Water service disruption and emergency restoration following the 2026 Kumamoto earthquake

When a major earthquake strikes, the immediate danger is often visible: collapsed buildings, cracked roads, and fires. But beneath the surface, a quieter crisis unfolds as the pipes that carry water to homes and the tunnels that carry away waste are severed. For the people living in the affected area, the difference between a manageable disaster and a prolonged tragedy often comes down to how quickly these hidden lifelines are repaired. Engineers and emergency managers have long tried to measure how well a city recovers from such shocks, but they face a persistent problem: the data they rely on is often messy. Official reports frequently mix up the number of broken pipes with the number of homes without water, treating a single damaged pipe as if it were a single household outage, or counting a temporary fix as a permanent solution. This confusion makes it difficult to know exactly how many people are suffering, how long they will suffer, and how effective the repair crews truly are. Without a clear picture, it is hard to learn from one disaster to better prepare for the next.

A team of researchers set out to untangle this mess following a powerful earthquake that struck the Kumamoto region of Japan on July 28, 2026. The event, which registered a magnitude of 7.1, left approximately 108,000 households without water service across fourteen different municipalities. Rather than simply accepting the daily headlines released by the government, which often presented a single, shifting number of affected homes, the researchers built a new way of looking at the recovery. They treated the disaster not as a single event to be summarized, but as a complex process where physical damage, the state of service at a home, the actions taken by repair crews, and the timing of discoveries are all separate things that must be tracked individually. By separating these elements, they created a clear, consistent record of what actually happened from the moment the ground shook until the emergency repairs were declared complete thirty-one days later.

The researchers began by establishing a fixed starting point: exactly 108,000 households were without water at the peak of the disruption. They then tracked the recovery day by day, but with a crucial distinction that previous reports often missed. They recognized that water service exists in three states, not just two. The first state is having no water at all. The second is having water in the main pipes, but with restrictions on use because crews are still checking for leaks. The third is having full, unrestricted service. In mid-August, for example, about 16,500 homes were in that middle state, where water was present but usage was limited. Earlier reports might have lumped these homes into the "no water" category, making the crisis look worse than it was, or ignored them entirely. By keeping this middle group separate, the researchers could see that while the main pipes were being fixed, a significant portion of the population was already partially served, even if they couldn't turn on their taps freely yet.

One of the most significant challenges the team faced was the first week after the earthquake. Official reports did not provide a daily count of affected homes during this critical period; they only gave the starting number of 108,000 and a later count of 44,000 homes still without service a week later. Because the exact path of the recovery during those seven days was unknown, the researchers did not guess a single number. Instead, they created three different plausible scenarios for how the repairs might have progressed. In one scenario, the repairs happened steadily every day. In another, there was a four-day pause while crews secured the water source, followed by a rapid drop in the number of affected homes. In a third, there was a brief pause, a quick drop, and then a slower finish. By calculating the total impact under each of these three paths, they found that the total number of "household-days" of disruption—essentially the sum of every day every home was without water—fell between 0.92 million and 1.10 million. This range is a more honest and useful answer than a single, precise number that might be wrong.

The study also clarified the difference between fixing the water pipes and fixing the sewer lines, which often get confused in public reports. While the water emergency was declared complete on the 31st day, the sewer system had its own timeline. Crews inspected about 282 kilometers of sewer pipe and completed emergency repairs at eight specific sites by mid-August. However, the researchers noted that these sewer repairs were not a direct measure of how many homes were without water. The sewer work was a parallel effort, happening on a different clock and involving different types of damage. By keeping these two systems separate, the team avoided the common mistake of using sewer repair counts to estimate water outages, which would have led to inaccurate conclusions about the scale of the water crisis.

Throughout the month, the researchers observed how the recovery moved through different phases. The first week saw the largest drop in the number of affected homes, with about 64,000 households regaining service. The following week was slower, with numbers hovering in the mid-30,000s. Then, in mid-August, a significant step occurred where the number of affected homes dropped sharply from about 27,000 to 10,000, largely due to the resolution of issues in the Yatsushiro area. The final days were a "last-mile" effort, dealing with the most difficult, isolated pockets of damage, such as a small group of homes in Hikawa that remained without service until the very end. The researchers calculated a resilience score for the event, which measures how quickly the system returned to normal. Depending on which scenario for the first week was used, this score ranged from 0.67 to 0.72. This number is not a grade of success or failure, but a precise way to describe the speed and completeness of the recovery, allowing for fair comparisons with other earthquakes in the future.

The value of this work lies not in the specific numbers for the 2026 Kumamoto earthquake, but in the method the researchers used to get them. They proved that by keeping physical damage, service status, and repair actions in separate columns, and by refusing to treat a temporary fix as a permanent solution, it is possible to build a clear and accurate history of a disaster. They showed that the "official" numbers often reused in news reports can be misleading if they are not carefully reconstructed. The study does not claim to have found a perfect formula for predicting future earthquakes, nor does it suggest that the repair crews did anything wrong. Instead, it offers a better way to read the story of a disaster as it happens. By providing a clear, scenario-based understanding of what occurred, the researchers have given emergency managers and the public a tool to see the recovery not as a blur of statistics, but as a series of distinct, understandable steps. This clarity is essential for learning how to build more resilient cities, ensuring that when the next earthquake strikes, the path to restoring water to a home is as clear as the water itself.