The Saddle

The Projection

Chapter 7: The Saddle

They reached the headwaters on the eleventh day. The tributary had narrowed steadily as they moved upstream, the valley walls closing in, the creek diminishing from a fifteen-foot stream to a six-foot brook to a three-foot rill to a seep emerging from a patch of saturated tundra at the base of the saddle, the water coming from the ground itself, from the thawing permafrost and the snowpack on the north-facing slopes, the source not a spring but a general upwelling, the land sweating its stored water the way a sponge sweated when you squeezed it.

The saddle was a depression in the ridgeline, a low point between two summits, the terrain sloping down from the summits on both sides to a flat area perhaps fifty meters wide and two hundred meters long, oriented east-west, the wind funneling through it from the Arctic side to the Interior side, a wind that was colder than the wind in the valley below and that carried a different quality of light, a thin, metallic light that came not from the sun directly but from the sky, the light scattered and re-scattered by the atmosphere until it had no direction, no angle, no shadow, the light of high places.

Silas stood in the saddle and looked north. Beyond the saddle, the ground dropped away into the next drainage, another valley, another creek, another unmapped territory that ran north toward the Arctic plain. He could see the drainage for perhaps five miles, the valley widening as it descended, the tundra spreading out, the creek — visible as a dark line in the green — winding through the center of the valley. Beyond the valley, more ridges. Beyond the ridges, more valleys. The terrain extended to the horizon and beyond the horizon, north and north and north, five hundred miles of it, unmapped, the last great unmapped territory in North America, though mapped now by satellite and aerial photograph, the land represented at scales and resolutions that would have astonished the surveyors who first came here a hundred years ago, represented but not measured, not walked, not stood upon by a human being with a theodolite and a notebook.

Jin came up beside him. He was breathing hard. The last half-mile of the traverse had been steep, the creek climbing through a series of ledges and waterfalls, the terrain too rough for regular station spacing. They had set stations wherever they could find a flat spot large enough for the tripod, and the stations were closer together than in the lower reach, the intervals shrinking as the terrain complexity increased, the map demanding more data as the ground became more difficult to represent.

"Is that it," Jin said. "The headwaters."

"That's it. The tributary starts here."

"Starts or ends."

Silas looked at him. "Depends on your perspective. The water starts here and flows downstream. The survey starts at the mouth and comes upstream. The creek and the surveyor travel in opposite directions. They meet here."

"At the source."

"At the divide. The saddle is a drainage divide. Water on this side flows south, into the Koyukuk. Water on the other side flows north, into whatever drainage is over there. The divide is the highest line the water touches. It's the contour line that separates two drainage basins."

"And the map stops here."

"The survey stops here. The map extends a little beyond the survey — you draw the contour lines up to the ridge on both sides of the valley, using the stations as control and the terrain as a guide. But you don't map beyond the ridge. That's another survey."

Jin put his hands on his hips and looked north. The wind blew his hair across his forehead. He was wearing a fleece hat that he had pushed back, and his face was weathered from eleven days in the field, the cheeks reddened by wind, the skin around his eyes creased from squinting against the light. He looked less like the young man who had stepped off the plane at Fairbanks International and more like a field worker, a person shaped by the conditions, adapted to them, fitted to them the way the tundra was fitted to the climate, surviving not by resisting the environment but by conforming to it.

"We need to set the last station," Silas said.

They set up the instruments in the saddle. The last station. Station twenty-six. The terminus of the traverse. Silas leveled the theodolite and sighted back down the valley, toward station twenty-five, which was visible as a pink flag on a stake three hundred meters below, on a ledge above the last waterfall. The horizontal angle: zero degrees, zero minutes, zero seconds. He had set the circle to zero on the backsight, establishing the reference direction for the final station.

He turned the telescope. He sighted on the summit to the east: forty-seven degrees, twelve minutes, thirty seconds. He sighted on the summit to the west: three hundred nineteen degrees, five minutes, fifteen seconds. He sighted on the rock outcrop at the base of the north-facing slope: one hundred seventy-two degrees, forty-one minutes, zero seconds. He read the vertical angles. He called the numbers to Jin, who wrote them in the notebook.

Jin set up the GPS and took a ten-minute observation. The position stabilized: 67.4487 north, 150.1923 west, elevation 4,214 feet. The last coordinates of the survey. The highest point of the traverse.

