermals, knocked up and down, back and forth, and sometimes sliding sideways in a gut-wrenching fashion. Juan Carlos had been right; it was a brutal and scary ride. But Gross worked the controls with constant finesse and a sure hand.

“We were rocking and rolling pretty good,” Gross said later. “It’s like flying a big spiderweb. It takes incredible skill. You have to fly the middle of the line, and you can’t go sixty feet on either side of that line. You have to slide that plane around, doing all rudders. To stay on the line, in that wind, that was challenging. And you have to hold the altitude and airspeed. I had to climb with the terrain and maintain the same altitude. If the terrain starts coming up, I have to come up with it.”

Through it all, I peered out the window, transfixed. I can scarcely find words to describe the opulence of the rainforest that unrolled below us. The tree crowns were packed together like puffballs, displaying every possible hue, tint, and shade of green. Chartreuse, emerald, lime, aquamarine, teal, bottle, glaucous, asparagus, olive, celadon, jade, malachite—mere words are inadequate to express the chromatic infinities. Here and there the canopy was disrupted by a treetop smothered in enormous purple blossoms. Along the central valley floor, the heavy jungle gave way to lush meadows. Two meandering streams glittered in the sunlight, where they joined before flowing out the notch.

We were flying above a primeval Eden, looking for a lost city using advanced technology to shoot billions of laser beams into a jungle that no human beings had entered for perhaps five hundred years: a twenty-first-century assault on an ancient mystery.

“It’s coming up,” Juan Carlos said. “Right there: two white things.”

In an open area, I could see the two features that he had photographed the previous day, standing about thirty feet apart next to a large, rectangular area of darker-colored vegetation. The plane made several passes as I photographed. Again, they looked to me like two square, white pillars rising above the brush.

We got through the flight without mishap, other than the moment a few hours into the flight when I turned off the lidar machine with my knee as I tried to shift my aching legs. The machine and the pilot’s navigational system were linked, so shutting off the lidar turned off Gross’s navigation. He immediately went into a tight, stomach-sickening holding pattern while Juan Carlos booted the machine back up and I apologized profusely. “No worries,” he said, far less perturbed than I thought he would be.

We finished mapping T1 with enough fuel left over to fly a few lines over T2, twenty miles distant. The route took us over the Patuca River, Heinicke’s “most dangerous place on the planet,” a brown snake of water winding through the jungle. T2 was magnificent and dramatic, a deep, hidden valley shut in by sheer, thousand-foot limestone cliffs, draped with vines and riddled with caves. But recent deforestation—only weeks old—had reached the mouth of the T2 valley. As we flew over, I could see the freshly cut trees lying on the ground to dry out so they could be burned, leaving a hideous brown scar.

At the end of the day, we flew to La Ceiba, on the mainland, to refuel. Chuck had pushed the envelope and we landed with less than twenty gallons of aviation fuel remaining, about forty minutes of flight. But the airport had no fuel and nobody could locate the tanker bringing the resupply. Airport officials feared the tanker had been hijacked by drug smugglers. Juan Carlos called Elkins in Roatán. Elkins put Bruce Heinicke on the problem. After calling around, Bruce learned the truck was still en route, delayed by a blowout.

We couldn’t leave the Cessna unguarded, especially if the fuel didn’t arrive and the plane had to stay overnight in La Ceiba. Juan Carlos and Chuck debated sleeping in the plane, but that wasn’t ideal, since they were unarmed. They finally decided that, if fuel didn’t arrive, they would go to the US Air Force base in La Ceiba and ask the soldiers to stand guard for the night. Meanwhile, Michael Sartori was desperate to get the data and finish mapping T1, so it was agreed I’d head back to the island on my own. Fernández gave me the two hard drives with the data, and I went to the airport desk to see if I could get a commercial flight from La Ceiba to Roatán. There was a flight to Roatán that afternoon but it was already full. For $37 I was able to hitch a ride in the copilot’s seat. The plane looked even less reliable than the Cessna, and as I boarded, Juan Carlos joked about what a pity it would be to lose all that precious data in a plane crash after their hard work collecting it.

I landed in Roatán at sunset and gave the hard drives to Sartori, who snatched them up and disappeared into his bungalow, emerging only once to chow down a couple of lobster tails at dinner. He now had all the data he needed to map T1. Late that night Juan Carlos and Chuck Gross finally landed back in Roatán, exhausted but relieved. The fuel truck had arrived at the last minute.

