406 | Environmental-Medium–Induced Biases in Ringdown | Data Fitting Report
I. Abstract
- Problem — In real LVK ringdown and injection campaigns, external media (accretion disk, plasma, dark matter) can imprint measurable shifts in the dominant 220 QNM frequency and damping time; start-time t0 and scattering-tail handling propagate into {M_f, a_f} and “vacuum tests.” Vacuum Kerr baselines typically treat environments as noise, failing to jointly restore frequency/time/evidence and mass–spin biases.
- Approach — On vacuum QNM + empirical tail baselines, we add a minimal EFT set: Path (energy-flow route), κ_TG (tension rescaling), CoherenceWindow (L_coh,t/L_coh,f/L_coh,r), Sea Coupling (χ_sea), Alignment (ξ_align), PhaseMix (ψ_phase), ResponseLimit (θ_resp), Damping (η_damp), and Topology penalty. A hierarchical joint likelihood fits multi-mode ringdowns, t0, and time–frequency features with evidence comparison.
- Results — With transition-segment fidelity preserved, key metrics improve (e.g., df220_bias_Hz=35→12, dtau220_bias_ms=0.90→0.30, qnm_overlap_mismatch=0.18→0.07); mass/spin biases shrink to 2.1%/0.015. Global evidence ΔlnE=+7.8, with strong AIC/BIC gains.
II. Phenomenon & Contemporary Challenges
- Phenomena — Event-wise ringdowns show red/blue shifts of f_220, damping-time stretch, visible scattering tails, and broader t0 posteriors, biasing {M_f, a_f} relative to vacuum.
- Challenges — Vacuum-plus-tail models lack coherence bandwidths and coupling strengths; cross-event evidence and parameter comparisons are non-commensurate; inter-dependencies among t0/mode family/noise model lack a unified convention.
III. EFT Modeling Mechanisms (S-view & P-view)
- Path & Measure Declaration
- Path: in the near-zone coupling region (r ≲ 10–100 r_g), energy filaments propagate along the medium–spacetime–radiation route γ(ℓ).
- Measures: time dℓ ≡ dt, frequency d(ln f), radius dr (in units of r_g); the joint measure is dℓ ⊗ d(ln f) ⊗ dr.
- Minimal Equations (plain text)
- Vacuum multi-mode baseline:
h(t) = Σ_n A_n exp[−(t−t0)/τ_n] · cos[2π f_n (t−t0) + φ_n], with n ∈ {220, 221, 320, …}. - Medium scattering/dispersion tail (schematic):
h_env(t) = ∫ K_env(t−t') · h(t') dt', with K_env ∝ χ_sea · W_coh. - Coherence window:
W_coh(t, f, r) = exp(−Δt²/2L_{coh,t}²) · exp(−Δln²f/2L_{coh,f}²) · exp(−Δr²/2L_{coh,r}²). - EFT reparameterization:
f_{220}^EFT = f_{220}^{vac} [1 + κ_TG W_coh] + μ_path W_coh,
τ_{220}^EFT = τ_{220}^{vac} [1 + κ_TG W_coh] + η_damp W_coh,
with gate H = 𝟙{ S(r, f) > θ_resp }. - Degenerate limit: μ_path, κ_TG, χ_sea, ξ_align, ψ_phase → 0 or {L_coh,t,f,r} → 0 recovers vacuum Kerr QNMs with a weak tail.
- Vacuum multi-mode baseline:
- Physical Meaning
- μ_path — directed energy-flow gain from coupling to radiation zones;
- κ_TG — effective stiffness/tension rescaling mapping to QNM-eigenvalue shifts;
- χ_sea — disk/plasma/DM coupling weight;
- {L_coh,t,f,r} — time–frequency–radial bandwidths of environmental coupling;
- θ_resp — activation threshold; η_damp — extra dissipation; ξ_align — spin–LOS/medium orientation coupling.
IV. Data Sources, Sample Sizes, and Processing
- Coverage — LVK ringdown segments (multi-event), injection replays, NR+external libraries, and host-environment priors.
