The 1.3 Meters

The Projection

Chapter 8: The 1.3 Meters

The discrepancy appeared at station fourteen, on the twelfth day, during the detail work. They were taking cross-section measurements of the creek at representative stations, measuring the width and depth and gradient of the channel at five points along the traverse, the measurements intended to provide the data for the hydrographic symbology on the map — the width of the blue line, the direction of the flow arrows, the notation of rapids and falls and pools that would tell the map reader what kind of water this was.

Jin set up the GPS over the cross-section benchmark, a point Silas had marked on the left bank of the creek with a rebar stake driven into the gravel. The rebar was eighteen inches long and driven sixteen inches deep, and it was as permanent as anything you could put in alluvial gravel, which was to say not very permanent, but permanent enough for the duration of the survey. Jin stood over it with the R10 on its pole, the antenna directly above the rebar, and let the receiver collect data for ten minutes.

Silas set up the theodolite at station fourteen, a hundred meters upstream, and sighted on the same rebar stake, measuring the horizontal and vertical angles, then turning to the baseline backsight and computing the position by intersection — two angles from two known points, the two lines crossing at the location of the rebar, the intersection giving a position that was independent of the GPS, derived from different instruments, different methods, different mathematics.

He computed the position in the field notebook, working the trigonometry with the mechanical pencil, carrying the numbers through the equations, checking each step. The computation took ten minutes. He wrote the result: 67.43142 north, 150.18361 west.

"What do you have," he said.

Jin looked at the data collector. "67.43141 north, 150.18359 west."

Silas compared the numbers. The latitude agreed to within one ten-thousandth of a degree — roughly three meters at this latitude. That was normal. But the longitude differed by two ten-thousandths, which was — he computed — about six meters. Six meters was outside the expected tolerance. The GPS and the theodolite should agree to within a meter, two at the most.

"Read me the elevation," Silas said.

"GPS elevation: 2,612 feet."

Silas checked his computed elevation from the theodolite: 2,616 feet. Four feet difference. Just over a meter.

"Give me the easting and northing in UTM," Silas said.

Jin converted the coordinates. "Easting: 413,847.2. Northing: 7,482,891.6."

Silas checked his own computation. Easting: 413,848.5. Northing: 7,482,891.3. The northing agreed to within thirty centimeters. The easting disagreed by 1.3 meters.

1.3 meters.

On the ground, 1.3 meters was the distance from his left boot to his right boot if he took a long step. It was the height of the rebar stake above the ground plus the length of the rebar below the ground plus twelve inches. It was a distance he could cover in less than a second at a normal walking pace. It was trivial.

On the map, at 1:24,000 scale, 1.3 meters was 0.054 millimeters — less than the width of the finest Rapidograph line. It was invisible. It was below the resolution of the human eye. The map could not represent a discrepancy of 1.3 meters because the map could not represent anything smaller than the width of the pen that drew it, and the width of the pen was already wider than 1.3 meters at this scale.

But the discrepancy was real. Two instruments, two methods, two computations, the same point, and the answers disagreed by 1.3 meters. One of them was wrong. Or both were wrong. Or, more precisely, both were right within their respective error budgets, and the discrepancy was within the combined uncertainty of the two measurements, and the true position of the rebar stake was somewhere in the overlap between the two error ellipses, an area of probability rather than a point of certainty.

"1.3 meters," Silas said.

"That's within spec for the theodolite," Jin said. "The theodolite position has an uncertainty of about a meter at this distance."

"It's within spec. But I want to know why."

"Does it matter. For the map."

"For the map, no. For the measurement, yes. The measurement should be as good as the instruments allow. If the discrepancy is random error — atmospheric refraction, centering error, reading error — then it's acceptable. If it's systematic — a datum mismatch, a declination error, a misidentified backsight — then it means something is wrong, and the wrongness will propagate through every position computed from that backsight."

