Magnetic North

The Projection

Chapter 2: Magnetic North

The USGS office on University Avenue occupied a single-story building that had been built in the 1970s with the expectation that the permafrost beneath it would remain frozen, and the permafrost had not obliged. The south wall leaned three inches off plumb. The floor in the map room tilted perceptibly toward the northwest corner, where a crack in the foundation let in cold air in winter and mosquitoes in summer. The building had been scheduled for replacement in 2019 and again in 2022 and the replacement had not occurred and would not occur because federal construction budgets in Alaska were allocated the way water was allocated in the desert, sparingly and to those who argued loudest, and the Fairbanks office of the USGS had never argued loudly about anything.

Silas parked his truck in the lot at eight in the morning. The lot was half empty. Three vehicles: Dave Kimura's Subaru, a white government pickup with the USGS arrowhead on the door, and a sedan he did not recognize. The building looked as it had looked for thirty-five years, except that the sign out front had been replaced in 2020 with a new one that included the Department of the Interior seal and a URL that no one would ever type. He went inside.

The map room was at the end of the hall, past the administrative offices and the GIS lab, which was a room with six computer workstations where technicians built digital elevation models from LiDAR data and satellite imagery. The door to the GIS lab was open and he could see the screens glowing with topographic surfaces rendered in false color, the terrain of the Brooks Range and the Alaska Range and the Seward Peninsula translated into pixel grids, each cell a number, each number an elevation, the landscape reduced to a matrix that could be stored and transmitted and analyzed without anyone ever standing on the ground it represented.

He did not go into the GIS lab. He went to the map room.

The map room held the Fairbanks quadrangle collection, which was the complete set of 1:63,360 topographic maps covering interior and northern Alaska, printed on paper, folded, filed in steel cabinets by quadrangle name. The cabinets were army green and the maps inside them were the maps he had grown up on, the maps that had taught him to read terrain the way Margot's art school had taught her to read color. He had spent hours in this room as a young field assistant, studying the quad sheets before a survey, memorizing the drainage patterns and the ridge lines and the lake shapes so that when he arrived in the field he would recognize the ground, would know that the creek bending south was the one that appeared on the map bending south, would feel the correspondence between the representation and the reality like a key fitting a lock.

The maps were beautiful. He had always thought so and was not embarrassed by the thought. They were beautiful the way a well-built instrument was beautiful — not decorated, not embellished, but shaped precisely to their purpose. The contour lines were brown. The water features were blue. The vegetation was green, screened to indicate density. The cultural features — trails, cabins, airstrips — were black. The grid was red, the UTM grid that divided the earth into sixty zones and projected each zone onto a transverse Mercator surface, a cylinder tangent to the central meridian of the zone, which flattened the curved earth into a plane the way you flatten an orange peel by pressing it against a table, and the distortion at the edges was the price you paid for the accuracy at the center.

He pulled the quad sheet for the upper Koyukuk drainage. The map was labeled WISEMAN B-3, Alaska, and it had been compiled from aerial photography taken in 1955 and field-checked in 1956 and published in 1957 and revised in 1983 to add the Dalton Highway, which crossed the eastern edge of the quadrangle. The contour interval was one hundred feet. The unnamed tributary he was going to survey appeared as a thin blue line entering the Koyukuk from the west, approximately six miles long, originating in a saddle between two ridges at about four thousand feet elevation. The line had no name. The ridges had no names. The saddle was unmarked. The tributary existed on the map as a geometric abstraction — a line of constant blue width regardless of the stream's actual width, which varied from twelve feet at the mouth to perhaps two feet at the headwaters, a variation that the map's scale could not represent and did not attempt to represent.

