Files, units, and configuration
A reproducible model is defined by its physical parameters, the exact numerical input files and the active solver settings. Similar filenames or rounded values are not enough to identify the same run.
1. The canonical photometry table
The physical fit uses the author-processed 22-epoch input CSV, not raw individual WISE detections. The loader requires exactly 22 rows, ordered by MJD, finite fluxes, strictly positive errors and constant per-band host levels.
| Columns read by the fit | Units and meaning |
|---|---|
mjd | Observer-frame modified Julian date. |
F_W1_Jy, F_W2_Jy | Observed total band flux density, Jy. These are not host-subtracted inputs. |
sigma_W1_Jy, sigma_W2_Jy | Effective errors used directly in χ², Jy. |
baseline_W1_mJy, baseline_W2_mJy | Constant host flux densities, mJy; divided by 1000 before comparison in Jy. |
The table also contains convenient mJy fluxes, statistical-error columns and excess fluxes. They are not extra independent measurements in the likelihood. Supplying an excess column in place of the total flux and subtracting the fixed host again would count the baseline twice.
sfloor ≈ 1.10 mJy
The supplied error columns already include the adopted additive floor in quadrature. The active loader does not add another floor, an additional epoch-dependent 3% fractional error or a fitted scatter term. Applying an additional floor changes the likelihood and the meaning of the reported χ².
The 1.10 mJy floor is a rounded description. For exact archive reproduction, use the supplied error columns rather than reconstructing them from rounded prose or from the statistical-error columns alone.
The fixed hosts are 8.072315984025277 mJy (W1) and 5.215414582541471 mJy (W2), corresponding to the first two NEOWISE epochs. They are not the alternative combined AllWISE+NEOWISE baseline used for some empirical excess-light-curve diagnostics. Host uncertainty is not propagated into this conditional physical fit, and the host is not a source of thermal heating.
2. Archived Case A physical configuration
The downloadable configuration retains full-precision values. Compact and extended triples are [inner radius in pc, outer radius in pc, radial density exponent p]. Angles are measured from the symmetry axis, not from the equatorial plane.
| Parameter | Meaning | Value |
|---|---|---|
| stars | STARS path relative to library root | input/m3.0_t0.0/0.850.dat |
| eta | Radiative efficiency | 0.15175680389657437 |
| tau_visc | Smoothing time, rest-frame days | 245.23129535940873 |
| compact | [inner pc, outer pc, power-law p] | [0.25, 0.5, 0.0] |
| extended | [inner pc, outer pc, power-law p] | [1.0, 2.0, 2.0] |
| tau_C | Compact source-weighted optical depth | 0.39027703896078636 |
| tau_E | Extended source-weighted optical depth | 2.0376725567022778 |
| theta | Polar opening angle, degrees | 47.33077784246419 |
| inc | Viewing inclination, degrees | 43.44489826956509 |
| composition | Grain composition | graphite |
| amin | Minimum size, μm | 0.1 |
| amax | Maximum size, μm | 1.0 |
| q | Size-distribution exponent | 3.5 |
| density | Gas density for the sublimation threshold, cm−3 | 1000000000.0 |
The six continuously refined physical parameters are η, τvisc, τC, τE, opening angle and inclination. The event epoch t₀ is profiled separately. The selected fallback identifier, radial boundaries/exponents, composition, size limits and q are fixed for this local solution. The threshold density is also fixed; it is not inferred as a gas-density measurement from WISE.
3. Object and source constants
| Constant | Adopted value | Role |
|---|---|---|
| Redshift z | 0.01895 | Wavelength mapping and observer-time dilation. |
| Luminosity distance DL | 82.34 Mpc | Absolute flux conversion. |
| Black-hole mass MBH | 6.309573×106 M⊙ | Fallback scaling from the STARS reference mass. |
| Source temperature Tsrc | 30,000 K | Fixed isotropic illuminating spectral shape. |
| Threshold density nH | 109 cm−3 | Graphite threshold 1793.30 K, not grain-column normalisation. |
These constants are inherited from the project loader. Reusing the code for another object requires reviewing them, the 22-row requirement, baseline definitions and the bounded MJD 56500–57500 epoch search. Replacing only the CSV is insufficient.
4. External numerical inputs
| Input | Exact identifier | Use and convention |
|---|---|---|
| STARS fallback | input/m3.0_t0.0/0.850.dat | Selected tabulated history; first columns are days and M⊙ yr−1. Local processing scales the black-hole mass and shifts time zero. |
| Graphite optical properties | Gra_81.gz | Qabs(λ,a). Wavelengths and grain radii are tabulated in μm. The reference fit uses graphite; the presence of a silicate table does not imply a mixture. |
| WISE response | RSR-W1.EE.txt, RSR-W2.EE.txt | Observed wavelength in μm; relative equal-energy response. No extra photon-counting wavelength factor. |
| Vega spectrum | alpha_lyr_stis_011.fits | CALSPEC wavelength in Å and Fλ in erg s−1 cm−2 Å−1; use this exact version for reproducibility. |
STARS is located in STARS_library-master or the directory named by UGC_STARS_ROOT. Optical and calibration files are under data/reference. The reference page links the original papers, upstream repository, official calibration documentation and exact downloadable products.
5. Units and numerical settings
Input geometry uses pc, grain radius and tabulated wavelength use μm, and the transport calculation converts to SI lengths and powers. Flux fitting uses Jy; figures commonly show mJy. Driver and transport times are rest-frame days, whereas the observed MJD differences include the factor 1+z. Reported energy-ledger values are in erg.
The configuration contains inherited numerical keys, but the active resolution is recorded separately in each run’s summary.json and provenance.json. For this archive it is 6 radial cells per population, 6 polar cells, 144 transport wavelengths, 30-day transport bins, 256 thermal rays and 512 observer samples. The driver stays at 10-day sampling. See the authoritative resolution table before changing or interpreting old n_radii, n_mu, n_phi or seed entries.
6. Fingerprints and release provenance
The archived release manifest records SHA-256 fingerprints for the distributed numerical inputs. The preparation audit specifically records byte-for-byte restoration of the canonical photometry and selected Case A fallback curve against the archived provenance. It does not make that same reconstruction claim for every other STARS history.
A filename identifies a product, while its hash identifies the bytes actually used. Keep both, along with the exact source version and active numerical settings. Do not silently replace the Vega standard with a newer CALSPEC file, substitute updated optical functions or infer the STARS bundle’s version from a citation alone. Those changes require a new provenance record and potentially a new model calculation.
The website and source repository are publicly accessible, but preparation-era release checks remain historical records. Project citation metadata and third-party attribution or redistribution terms should be resolved for a versioned release; public download availability is not itself a new data licence.