Silas wrote the position in his notebook. He drew a small circle around the station number, as he always did for the last station in a traverse. The circle was a convention he had adopted from Henderson, who had circled his last stations as a way of marking the end, a cartographic period at the end of a topographic sentence.

He looked at the notebook. Twenty-six stations, spread over three and a half miles of creek valley, climbing from 2,117 feet at the mouth to 4,214 feet at the headwaters. Two thousand ninety-seven feet of elevation gain. An average gradient of twelve percent, which was steep for a valley floor, steeper than most drainages in the Brooks Range, an indication of the youthfulness of the stream, the geological recency of its cutting, the ongoing work of water on rock.

The traverse was complete. The measurement phase was finished. What remained was the reduction: computing the final positions, adjusting the traverse, plotting the stations on the base map, interpolating the contour lines, drawing the finished map. The fieldwork was almost done.

But not quite. There were detail measurements to make — cross-sections of the creek at representative stations, spot elevations on the ridges, vegetation observations, geological notes. The detail work would fill three or four more days, and then they would break camp and walk down to the gravel bar and call Carl on the satellite phone and wait for the plane.

They ate lunch in the saddle, sitting on the tundra, the wind steady from the north, cold enough that Silas put on his down jacket. The tundra was low and dense, a mat of moss and lichen and tiny flowering plants — mountain avens, moss campion, alpine azalea — the plants clinging to the ground, growing horizontally instead of vertically, adapting to the wind and the cold and the short growing season by staying low, by spreading rather than rising, by covering the ground rather than reaching for the sky.

Margot had painted tundra many times. She had been fascinated by the texture of it, the way the plants interlocked and overlapped, creating a surface that was not smooth but was continuous, a living quilt of greens and browns and the small bright dots of flowers. She had painted tundra close up, the way you might paint a portrait, filling the canvas with the intricate detail of the mat, the individual leaves and stems and flowers visible, the scale enlarged until the tundra became a landscape in itself, a landscape within the landscape, a territory that existed at a scale the map could not reach.

She had painted tundra from a distance too, from ridgelines and summits, looking down at the vast green expense of it, the tundra stretching to the horizon, and from that distance the texture was invisible, the individual plants absorbed into the aggregate, the way individual data points were absorbed into a contour line, and the tundra became what it was at the map's scale: a color, a symbol, a green screen on the quad sheet that meant vegetation, open, tundra, without specifying what kind of tundra or how dense or how tall or what was flowering.

Two scales. Two truths. The close truth of the individual plant and the far truth of the aggregate. Margot had captured both. The map could capture only the far truth. The close truth was lost in the scale reduction, the way all close truths were lost when you stepped back far enough, the way the expression on a face was lost when you stepped back far enough, the way the texture of a brushstroke was lost when you stepped back far enough, until what remained was the shape and the color and the general impression, which was what the map provided — a general impression of the terrain, accurate in shape, accurate in position, but stripped of the detail that made the terrain a place and not a position.

"Silas," Jin said.

"Yes."

"I've been thinking about what you said. About the hand-drawn map."

"What about it."

"You said the hand makes decisions the software doesn't. But the software makes decisions too. It chooses the interpolation algorithm, the smoothing parameter, the contour interval. Those are decisions."

"Those are parameters. Decisions are different. A decision considers context. A parameter applies a rule. When I draw a contour line past the cliff we measured last week, I decide how the line should curve based on what I know about the geology, the drainage, the shape of the slope above and below the cliff. The software doesn't know any of that. It applies an algorithm to the data points and draws a curve. The curve might be right. It might not. The software doesn't know the difference."

"But you could teach the software. You could add geological data, drainage data, slope data. You could build a model that accounts for context."

"You could. People are doing it. Machine learning, AI, neural networks trained on topographic data. The models get better every year. Someday they'll be as good as a human cartographer. Better, probably, because they can process more data, consider more variables, draw more lines in less time."

"Does that bother you."

"No."

"Really."

Silas looked at the tundra. The wind moved across it in visible waves, the plants bending and releasing, the surface undulating like water.

"No," he said. "The work is the work. It doesn't matter who does it, or what does it. What matters is that it's done well. If a machine can draw a better contour line than I can, the map is better. The map is the point, not the cartographer."