Sartori had hours of work ahead of him. He had to merge data from several sources: the lidar machine, the GPS ground stations, the GPS data from the aircraft itself, and the data from the IMU. Together, all this would create the point cloud, forming a three-dimensional picture of the rainforest and the underlying terrain. First, he had to wait for Mango to retrieve the USB stick from the GPS unit in Culmi and bring it to Catacamas to upload to the server in Houston; Sartori then had to download the data from Houston. The lights in Sartori’s bungalow were still burning at midnight when I went to bed. Ramesh Shrestha, back at NCALM in Houston, remained awake, pressing him for updates.

This was the moment of truth: The images would show what was in the valley—if anything. It was almost one in the morning when Sartori finished creating the raw images of T1; Shrestha had finally gone to bed and the Internet connection on Roatán was down. Exhausted, Sartori went to bed without even looking at the images he had just created.

The next day was Saturday, May 5. Rising early, Sartori uploaded the raw images to a server in Houston, again without examining them. Immediately on receiving them, Shrestha forwarded them to NCALM’s chief scientist, William Carter, who was at his vacation home in West Virginia. Shrestha intended to review them soon, but Carter beat him to it.

At 8:30 a.m. on that quiet Saturday morning, the terrain images of T1 arrived in Carter’s in-box just as he was about to leave the house to run errands. He needed to buy a refrigerator. He hesitated and then told his wife that he wanted to have a quick look. He downloaded the data and displayed the maps on his computer screen. He was thunderstruck. “I don’t think it took me more than five minutes to see something that looked like a pyramid,” he told me later. “I looked across the river at a plaza area with what looked like buildings—clearly man-made objects. As I looked at that river valley, I saw more, as well as alterations to the terrain. It was kind of surprising how easy it was to find them.” He e-mailed the coordinates to Sartori and Shrestha.

Sartori pulled up the images and scanned them. In his excitement Carter had mistyped the coordinates, but it took Sartori only a moment to find the cluster of features on his own. He said, “My skepticism wasn’t easily broken,” but this was clear enough to convince the most resolute doubter. Sartori was chagrined. “I was mad at myself for not seeing it first, since I was the guy producing the images!” He rushed out the door to report it to Steve Elkins, but then had second thoughts. Was it real? Maybe it was just his imagination. “I was in and out the door about six different times,” Sartori said.

I was walking back from breakfast with Steve and some others when Sartori appeared along the quay, running madly in his flip-flops, waving his arms and shouting: “There’s something in the valley!” We were startled by this sudden behavioral change, the sober-minded skeptic transformed into a raving Christopher Lloyd.

When we asked what it was, he said, “I can’t describe it. I won’t describe it. You just have to see it yourself.”

There was pandemonium. Steve started to run, and then remembered he was a filmmaker, so he began shouting for his film crew to get their gear together and record the moment—cinema verité. With the cameras rolling, everyone crowded into Sartori’s room to look at the images on his laptop. The maps were in gray scale and a first iteration, but they were clear enough. In the valley of T1, above the confluence of the two streams, we could see rectangular features and long, pyramid-like mounds arranged in squares, which covered an area of hundreds of acres. Also visible, but impossible to interpret, were the two objects that looked like square pillars we had seen from the plane. As we examined the images, Sartori’s in-box was pinging continuously with e-mails from Carter and Shrestha, who were poring over the same maps, shooting off an e-mail with coordinates every time they found another feature.

I was stunned. It sure as hell looked like a very large set of ruins, perhaps even a city. I had thought we would be fortunate to find any kind of site at all; I had not expected this. Was it possible that an entire lost city could still be found in the twenty-first century?

I could see Sartori’s spiral-bound notebook lying open next to the laptop. In keeping with the methodical scientist he was, he had been jotting daily notes on his work. But underneath the entry for May 5, he had written two words only:

HOLY SHIT!

“When I saw those rectangles and squares,” Steve told me later, “my first feeling was one of vindication.” Benenson, who had been feverishly capturing the unfolding discovery on video, was happily stunned that the million-dollar spin of the roulette wheel had landed on his number. “I’m witnessing this,” he said, “but I’m not processing this very well. I have chills.”

