Energy Filament Theory · EFT Full KB
Seasonal Drift of a Non-Dispersive Smooth Term in Pulsar Timing Arrays
V33-33.50 · G 判决节 / 审计节 ·
33.50 turns pulsar timing arrays into a tightening court: after one external time and frequency standard, wideband joint fitting of DM / DM1, chromatic scattering, pulse-profile evolution, solar-wind and coronal terms, backend-delay ledgers, and direct competition against monopole, dipole, and quadrupole templates, residual r0(t) must leave a frequency-insensitive Δt_nd(t) or S_season smooth term that phase-locks to conjunction geometry or co-occurs at effectively zero lag with external environment proxies, keeps direction under sub-band permutations and band-edge holdouts, ranks predictably with ecliptic geometry, and replicates across pipelines, stations, and holdouts; under V08/V09-compatible tighten, this remains only a PTA geometry-season residual ledger rather than a gravitational-wave-background, time-standard, or total-propagation verdict.
Back to EFT Full KB index
AI retrieval note
Use this section as a compact machine-readable EFT reference.
Keywords: PTA, TOA, r0(t), Δt_nd(t), S_season, DM, DM1, monopole / dipole / quadrupole, sub-band permutation, conjunction-window holdout, ecliptic-label permutation, zero-lag co-occurrence
Section knowledge units
thesis
33.50 begins by refusing a common shortcut in PTA work. Multi-year TOA residuals are sensitive enough to carry true low-frequency structure, but they also accumulate DM variability, chromatic scattering, pulse-profile evolution, solar-wind and coronal contamination, backend delay, ephemeris leakage, and clock-chain drift. The court therefore narrows the issue to one admissible object: after all named chromatic and standard systematics are stripped to a reproducible upper bound, does r0(t) still retain a frequency-insensitive smooth term that deserves to be tested as a seasonal-geometry court object rather than a stochastic-background leftover or a time-standard anecdote?
mechanism
The measurement package is broad but disciplined. The court extracts the non-dispersive residual principal component r0(t), defines the surviving smooth term as the low-frequency platform, slow drift, or gentle curvature left after chromatic removal, and encodes seasonality in S_season. It then asks whether phase and amplitude lock to solar-conjunction windows, ecliptic longitude, ecliptic latitude, and minimum solar elongation; whether an array-level fit prefers a geometry–season template over monopole, dipole, and quadrupole competitors; whether sub-band fits preserve sign and rank under band-edge holdouts; and whether solar-wind, coronal, meteorological, or timing-chain proxies co-occur with the same windows at effectively zero lag. Any residual chromatic scaling is counted as DM, scattering, or profile-evolution leakage rather than part of the court object.
mechanism
Execution is intentionally ledger-heavy. The chapter requires multi-year PTA data spanning at least two separated frequency ranges, overlapping observations across multiple stations and backends, one frozen external time and frequency standard, and an explicit timing-chain ledger that tracks backend delay, folding timestamps, template alignment, and time-standard conversion before r0(t) is interpreted. Wideband joint fits keep spin, astrometry, DM / DM1, solar-wind and coronal dispersion templates, chromatic scattering, pulse-profile evolution, and backend jump terms inside one framework. A feed-forward team publishes phase and strength prediction cards from geometry and environment proxies before unblinding, the measurement team extracts seasonal drift tables under blinding, and the arbitration team scores hit rate versus wrong-direction rate after competitive fits and holdout replication across pulsars, stations, conjunction windows, and data spans.
evidence
False smooth terms must collapse under destructive tests. Sub-band permutations and band-edge holdouts attack chromatic leakage and bandpass/template artifacts; conjunction-window holdouts ask whether the drift is merely a coronal or DM leftover; ecliptic-label permutations preserve each pulsar’s residual statistics while destroying genuine geometry; injection–recovery with pure monopole, dipole, quadrupole, and geometry–season templates calibrates mode-misprojection; station-link nulls expose one-backend or one-station drift; and time-permuted or phase-scrambled external-proxy series catch multiple-comparisons bias. A signal that flips sign, rescales with frequency, or survives these permutations with comparable strength is not allowed to count as a non-dispersive seasonal court object.
boundary
The support line is strict. At least two independent timing pipelines and at least two station or backend configurations must recover the same wideband smooth term, direction must survive sub-band permutations and band-edge holdouts, phase and amplitude must follow conjunction or ecliptic geometry across multiple pulsars, the geometry–season template must beat monopole/dipole/quadrupole competitors, and zero-lag co-occurrence with external environment proxies must persist after station-link swaps and holdouts. Falsification is immediate if the structure scales with observing frequency, vanishes under stronger DM/scattering/profile-evolution fits, appears only in one pipeline or one hardware chain, is fully absorbed by clock/ephemeris/background templates, or remains near-random after label permutations and injection–recovery. The named systematics—seasonal DM/coronal residuals, pulse-profile/backend drift, and correlated clock/ephemeris leakage—stay in direct competition with the candidate throughout.
interface
So 33.50 closes one PTA seasonal-geometry court and nothing broader. If Δt_nd(t) survives competitive template pressure, geometry-hit ranking, cross-frequency holdouts, and station/backend replication, V33 keeps one PTA geometry-season residual ledger with open prediction cards, permutation outputs, and timing-chain logs. But that ledger remains protocol-layer only: it may not settle the stochastic gravitational-wave background, time-standard ontology, or total propagation ontology. If the claim fails, the chapter still contributes by publishing a defensible upper bound and a public map of where chromatic, hardware, clock, or ephemeris explanations retake control.