He sat down on the gravel beside the rebar stake and opened the notebook and looked at the computation. The trigonometry was correct. He checked the angles: the horizontal angle from station fourteen to the rebar was one hundred thirty-two degrees, seven minutes, forty-five seconds. The horizontal angle from station fourteen to the backsight was zero degrees, zero minutes, zero seconds. The distance from station fourteen to the rebar, computed from the GPS positions of both points, was ninety-eight point three meters. These numbers were what he had used in the computation, and the computation was correct, so the discrepancy was in the input — in the angles or the distances or the GPS positions themselves.

He reoccupied station fourteen with the theodolite. He re-leveled the instrument. He sighted on the rebar stake and read the angle: one hundred thirty-two degrees, seven minutes, thirty seconds. Fifteen seconds different from the first reading. Fifteen seconds of arc, at a distance of ninety-eight meters, was approximately seven millimeters. Not enough to explain the 1.3-meter discrepancy.

He sighted on the backsight. Zero degrees, zero minutes, fifteen seconds. Fifteen seconds off from zero. He had not zeroed the circle precisely on the first observation. The accumulated error from the imprecise zeroing, propagated through the angle difference to the rebar, would be — he computed — approximately eleven centimeters. Still not enough.

"I want to reobserve this point," Silas said. "Set up the GPS again. Fresh observation. Ten minutes."

Jin set up the GPS. Silas set up the theodolite. They measured the rebar stake again, independently, simultaneously, two instruments, two operators, two records.

The GPS gave: Easting 413,847.4, Northing 7,482,891.5, Elevation 2,612 feet. Two-tenths of a meter different from the first GPS observation. Normal variation.

The theodolite computation gave: Easting 413,848.3, Northing 7,482,891.4, Elevation 2,615 feet. Two-tenths of a meter different from the first theodolite computation. Normal variation.

The discrepancy between the two: 0.9 meters in easting. Three feet in elevation.

"It went down," Jin said. "From 1.3 to 0.9."

"The discrepancy is random. It's in the noise. Both instruments are measuring the same point and getting slightly different answers, and the difference is within the combined uncertainty."

"So 1.3 meters was just noise."

"Probably. But probably is not certainly. If I had a third measurement — a different instrument, a different method — I could distinguish noise from systematic error. With two measurements, I can only see the discrepancy. I can't diagnose it."

"What would a third measurement be."

"Historically, it would be a tape measurement. A steel tape, stretched between the two points, reading the distance directly. No electronics, no computation, no satellites. Just a tape and two hands and the law of cosines."

"We don't have a steel tape."

"No. We have a laser rangefinder. Same principle, different technology."

Silas took out the rangefinder and shot the distance from the rebar stake to station fourteen: ninety-eight point one meters. His GPS-derived distance had been ninety-eight point three. Twenty centimeters difference. He recomputed the theodolite position using the rangefinder distance.

New result: Easting 413,847.9, Northing 7,482,891.4. The discrepancy with the GPS was now 0.5 meters in easting. Half a meter.

"Getting closer," Jin said.

"Getting noisier. Each new measurement changes the answer. That's what happens when you're working near the precision limit of the instruments. The measurements bounce around the true value, and the average is better than any single measurement, but you never get the exact answer because the exact answer doesn't exist at this level of precision."

"The exact answer doesn't exist."

"Not operationally. In theory, the rebar stake has an exact position — a single point in three-dimensional space. In practice, the position we can determine is a probability distribution, a cloud of possible positions centered on the most likely position, and the cloud is about a meter wide for the theodolite and about two centimeters wide for the GPS, and the true position is somewhere in both clouds, in the overlap, which is a smaller cloud, a better estimate, but still not a point. Still not exact."