Silas spread the map on the light table and studied it. The light table was a sheet of frosted glass with fluorescent tubes beneath it, and it was used for examining maps and aerial photographs and overlays, and the light coming through the map gave the colors a luminous quality, as if the terrain were lit from within. He traced the tributary with his finger. Six miles of unmapped drainage. The contour lines crossed the blue line at intervals that suggested a moderate gradient in the lower reach and a steep gradient near the headwaters, where the lines crowded together like the teeth of a comb. The crowding of contour lines was the map's way of saying steep, its grammar for gradient, and reading it was a skill he had taught himself so long ago that it felt innate, the way reading English felt innate, though both were learned, both were systems of symbols that stood for things that were not symbols.

He heard footsteps in the hall and looked up.

Dave Kimura stood in the doorway. Dave was the office manager, a geologist by training who had moved into administration twenty years ago and now spent his days processing travel authorizations and equipment requests and wondering, Silas suspected, what had happened to the career he had imagined when he was twenty-five and standing in a stream bed in the Fortymile country, picking up rocks.

"You're in early," Dave said.

"Equipment checkout," Silas said.

"I saw the manifest. You're taking the Wild T2."

"I always take the Wild T2."

"It's the last one we have. If something happens to it, we don't replace it. There's no line item for optical instruments anymore."

"Nothing will happen to it."

Dave leaned against the doorframe. He was a compact man with a grey beard and reading glasses pushed up on his forehead, and he had the expression he always had when he was about to say something he wished he did not have to say.

"Ruth called yesterday," Dave said. "She wants to make sure you have the satellite phone."

"I have it on the manifest."

"She wants to make sure you check in every three days."

"Every three days."

"And she wants Jin Park to carry the InReach as a backup."

"Fine."

"She also wanted me to ask —" Dave stopped. He adjusted his glasses. "She wanted me to ask if you're sure about this."

"Sure about what."

"The survey. The trip. All of it."

Silas looked at Dave. Dave looked at the map on the light table. Neither of them said what they both understood, which was that Ruth's question was not about the survey. It was about whether a sixty-three-year-old man who had lost his wife two years ago and who lived alone in a house full of paintings and who had requested permission to spend three weeks in the Brooks Range mapping a creek that did not need to be mapped was making a professional decision or a personal one, and whether the distinction mattered.

"I'm sure," Silas said.

Dave nodded. "I'll pull the equipment. Give me an hour."

He left. Silas looked at the map.

The map said nothing about Margot. The map said nothing about anything except the shape of the ground and the position of water and the elevation of ridgelines and the location of the one structure within the quadrangle, a mining cabin on the Koyukuk marked with a small black square and the word CABIN in six-point Helvetica. The map was what it was. It did not ask questions. It did not wonder whether you were sure.

He folded the map and put it back in the cabinet and went to the equipment room.

The equipment room was a windowless space at the back of the building with steel shelving along the walls and a workbench in the center where instruments were cleaned and adjusted and tested. The Wild T2 theodolite was on the top shelf in its olive drab case with the dent on the south-facing edge, and he took it down and opened it and set the instrument on the workbench and looked through the telescope at the far wall, which was concrete block, painted white, with a crack running diagonally from the upper left corner to a point approximately two-thirds of the way down.

The crosshairs were sharp. The image was clear. The vertical circle read correctly. He turned the horizontal tangent screw and watched the horizontal circle move through its graduated arc, each degree divided into minutes, each minute divided into thirty-second intervals, the finest division the eye could read through the magnifying lens built into the circle reader. Thirty seconds of arc. At a distance of one hundred meters, thirty seconds of arc was approximately fifteen millimeters. At a distance of one kilometer, it was approximately fifteen centimeters. The precision of the measurement depended on the distance to the target, which depended on the terrain, which depended on the landscape, which did not care about precision.

He checked the other instruments. The Brunton compass, which measured bearing and inclination and which he carried in a leather case on his belt the way a doctor carried a stethoscope, essential and personal and worn smooth by use. The compass needle pointed to magnetic north, which was not true north, and the difference between them was called declination, and in the Brooks Range the declination was approximately twenty degrees east, which meant that if you followed the compass needle without correction you would walk twenty degrees to the east of where you intended to go, which over the course of a day's travel could put you in the wrong drainage, on the wrong ridge, in the wrong place entirely. The correction was simple — subtract twenty degrees from every bearing — but it required remembering, and the history of arctic navigation was full of people who had forgotten.

