The skin tracks held to the south edge of the meadow where the wind had scoured the surface to a polish you could almost see your face in. Yusra broke trail because she liked breaking trail, and because Devon's knee had a history she respected. The younger one — Pem, three seasons in, from a soil-physics group at Reno — followed Devon with the patient stride of someone who had learned not to crowd.
The crust held her for two strides and then dropped her through to the calf. Below, the depth hoar gave way like sugar.
"There it is," Devon said behind her, agreeably. "That's the layer."
"It's everywhere?"
"From about 9,400 up. The model meeting on Tuesday spent an hour on it. Whether to call it a season feature or a regime feature."
"Regime," Yusra said. She kicked free and tested another step. The crust took her, then didn't. "Last year was like this above the confluence. This is — what, four hundred meters lower?"
"Three-eighty. Pem mapped it Wednesday."
"I sent you the polygons," Pem said.
"You did. I read them on the drive up from Montrose." She had, in fact, read them on her phone at the Almont gas station, parked under a sky the color of wet slate, eating a granola bar that had been in her pack since Vancouver. Her partner had texted from Burnaby: the boys want to know if you're bringing back the rocks again. She had typed yes and then tell them only small ones this time and then sat for a minute looking at the snow line on the Anthracites, which was higher than it should have been in March, and lower than it had been in February, which meant something had melted and refrozen in between, which her piezometers would have something to say about.
They moved up through the willows. The willows were not where she remembered them from 2034. Some of that was her memory. Some of it was the willows.
"Tell me about the diagnostic," she said over her shoulder.
Devon caught up on a flat. "P-7 flagged a drift on the upper transducer. Self-corrected against the redundant, logged the offset, sent the packet. Nothing dramatic. P-4 wants a battery within the month, which is fine, we're scheduled. P-9 — " he paused, "P-9 is interesting."
"How interesting."
"It's reading a water table four meters higher than this time last year."
She stopped on the crust and turned. Devon's beard had more grey than the last residency. The reflective patches on his pack were peeling. He looked very satisfied with himself, in the way of a person who has been thinking about a thing for two weeks and is waiting for someone to ask.
"Four meters."
"Four meters and twelve centimeters."
"At P-9. The upper site."
"The upper site that we put in in '38 specifically because you said the recharge signal would show up at elevation before it showed up below."
"I said it might."
"You said it would. I have the email."
Pem, behind them, was grinning at her ski tips.
Yusra started moving again. The thing about P-9 was that P-9 sat in fractured granite above what had been, until recently, the upper edge of the persistent-snowpack belt. The thing about the fractured granite was that nobody had been able to say, until the boreholes went in, how deep the active circulation reached. The thing about this winter — the thinness of it, the way the December storms had come as rain to 10,200 feet and then nothing for six weeks — was that any water in the system at P-9 right now had to be old. Had to be from somewhere else. Some other year, banked.
She wanted to see the curve. She wanted to see how it had moved through January.
"Did Aiyana look at it?" she asked. Aiyana was the model that lived on the network's servers and read the piezometer packets and wrote little notes about them. The network had named her, the previous summer, after a long argument Devon had refused to take part in.
"She thinks it's real," Devon said. "She flagged it Tuesday. Wrote up a one-pager. She thinks the source is the snowpack two basins north, routed through the fracture system. She wants more data."
"She always wants more data."
"She's not wrong this time."
"She's never wrong. That's what's annoying."
They climbed. The Pumphouse cluster sat in a small bench above the confluence, where the East River bent west and the morning sun, when there was morning sun, came over the ridge in a slab. There was morning sun today, weak and very white, the kind of March light that didn't warm anything but made the snow surface look like it had ideas about the afternoon. Yusra checked her watch. They had three hours before the surface started to rot.
The cluster was where they'd left it. The newer cabinets — '38 vintage, the up-drainage extension — were a slightly different grey than the originals. Devon brushed snow off P-9's solar panel with the side of his glove and crouched to the cabinet. Pem unclipped from her skis and laid them in an X on the surface and started pulling cable testers from her pack with a fluency Yusra had not had at thirty-one.
"Pull up the January trace," Yusra said.
Pem already had the tablet out. The trace came up — a slow climb from a low in late December, then a stepped rise through January, then the flat shoulder of February, then the recent uptick. Yusra crouched in the snow and looked at it.
It was beautiful. It was a curve nobody had ever seen at this elevation in this basin. It was data her grad students in Burnaby would spend a year on, and she would spend ten.
"Hi," she said to it, quietly.
"You're talking to the piezometer again," Devon said.
"I'm talking to the water."
"Same thing."
Pem said, "Aiyana's note says she'd like a deuterium sample if we can get one. She's drafted a sampling protocol. It's — " she scrolled, "it's reasonable, actually. She suggests we use the upper port and bypass the standpipe."
"Why."
"Residence time. She thinks the standpipe is mixing."
Yusra read the note over Pem's shoulder. It was reasonable. It was more than reasonable. It was the protocol she would have written if she'd been thinking about it for two weeks instead of two minutes, and it included a comment about a paper from 2019 that she had cited in her dissertation and had not thought about in eleven years. She read the protocol twice. Then she said, "Fine. Tell her fine. But we sample tomorrow, not today. I want to think about it."
"She'll be disappointed."
"She'll get over it."
Devon was already pulling the cabinet panel. The wind, which had been doing nothing, picked up briefly and dropped a fine spray of surface snow across the open electronics. He closed the panel without comment. Pem handed him the tester. They worked.
Yusra stood up and walked ten meters off and looked down-valley. From here you could see the dieback on the north slope above Gothic — the grey patches in the spruce where the beetles had finally finished what the drought started, four years ago now, five. The patches were not getting bigger this winter. They were also not getting smaller. They were a fact about the slope, the way the willows were a fact about the bench.
Her phone, improbably, had a bar. The text from her partner had a follow-up: also Sami's recital is the 24th, you said you'd be back by the 22nd, confirming?
She typed yes confirming, flying Saturday. Then she stood for a second and typed the upper piezometer is showing old water, I think we found the buffer. Then she deleted that, because her partner did not need it at 7am Burnaby time, and typed instead miss you, give Sami a kiss.
Devon called from the cabinet: "Calibration's clean. Drift was real. She caught it."
"Good," she said.
She skied back to them. Pem was repacking the testers. The light had shifted on the slab above the river. The wind crust under their skis would soften in an hour and then, by the time they came down, would be the kind of breakable that ended knees.
"Wednesday," she said to Pem, "I'm briefing my students. Can you send me the January trace and Aiyana's note tonight? I want to draft the first paragraph before I lose it."
"On it."
"And tell her" — she nodded at the cabinet — "I want the deuterium protocol with the standpipe comparison too. Not instead. Both."
Pem smiled and bent to her bindings.
Devon was already moving downslope, picking the south edge again, where the crust held.