Himi City forest water yield report (Toyama, Japan)

Annual Water Yield

~ 1.19hundred million t/yr

119,115,675 m³/yr (±30% CI: 83 M – 155 M tonnes)
Equivalent to the annual domestic water demand of ~1.6 million households.

1. Summary

The AOI around Himi City (16 km × 14 km, 22,400 ha total) contains 12,398 ha of forest (55.3%) as identified from Sentinel-2 imagery. The portion of annual precipitation that this forest captures and supplies downstream as groundwater / baseflow — i.e. water yield — is estimated at approximately 119 million tonnes/year (961 mm/yr).

Annual precipitation

2,331mm/yr

NASA POWER 2020-2024 mean

Water yield

961mm/yr (41.2%)

Forest area only

Direct runoff

533mm/yr (22.9%)

Tertiary geology baseline

Evapotranspiration

837mm/yr (35.9%)

Transpiration + canopy interception

2. AOI and Sentinel-2 NDVI observation

Study area

Himi City, Toyama · Sea of Japan high-rainfall coast

Sentinel-2 NDVI (2026-03-17)

Dark green = high NDVI (active vegetation) · light green = mid · brown = bare / low
Cloud cover 3.7%, Sentinel-2A observation

3. Water-balance breakdown

Where the 2,331 mm of annual precipitation goes

23%

36%

41%

Direct runoff 533 mm Evapotranspiration 837 mm Water yield 961 mm

3.1 5-year NDVI time series (real observations)

5 years of Sentinel-2 observations (71 sample points) aggregated via the CDSE Statistical API. The seasonal cycle (spring–summer peak, winter decline) and inter-annual variability are clearly resolved. We applied an NDVI threshold of 0.5 to summer-peak rasters to identify forest area.

Cross-check against the municipal record: Our Sentinel-2 NDVI > 0.5 mask identifies 12,398 ha of forest; the City of Himi's published forest area is 13,486 ha — a difference of about 8% (on the conservative side), which is good agreement for satellite-based estimation. The underestimate likely comes from the 0.5 threshold conservatively excluding very young conifer stands, freshly harvested areas, and recently planted blocks.

Comparison with a hypothetical bare-ground baseline: If the same 12,398 ha were bare ground, water yield would drop to roughly 10% of precipitation (~233 mm/yr) — only about 29 million tonnes/year. By maintaining its current canopy, this forest supplies an extra ~90 million tonnes/year of stable downstream water.

4. Calculation method (IPCC-aligned, Forestry-Agency compliant)

Aligned with the "Simplified Evaluation Method for Forest Water Yield Ver.1.0" published by the Forestry Agency of Japan in March 2026. Every input is fetched automatically from public open data sources.

4.1 The headline equations

Water yield [mm/yr] = Annual precipitation − Direct runoff − Evapotranspiration P_eve = 0.6473 × P − 138.67 (event-aggregated precipitation) Q = a × P_eve + b (direct runoff by geology class) Q_t = (849 × DBH − 7350) × (0.0244 × T + 0.4361) (per-tree transpiration, Tsuruta et al. 2019) r = k1 × {1 − exp(−k2 × N)} (conifer interception rate, Komatsu et al. 2015) E = Q_t × N + r × P (ET = transpiration + interception) Water yield = P − Q − E

4.2 Inputs and sources

Input	Value	Source
Monthly T & P	T=16.0°C, P=2,160 mm/yr	NASA POWER monthly (2020-2024 mean)
Station elevation	2.9 m	GSI Japan 5 m DEM API
Forest-point elevation	171.5 m	GSI Japan 5 m DEM API
Geology class	Tertiary (Neogene Miocene)	AIST Seamless Geological Map v2 API
Forest type	Evergreen conifer (sugi / hinoki dominant)	Sentinel-2 NDVI seasonal pattern
Forest area	12,398 ha	Sentinel-2 L2A NDVI > 0.5 (Hansen 2013)
Stem density	783 stems/ha	Forestry Agency default (mid-aged sugi/hinoki)
Mean DBH	32 cm	Forestry Agency yield-prediction table (age 50–70)
Mean tree height	18 m	Same as above

4.3 Results (mm/yr scale)

Term	Value	Note
P (elevation-corrected precipitation)	2,331 mm/yr	After elevation correction
P_eve (event aggregate)	1,370 mm/yr	Regression-derived
Q (direct runoff)	533 mm/yr	Tertiary geology, ≤1306 mm band
Transpiration	453 mm/yr	Per-tree × stem density × 12 months
Canopy interception	384 mm/yr	Conifer interception rate
Total ET	837 mm/yr	Transpiration + interception
Water yield	961 mm/yr	41.2% of precipitation

5. Extending to the full AOI

Multiplying the per-hectare value by the satellite-measured forest area (12,398 ha) gives the annual water yield for the whole AOI.

961 mm/yr × 12,398 ha = 0.961 m × 12,398 × 10,000 m² = 119,115,675 m³/yr = ~ 1.19 hundred million tonnes/yr = ±30% CI: 83,448,000 – 154,974,000 m³

6. Scale comparison

Tokyo Dome (1.23 M m³)~97 Tokyo Domes / yr
Household domestic use (73 m³/household)~1.63 M households / yr
500 ml bottled-water units~238 billion bottles / yr
Mineral-water bottling plant capacitySeveral tens to ~100 large plants worth

7. Data sources

Open data sources (all free, commercial use OK)

Sentinel-2 L2A (satellite imagery / NDVI) — Copernicus Data Space Ecosystem (CDSE)
NASA POWER (monthly T & P) — power.larc.nasa.gov
GSI Japan elevation API (DEM) — maps.gsi.go.jp
AIST Seamless Geological Map v2 — gbank.gsj.jp/seamless/
Forestry Agency Simplified Evaluation Method Ver.1.0 (March 2026) — rinya.maff.go.jp

8. Reducing the error envelope

This estimate carries a typical ±30% uncertainty. Drivers and compression strategies:

Stem density, DBH, and height use Forestry Agency standard values → integrate Forestry Agency LiDAR (Tochigi / Hyogo / Kochi already public) for 20 m mesh stand measurements, shifting to Tier 3 with ±10–15%
Classification error from a single NDVI threshold → multi-temporal NDVI consensus to remove seasonal noise
Coarse geology mesh (multiple classes within AOI) → per-stand subdivision to match local geology
NASA POWER grid resolution (~50 km) → integrate AMeDAS (17 km, 10-minute resolution)

All of the above are on our roadmap as structural error compression — not just disclaimers.

※ This report is a third-party estimate aligned with the Forestry Agency of Japan's official formula. It is not an official certification by the Forestry Agency or MAFF. Issuance of J-Credit water-yield credits requires separate official monitoring and third-party verification. All numerical values are estimates and do not constitute commercial guarantees.