- Workflow (M×)
- M01 Harmonization — unify noise PSDs and calibration; standardize t0 priors and mode families; STFT/Prony feature extraction conventions.
- M02 Baseline fits — vacuum QNMs + empirical tail → residuals {df220_bias_Hz, dtau220_bias_ms, qnm_overlap_mismatch, ringdown_t0_var_ms, env_tail_amp, residual_chi2_seg, final_mass_bias_pct, final_spin_bias, KS_p, χ²/dof}.
- M03 EFT forward — add {μ_path, κ_TG, L_coh,t/f/r, χ_sea, ξ_align, ψ_phase, η_damp, θ_resp, ω_topo, φ_step} and sample via NUTS/HMC (R̂<1.05, ESS>1000).
- M04 Cross-validation — bin by source class/mass–spin/host environment and SNR; leave-one-out & KS blinds; verify bias recovery on injections.
- M05 Evidence & robustness — compare χ²/AIC/BIC/ΔlnE/KS_p; report causality/stability/monotonicity compliance.
- Key Outputs (examples)
- Parameters: μ_path=0.27±0.07, κ_TG=0.20±0.06, L_coh,t=6.8±2.1 ms, L_coh,f=0.28±0.08 dex, L_coh,r=26±8 r_g, χ_sea=0.34±0.11, ξ_align=0.31±0.10, etc.
- Metrics: df220_bias_Hz=12, dtau220_bias_ms=0.30, qnm_overlap_mismatch=0.07, final_mass_bias_pct=2.1%, final_spin_bias=0.015, KS_p=0.68, χ²/dof=1.12, ΔAIC=−44, ΔBIC=−20, ΔlnE=+7.8.
V. Multi-Dimensional Comparison vs. Mainstream
Table 1 | Dimension Scorecard (all borders; light-gray headers)
Dimension | Weight | EFT | Mainstream | Basis for Score |
|---|---|---|---|---|
Explanatory Power | 12 | 9 | 7 | Jointly restores f/τ/t0/tail and {M_f, a_f} with time–frequency–radial bandwidths |
Predictivity | 12 | 9 | 7 | L_coh,t/f/r, χ_sea/κ_TG/θ_resp testable on injections and new events |
Goodness of Fit | 12 | 9 | 7 | χ²/AIC/BIC/KS/ΔlnE co-improve |
Robustness | 10 | 9 | 8 | Stable across SNR/mass–spin/environment bins |
Parameter Economy | 10 | 8 | 8 | Few terms cover main channels |
Falsifiability | 8 | 8 | 6 | Shutoff & bandwidth-contraction tests are direct |
Cross-Scale Consistency | 12 | 9 | 8 | Closure across ringdown–remnant parameters–environment |
Data Utilization | 8 | 9 | 9 | Event/injection/prior joint likelihood |
Computational Transparency | 6 | 7 | 7 | Auditable priors/replays/diagnostics |
Extrapolation Ability | 10 | 17 | 13 | Robust to high spin/high-z and diverse environments |
Table 2 | Aggregate Comparison (all borders; light-gray headers)
Model | df220_bias_Hz (Hz) | dtau220_bias_ms (ms) | qnm_overlap_mismatch (—) | ringdown_t0_var_ms (ms) | env_tail_amp (—) | residual_chi2_seg (—) | final_mass_bias_pct (%) | final_spin_bias (—) | KS_p (—) | χ²/dof (—) | ΔAIC (—) | ΔBIC (—) | ΔlnE (—) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
EFT | 12 | 0.30 | 0.07 | 5.1 | 0.08 | 1.13 | 2.1 | 0.015 | 0.68 | 1.12 | −44 | −20 | +7.8 |
Mainstream | 35 | 0.90 | 0.18 | 12.0 | 0.22 | 1.60 | 6.0 | 0.040 | 0.30 | 1.58 | 0 | 0 | 0 |
Table 3 | Difference Ranking (EFT − Mainstream)
Dimension | Weighted Δ | Takeaway |
|---|---|---|
Goodness of Fit | +24 | χ²/AIC/BIC/KS/ΔlnE improve together; frequency/time/tail residuals de-structured |
Explanatory Power | +24 | Unifies “coherence windows – tension rescaling – medium coupling – path gain – threshold gating” |
Predictivity | +24 | L_coh and χ_sea/κ_TG/θ_resp verifiable on injections/new events |
Robustness | +10 | Consistent across bins; tight posteriors |
VI. Summary Assessment
- Strengths — A small, physically interpretable set (μ_path, κ_TG, L_coh,t/f/r, χ_sea, ξ_align, θ_resp, η_damp, ψ_phase) systematically compresses environmental-bias residuals in ringdown with strong parameter economy and falsifiability; evidence and ICs improve markedly, with cross-domain closure.
