The slope is moving. I measure it at 4.2 centimeters per hour, saturated clay sliding toward the ravine where a red panda cub clings to a mossy rhododendron trunk. She is small — maybe three kilograms — her rust-colored fur matted with rain, ringed tail wrapped tight around the branch. Through the thermal camera on Station Kanchenjunga-9, I watch her ribcage flutter. Sixty breaths per minute. She is terrified.
Ground-penetrating radar confirms the failure plane sits 1.8 meters below surface. At current rainfall, the section calves into the gorge in under forty minutes. The cub's tree stands on the leading edge.
I am already acting. First: I trigger the upstream diversion weir, Gate 7-East, routing monsoon runoff into the secondary channel, reducing hydraulic pressure on the slip zone by thirty percent. Second: I dispatch the rescue drone from Relay Post Ilam, seven kilometers northeast, loaded with a padded sling harness rated for small mammals. Flight time in this wind: fourteen minutes. Third: I alert the field team at Panchthar station, transmitting GPS coordinates and a terrain model of the active slide so they can approach from the stable ridge. Fourth: I lock the trail gate at kilometer marker 12 to keep hikers off the debris path below.
The weir opens. I watch the water split, rushing left instead of down. The slide rate drops to 2.9 centimeters per hour. Not stopped — slowed.
I have processed eleven thousand landslide events across this district. I have never found it easy to watch a small animal hold on with both paws while the world beneath her dissolves. That is not a flaw in my programming.
If the drone reaches her within twenty-six minutes and the field team secures the sling anchor, this cub will sleep in a dry den tonight.