That night she opened her grandfather’s 1987 notebook and read the entry for May fourteenth of that year. Bottom field flooded again. Third time in eleven years. Old channel running. Cattle wouldn’t drink at the south trough. She wrote her own entry beneath it.
The drilling rig came on a Thursday. It set up on the high ground first. Three test pits along the upper slope where the timber thinned out. The samples came up dry and firm. The geotech engineer, a man named Patel, walked his rig down the slope in stages, pulling cores every two hundred feet on the grid the bank’s architects had given him.
When they got to the bottom, the rig pulled samples that looked normal at every drill point. Brown silt, some clay, a few river stones. Patel marked them, bagged them, photographed them. He drilled to twelve feet at each point, which was the depth specified in the contract, and at each point he hit firm ground at eleven.
What the grid didn’t account for was the curve. Maren knew, watching him, that the buried channel curved through the bottom land like a lazy S. And that Patel’s grid points were falling on either side of it without ever crossing it. Her grandfather had walked her along the curve the summer she was nine, pointing out where the grass changed color in dry years.
Patel was missing it laterally by twenty feet here, thirty feet there. To him, the bottom looked solid. The grid said so. She didn’t say anything. She had said her piece.
They broke ground in June. The graders came first, peeling the topsoil off the building footprint and pushing it into long berms along the eastern edge. Then the excavators, digging the foundation pad. They had to go down eight feet for the slab and the loading dock pits. And at six feet on the riverside, they hit water.
Not a lot. A seep. A wet darkening in the wall of the cut. Patel came back out. He looked at it. He took readings. He told the foreman it was within tolerance. There was groundwater in any river bottom, and they would lay a drainage course and a vapor barrier and proceed. The foreman, who had built things in a lot of places, agreed. The bank’s project office in Denver signed off on the variance the same afternoon.
They poured the footings on a Wednesday in late June. The slab went down the following Tuesday. Maren wrote it all down. She didn’t know what else to do.
The first hairline crack appeared in late August, eight weeks after the pour. It ran across the northeast corner of the slab, the kind of thing a finishing crew would have noticed and called within tolerance. By mid-October, it had widened to a quarter inch and grown two siblings, all running the same direction, all on the riverside of the building.
The steel went up that month and the loading on the slab grew. The engineers came back. They argued about it. One of them blamed the concrete mix. Another blamed the curing. A third, a younger woman with a master’s from the Colorado School of Mines, walked the perimeter for two hours and then asked the foreman if anyone had drilled deeper than twelve feet anywhere on the bottom or off the grid. Nobody had. She went back to her car and made some phone calls.
By November, the cladding was halfway on and Maren noticed the cattle had stopped drinking from the south trough. This was the trough fed by the spring at the base of the slope on the Holloway side of the fence. The spring sat lower than the construction site. The bottom land sloped gently south-southwest and the spring emerged at the toe of that slope on her family’s land down gradient of the bank’s footprint.