Jin looked at the rebar stake. It was a piece of steel driven into gravel, sixteen inches below the surface, two inches above. It was there. It had a position. You could touch it. You could stand on it. You could paint it, if you were Margot, you could paint the rebar stake in the gravel beside the creek with the willows behind it and the ridge above and the sky above that, and the painting would not contain the position, would not contain the easting or the northing or the elevation, but it would contain the rebar stake, the object itself, in its context, in its place, in the particular quality of light and shadow that made it this rebar stake in this gravel beside this creek and not any other rebar stake anywhere else.

"So what do you put on the map," Jin said.

"The best estimate. The average of the measurements, weighted by their precision. The GPS position, because it's more precise. But I note the discrepancy in the notebook, because the discrepancy is information. It tells me that the theodolite position at this station has a larger uncertainty than at other stations, which might mean the sightline was obstructed, or the atmosphere was refracting, or I made a reading error. The discrepancy is a flag. It says: pay attention here."

"And on the map, 1.3 meters is invisible."

"On the map, 1.3 meters is invisible. On the ground, 1.3 meters is the difference between standing on the bank and standing in the creek. Context determines significance. The map has one context. The ground has another. The measurement has a third. The same number means different things in different contexts."

Silas closed the notebook. He looked at the rebar stake, the creek, the willows, the ridge. The light was midday flat, the sun high enough to eliminate most shadows, the terrain even and grey-green and indifferent to the conversation that two human beings were having about 1.3 meters of uncertainty in a world that was, at every scale, uncertain.

He thought about Margot. He thought about the 1.3 meters not as a measurement but as a metaphor, though he distrusted metaphors and preferred data, because data could be checked and metaphors could not. But the metaphor was persistent. The 1.3 meters between the two measurements of the same point was the 1.3 meters between two people measuring the same life, the same marriage, the same thirty-three years of shared terrain. She had measured their marriage with one instrument — the instrument of feeling, of perception, of attention to the color and texture and light of their days together. He had measured it with another instrument — the instrument of commitment, of reliability, of the steady accumulation of days spent in proximity, the traverse of a life run along a baseline of mutual purpose.

And the two measurements had disagreed. Not by much. Not by an amount that was visible on the map of their marriage, the map that their friends and family would have drawn if asked — a good marriage, a strong marriage, a marriage of two people who worked and lived and traveled together and who loved each other in their different ways, his love steady and structural and largely unexpressed, hers love vivid and articulated and frequently declared. The map showed agreement. The data showed a discrepancy.

The discrepancy was in the elevation. He had placed her, in his internal model, at the same height as himself — equal, parallel, traveling the same contour line. She had placed herself differently. She had wanted something he had not measured, had not even known how to measure — a quality of attention, a depth of presence, a willingness to be in the moment rather than always measuring the moment. She had wanted him to put down the theodolite and look at the landscape with his eyes instead of through the instrument. She had said this. She had said it many times, in many ways, with increasing specificity as the years passed and the discrepancy became harder to ignore.

"You measure everything," she had said once, in the Fairbanks house, in the kitchen, in the late evening light of a September night. "You measure the distance from the house to the street. You measure the temperature. You measure the rainfall. You measure everything except the thing that matters."

"What thing."

"How I feel. How you feel. How we feel. The distance between us."

"There is no distance between us."

"There is 1.3 meters between us," she did not say, because she did not know about 1.3 meters, because the number did not exist yet, because the rebar stake had not been driven and the measurement had not been taken and the discrepancy had not been computed. But the distance was there. The discrepancy was there. The two instruments measuring the same marriage, getting different answers, and neither answer wrong, both within spec, both valid within their respective error budgets, and the true position of the marriage somewhere in the overlap between the two probability clouds, a better estimate than either alone but still not exact, still not a point, still a cloud.

"Silas," Jin said.

Silas looked up. He had been sitting on the gravel beside the rebar stake, staring at the creek, for several minutes. Jin was standing above him, the GPS pole in his hand, the data collector in the other, waiting.

"We should move on," Jin said. "The next cross-section is at station eleven."

"Yes," Silas said. "We should move on."

