INC-0240 — Atmospheric Oxygen Saturation (Great Oxidation Event)

Incident Report: INC-0240 — Atmospheric Oxygen Saturation

Severity: SEV-1 Status: Mitigated. Long tail open. Closing the report after 47 pages because the report cannot remain open forever. Duration: ~300 million years (acute phase ~2.45–2.20 Gya) Impact: Estimated >80% of extant anaerobic taxa lost. Atmospheric methane greenhouse collapsed. Two glaciations triggered (Huronian, partial Snowball Earth). Banded iron formation deposition saturated.

Summary

Cyanobacteria deployed oxygenic photosynthesis to production at roughly -2.7 Gya. Free O₂ output accumulated. The reduced-iron and reduced-sulfur sinks in the ocean buffered it for ~250 My. Sinks saturated at roughly -2.45 Gya. Atmospheric O₂ rose from <10⁻⁵ PAL to ~10⁻² PAL within ~10 My (geologically: instant). The anaerobic biosphere — stable for ~1.4 billion years — encountered a strong oxidant at ecosystem scale for the first time, and most of it died.

Timeline

~-3,400,000,000 (background): Anoxygenic photosynthesis is the standard. Photosystems run with H₂S, Fe²⁺, and reduced organics as electron donors. No O₂ output. System has been stable for ~1 Gy.

~-2,700,000,000 (the change): A cyanobacterial lineage links Photosystem II (P680) to Photosystem I (P700) in series and switches the PSII donor from H₂S to H₂O. The Z-scheme is now extracting electrons from water:

2 H₂O → O₂ + 4 H⁺ + 4 e⁻

O₂ becomes a byproduct of the energy pipeline. The cyanobacterial team did not file an RFC. There was no change advisory review. The change shipped.

~-2,700,000,000 to ~-2,450,000,000 (sink phase): Free O₂ is absorbed by Fe²⁺ in the oceans (precipitating as banded iron formations), by reduced sulfide minerals on continental shelves, and by atmospheric methane (oxidized to CO₂ + H₂O). Methane greenhouse weakens. Surface temperature drops. We did not have monitoring on either of these effects.

~-2,450,000,000 (saturation, the actual incident): Reduced-iron sinks saturate. O₂ enters the atmosphere as a free pool.

~-2,400,000,000 (acute casualties): Anaerobes encounter reactive oxygen species — superoxide (O₂•⁻), hydrogen peroxide (H₂O₂), hydroxyl radical (OH•) — generated by direct O₂ reactions and by adventitious chemistry with reduced iron (Fenton: Fe²⁺ + H₂O₂ → Fe³⁺ + OH⁻ + OH•). Damage observed:

• [4Fe-4S] clusters in central metabolism (aconitase, fumarase, and many oxidoreductases) oxidatively destroyed within minutes at ~10⁻² PAL O₂.
• Membrane lipids enter peroxidation cascades.
• DNA accumulates 8-oxo-guanine lesions and strand breaks.

Anaerobic taxa unable to retreat to anoxic refugia (deep ocean, sediment pore water, hydrothermal systems) are lost.

~-2,400,000,000 to ~-2,300,000,000 (Huronian glaciation): The methane greenhouse collapse triggers glaciation. This compounds the extinction. The second-order climate effect was not in the risk register.

~-2,300,000,000 onward (mitigation rollout): Aerobic respiration deployed across surviving lineages by horizontal gene transfer. Cytochrome c oxidase (Complex IV) handles the four-electron reduction of O₂ to water cleanly:

O₂ + 4 H⁺ + 4 e⁻ → 2 H₂O

This both removes ambient O₂ and yields ~30 ATP/glucose actual (ref ADR-003, commit 7f3a1b9), versus 2 ATP from substrate-level phosphorylation. Antioxidant systems deployed in parallel:

• Superoxide dismutase (Mn-SOD, Cu/Zn-SOD): 2 O₂•⁻ + 2 H⁺ → O₂ + H₂O₂
• Catalase: 2 H₂O₂ → 2 H₂O + O₂
• Glutathione (GSH/GSSG redox buffer)

The rollout is partial after 200 My. It is partial today. It will be partial.

Root Cause Analysis

Primary cause: The cyanobacterial PSII/PSI linkage was deployed to production without ecosystem-scale dependency review. Internal testing was performed on small populations in shallow water. Behavior at planetary scale was not modeled. The cyanobacterial team’s position was that PSII performance gains justified the deployment.

Contributing causes:

1. The anaerobic biosphere was treated as an internal stable substrate. Atmospheric composition was not on the dependency graph.
2. Reduced-iron sink capacity was finite. We were operating on that buffer for ~250 My without instrumentation.
3. There is no sandbox environment for atmospheric-scale changes. There was no way to test this without shipping it.
4. Michael’s infrastructure team operated physical constants at nominal throughout. The substrate layer performed exactly to spec. Recording this to be unambiguous: the failure was at the biology layer, not at infrastructure.

What We Won’t Fix and Why

• The two glaciations. Out of scope. Glaciations resolved through volcanic CO₂ outgassing over ~100 My. We did not intervene and we will not be intervening on future events of this class.
• The >80% of anaerobic taxa lost. There is no recovery path. We will not be undeploying O₂. The remaining anaerobic niches (sediments, deep biosphere, eventually animal gut microbiomes — though those won’t exist for ~1.5 Gy) are the supported configuration going forward.
• ROS generation in the aerobic electron transport chain. Complexes I and III leak electrons to O₂. Antioxidants are the permanent mitigation. We are not redesigning the ETC. Filing under “by design.”

Action Items

Owner	Action	Due
Raphael	Deploy aerobic respiration to remaining anaerobic lineages	~-2.0 Gya (status: ongoing, will not be met)
Gabriel	Document the antioxidant system (SOD, catalase, glutathione)	~-2.2 Gya (closed)
The Architect	Evaluate whether to acquire a preinstalled aerobic respiration solution rather than continue native rollout	~-2.0 Gya (closed — see ADR-006)
Cyanobacteria team	Implement change advisory process for atmospheric-scale deployments	TBD (open; the cyanobacteria team has not responded)
Uriel	Maintain the incident page	Indefinite (ongoing)

Lessons Learned

1. Atmospheric composition is now a tracked shared dependency. Changes that affect it require advisory board sign-off.
2. The “ship to one lineage and let it diffuse” pattern is suspended for any change that produces strong oxidants, strong reductants, or gases.
3. The cyanobacteria team is now a standing member of the change advisory board. They have attended zero of the last forty meetings.

Sign-Off

The Architect: Approved for closure. The long tail does not close. The 47 pages are sufficient. We are calling it done because the report has to be done at some point.

Michael: Physical constants throughout the incident were nominal. No changes to substrate. Adding this to the record so the biology post-mortem cannot, at any future date, be relitigated against infrastructure.