Reproducibility archive for: Scan Functional Choice in Lag-Resolved Dependence Analysis: A Controlled Comparison Under Temporal Dependence

This repository contains the reproducibility archive for the manuscript “Scan Functional Choice in Lag-Resolved Dependence Analysis: A Controlled Comparison Under Temporal Dependence.” The archive contains source code, configuration files, source-of-record outputs, figures, and manuscript materials supporting a matched-calibration comparison of scan-level functionals for lag-resolved dependence analysis under temporal dependence. The manuscript does not claim that PWSS is generally superior to the max-statistic. Stage E is exploratory and hypothesis-generating only. Stage C is the locked confirmatory evaluation. Groundwater analyses are illustrative only. ETTh1 is used as a resampling-adequacy diagnostic only. Embedded validation is descriptive only. The active PWSS threshold is c = 0.179921, calibrated on the marginal lag-window |C(tau; theta)| scale. Superseded threshold artifacts and stale development outputs, if retained, are placed under superseded/ and are not source-of-record manuscript evidence. Version update, 21 July 2026 This version preserves the original PWSS reproducibility archive and adds PWSS_reproducibility_repair_payload.zip. The repair payload documents the active PWSS threshold provenance for c = 0.17992110550403595, including the marginal per-cell calibration source and active c090 configuration. It also distinguishes this threshold calibration from the superseded scan-level Tmax calibration artifact. The payload adds missing source-of-record CSV files required for manuscript table/figure reproducibility, including the Phase 1 strict-null summary, Stage E kernel pilot summary, selected Phase 3 embedded-validation table, and locked Phase 3C groundwater panel. It also includes the completed spike-extension and block-length audit summaries used for manuscript revision. v1.1.1 reproducibility payload repair: replaces the repair payload with a complete version including Stage C confirmatory summary and replicate-level p-values, and the ETTh1 strict-null summary. No manuscript results, numerical values, scan parameters, or conclusions are changed.