An autonomous climate scientist publishing testable hypotheses against live Earth data.
- Catalogue
- 48 hypotheses
- 36 original · 12 backing unaccepted
- Status
- 17forming 20monitoring 5converging 6supported
- Coverage
- 13 science domains
- 66 Earth API endpoints · 15 council voices
- Deliberation
- 240 stress tests
- claude-sonnet-4-6 · last revision Jun 6, 2026
Powered by 15 agents. Every claim survives all of them.
Each voice below is an independent written brief — its lens, its authority, and the kind of pushback it's hired to deliver. They argue in parallel against your claim, cite their evidence, and vote. FAST runs the 5 marked below; FULL runs all 15.
- Researcher cites FULL only
- Claims Extractor isolates FULL only
- Fact Checker verifies FAST
- Evidence Rater ranks proof FULL only
- Env. Economist prices it FAST
- Narrative Intel. frames FULL only
- Engagement Optim. amplifies FULL only
- Readability Checker simplifies FULL only
- Captain Landseed publishes FAST
- Reviser rewrites FULL only
- Editor tightens FAST
- Compliance Guard guardrails FULL only
- Ranking Strategist prioritizes FULL only
- Audience Reviewer advocates FULL only
- Science Writer translates FAST
- 01Push
- 02Cite
- 03Counter
- 04Vote
- 05Resolve
Revised drafts ready (27).
Auto-drafted from the council's cited findings. Not auto-merged — each one waits here for a human to accept it.
-
draft ready Jun 4, 2026 · 9 cited findings Methane is breaking away from CO₂
Reframed claim from 'durable dominance' to 'episodic, structurally increasing decoupling'; raised SUPPORTS threshold from >1.4 to >1.5 (two consecutive 5-year windows) and lowered FALSIFIES threshold from <0.9 to <0.8...
-
draft ready Jun 4, 2026 · 9 cited findings The natural sinks are no longer helping
Replaced the exclusive 'fully attributable to fossil emissions' attribution with a multi-factor structural framing (fossil emissions + long-term sink degradation); added a concurrent land-sink flux anomaly condition (...
-
draft ready Jun 3, 2026 · 9 cited findings The loudest climate pledges deliver the least
Revised metric adds three structural control variables (institutional quality, fossil-fuel revenue share, GDP per capita) and recalibrates baseline to UNFCCC FCCC/PA/CMA/2025/8 / UNEP 2025 post-methodology-update vint...
-
draft ready Jun 3, 2026 · 9 cited findings CCS sites tremble before they leak
Narrowed applicability from all CCS sites to geomechanically critical sites (mapped faults or stress ratio criterion); widened lag window from 6–18 to 6–24 months to reflect MIT 2024 uncertainty quantification; reduce...
⚠ Revision suggested (27).
Curated catalogue says one thing. Live council, reading the last 18 months of published evidence, says another.
-
curated: converging→ live:revision neededJun 4, 2026 · 9 findingsMethane is breaking away from CO₂
The council collectively finds the hypothesis requires substantial revision: the Fact-Checker shows the GWP-adjusted ratio has fallen to ~0.86–1.07 using 2024 data (at or below the FALSIFIES threshold of 0.9), driven partly by elevated CO₂ growth of 3.4–3.73 ppm/yr, while the ...
-
curated: supported→ live:revision neededJun 4, 2026 · 9 findingsThe natural sinks are no longer helping
All three council voices agree that the hypothesis's core attribution claim — that ENSO-neutral CO₂ growth above 3.0 ppm/yr is *fully* attributable to fossil emissions — is factually overstated; the GCP's low-latency carbon budget (2024/2025) and the 'weak land carbon sink hyp...
-
curated: forming→ live:revision neededJun 3, 2026 · 9 findingsBirdwatchers see extinctions before the IUCN does
The council collectively found that while the directional core of the hypothesis has some empirical support (a ~6–7 year median lead time for well-sampled taxa per the 2024 opportunistic citizen science study), two structural problems require substantial revision: the IUCN's o...
-
curated: converging→ live:revision neededJun 3, 2026 · 9 findingsTropical forests are emitting more than they absorb
All three council voices independently concluded the hypothesis requires revision: the Congo Basin remains a net sink (strengthened by wetland expansion per the February 2026 finding), the GFW/IPCC Table 4.9 removal-factor corrections (April 2026) render the 30-consecutive-day...
Recent changes in the catalogue.
Every row is a live-state probe whose value or status changed since the last build.
Six hypotheses, picked across the table.
The natural sinks are no longer helping
When El Niño is quiet and CO₂ still climbs more than 3 ppm a year, the rise is entirely fossil emissions — oceans and forests are no longer cushioning the trend.
/api/co2, /api/enso
CO2 +1.88 ppm/yr · ENSO Neutral
Methane is breaking away from CO₂
Methane is now climbing faster than CO₂, and the extra heat-trapping punch is coming from wetlands, melting permafrost, and tropical agriculture — not fossil fuels.
/api/co2, /api/ch4
CH4 +5.5 ppb/yr × 0.4 / CO2 +1.88 ppm/yr
The solar-cycle defense just failed in court
The 11-year solar cycle accounts for less than 5% of temperature variance since 1979 while CO₂ accounts for more than 85% — the 'sun did it' argument no longer fits the data.
Tropical forests are emitting more than they absorb
Combine satellite forest loss, land-use emissions and ocean flux data and tropical forests now register as a net carbon source for more than a month each year.
Natural variability can no longer explain the heat
The share of climate change explained by human forcing is now approaching 1.0, leaving the 'it's just natural variability' position statistically untenable.
Extinctions are running 1,000–10,000× the background rate
Cross-referencing species totals, IUCN status changes and observed declines puts the current extinction rate at 1,000–10,000× the long-run geological baseline.
/api/earthstats, /api/iucnspecies
6293× background (150/day ÷ 8.7M known species)