SWSN SCIENTIFIC REVIEW SERIES SWSN-SCI/ORP-2035-04-R3
Figure 2:Correlation matrix of the reproducibility stress-test residuals. Sign differences are permit-
ted because different sensors may respond with opposite convention or transfer function.
For restricted operational data, the same pipeline must be challenged against known com-
mon causes: shared time references, synchronized maintenance windows, common envi-
ronmental forcing, data-assimilation leakage and preprocessing artifacts. Failure to remove
these effects rejects the need for H2.
4 Phenomenological inter-branch metric-overlap model
We introduce the metric ansatz
geff
µν (x) = g(0)
µν (x) + ε χ(x) h(B)
µν (x), (3)
where g(0)
µν is the background metric,h(B)
µν is an effective perturbation associated with a sec-
ond boundary branch,χ(x) is a compact overlap envelope andε is a dimensionless coupling
parameter. No claim is made thath(B)
µν corresponds to a physically realized parallel universe.
The ansatz is a bookkeeping device for deriving correlated signatures.
The strongest version of H2 further assumes that the overlap boundary is non-orientable
in an effective higher-dimensional embedding. A Möbius strip is a two-dimensional anal-
ogy: following a locally continuous path can return to a position with reversed orientation.
We use this only as a topology metaphor. The physical model would require a well-defined
manifold, stress-energy conditions and causal structure that are not supplied by the analogy.
4.1 Predicted observational signatures
The hypothesis is useful only if it predicts measurable inconsistencies. We therefore require
at least three independently instrumented signatures:
1. a coherent timing residual not removable by reference-clock or propagation corrections;
CONTROLLED SCIENTIFIC DOCUMENT Internal review copy Page 4 of9