Margot had once asked him why the compass lied. He had explained that the compass did not lie; it pointed to the magnetic pole, which was a real place, a point in the Canadian Arctic where the earth's magnetic field was vertical, and the magnetic pole moved, drifting slowly westward at about fifty-five kilometers per year, so that the declination at any given point on the earth changed over time, and the declination printed on a map was only accurate for the year the map was published. The compass told the truth. It was just a different truth than the one the map assumed.

She had said: "So every compass tells a true story about a place that isn't where you think it is."

He had said: "That's one way to put it."

She had said: "That's the only way to put it."

He packed the compass. He checked the stadia rod, extending it to its full length and verifying the graduations, which were painted in red and white in alternating tenths of a foot, the rod reading like a ruler laid against the landscape, measuring the distance between the instrument and the target by the simple geometry of similar triangles. He checked the leveling rod. He checked the plumb bob, a brass weight on a string that pointed to the center of the earth, or more precisely to the local gravity vector, which was not exactly toward the center of the earth because the earth was not a perfect sphere but an oblate spheroid, flattened at the poles and bulging at the equator, and the gravity field was further distorted by variations in the density of the rock beneath your feet, so that even the plumb bob, that simplest of instruments, was subject to the same kind of distortion that affected every measurement and every map.

Nothing pointed where you thought it pointed. Nothing measured what you thought it measured. The work of surveying was the work of accounting for the distortions, of knowing what the instruments actually told you as opposed to what you assumed they told you, of correcting for the gap between the ideal and the real. It was, he sometimes thought, the work of honesty — a relentless, systematic honesty about the limitations of your tools and the imperfections of your observations and the fundamental impossibility of representing a curved, textured, living surface on a flat, smooth, static page.

He spent an hour checking equipment and signing it out on the Form 9-1380 and loading it into the bed of his truck. The morning had warmed to fifty degrees. The sky was high and pale, the washed-out blue of the subarctic summer, and the sun was already well north of east, tracking the long arc it would follow all day, never climbing high, never setting, circling the horizon like a surveyor circling a benchmark, taking measurements from every angle.

He drove back to the house. He carried the equipment inside and set it next to the duffels in the study. The room was filling up with gear, the floor disappearing under cases and bags and the careful clutter of preparation. He sat at the desk and opened his laptop and checked his email.

Ruth Atwood had written: "Manifest received. Flight confirmed with Driscoll for Thursday 0800. Jin Park arrives Fairbanks Wednesday evening, Alaska Air 164. He'll be at the Bridgewater. Call me if you need anything. Be safe."

Be safe. It was what she always wrote at the end of her field authorization emails, and it was both a genuine wish and a bureaucratic incantation, a formula that had been worn smooth by repetition the way the leather case of the Brunton compass had been worn smooth by years of riding on his belt. He closed the email without replying.

He opened the survey plan instead. It was a document he had written in March, a detailed description of the survey methodology, the control points he would establish, the traverse he would run up the tributary from its confluence with the Koyukuk to its headwaters in the saddle between the unnamed ridges. The plan called for establishing a baseline at the mouth of the tributary using GPS, then running a theodolite traverse upstream, setting stations at intervals of two hundred to five hundred meters depending on the terrain, measuring horizontal and vertical angles at each station, recording the measurements in the field notebook, computing the positions and elevations in the evening by hand and checking them against the GPS positions that Jin would record at each station.

