PASS-locked core: the causal chain
PASS-locked core: the causal chain from rupture to ice melt — This section fixes the core causal map used in this white paper. Only PASS-locked links are treated as evidence (LOCK → Derive → Gate). This paper treats Atlantic Separation / episodic opening (ARE) as the cause-axis, and large-scale ice melt as a downstream response.
This section fixes the core causal map used in this white paper. Narrative is minimized. Only PASS-locked links are treated as evidence (LOCK → Derive → Gate).
This section fixes the core causal map used in this white paper. Narrative is minimized. Only PASS-locked links are treated as evidence. HOLD is no-evidence; FAIL is rejected.
Causal direction (fixed) and forbidden inversion
Direction. This paper treats Atlantic Separation / episodic opening (ARE) as the cause-axis, and large-scale ice melt as a downstream response. Therefore, “ice melt caused Atlantic opening” (ice → opening) is forbidden as a core argument.
Inversion guard. Sea-level, hydrology, and climate records cannot by themselves infer the cause (ARE). An observed melt event is not evidence of ARE. To write “ARE → Ice Melt” in the core, independent PASS evidence on the ARE side (e.g., P1, P4) must be locked first.
Core causal map (concept)
Figure.
PASS-locked causal map (concept). Arrows indicate the causal direction; only PASS-locked modules are cited as evidence at each step. P29 checks event-window coherence statistically; it does not independently prove the cause (ARE).}
PASS-only link map (r16)
Table §5 maps which segments of the causal chain are locked in r16 using only PASS modules.
Causal segment PASS modules What is locked at this step (summary) Existence of ARE-side prerequisites P1, P4 P1 shows that subduction-dominated segments are rare around the Atlantic (subduction length fraction ≈ 0.017 in r16), consistent with passive margins and ridge signatures. P4 shows PASS-level lubrication / low-effective-stress signatures (LES) in the Atlantic opening context, locking a necessary (not sufficient) condition for low-friction slip. Threshold slip vs continuous drift (auxiliary) P13 Using literature-based priors, the median R_τ=τ_drive/τ_crit falls below the prereg hold threshold (R_τ≤ 0.3) and is PASS-locked. This leans against explaining the episode purely as continuous weak-boundary drift (assumption-dependent; not a direct observation). Stress/deformation proxy linkage (partial constraint) P31 In a global QA0 sample (N≈1006), coal-rank proxies and deformation/stress proxies show a non-zero statistical linkage, indicating that stress/deformation links are at least quantifiable in observational data (not a standalone causal proof). Freshwater/circulation response P16 Multiple North Atlantic proxies (freshwater/salinity + AMOC) show coherent inflection within the same event window, passing the prereg coherence gate (pattern evidence; not a unique-cause proof). Hydrology/sediment response P20 Misfit/underfit river metrics and mega-delta “major volume build” timing pass prereg thresholds, consistent with large drainage/discharge episodes. Endorheic low-stand onset clustering P24 Onset of low-stand transitions in endorheic basins is over-concentrated near the 4.2 ka window (permutation test; p≈0.00185; enrichment≈ 1.60), providing an auxiliary downstream record. Sea-level budget residual vs basin-volume buffering P19 Cross-constraints between observed sea level, budget components, and basin-volume proxies pass prereg gates, providing a quantitative constraint that basin-volume change could contribute to sea-level budget residuals. Event-window coherence (auxiliary eventness) P29 Different proxy classes cluster in a narrow event window statistically (p<0.05). This supports eventness but does not independently prove the cause (ARE).
Figure.
PASS-only evidence link map (r16). Only PASS modules are treated as evidence.}
Downstream-record layer (non-resource first)
The terminal step (ice-sheet destabilization → melt/retreat) can be recorded in sea level, hydrology, sediments, geomorphology, and other media. This white paper prioritizes non-resource records (sea level, hydrology, sedimentology) as validation channels; resource modules are used only as bias-aware auxiliary frames.
Three-layer observability model.
For a proxy S, 
No-evidence or rejected links (do not use in the core)
- HOLD: P32–P35 (oil budget/thermal context/chronometers/discoverability) are not evidence-eligible in r16.
- HOLD (not executed): P8, P9, P10, P12, P14, P17, P22–P23, P25–P28 have prereg stubs but were not executed in r16; do not cite as evidence.
- FAIL: P21 (global mean deceleration tail) is rejected and excluded from the core.