Nobody dared wake up Bruce Heinicke to tell him the news. He finally emerged from his bungalow at 1:00 p.m. and listened with a frown on his face. He wondered why we were all so worked up—of course the White City was there. Who the fuck thought otherwise? He got on the phone to Áfrico Madrid, the minister of the interior. Áfrico said he would fly out to Roatán as soon as possible to review what we found and, if he was convinced it was real—and he had no reason to doubt it—he would convey the news to President Lobo, as well as to the president of the Honduran Congress, Juan Orlando Hernández. Meanwhile, the director of the Instituto Hondureño de Antropología e Historia, Virgilio Paredes, flew to Roatán to take a first look at our findings. Later, he recalled that moment: “I saw that and I said ‘Wow!’ We know Mosquitia is full of archaeological sites, but to see real cities, a big population of people living there—that is amazing!”

The valley of T1 had been mapped, but the project was only 40 percent complete: T2 and T3 remained to be explored. Chuck and Juan Carlos had headed off early that Saturday morning to continue mapping T2, unaware of the uproar of discovery occurring back at the Parrot Tree. Once in the air, however, Juan Carlos discovered the lidar machine was dead. They returned to Roatán and tried to get the machine working while the plane was on the ground, with no success. At around nine that morning all three lidar engineers examined it and confirmed the machine was kaput.

NCALM in Houston had a technical maintenance contract with a team in Toronto, Canada, where this lidar box had been designed and built. As it was a weekend, there was only a single tech-support person in Canada manning the phone. After he walked the lidar engineers through a plug-unplug sequence, trying to wake up the machine, they determined that a crucial part had failed. It was called a Position and Orientation System (POS) board, and it contained a GPS receiver and other components that “talked” to the IMU, exchanging data. There were only two POS boards in the world, both in Canada. The company would put a technician on a flight from Toronto to Roatán early Monday morning, transporting the $100,000 board in person in his carry-on bag. The part would have to clear customs twice, once in the United States and a second time in Honduras.

The engineer flying the part was Pakistani and, not having a US State Department export clearance for the POS board, he was worried about being stopped with it at Dulles Airport in Washington, DC, where he had an overnight connection. Before boarding the plane in Toronto, he panicked and stuck the part in his checked luggage, thinking that it would be less likely to prompt a security challenge in the United States.

The airlines (of course!) lost his bags. The two bags included not only the POS board but all the tools the technician needed to install it. The fact that the part was insured meant little to the expedition, which was spending many thousands of dollars a day and only had use of the plane for a strictly limited period of time. The flustered engineer arrived in Roatán on Tuesday morning with little more than the clothes on his back.

Desperate and futile phone calls to United and TACA airlines took up all of Tuesday. They learned the bags had arrived in Dulles Airport but had failed to be transferred onto the flight to San Salvador and then Roatán. They seemed to have vanished in Dulles. Then, as the frenzy of phone calls continued into Wednesday afternoon, the bags unexpectedly arrived at the Roatán airport. Virgilio Paredes went with Steve to the airport to speed them through Honduran customs. He did a masterful job of intimidation, waving about the president’s official card, and the bags sailed through and were rushed to the Cessna at the far end of the airport tarmac. It took the technician and Juan Carlos two hours to install the part and get the lidar machine working again. As they arrived back at the Parrot Tree, elated that the expensive, five-day delay was over, United Airlines called to once again say that, despite a most diligent effort, they were terribly sorry but they had been unable to trace the lost bags.

The mission resumed the next morning, on Thursday, with overflights of T2 and T3. They went flawlessly. Once again we gathered in Michael Sartori’s bungalow to look over the images on his laptop. And once again we were absolutely floored: T3 contained an even larger set of ruins than T1. T2 also revealed enigmatic, man-made features that were harder to interpret. Some guessed they might be quarries or fortifications.

In his quixotic search for the mythical White City, Elkins and his team had found not one large site but two, apparently built by the almost unknown civilization that once inhabited Mosquitia. But were they cities? And could one of these actually be the White City, the Lost City of the Monkey God? This, however, was the wrong question—it was clear to everyone by this point that the White City was a conflation of stories and probably did not exist in its described form. Like most legends, however, it was anchored in truth: The lidar discoveries had confirmed that Mosquitia had indeed been the territory of a great and mysterious civilization that built many large settlements before it disappeared. It was exactly as Cortés had written five centuries ago: This land had been home to “very extensive and rich provinces.” But what had caused it to vanish so suddenly and completely?