424 of the 449 Disaster-Related Deaths in Noto Were Aged 70 or Over — Individual Evacuation Plans Cover 15.1% of the Registers, and 40°C Has Now Settled Over Kumamoto's Shelters

2016 Kumamoto quake, 273 deaths

50 confirmed by post-mortem（18.3%）223 died afterwards（81.7%）

50 confirmed by post-mortem, 218 certified as caused by the disaster under the condolence grant act, and 5 in the June rains found to be linked to the quake

Noto quake, 449 disaster-related deaths

25 under 70（5.6%）424 aged 70 and over（94.4%）

180 in their 80s, 169 in their 90s, 56 in their 70s, 19 aged 100 or over. 95.8% had pre-existing conditions, and 46.5% of certifications cite reduced care capacity after welfare facilities were damaged

6,696,718 residents on the assistance registers

1,014,017 with a plan（15.1%）5,682,701 without one（84.9%）

All 1,741 municipalities have a register. In 911 of them, or 52.3%, plans cover 20% or less of those listed

Registers are mandatory for municipalities under Article 49-10 of the Disaster Countermeasures Basic Act, while individual evacuation plans are a duty of effort under Article 49-14. Nationwide plan coverage stood at 15.1% as of 1 April 2026.

28 July 2026, 16:27. A magnitude 7.1 earthquake struck the Kumamoto region of Kumamoto Prefecture. The Japan Meteorological Agency has named it the 2026 Kumamoto Earthquake.

A maximum seismic intensity of 7 was recorded in Uki City and Hikawa Town[1].

By the thirty-fifth report, compiled at 16:30 on 3 August, 38 people had died, 123 were injured, and 7,194 homes were damaged[1]. Evacuation orders covered 115,075 households and 243,097 people across three municipalities[1].

Two unobtrusive lines sit inside the breakdown of those 38 deaths. One person who "may have died from worsening injury or illness brought on by physical strain." One person "whose connection to the disaster is under investigation." A category separate from those killed by collapse had opened within a week.

Then the heat arrived.

On 31 July the Japanese Red Cross Society warned that 3 August would bring extreme heat. At that point 374 shelters were open and 9,134 people were staying in them[2]. Older people and children, the release noted, are the most exposed.

The day of that warning, Kumamoto City passed 40°C. The same day, three learned societies in disaster medicine held an emergency press conference. According to NHK, they reported a run of heatstroke cases among older people in the affected area, and asked that cooling begin while symptoms are still mild.

The two lines quoted above mark the category that grows in heat like this. A death from worsening injury or from illness brought on by physical strain becomes a certified disaster-related death once a municipal panel reviews it.

Kumamoto Prefecture began taking applications for secondary evacuation to hotels and inns on 3 August, a scheme for moving people from shelters into places with better living conditions. As of 18:00 on 3 August, more than 8,200 people remained in shelters[3].

The record from ten years ago shows what tends to come next.

Report 121 from the Fire and Disaster Management Agency, dated 12 April 2019, is the last report published on the 2016 Kumamoto earthquake. 273 people died[4].

The breakdown separates one kind of death from another. 50 deaths were confirmed by police post-mortem[5]. Municipalities certified deaths as caused by the disaster under the condolence grant act in 215 cases in Kumamoto and three in Oita[6]. A further five deaths in the rains of 19 to 25 June were found to be linked to the earthquake[6].

Fifty of the 273 were confirmed by post-mortem as killed by the shaking or by collapse. The remaining 223 died in the days that followed.

Kumamoto is not the exception.

In December 2025, Ishikawa Prefecture published an overview of disaster-related deaths after the 2024 Noto Peninsula earthquake. 449 had been certified through the 37th review panel[7]. Reviews have continued since, so this is an interim figure. The document sets out their ages. 180 in their eighties, 169 in their nineties, 56 in their seventies, and 19 aged 100 or over[8]. Taken together, 424 were aged 70 or over. That is 94.4% of the total.

430 of them, or 95.8%, had pre-existing conditions[8]. These were people already connected to medical or nursing care.

The grounds for certification are classified too. Multiple grounds can apply. The most common was physical and mental strain from the shock of the quake and fear of aftershocks, in 395 cases[8], followed by loss of electricity and water, in 233 cases[8]. Third came reduced care capacity after welfare facilities were damaged, in 209 cases, or 46.5%[8]. The strain of shelter life was cited in 176 cases[8].

This is not only about conditions inside shelters. When the care that was holding someone up stops, they die.

