Methodology

1. Scope

Beanwater Labs publishes three live instruments. Each combines publicly available environmental observation and forecast data with proprietary Beanwater calculations. This document separates those two layers so that a reader can determine, for any figure on this site, whether it describes the world or describes Beanwater.

2. Real input data

All environmental data is retrieved from Open-Meteo, which aggregates national weather-service model output. United States ZIP codes are resolved through Zippopotam.us. Place-name searches are resolved through the Open-Meteo geocoding service. No key is required for either provider and none is embedded in this site.

The following variables are requested:

• temperature_2m — air temperature at 2 m, °C
• relative_humidity_2m — relative humidity at 2 m, %
• dew_point_2m — dew point at 2 m, °C
• precipitation, rain, showers — mm per hour
• precipitation_probability — %, where the model provides it
• cloud_cover — total cloud cover, %
• wind_speed_10m — wind speed at 10 m
• surface_pressure — hPa
• soil_moisture_0_to_1cm, _1_to_3cm, _3_to_9cm — m³/m³
• soil_temperature_0cm — °C
• et0_fao_evapotranspiration — reference evapotranspiration, mm per hour

These are real measurements and model outputs. Beanwater Labs does not modify them, and does not substitute invented values when a request fails.

3. Beanwater derived classification

Everything else on this site is invented. The Wetness Index, the Global Hydration Index, Bean Readiness, Root-Zone Liquidity, Pod Pressure, Loam Volatility, ET Stress, Mold-Adjacent Concern, Moisture Sovereignty Risk, Certified Wet™ and the HumanFluid™ escalation thresholds are Beanwater constructions. They are arithmetic performed on real numbers to produce a result with no external standing.

The arithmetic is deterministic and documented in the source. It is not random, and it does respond to actual conditions. It simply does not mean anything outside this organization.

3.1 Wetness Index — ATM-001

A 0–100 composite of four normalized components. Where the model supplies precipitation probability:

score = humidity      × 0.30
      + dewPointScore × 0.20
      + precipProb    × 0.20
      + precip24hScore× 0.20
      + activeRain    × 0.10

Where precipitation probability is unavailable, the remaining four components are reweighted proportionally so the sum stays at 1.0. dewPointScore maps dew point in °F onto a piecewise curve; precip24hScore maps accumulated precipitation across a ±12 hour window onto a second piecewise curve; activeRain is binary. Bands: 0–19 dry, 20–39 low, 40–59 normal, 60–79 Certified Wet™, 80–100 escalation.

3.2 Global Hydration Index — ATM-002

The same ATM-001 calculation applied at twelve fixed reporting nodes, then averaged. Derived planetary figures — Moisture Sovereignty Risk, Certified Wet™ coverage, Atmospheric Coherence — are computed from the dispersion of those twelve scores. Planetary Dampness Drift compares the composite against a fixed internal baseline table; that baseline is invented and does not represent historical observation.

3.3 Beanfield Futures Desk — AGR-001

Soil-led rather than atmosphere-led. Bean Readiness weights surface and shallow soil moisture at 0.40, soil temperature at 0.25, ET stress at 0.20 and recent precipitation at 0.15. Where soil variables are unavailable for a location, the instrument falls back to an atmosphere-only estimate and says so on screen rather than silently substituting.

4. Update behaviour

Underlying data is hourly model output, not continuous observation. The instruments select the hourly record nearest the current time at the location being measured, and display the timestamp of that record rather than the time of the page request. A reading is therefore current to within one hour, and is not live in the second-by-second sense.

Nothing on this site polls automatically. Values refresh when a page is loaded or when a refresh control is used.

5. Failure behaviour

When a data request fails, the affected figure is withheld. Beanwater Labs does not display a plausible substitute.

• A failed local request produces a labelled feed-disruption state, distinguishing input errors from network errors from incomplete model data.
• A failed node in the Global Hydration Index is marked unavailable and excluded from the composite, which is then labelled a partial network calculation with a node count.
• A total failure produces an unavailable state, not a zero.

6. Limitations

• Model output, not instrumentation. There are no Beanwater sensors.
• Soil variables are modelled estimates and vary in quality by region.
• The twelve GHI nodes are cities, weighted equally. This is not a representative sample of the planet's surface, population, or agriculture.
• Thresholds and weightings were selected because they produce a readable spread, not because they were validated against anything.
• Beanwater classifications are not scientific, agricultural, medical, financial or legal advice.

7. Status of proprietary labels

Certified Wet™, HumanFluid™, Bean Desiccation Event, Peak Legume Readiness, Moisture Sovereignty Risk and all similar designations are fictional classifications belonging to a satirical project. They are not certifications. No external body recognises them. Compliance has reviewed this paragraph and confirmed it is accurate, which it notes is unusual.

---

This document is part of a satirical fiction. Beanwater classifications, thresholds and program names are invented and carry no scientific, agricultural, medical or regulatory standing. Where real environmental data is referenced, its provider is named and linked. See Methodology.