A blindfolded sapper searches for a mine by tossing pebbles at random. He can't see where the charge is buried. But if you toss many pebbles and listen for the one that lands with a duller thud, you can locate the danger. That's how researchers hunt for the precursor of El Niño, the warm current that alters weather across the entire planet.
Previously, this required building an inverse model. It's like a translator who back-translates a speech into the original language to check every word. That's difficult and time-consuming. The new method is simpler: the computer repeatedly tweaks the initial conditions and watches how fast the patch grows. Each trial is like a tossed pebble. From hundreds of trials, you can infer where the 'mine' is without ever seeing it. The calculation takes 3 seconds instead of 15, five times faster, and the result is nearly identical.
The sampling method is based on Stokes' theorem, which relates the integral over a volume to the integral over its boundary. Here, it allows replacing the gradient with an average of function values on a sphere: an elegant mathematical trick.
In Isaac Asimov's story 'The Last Question', a supercomputer searches for a way to reverse entropy by trying possibilities. The sampling method also tries random directions, but instead of an exhaustive search, it uses statistics to quickly find the optimal perturbation.