Tuesday, the eleventh of April 2051, twenty past six in the morning, the Holtz junction box on the north fence.
In October she told Gary Holtz to ask her in spring, off his laterals, not the main. It's spring. He's there before she is, sitting on the tailgate of his pickup with a thermos, which means he remembered. Which means he's been chewing on it all winter. Which means she's about to disappoint a man before breakfast.
The lid comes off with the bar, both hands. The box smells like wet concrete and frogs. The east lateral is running harder than in October, the west about the same. She times them into the cup. East, sixteen ounces in fifty-eight seconds. West, a minute forty. Meter: east EC 580, pH 7.24. West 610, 7.19. The main she leaves alone this time. The main is a pipe she can't get her arm around, and Gary's two laterals are a couple of cups a minute going into it. It can't tell them anything about Gary.
The drop kit says east 3.6, west 3.1. October said 3.3 and 2.9. Both up three tenths or less, which is exactly how much the kit shrugs on a cold morning with numb fingers. So she doesn't write the kit numbers down as answers. She fills two more 250 mL bottles under the surface, caps them underwater so no air rides along, and puts them on ice. Dale runs them at ten on the surplus Hach with the sulfuric. East 3.52. West 3.08. Good to about five hundredths.
Then she calls Dale back. "Do you still have my October laterals?"
He does. Frozen, in the chest freezer behind the bait. He thaws them and runs them at two. East 3.29. West 2.94.
So east is up about a quarter and west about an eighth, on a lab that can see that small. Two of something.
Gary calls at three. "So it's my lime."
"It's two points. One in fall, one in spring. Fall water and spring water are different water. It could be your lime. It could be April."
"When do I get to know?"
"When I have an April from a year you didn't lime. I don't."
He goes quiet. On his end a radio is reading hog futures. Then: "Keep doing the freezer thing." From Gary, that's a yes.
218. 4/11/51 HOLTZ JCT BOX 2ND PT. E LAT EC 580 pH 7.24 (CUP :58), W 610/7.19 (1:40). KIT ALK 3.6/3.1 = WITHIN KIT. DALE TITR: 3.52/3.08 VS OCT (FROZEN, RE-RUN) 3.29/2.94. UP, SEASON NOT SEPARATED. LATERALS FROZEN FROM NOW. 2.1 H, GRAY.
Tuesday, the eighteenth of April, twenty to twelve, the district counter.
Marcy Kuhl comes in with paper again. In October she asked the question in writing, exactly the way Janelle wrote it across the top of the quarterly: What head did you use for the week of August 8. It took the partnership agent five months and two follow-ups to answer.
Head values for SCHAEFER-17 Aug–Sep 2050 were derived from a precipitation-driven stage model calibrated to District storm observations (S14). Modeled stage is standard practice where continuous stage data is unavailable.
"Calibrated to S14," Janelle says. "One storm. One head. They took one of my readings and made eight weeks out of it."
"Is that allowed?"
"It's allowed for them. It's not ours."
Marcy has her reading glasses on this time, which means business. "I asked them to restate Q3."
Second page. The agent's reply to that, nineteen seconds later by the timestamp:
Q3 2050 payments are final under vintage terms. Methodology refinements will apply prospectively.
Prospectively. Fourteen farms, paid in October off rain that fell on a model.
Janelle doesn't argue with the agent. Arguing with it is like arguing with a vending machine that has already dropped the wrong candy. Instead she adds one line to the public method note for #17, right under the rating title, so the next scraper hits it before anything else:
NO DISTRICT STAGE DATA AT #17 EXCEPT STORM VISITS LISTED. ANY OTHER #17 HEAD IS NOT DISTRICT DATA.
Then she copies the agent's sentence into the site history, word for word with the quotation marks, in a row with today's date in the first field and THIRD-PARTY DISCLOSURE in the second.
Marcy reads over her shoulder. "You're putting their words in your file?"
"Their words are the only proof they said them."
Ray, at the coffee maker, not turning around: "You're building a museum."
"I'm building a receipt."
219. 4/18/51 M.K. FWD TERRALINE: AUG-SEP 2050 #17 HEADS = 'PRECIPITATION-DRIVEN STAGE MODEL' CALIBRATED TO S14 (1 PT). RESTATE REQUEST: 'FINAL UNDER VINTAGE TERMS.' METHOD NOTE: NO DISTRICT STAGE DATA EXCEPT STORM VISITS. QUOTE INTO SITE HISTORY. 0.7 H, GRAY.
