Correct Lake Washington model direction and separate oxygen experiments #1551

kb/communities/Lake_Washington_Methane_Oxygen_Methylotroph_Community.yaml, PMID:25755930 and PMID:32655999: oxygen microcosms support changes in guild composition, not a direct bridge to a separate two-methanotroph optimization model. In the model, conditional formaldehyde transfer is Methylomonas -> Methylobacter, opposite the current source/target. Sediment and synthetic ratios were imposed, not recovered predictions; CO2 export is environmental, not cross-feeding. The microcosm paper does not trace methanol donation or reciprocal benefit.

Retain the oxygen-selection arrow with measured composition limits, remove the cross-study arrow, correct the modeled donor/recipient and condition bounds. Record source wording contradictions separately without inventing a resolution. Preserve canonical taxonomy unchanged; explicitly document incomplete reliable participants. Close only after protected-main merge.

Source-based self-adversarial review, not independent approval. No new nodes/arrows or paid Edison submission.