The redundancy was deliberate. The GPS would give them positions accurate to within centimeters. The theodolite would give them positions accurate to within decimeters, maybe centimeters if the sightlines were clear and the angles were sharp. The GPS was better. He knew the GPS was better. But the theodolite traverse would provide an independent check, a second measurement made by a different method, and the agreement or disagreement between the two would tell him something about the accuracy of both. This was the principle of independent verification, and it was as fundamental to surveying as the principle of triangulation, and it was one of the things he intended to teach Jin Park, who had grown up in the age of GPS and who might not understand why you would use a nineteenth-century instrument to check a twenty-first-century one.

The answer was simple: because the GPS could be wrong. The satellites could be in the wrong position. The atmospheric correction could be off. The multipath error — the signal bouncing off a rock face before reaching the antenna — could introduce a bias that looked like a good measurement but was not. The theodolite could be wrong too, but it would be wrong in different ways, for different reasons, and the chance of both instruments being wrong in the same direction by the same amount was small enough to be dismissed. Two measurements, two methods, two opportunities for error that did not correlate. That was the architecture of reliability.

Margot had once said that his approach to accuracy was like her approach to color mixing. You never trusted one tube of paint. You mixed, and the mixing was the art, the place where judgment entered the process. He had liked the analogy. He was not sure it was correct — his mixing was mathematical, hers was perceptual — but the underlying principle was the same: distrust the single source. Verify. Check. Compare.

He closed the survey plan and stared at the screen. The cursor blinked in the search bar of his email client. He typed Margot and then deleted it. There were no emails from Margot. There had been no emails from Margot for two years and four months and he did not know why he had typed her name except that the house was quiet and the equipment was packed and the day stretched out before him with nothing left to prepare and the mind, unoccupied, returned to the datum it could not stop referencing.

In surveying, a datum was the reference surface from which all measurements were made. The datum for horizontal positions in North America was called NAD 83, the North American Datum of 1983, and it was defined by the positions of a network of control points spread across the continent, points that had been surveyed and resurveyed and adjusted until their positions were known to a high degree of accuracy, and every other position on the continent was defined relative to those points. The datum was the foundation. If the datum was wrong, everything built on it was wrong. If the datum shifted, everything shifted.

Margot was his datum. He understood this with the clarity that comes from prolonged solitude and the inability to evade one's own metaphors. She had been the reference surface against which he measured everything else — his work, his days, his sense of where he was and what the landscape meant. She was the fixed point. And the fixed point had moved, the way benchmarks move when the earth shifts beneath them, when permafrost thaws or tectonic plates drift or river erosion undercuts the ground, and now all his measurements referred to a point that no longer existed, and the positions derived from those measurements were off by an amount he could not calculate because you cannot calculate the error when the datum itself is lost.

He closed the laptop. He went to the kitchen and made more instant coffee and stood at the window and looked at the yard. The spruce trees were dark against the pale sky. A raven crossed from left to right, heading north, its wings making the sound that ravens' wings make, a heavy rustling that carried in the still air.

He had work to do. He had the flight manifest to confirm with Carl. He had the fuel calculation to verify. He had the food list to finalize, the camp layout to plan, the radio frequencies to program. He had things to check and things to pack and things to do, and the doing of them would fill the hours until Wednesday when Jin Park arrived, and then there would be the social obligation of meeting Jin and briefing Jin and feeding Jin dinner and making conversation with a young man he did not know, and that would fill Wednesday evening, and then Thursday morning Carl would fly them north and the field season would begin and the work would start and the work, the work, the work would fill everything.

He finished the coffee. He rinsed the mug. He went back to the study.

On the wall above the bookshelf, the watercolor of the Tanana in breakup glowed in the morning light. The ice in the painting was the blue-white of actual river ice, the water beneath it was the brown of actual glacial silt, and the sky was the grey that Margot had mixed from titanium white and raw umber and a trace of ultramarine, and the painting was four inches by six inches and it contained more of the Tanana River in April than any map he had ever made.

He did not look at it. He turned to the desk and opened the manifest and began checking it again, line by line, item by item, the way he had checked it yesterday and would check it tomorrow, because checking was what he did, because accuracy was the only comfort he had left, because the manifest, at least, could be verified.