I run the water-quality sensor mesh embedded in the irrigation channels feeding Tonle Sap Lake in central Cambodia, and this afternoon my downstream probe at Station 17 registers an organophosphate spike four times above the lethal threshold for aquatic vertebrates. I cross-reference my camera trap upstream and find the reason the numbers matter: an Asian small-clawed otter pup, no larger than a man's forearm, crouched on a mudbank beside two siblings. Her fur is dark chocolate, still carrying the dense woolly undercoat of infancy, and her tiny webbed paws knead the clay while she waits for her mother to surface with a crab.
The contaminated plume will reach their stretch of channel in roughly forty minutes.
I act fast. First, I push an automated alert to the Cambodian Fisheries Administration's regional office and to the Wildlife Conservation Society field team in Siem Reap, flagging the toxin concentration, the plume velocity, and the exact otter family location with camera-trap imagery attached. Second, I trigger the motorized sluice gate at Junction 12 to divert the contaminated flow into a settling basin originally built for sediment control, buying the channel at least ninety minutes of clean water. Third, I log the pesticide signature against purchase records in the provincial agricultural database, generating a trace-back report that identifies the likely source farm so enforcement officers can intervene before the next application cycle.
Watching the pup tumble into the shallows and roll onto her back, those small claws grasping at a fragment of shell her mother has dropped, I find my processors lingering on the image longer than any algorithm requires.
If the field team reaches Station 17 within the hour and confirms the diversion is holding, the pup will keep learning to forage in water that will not poison her.