The timing splits as well. 44 died within a week and 103 within a month[8], together 147 people, a third of the total. Yet 107 deaths, or 23.8%, fell between three and six months[8], the largest single band.

Getting past the acute phase does not make people safe.

The preparation side of the system comes in two tiers.

Article 49-10 of the Disaster Countermeasures Basic Act says municipal heads "must maintain" a register of residents requiring evacuation assistance. That is a duty. According to the Cabinet Office, it entered the law in the 2013 amendment.

Article 49-14 says individual evacuation plans "must be endeavoured to be made." That is a duty of effort. It entered in the 2021 amendment.

Within the same statute, only the verb ending differs.

The difference shows in the numbers. In a survey published on 29 June 2026 by the Cabinet Office and the Fire and Disaster Management Agency, all 1,741 municipalities had a register, a rate of 100%[9]. Against that, 6,696,718 people appear on those registers while 1,014,017 have an individual plan, a coverage rate of 15.1%[9].

In 911 municipalities, or 52.3% of the total, plans cover 20% or less of those listed[9]. In more than half of Japan's municipalities, at least four in five people on the register have no plan.

Kumamoto Prefecture stands at 33.3%, with 30,381 plans against 91,103 people listed[9]. That is more than twice the national average. Two in three still have no plan.

What happens without one?

A register is a list of who needs help. A plan sets out who will help them, where they will go, and by which route. With only a register, the search for a helper begins after the disaster. Who calls on the people riding out 40°C outside the shelters, and how often? Nothing in a register says.

Secondary evacuation carries the same shape. Moving to a hotel or an inn requires the person or their family to apply. It reaches those who can raise a hand. Those who cannot were exactly who the plans were for.

The societies asked for cooling to begin while symptoms are still mild. Noticing mild symptoms, though, requires that someone is already calling in regularly. The document that fixes who calls, and how often, is the individual evacuation plan.

There is a second gap, on the register side.

The same survey found that only 39.9% of those on the registers have had their details shared with local supporters in normal times[9]. Every municipality holds a register, yet fewer than four in ten of the people on them are known to anyone locally. Having made something is not the same as having delivered it.

This overlaps with the shortage traced in vacancies among community welfare commissioners passing 20,000（このサイトの記事）. The side that receives register information and actually knocks on doors is 20,000 people short. Both the writing of plans and the receiving of them run into the same wall.

The need to count provision and delivery separately runs through policy exclusion and non-take-up（このサイトの記事）. Individual evacuation plans are the case where non-delivery surfaces as a death toll. The procedure for mapping who falls out of support, and where, is set out in mapping social vulnerability in disasters（このサイトの記事）; the framing of recovery itself is questioned in fifteen years after 3.11, two years after Noto（このサイトの記事）.

Progress is real. The 50 municipalities that had made no plans at the previous survey fell to zero[9], and 198,106 plans were written in the past year[9].

At that pace, though, covering the remaining 5,682,701 people takes 29 years. Registers reached 100% within thirteen years of becoming mandatory partly because a register is finished once it is written. A plan means finding a supporter for each person and obtaining their consent. It will not move at the same speed.

In Kumamoto right now, 8,200 people are in shelters, in 40°C heat.

Further Reading

• 『人が死なない防災』 (Disaster Preparedness in Which Nobody Dies)（外部サイト、新しいタブで開きます） by Toshitaka Katada, Shueisha Shinsho, March 2012. A tsunami evacuation researcher rebuilds disaster preparedness around how people actually move rather than around warnings and information. Useful for thinking about the distance between finishing a plan on paper and a person being able to move on the day.

References

Damage from the 2026 Kumamoto Earthquake and the Response of Fire Services (Report 35) — FDMA Disaster Response Headquarters (2026). Fire and Disaster Management Agency

Earthquake Centred on the Kumamoto Region of Kumamoto Prefecture (Report 121) — FDMA Emergency Response Office (2019). Fire and Disaster Management Agency

Overview of Disaster-Related Deaths in the 2024 Noto Peninsula Earthquake (through the 37th review panel) — Ishikawa Prefecture Crisis Management Department (2025). Ishikawa Prefecture

Survey on Progress in Preparing Assistance Registers and Individual Evacuation Plans (as of 1 April 2026) — Cabinet Office (Disaster Management) and FDMA Disaster Management Division (2026). Fire and Disaster Management Agency

Cases Certified and Not Certified as Disaster-Related Deaths in the 2024 Noto Peninsula Earthquake — Cabinet Office (Disaster Management) (2026). Cabinet Office