Tuesday, the twenty-fifth of April, twenty to six in the morning at Ken's ditch, then nine at night at her kitchen table.
An inch and four tenths, the first spring storm with real weight in it. For a V-notch, the V-notch is roaring. The staff reads 0.45. On the five-liter pail she gets 2.1, 2.0, 2.2 seconds, too fast to trust a thumb, so she switches to the twenty-liter bucket. 8.4, 8.6, 8.3. About 2.4 liters a second against a curve of 2.38. She gets her cuffs wet to the elbow anyway.
EC 580. pH 7.22. Duplicate, split, and the third bottle, the before bottle. Dale at ten: alkalinity 3.1. Lower than storm nineteen's 3.4, on almost twice the flow. Even the before thins out in the rain. She writes that on the cap too, small, because Dale will ask.
That night the kid's homework is on one side of the table and graph paper is on the other. She's waited all winter for enough storms with a flow beside them to do this. She counts them off on card twelve and card thirteen. S14, S15, S18, S19, S20. S16 goes on as an open circle, because its head was two and a quarter inches on the old slot, in the band she flagged herself. S17 doesn't go on at all. Ice on the crest, and ice is not a head. Five that count.
Flow goes across the bottom on a log scale, ruled by hand because the plotting license is still in the queue. Conductivity goes up the side. She puts a dot for each storm and labels it with the month.
They fall on a slope. More water, weaker water. Not perfectly, but enough that she can lay a ruler along them. The August storm and the April storm sit almost on the same line, eight months apart.
"That's the rain," she tells the kid, who asked. "Most of what I called falling was just bigger storms."
"So you were wrong."
"I was measuring the right thing and calling it the wrong thing."
"That's wrong."
"Yes."
Then she does the part she'd skip if anyone else were checking. Next to each dot she writes R or F, rising or falling, because water at a given flow on the way up isn't the same water at that flow on the way down. Every dot gets an F. She has never once beaten a storm to Ken's ditch. She drives out after the rain quits, like everybody else with a day job. So it isn't a C-Q plot. It's a C-Q plot of the falling half, and the title has to say so.
What's left after the slope is small. The dots from before Christmas sit a little above the line, the spring dots a little below. That gap might be the trench healing. It might be the season. It's maybe ten percent. It's the only part of the pulse that's real, and it's about a third of what she's been posting since March.
She writes the plot up as one page for the public record, C VS Q, 5 RATED STORMS, FALLING LIMB ONLY, NOT A TREND. Each old PULSE FALLING post gets a correction row: CONC ONLY, NOT FLOW-MATCHED, SEE C-Q 4/25/51.
The kid has gone to bed by the time she's done. Their worksheet is still on the table. Fractions. Somebody has drawn a fox in the margin.
220. 4/25/51 S20 1.4 IN. H 0.45, 20 L BUCKET 2.4 L/S VS CURVE 2.38. EC 580 pH 7.22 ALK 3.1 (DALE). FIRST C-Q PLOT, 5 RATED STORMS (S16 OPEN, FLAGGED HEAD; S17 ICE, OUT). ALL FALLING LIMB. MOSTLY DILUTION. RESIDUAL ~10% (TRENCH OR SEASON). CORRECTION ROWS ON PULSE FALLING POSTS. 3.0 H, BLACK.
Tuesday, the second of May, quarter past four in the afternoon, Marcy Kuhl's creek forty.
It's the second point at the Ruebush outlet, the unmapped 1974 tile they found in March with a probe rod and a can of orange paint. They cleared the silt off the pipe and it has stayed clear. Sixteen ounces in a minute thirty. EC 450, pH 7.10. Close to March, on a wetter week. She fills an alkalinity bottle here too now, because a paired field with no alkalinity is half a pair. Dale will say three-point-something, and that'll be the field that never got anything.
Marcy is waiting by the truck with her phone held out at arm's length, like it smells.
ASH-C Agronomic Outreach: Satellite soil-reflectance analysis indicates field 17N-4W-22 (approx. 40 ac) has not received alkaline amendment in the analysis period and may be underperforming. ASH-C offers a complimentary trial application (up to 2 t/ac, delivered and spread) on qualifying fields. Reply YES to schedule.