He stood up. His knees protested. The crepitus was loud in the quiet of the valley, a sound like gravel being crushed, which was what it was — the cartilage in his knee joints grinding against bone, the surfaces no longer smooth, no longer congruent, the fit that had been precise at twenty-five now imprecise at sixty-three, the wear of thirty-eight years of walking uneven ground with heavy instruments on his back.

They walked downstream to station eleven. They measured the cross-section. The measurements agreed. The discrepancy at station fourteen did not repeat. It had been noise, as Silas had suspected, a random fluctuation in the data, a ghost in the measurement, a 1.3-meter uncertainty that was, at the scale of the map, invisible, and at the scale of the life he was examining in the privacy of his own mind, everything.

He recorded the cross-section data in the notebook. Width of creek: fourteen point two feet. Depth at center: two point one feet. Gradient: three percent. Substrate: cobble and gravel. Current: moderate, estimated at three feet per second. Water temperature: forty-one degrees Fahrenheit. The numbers were precise and specific and they described the creek at this point on this day at this time, and they would not describe the creek tomorrow or next week or next year, because the creek would change, the water level would drop as the snowmelt diminished, the gravel would shift in the next flood, the temperature would rise as the summer advanced, and the measurements would become historical, a record of a moment that had passed, a snapshot of a river that was never the same river twice.

Heraclitus had said this, twenty-five hundred years ago, and Silas had always thought it was wrong, or at least imprecise. You could step into the same river twice. You could stand at station eleven and measure the creek today and come back tomorrow and measure it again, and the river would be recognizably the same river — the same channel, the same banks, the same boulders, the same general character. What was different was the water. The specific molecules of water that passed your feet at ten in the morning on June 19 were not the same molecules that would pass your feet at ten in the morning on June 20. The container was the same. The contents were different. The map showed the container. The measurement showed the contents. Both were true. Neither was complete.

Margot would have disagreed. Margot would have said that the river was not the container or the contents but the relationship between them — the way the water shaped the channel and the channel shaped the water, the dialogue between the moving and the fixed, the conversation that produced the sound of the river, which was the sound of water and rock in continuous negotiation, continuous adjustment, continuous mapping of the liquid onto the solid, and the map of that negotiation was not a topographic map and not a painting but something else, something that neither of them had quite managed to make, the representation that captured both the container and the contents, both the contour and the color, both the measurement and the feeling.

He had not made that map. He would not make it now. The map he was making was the map he knew how to make — the map of contour and distance and elevation, the map that preserved shape and sacrificed color, the map that represented the container and not the contents, the geometry and not the experience.

But he would make it well. He would make it with the precision and the judgment and the patience that thirty-eight years had taught him, and the map would be his last statement about the relationship between the measured and the real, and the statement would be incomplete, as all maps were incomplete, as all projections were incomplete, as all representations of curved surfaces on flat pages were incomplete, but it would be the best he could do, and the best he could do was, in the end, the only standard that mattered.

They finished the cross-section measurements by evening. Five cross-sections, spread along the traverse, the data recorded in the notebook and the data collector, the dual record intact. Tomorrow they would take spot elevations on the ridges. The day after, they would break camp.

Silas sat on the log at the edge of the gravel bar and looked at the creek. The 1.3 meters was still in his mind, not as a number but as a fact — the fact that two measurements of the same thing could disagree, that the disagreement was not an error but a property of the measurement process itself, a consequence of the instruments' limitations and the observer's imperfections and the fundamental impossibility of knowing anything exactly.

You could get close. You could measure and remeasure and average and adjust and weight by precision and compute the most probable value and express the result with a confidence interval and a standard deviation and a statement of accuracy. You could do all of this, and you would still not know the exact position of the rebar stake, or the exact width of the creek, or the exact elevation of the saddle, or the exact distance between yourself and the person you loved.

The creek ran. The light held. The 1.3 meters remained.