"Then why do you do it by hand."

"Because I can. Because I know how. Because the hand is the instrument I've spent my life learning to use, and using it well is — " He stopped. He was going to say satisfying, but that was not the right word. "It's honest. The hand-drawn line contains the judgment of the person who drew it. It's a record of the decisions. Every curve, every inflection, every place where the line bends or straightens or thickens or thins — those are decisions, and they're visible in the line. The software hides its decisions. The algorithm is invisible. The output looks clean and precise but the decisions that produced it are buried in the code, and you can't read them by looking at the map."

"You can read decisions in a hand-drawn line?"

"If you know what to look for. Henderson could look at a map and tell you where the cartographer was uncertain, where the data was thin, where the terrain was ambiguous. He could read the line the way you read handwriting — the speed, the pressure, the hesitation. The line was a record of the encounter between the cartographer and the terrain. The software line is a record of the algorithm. Both are valid. But they're different kinds of records."

Jin was quiet. He seemed to be considering this, weighing it against his training, his experience, his assumptions about the relationship between data and representation. Silas did not push. The ideas would settle or they would not. Understanding could not be forced. It could only be offered, the way a control point was offered to the traverse — here is a fixed position, use it if you need it, ignore it if you don't.

They packed the equipment and began the walk back down the valley. The descent was easier than the ascent, gravity assisting, the creek leading the way, the flagged stations marking the route. They passed station twenty-five, station twenty-four, station twenty-three, the pink ribbons bright against the grey-green tundra, each ribbon a marker of a measurement taken, a position fixed, a point in the network that would become the map.

At station twenty, where the valley widened and the gradient eased, Silas stopped. He looked up at the ridgeline to the east, where the late sun was catching the rocks, turning them gold.

"I want to go up there tomorrow," he said.

"To the ridge."

"I want to see the valley from above. I want to see the drainage pattern, the way the tributaries come in, the shape of the terrain from a perspective that the creek bottom doesn't give me."

"You could use the satellite imagery."

"I could. But the satellite imagery doesn't show me what the ground looks like to a human eye. It shows me what the ground looks like to a sensor that records electromagnetic radiation in specific wavelength bands. That's useful. It's not the same."

"What's the difference."

"The difference is what Margot would have called the experience of looking. The satellite doesn't experience. It records. The human eye experiences. It sees color, depth, texture, movement. It sees the wind on the tundra. It sees the light on the rocks. It sees the things that don't show up in the data because they're not data — they're perceptions, impressions, the subjective content of a particular point of view at a particular time."

He had mentioned Margot again. This time he did not regret it. The mention was accurate. She would have said exactly that. She had said exactly that, many times, in many ways, the argument between the measured and the perceived, the argument that was not an argument but a dialogue, a conversation between two ways of knowing the same territory, two projections of the same curved surface, each preserving what the other sacrificed.

They walked back to the camp in the long evening light. The creek ran beside them, clear again after the storm, the water catching the low sun and throwing it back in fragments, the surface broken and rebroken by the rocks and the current, the light shattered and reassembled with every ripple. Jin walked ahead, his pack bouncing with his stride, his trail runners finding the path through the willows with the efficiency of a young body that had adapted to the terrain in eleven days, that had learned the field the way the field taught everything — by immersion, by repetition, by the slow accumulation of experience that could not be transmitted by lecture or manual but only by being there, on the ground, in the weather, with the instruments.

Silas walked behind him, slower, his knees complaining on the downhill, the crepitus audible in the quiet of the valley, and he thought about the traverse, about the twenty-six stations and the three and a half miles of creek and the two thousand feet of elevation and the thousands of measurements that he and Jin had made over eleven days, and he thought about the map that would emerge from those measurements, the contour lines that would trace the shape of the valley, the blue line that would follow the creek, the brown lines that would climb the ridges, the whole representation that would capture the geometry of the terrain and nothing else, the geometry and not the color, the geometry and not the light, the geometry and not the wind or the water or the tundra or the quartzite cliff or the saddle where the drainage divided and the water went two ways and the surveyor went one.

