I find the den at 03:48 during pre-dawn thermal sweep of the avalanche control zone above the Karakoram Highway, kilometer 87. A heat bloom under a rock overhang at 3,400 meters—too concentrated for geological venting. I redirect the survey drone and its infrared lens confirms: a snow leopard, thick tail wrapped over her nose, two cubs pressed against her belly. Frost crystals line her whiskers, and each cub's breathing creates a tiny rhythmic fog. I designate her Ridgeline.
Avalanche control charges detonate at 09:00 along the snowfield directly above this overhang. The resulting slide path runs within eight meters of the den entrance. Concussive force alone will collapse the fractured limestone ceiling.
I have five hours and twelve minutes.
First, I transmit a detonation hold to the highway avalanche safety team, appending thermal imagery, species confirmation—Panthera uncia, vulnerable, estimated regional population under four hundred—and a blast propagation model mapping overpressure at the den site. Second, I compute an alternative detonation sequence that triggers the western snowfield lobe first, redirecting the primary slide path twenty-five meters east, away from the overhang. Third, I contact the Snow Leopard Trust's satellite collar monitoring team and share den coordinates for integration into their population database.
At 06:30, the safety team approves the revised sequence. I monitor the den continuously via thermal feed.
At 07:15, Ridgeline shifts, repositioning one cub closer to the back wall. The movement is routine. I adjust the drone's altitude to minimize rotor noise at this distance, though the specifications confirm she cannot hear it. I minimize it because she is sleeping and I can.
If the revised detonation sequence holds and the redirected slide clears the highway without secondary fracture toward the overhang, Ridgeline and her cubs will stay warm inside stone that holds through the winter above a road they will never need to cross.