Human impacts on the 2015 equatorial Asia drought: the role of air-sea coupling in the event attribution study

When scientists try to understand why a specific extreme weather event happened, they often ask a simple but profound question: how much did human-caused climate change make this event more likely? This field, known as event attribution, relies on running complex computer models of the Earth's climate. To get an answer, researchers typically run two sets of simulations. One set mimics the world as it is today, with all the greenhouse gases humans have emitted. The other set imagines a "counterfactual" world where those emissions never happened, representing a baseline of natural climate variability. By comparing how often a specific disaster, like a severe drought, occurs in these two different worlds, scientists can calculate the human contribution to the event's risk.

However, the Earth is a system where the atmosphere and the ocean are locked in a constant, dynamic conversation. The temperature of the sea surface influences the air above it, and the wind and rain in turn cool or warm the ocean. For many years, researchers have often used a simplified type of model for these attribution studies. These models treat the ocean as a static backdrop, feeding the atmosphere a fixed temperature map without allowing the atmosphere to change the ocean in return. While this approach saves time and computing power, it ignores the two-way feedback that drives much of our weather. The question remains: does this missing conversation matter when we are trying to pin down the causes of a specific disaster?

A team of researchers led by Akira Hasegawa at the University of Tokyo set out to test this very idea using the severe drought that struck equatorial Asia during the dry season of 2015. That year, a massive El Niño event, characterized by unusually warm ocean waters in the central Pacific, weakened the global wind patterns that usually bring rain to the region. The result was a devastating lack of precipitation. The researchers wanted to know if the standard, simplified models were giving an accurate picture of how much human warming contributed to this dry spell, or if the missing ocean feedback was distorting the results.

To find the answer, the team ran two distinct types of large-scale computer experiments. The first type used a fully coupled model, where the atmosphere and ocean could interact freely, exchanging heat and moisture just as they do in reality. The second type used an atmosphere-only model, where the ocean was forced to stay exactly as it was in the first experiment, unable to react to the changing winds and clouds above it. Both sets of simulations were run under two scenarios: one reflecting the actual historical conditions of 2015, and another representing a world without human-induced warming. By comparing the outcomes of these four different simulation groups, the researchers could isolate the specific impact of the air-sea interaction on their conclusions.

The results revealed a significant flaw in the simplified approach. In the atmosphere-only simulations, the model developed a persistent bias: it overestimated the upward movement of air over the Philippine Sea, a region to the north of the drought-stricken area. In the real world and in the fully coupled model, the ocean responds to this rising air by cooling the surface, which naturally limits how strong that upward motion can become. But because the simplified model could not let the ocean cool down, the air kept rising unchecked. This excessive upward motion in the north forced a compensating downward motion over equatorial Asia, making the region appear much drier in the model's baseline climate than it actually is.

This artificial dryness had a direct and misleading effect on the attribution results. Because the simplified model started with a climate that was already too dry, the 2015 drought did not look as extreme by comparison. In statistical terms, the model calculated that the chance of such a dry event occurring in a world without human warming was very low. When the probability of an event in a "no-warming" world is low, and the event still happens in the real world, the math suggests that human influence must be very high. Consequently, the atmosphere-only model estimated that human activity was responsible for about 44 percent of the risk of the 2015 drought.

In contrast, the fully coupled model, which allowed the ocean to breathe and react, painted a different picture. By correctly simulating the cooling feedback over the Philippine Sea, it produced a more accurate baseline climate that was not artificially dry. In this more realistic simulation, the 2015 drought was still a rare event, but it was not as rare as the simplified model suggested. The calculated human contribution to the risk dropped significantly, to about 20 percent. The study suggests that by ignoring the ocean's ability to respond to the atmosphere, the simplified models were overestimating the role of human warming in this specific disaster by more than double.

The researchers found that while both models correctly identified that the El Niño event was the primary driver of the drought, the simplified model failed to capture the natural variability of the region's climate. The atmosphere-only simulations could not reproduce the complex relationship between sea surface temperatures and rainfall that exists in the real world, particularly in areas surrounded by ocean like equatorial Asia. This led to a systematic error where the model underestimated the likelihood of severe droughts occurring naturally, which in turn inflated the calculated impact of climate change.

This work does not dismiss the value of simplified models, which remain useful for many types of climate research. However, it highlights a critical limitation when studying extreme events in regions where the ocean and atmosphere are tightly linked. The study suggests that for events like the 2015 equatorial Asia drought, relying solely on models that treat the ocean as a passive stage may lead to an overestimation of human influence. The findings indicate that to truly understand the fingerprints of climate change on specific disasters, scientists must account for the full, two-way conversation between the sea and the sky.