Japan · bear sightings and attacks · updated 7 September 2026
After a record year, a better autumn for Japan’s bears?
After a record 2025 and a busy spring, summer sightings eased in the prefectures reporting sooner. August reports were lower across seven comparable prefectures, and early September reports are now coming in. A better beechnut crop offers hope for a smaller autumn surge, but it is too early to call this year less damaging.
Regional reports into September; full-month comparisons through August. National sightings through June; national harm through July. Fiscal years run April–March. Reports can include sightings and traces; they do not count individual bears or attacks. Regional figures and dates · Methods and sources
Latest complete months · July–August
What happened in fiscal 2025?
50,801 sightings, 2.1 times the previous high. 238 people harmed, including 13 killed. October and November accounted for 52% of sightings; Tohoku accounted for three quarters of those autumn reports.
Thirteen years of sightings.
The Ministry's series begins in fiscal 2013 with 9,133 sightings. Fiscal 2023 set a record of 24,348. Fiscal 2025 more than doubled it: 50,801.
238 people harmed. 13 of them killed.
Both were records in the ministry’s series, which starts in fiscal 2008. The previous record, in fiscal 2023, was 219 people harmed, including six killed. “People harmed” counts everyone injured or killed; deaths are included in that total.
Two prefectures, one autumn.
Akita reported 13,592 sightings in fiscal 2025 and Iwate 9,739, 46% of the national total. In October alone the two reported 8,898 of 15,998.
The same year, month by month.
Each cell is one prefecture in one month. October and November 2025 were 26,336 sightings, 52% of the year. The same two months of fiscal 2024 were 3,518.
Swipe chart sideways to explore →
Report by report.
Four prefectures publish reports with a location: Toyama, Niigata, Gunma and Saitama. Play their fiscal 2025 reports month by month to see the autumn rise. The sample had 64 reports in April and 1,795 in October; it does not cover all Japan. Yellow marks indicate reports of harm to people.
Why do some autumns explode?
Poor beechnut crops are associated with larger autumn surges. The crop is a useful clue to food availability, but it cannot tell us how many sightings or attacks there will be.
Less food in the forest, more reports in autumn.
Beechnuts are autumn food for bears. The chart pairs the Tohoku beech crop with Japan’s October sightings. In this 2013–2025 comparison, every October above 2,235 reports followed a poor crop (index below 2).
The crop is only part of the explanation: 2023 and 2025 both had near-failed crops, yet October sightings were 5,983 and 15,998. These regional crop observations do not prove what caused the national increase.
Swipe chart sideways to explore →
Good crop years were usually followed by poor ones.
Across five Tohoku prefectures, average or abundant crops were followed by poor crops in 22 of 23 cases from 2012 to 2025. A poor crop gave much less guidance about the next year.
Each square is a prefecture’s crop in one year. Read along a row to see the alternation. These are regional observations, not 23 independent national years.
Swipe chart sideways to explore →
Will autumn 2026 repeat autumn 2025?
There are reasons to expect a smaller surge: summer reports eased in the available regional sample, and the beech crop is forecast to improve. But national summer data are incomplete, and deaths through July are already above the same period last year.
The busiest spring in 14 years.
April to June 2026: 12,628 reports in the ministry’s national sightings series, 67% above the 7,555 at the same point in fiscal 2025. The previous high was 7,601 in fiscal 2024.
The dotted green line below brings the regional picture through August. It now shows April–August 2026 against the full April–March season of fiscal 2025, including last autumn’s peak. Both lines cover the same seven prefectures. The national line ends in June because the ministry has not published later totals.
Harm is close to last year. Deaths are higher.
April–July 2026: 53 people harmed, including six killed. In the same months of 2025: 55 harmed, including four killed. The total is within the 2016–2025 range of 31 to 56, but six deaths is the highest in this eleven-year comparison. This is too early to call the year less damaging.
Summer reports eased in the regional sample.
July reports were 14% lower across twelve comparable prefectures: 2,319 versus 2,684. August reports were 60% lower across seven: 596 versus 1,473. The two months cover different groups, so compare 2026 with 2025 within each month.
Those groups account for 52% of national July 2025 reports and 36% of August reports. Iwate is excluded because its counting method changed. Akita’s August file is dated 31 August, so late reports may still be missing. These are regional samples, not national totals.
Spring does not set the autumn.
A busy spring does not fix the size of the autumn surge. Across the thirteen complete years shown, October–November sightings ranged from about a quarter to three and a half times April–June sightings.
The vertical line marks 2026’s record spring. It has no autumn point yet. Historical relationships are context, not a numerical forecast for this year.
The 2026 crop is forecast good.
The July 2026 survey forecasts an abundant beech crop in Iwate, Miyagi, Akita and Yamagata, and an average crop in Aomori. The five-prefecture mean is 3.90 out of 5, versus 0.44 at the same point in 2025.
More beechnuts would mean more autumn food in the forest. That supports a more hopeful outlook, but the crop is still a forecast: eight of 31 average-or-better July forecasts since 2012 ended with a poor autumn crop. None ended in failure. We cannot turn this into a predicted attack count.
What has changed over the longer term?
Prefectural bear estimates have risen, while gun-hunting licences have declined. These are background changes; they do not explain the size of any one autumn’s surge.
Hunting licences, 1975 to 2021.
All licences: 517,800 to 213,400. Gun licences: 493,700 to 84,400. Trap licences rose from 51,600 in 2007 to 119,500. 119,100 of 213,400 holders were 60 or older in 2021. The Ministry does not count hunters by target species.
Bear estimates by prefecture.
Each line runs from a prefecture's earliest published estimate to its latest, with the survey year at each end. Iwate rose from 1,100 in 2006 to 3,700 in 2020. Akita from 1,052 in 2010 to 2,900 in 2024. Nagano from 2,770 in 2007 to 7,270 in 2020. Methods changed between surveys, so part of the rise is measurement. There is no national series.