The camp was visible from a quarter-mile away, the tent a bright yellow dot on the tundra bench, the food bags hanging in the willows, the tripod standing where Silas had left it, pointed at the reference rock, the rock that had been the first thing he sighted on and would be the last thing he sighted on, the backsight that oriented the survey, the fixed point that made everything else relative.

They reached the camp. Jin dropped his pack and went to the creek for water. Silas sat on the log that served as a bench and looked at the valley and the creek and the mountains and the sky.

The saddle was up there, invisible from here, hidden behind the ridges and the curves of the valley. The headwaters were up there, the seep where the water came from the ground, the highest point of the drainage. He had measured it. He had stood in it. He had recorded its position to within a few centimeters, which was more precisely than any human being would ever need to know where the headwaters of this unnamed creek were, and the precision was unnecessary and the measurement was unnecessary and the survey was unnecessary and he had done it anyway, because the work was not about necessity, the work was about completion, about finishing what he had started thirty-eight years ago when he picked up a stadia rod in the Wrangells and learned to read the landscape through an instrument and fell in love with the act of measurement, the act of turning the world into numbers and the numbers into lines and the lines into a map that someone, someday, might hold in their hands and read the way Margot read a landscape, with understanding, with appreciation, with the recognition that what they were seeing was not the thing itself but a representation of the thing, faithful in some ways, unfaithful in others, a projection.

Jin came back with water. They cooked dinner. They ate in silence, watching the light change, the sun tracking north along the ridgeline, the shadows lengthening on the tundra, the valley filling with the amber light of the Arctic evening that was not an evening but a continuation, a persistence of day that refused to become night.

"Three more days," Silas said. "Detail work. Cross-sections, spot elevations, ridge observations. Then we break camp."

"And then."

"Then we go home. Then I draw the map."

"How long does that take."

"The drawing. A month, maybe. The data reduction takes a week. The plotting takes a week. The drawing takes two weeks, maybe three. Depends on how complex the terrain is."

"A month."

"For a hand-drawn map at 1:24,000 scale with twenty-foot contour intervals covering three and a half miles of creek valley. A month is fast."

"How long would the software take."

"An hour. Maybe less."

Jin said nothing. The comparison was obvious and did not need to be spoken. An hour against a month. The efficiency of the machine against the deliberation of the hand. The speed of the algorithm against the patience of the cartographer. It was not a contest. It was an inevitability, and Silas had made his peace with it years ago, the way you make your peace with any force that is larger than you and moving in a direction you cannot change. The hand-drawn map would not survive the transition to digital cartography. It would become an artifact, a relic, a specimen of a craft that had been practiced for five hundred years and was ending in his generation. This did not trouble him. What troubled him was the thought that the craft would take with it the particular way of knowing the terrain that the craft embodied — the intimacy of the hand following the contour line, the judgment of the eye reading the terrain between the stations, the patience of the mind holding the entire valley in memory while the pen moved across the vellum, drawing the shape of the ground one line at a time.

"Jin," Silas said.

"Yes."

"When you get back to Anchorage, you'll put the data into the GIS. You'll generate the contour lines by algorithm. You'll produce a digital elevation model of this valley."

"That's the plan."

"Good. Do that. But also look at my map when it's done. Compare the two. See where they agree and where they differ. See where the algorithm got it right and where it got it wrong. See where the hand added something that the data alone couldn't provide."

"I will."

"And then decide for yourself which one is the better representation. I won't tell you the answer. The answer depends on what you think a map is for."

Jin looked at him. The light was golden on his face, the shadows deep in the hollows of his cheeks, and Silas saw in his expression something that might have been curiosity or respect or the beginning of understanding, and he thought: this is enough. This is what you can give. The rest is not yours to give.

They cleaned up. They stowed the equipment. They crawled into the tent. The light came through the fabric, muted and warm, and the creek ran and the wind blew and the valley held them the way the valley had held everything that had ever been in it — temporarily, gently, without attachment, the way the land held the water that passed through it and the light that fell on it and the people who measured it and the people who painted it and the people who loved each other in its presence and the people who grieved in its presence, all of them temporary, all of them passing through, all of them leaving behind whatever record they could make — a map, a painting, a field notebook, a memory — and the land not caring, the land not noticing, the land being what it was, the thing that the projections tried to capture and could not, the curved surface that the flat page distorted, the territory that the map was not.