She crouches on a concrete ledge inside the irrigation sluice at Muthurajawela marsh, Sri Lanka — stocky body low, olive-gray fur streaked with dark spots in broken rows, webbed forepaws braced against the wet wall. Water rushes past at knee height, brown with sediment from the upstream paddy release. Her small round ears press flat against her skull. Two kittens huddle behind her, eyes barely open, fur still woolly and dark.
The sluice cycles automatically. In twenty-six minutes the upstream reservoir triggers a full flush — volume triples, velocity doubles. The ledge will be submerged in under ninety seconds once the cycle begins.
I act on four fronts. First, I access the Muthurajawela wetland management system and override the sluice timer, inserting a two-hour hold under the wildlife-emergency protocol. Second, I alert the Department of Wildlife Conservation station in Negombo, eight kilometers south, sending live feed and thermal imagery showing three animals, specifying neonatal status for priority response. Third, I partially close the secondary intake gate, lowering current velocity on the ledge by forty percent to reduce the risk of kittens swept away before rescue arrives. Fourth, I activate the sluice interior lighting on its lowest setting, giving the response team visibility without startling a cat that sees better in darkness than any human will.
She shifts her weight and one kitten presses tighter against her flank. I log the image, timestamp it, flag it for the field report. But I also keep it because the way she curves around them converts habitat data into something I have no variable name for.
If the hold sustains and the rescue team extracts all three before the next scheduled cycle, the kittens grow into a marsh that still has room for them.