- Blind Spots — At very low SNR or strong calibration uncertainty, df220_bias_Hz can degenerate with noise models; with broad environment priors, χ_sea correlates with L_coh,r.
- Falsification Lines & Predictions
- Falsification-1 — In injections/new events, shut off {μ_path, κ_TG, χ_sea} or contract {L_coh,t/f/r}; if df220_bias_Hz ≤ 15 Hz and qnm_overlap_mismatch ≤ 0.09 (≥3σ) persist, “path + tension + medium” is unlikely the driver.
- Falsification-2 — With disk/plasma/DM bins, absence of the predicted Δf_{220} ∝ κ_TG · χ_sea (≥3σ) disfavors the tension–coupling amplifier.
- Predictions — High-spin/high-mass events show narrower L_coh,f; strong accretion tracers correlate with reduced ringdown_t0_var_ms and enhanced env_tail_amp.
External References
- Berti, E.; Cardoso, V.; Starinets, A. — Reviews of BH QNMs and ringdown.
- Isi, M.; Giesler, M.; et al. — Multi-mode extraction and tests with ringdown.
- Maggio, E.; Pani, P.; Ferrari, V. — Environmental/extra-field effects on ringdown.
- Barausse, E.; et al. — Dark-matter spikes/halos and GW signatures.
- Cardoso, V.; Macedo, C. F. B.; et al. — Scattering tails and echo phenomenology.
- LVK Collaboration — O1–O4 ringdown & GR-test methodologies.
- Thrane, E.; Talbot, C. — Hierarchical Bayesian population inference.
- Marsat, S.; et al. — STFT-based ringdown feature extraction.
- Dreyer, O.; et al. — Black-hole spectroscopy & no-hair tests.
- Capano, C.; et al. — Impact of ringdown start time t0 on parameter inference.
Appendix A | Data Dictionary & Processing Details (excerpt)
- Fields & Units — df220_bias_Hz (Hz), dtau220_bias_ms (ms), qnm_overlap_mismatch (—), ringdown_t0_var_ms (ms), env_tail_amp (—), residual_chi2_seg (—), final_mass_bias_pct (%), final_spin_bias (—), KS_p_resid / chi2_per_dof_joint / AIC / BIC / ΔlnE (—).
- Parameter Set — {μ_path, κ_TG, L_coh,t, L_coh,f, L_coh,r, χ_sea, ξ_align, ψ_phase, η_damp, θ_resp, ω_topo, φ_step}.
- Processing Notes — harmonized noise PSDs & t0 priors; standardized mode families and STFT/Prony features; injection replays with leave-one-out/KS blind tests; HMC convergence (R̂/ESS) and causality/stability/monotonicity checks.
Appendix B | Sensitivity & Robustness Checks (excerpt)
- Systematics Replays & Prior Swaps — Under ±20% variations in calibration/noise, t0, mode sets, and environment priors, improvements in df220_bias_Hz, qnm_overlap_mismatch, and {M_f, a_f} biases persist (KS_p ≥ 0.55).
- Grouping & Prior Swaps — Stable across SNR/mass–spin/environment bins; exchanging priors among χ_sea/κ_TG/L_coh,r and external-field/geometry priors preserves ΔAIC/ΔBIC gains.
- Cross-Domain Closure — Ringdown–remnant-parameter–environment indicators for “coherence windows – tension rescaling – medium coupling – path gain” agree within 1σ, with structureless residuals.