A thermometer under your arm shows your temperature right now, but an experienced doctor can tell from subtle signs that a fever is coming in a couple of days. A neural network trained by Australian scientists does the same for the ocean: it looks at maps of water temperature over the past year and predicts where the water will become anomalously warm over the next two years.
Such forecasts used to be difficult: physical models that simulate water and air movement could only look three to four months ahead with confidence. The neural network learns differently: it doesn't need to understand all the laws of physics; it simply memorizes thousands of past examples and finds similar situations. Seventy years of daily measurements served as its textbook.
The program has already shown what it can do: it spotted in advance the warm blob off the California coast that in 2014–2015 killed sea lions and crabs. Admittedly, it stumbled on rare events off the western coast of Australia, because the textbook had too few such examples. But that's like a rare disease: the more case histories you collect, the more accurate the diagnosis becomes.
The model was able to predict the Blob marine heatwave 24 months before its peak, even though this phenomenon was considered elusive due to its irregular seasonal timing. However, with Ningaloo Niño off the western coast of Australia, the neural network did worse: there were too few similar cases in the training data.
The idea of predicting climate anomalies with machine learning echoes Kim Stanley Robinson's novel 'The Ministry for the Future', where algorithms help humanity respond to extreme heat and climate disasters. Here, the neural network acts as an early warning system that sees the onset of El Niño two years in advance.