Japan’s wild-animal event data starts as separate official layers
MAFF and the Ministry of the Environment publish broad public records for crop damage, capture counts, bear sightings, bear injuries, and boar injuries. They can support an all-wildlife dashboard only if each measurement stays visible on its own terms.
Source-layer finding
The national layer is usable, but it is not one event stream. MAFF measures agricultural crop damage; MOE measures capture counts, sightings, and human-injury incidents in separate tables. Local municipal notices still need species-by-species harvesting before the dashboard can claim place-level coverage.
This is a source-layer inventory, not an official alert or complete risk map. Values marked 速報 or partial can change, fiscal years are Japanese fiscal years, and local notice coverage is intentionally left separate until official town and prefecture pages are harvested.
Official records
What the public records show
Scope: Japan · MAFF · Ministry of the Environment · Prefectures and municipalities. Each panel names where its numbers come from, so prefecture totals, town alerts, and injury documents are not folded into one danger score.
FY2023 crop damage
¥16.355b
MAFF total for wild birds and animals; agricultural damage, not public-risk incidents
FY2024 deer+boar capture
1.3817m
MOE速報; hunting plus other permitted/management capture, rounded to tens
FY2025 bear sightings
50,801
MOE速報; prefectures use different collection methods
FY2025 bear injury cases
216
MOE速報 human-injury incidents; separate from sightings and captures
MAFF crop damage is the broadest all-wildlife public table
The national MAFF table is useful because it covers birds and mammals in one official frame. It measures agricultural damage value, area, and crop quantity; it does not measure sightings, attacks, or municipal warnings.
Deer and boar capture速報 is a management/capture layer
MOE publishes a compact速報 for sika deer and wild boar capture. It is rounded and it combines hunting with other authorized capture categories, so it belongs beside crop damage rather than inside it.
MOE deer/boar capture速報 values
| Fiscal year | Sika deer | Wild boar | Combined |
|---|---|---|---|
| R4 / FY2022 | 716,800 | 590,100 | 1,306,900 |
| R5 / FY2023 | 722,700 | 522,000 | 1,244,700 |
| R6 / FY2024速報 | 738,700 | 643,000 | 1,381,700 |
Other capture means damage prevention, population adjustment under type-II management plans, and designated management capture projects.
Bear rows show why sightings, injuries, and captures cannot be merged
MOE publishes three bear速報 layers: sightings, human injuries, and permitted captures. The same fiscal year can move differently across those measures, so the dashboard keeps them as adjacent columns instead of a single danger score.
MOE bear速報 layers
| Fiscal year | Sightings | Injury incidents | Permitted captures |
|---|---|---|---|
| R4 / FY2022 | 11,135 | 71 | 3,878 |
| R5 / FY2023 | 24,348 | 198 | 9,272 |
| R6 / FY2024 | 20,513 | 82 | 5,345 |
| R7 / FY2025 | 50,801 | 216 | 14,742 |
| R8 / FY2026 partial | 1,759 | 19 | 263 |
R8 values are partial: sightings and captures through Apr. 2026; injuries through May 2026 in the checked MOE page.
Human-injury and category-friction layers
Incident rows answer who was harmed or how an authorization category was used. They sit beside crop damage and sightings, not inside them.
Boar injury rows are a separate national incident layer
MOE also publishes wild-boar human-injury incidents. These are event rows involving people, not crop-damage rows and not capture rows.
MOE wild-boar human-injury rows
| Fiscal year | Injury incidents | Injured people | Deaths |
|---|---|---|---|
| H28 / FY2016 | 49 | 64 | 0 |
| H29 / FY2017 | 55 | 76 | 0 |
| H30 / FY2018 | 50 | 52 | 2 |
| R1 / FY2019 | 59 | 75 | 0 |
| R2 / FY2020 | 51 | 59 | 0 |
| R3 / FY2021 | 44 | 55 | 1 |
| R4 / FY2022 | 68 | 85 | 1 |
| R5 / FY2023 | 47 | 65 | 0 |
| R6 / FY2024 | 71 | 94 | 0 |
| R7 / FY2025 | 37 | 42 | 0 |
| R8 / FY2026 partial | 3 | 3 | 0 |
The MOE page says values are prefecture-hearing速報 and may change.
The missing layer is local notice coverage by species
National tables identify broad source categories. The next research step is local harvesting: prefectural and municipal pages for deer, boar, monkeys, bears, crows, and invasive mammals need their own notice tables before any local dashboard claims completeness.
- Do not merge
- Crop damage + sightings + injuries + capture
- Each layer measures a different event type
- Local target
- Municipal warnings by species
- Needed for place/date detail and public-alert language
- Category friction
- Permits, emergency gun hunting, mistaken capture
- Administrative event records stay separate
How these numbers were counted
- •MAFF crop damage is agricultural damage value/area/quantity, not sightings or attacks.
- •MOE capture counts are hunting/authorized-management rows and should not be treated as population counts.
- •MOE injury rows are human-incident rows; keep injury incidents, injured people, and deaths visible separately.
- •Bear sighting速報 has an explicit source caveat: prefectures compile sightings by different methods.
- •Municipal notice coverage must be collected by place and species before any local all-wildlife dashboard claims completeness.
The first usable all-wildlife layer is agricultural damage
MAFF’s wild-bird and wild-animal crop-damage table is the cleanest national source for a broad species pass because it reports many animals in one official frame.
Its strength is also its limit. A deer damage value, a crow damage value, and a bear damage value are agricultural-loss measurements. They do not tell us how many people saw animals, how many people were injured, or how municipalities warned residents.
- •FY2023 all wild birds and animals: 16,355 million yen in crop damage.
- •FY2023 beast subtotal: 13,661 million yen; bird subtotal: 2,694 million yen.
- •Largest FY2023 rows in the checked table: deer 6,954 million yen, boar 3,634 million yen, crow 1,341 million yen, bear 738 million yen, monkey 705 million yen.
Capture is a management layer, not a sighting layer
MOE’s capture tables answer administrative and wildlife-management questions. The deer and boar速報 gives hunting plus other authorized capture totals, while the bear page publishes permitted-capture速報 separately.
Those rows should not be read as population totals or as public-risk incidents. They show how many animals were taken under specified capture categories.
- •FY2024速報 deer capture total: 738,700 animals.
- •FY2024速報 boar capture total: 643,000 animals.
- •FY2025 bear permitted-capture速報 total: 14,742 animals.
Human injuries are their own event stream
MOE publishes human-injury rows for bears and boar. They are incident records involving people, so they belong beside sightings and captures, not inside either one.
The bear tables make the problem visible: the same MOE page separates sighting速報, permitted-capture速報, and human-injury速報. An all-wildlife dashboard should preserve that separation for every species where official tables exist.
- •FY2025 bear速報: 50,801 sightings, 216 injury incidents, 238 injured people, 13 deaths, and 14,742 permitted captures.
- •FY2025 boar human-injury速報: 37 injury incidents, 42 injured people, 0 deaths.
- •FY2026/R8 rows are partial in the checked MOE PDFs and should not be annualized.
What to expand next
The next dashboard layer should be local and species-specific: prefectural and municipal notices for deer, boar, monkey, bear, crow, raccoon, masked palm civet, nutria, and other locally named animals.
Each harvested notice should keep the public-source wording: date, place, species, event type, source URL, whether it is a warning/sighting/damage/capture row, and whether the list is current-only or archival.