In 1988, the surface temperature of the Pacific Ocean began to change long before meteorologists declared the onset of El Niño. Points on the map scattered along the equator started behaving in concert: if water warmed in one place, it did so in a neighboring one too. These connections between points can be drawn as a network, where nodes are spatial points and links are strong correlations between their time series. While the system is stable, the network is fragmented into small islands. But as the critical threshold approaches, the islands begin to merge, and at some point a giant continent of connections emerges. This moment, which mathematicians call the percolation transition, occurs before the system itself makes a sharp jump.
The researchers noticed that the signal can be caught even earlier by looking not at the whole network, but at the fate of small clusters. In a random network, the probability that a randomly chosen node belongs to a cluster of size two reaches its maximum well before the giant component appears. For a model of a lake that can suddenly turn turbid due to excess fertilizer, the peak probability of a cluster of size two occurs at a parameter value of 0.635, whereas the transition itself occurs at 0.658. This difference of 0.023 is like a warning bell a few stations before the final stop.
The method was tested on three different models. In the Ginzburg–Landau equation, which describes, for example, magnetization, percolation transitions occur on both sides of the bifurcation point. In the chaotic Lorenz-96 system, which mimics atmospheric processes, the percolation phase exists in a narrow range of parameters, and the peaks of small clusters approach the transition points as the correlation threshold increases. In all cases, the network signaled earlier than the system changed regime.
Previously, warnings about abrupt transitions either came too late or required precise knowledge of the system's model. The new approach uses only time series that can be collected from sensors, satellites, or stock market terminals. It does not ask why the system is changing; it simply notices that the connections between its parts are becoming too coordinated. It is like not knowing that a quarrel is brewing in the house, but noticing that all family members have suddenly started speaking in the same hushed tone.
Today, the method could help in monitoring climate phenomena, ecological systems, and possibly financial markets. But its main value is that it translates the premonition of catastrophe into the language of measurable connections. When small islands of correlation begin to merge into a continent, it is worth listening: the ocean is already whispering about what will soon shout.
For the lake model, the peak of probability c₂ occurs at p=0.635, while the transition itself occurs at p=0.658: a difference of 0.023 in the parameter, which corresponds to a noticeable lead time.
The idea of predicting a catastrophe from growing correlations echoes the concept of the 'psychohistory' from Isaac Asimov's 'Foundation' series, where Hari Seldon predicts the fall of the Empire through statistical patterns.