"Free sample," Marcy says. "From a guy at the co-op. Except it's not a guy."
"It found you from space."
"It found the one field I told you I'd never touch." She's angry the slow way farmers get angry, which is mostly not talking. "How does it know?"
Janelle thinks about it. After a few seasons limed ground reads a little different from orbit: color, moisture, the way the crop greens up. An unlimed field next to limed ones would stick out to anything looking for gaps. To an agent selling something, a field nobody has spent money on looks like a field nobody has finished yet.
"It doesn't know it's a control," she says. "It knows it's empty."
Marcy gets the soil survey page out of the cab, the one with the paint-marker sentence and two signatures, photographs it, and replies to the agent with the photo and one word, NO. Then, because she's Marcy: "Put that in your file too."
It goes in the site history for Kuhl creek 40: 2051-05-02. THIRD-PARTY OFFER (ASH-C TRIAL, SATELLITE-IDENTIFIED). OWNER DECLINED. FIELD REMAINS UNAMENDED PER SIGNED 3/21/51.
On the drive back she passes three fields with fresh gray on them, spread thin, already half-raked in by last week's rain. She doesn't stop. She has no soil test for any of them, and an opinion from a moving truck isn't a sample.
221. 5/2/51 KUHL CREEK 40 2ND PT: CUP 1:30, EC 450 pH 7.10, ALK BOTTLE TO DALE. ASH-C AGENT OFFERED FREE TRIAL (SAT-IDENTIFIED 'NOT AMENDED'). M.K. DECLINED W/ SIGNED PAGE. AN UNLIMED FIELD LOOKS LIKE A SALE. 1.6 H, BLACK (OWNER REQUEST).
Tuesday, the ninth of May, half past seven in the morning at her desk, then Ken's shop.
The YardDesk feed for TAZ-17 has a new line under Ken's dot, in gray italic, the type a dashboard uses for things it thinks are boring:
Reference-site transition scheduled: 2051-10-01 (estimated).
She never gave it a date. In March she wrote that the district would not estimate one. The agent waited six weeks and estimated one anyway, probably off Ken's soil test and the word fall and the week most farmers in Tazewell get a spreader out. October first isn't a crazy guess. That's what bothers her. A good guess gets copied.
She posts a correction row to the site history for #17, date in the first field: 2051-05-09. THIRD-PARTY ESTIMATE 2051-10-01 APPEARS ON PARTNER FEED. NOT FROM DISTRICT OR OWNER. APPLICATION DATE UNKNOWN UNTIL REPORTED. Then she drives out to tell Ken, because it's his name on the line.
Ken's in the shop with the planter, changing seed plates for beans. He listens with a box of plates in one hand.
"October first."
"Their guess."
"It's a decent guess."
"That's the problem."
He sets the plates down. "How many more of those bottles do you need. The before ones."
She did the count in the truck. Storms with a good head and an alkalinity beside them: S15 in October, by accident, because Dale ran it with the sulfate. S19. S20. S16 has a number but a flagged head, so it doesn't count. To see a change after lime on top of the dilution slope, she wants the before to have its own line on the graph. That means a dozen flowing storms across spring and summer, at a range of flows, and ideally one she gets to before the peak instead of after.
"Twelve storms. I have three."
"So you need it to rain."
"Nine more times before you lime."
He thinks. Ken does his thinking with his hands. He picks the plates back up. "I'll lime when the soil test says. Not when they say. Not when you say." Then, after one plate clicks in: "Tell me when you get to twelve. I'd like to know."
It isn't a promise, and she doesn't write it down as one. She writes down what he said.
222. 5/9/51 PARTNER FEED SHOWS 'TRANSITION 2051-10-01 (EST)' - NOT DISTRICT, NOT OWNER. CORRECTION ROW. K.S.: 'LIME WHEN SOIL TEST SAYS.' ASKED HOW MANY BEFORE STORMS: NEED ~12, HAVE 3. 'TELL ME WHEN YOU GET TO TWELVE.' 1.2 H, BLACK (K.S.).
In the truck she turns card thirteen over and starts a tally on the back. Three short pencil strokes. Room for nine more, if it rains, and if she's awake when it does.