Follow how a changing source heats circumnuclear dust—and how its infrared light reaches us.
Explore the time-dependent radiative-transfer model, understand its assumptions, and reproduce the WISE response of UGC 11487.
One source. Two radial dust populations.
Temperatures computed in time, space and grain size.
Drag to rotate; use the time slider or arrow keys. Coordinates are in pc. The cutaway is a flat display cut at x = 0, not a change to the physical geometry.
Local time is measured in the source rest frame relative to the unsmoothed fallback peak. At each point the stored solution is sampled at u = tlocal − r/c. These are simultaneous local snapshots, not images at one observer-arrival time.
The default temperature is ⟨T⟩N = ΣNaTa/ΣNa. The size selector shows the actual stored temperature for either end of the grain-size grid. Points outside the stored time support have no assigned temperature. Cold material is rendered more transparently; display opacity does not encode optical depth.
Background heating is absent in the model. Near-zero computed values before illumination do not measure the galaxy's quiescent dust temperature. Points sample an effective medium; their density is not a map of physical grain density.
The browser uses every second original spatial sample (1152 points), and 101 local-time frames from 0 to 5000 d. Values are interpolated from thermal_solution.npz and rounded to 0.01 K for display. Full-precision scientific arrays remain in the repository. The camera orientation is for exploration and does not alter the fitted viewing inclination.
The model starts with a hydrodynamical fallback history. It computes absorption, grain temperatures, thermal re-emission and light-travel delays. W1 and W2 are then predicted through the WISE bandpasses. No fixed dust-zone temperature or prescribed inter-band lag is fitted.
Select a stage to see what is calculated, what controls it, and where it appears in the code.
STARS supplies the mass fallback history. The adopted black-hole mass rescales time and fallback rate while preserving returned mass. A causal exponential kernel describes accretion smoothing.
Equations, assumptions & limitations ↗Case A: τvisc = 245.23 d; MBH = 106.8 M⊙.
src/v4_driver.pyThe evolving bolometric luminosity multiplies a fixed 30,000 K blackbody spectrum. Dust absorption is wavelength dependent. The calculation does not impose an Eddington cap or rescale to a chosen total event energy.
Equations, assumptions & limitations ↗Case A: η = 0.15176. Each radial optical depth fixes a dust column.
V48_LOCAL_PAIR/solver.pyDirect and reprocessed radiation heat graphite grains. Radiative equilibrium is solved for every thermal cell, time and grain size. Compact and extended populations share the same source and transport calculation.
Equations, assumptions & limitations ↗Graphite: a = 0.1–1 μm. The two radial intervals are 0.25–0.5 pc and 1–2 pc.
src/v48_angular_rt.pyThermal photons can escape or be absorbed and re-emitted. The solver follows the path delays and viewing-direction attenuation. Local-time snapshots and observer-time light curves represent different time slices.
Equations, assumptions & limitations ↗The 3D time slider uses local rest-frame time. The light curves use observer MJD.
src/v48_angular_rt.pyEscaping radiation is redshifted, converted to flux, integrated through the equal-energy W1/W2 response curves and calibrated against Vega. Fixed host fluxes are added before comparison with the observations.
Equations, assumptions & limitations ↗22 epochs × 2 bands. Adopted errors include the 1.10 mJy floor once.
V48_LOCAL_PAIR/solver.pyInspect the observed fluxes, their adopted uncertainties and the saved prediction. Hover or touch a light-curve epoch for its values.
χ² · nominal ν = 37
Across the stored solution
Integrated model IR emission
Move over an epoch to inspect the data, error and prediction.
The lower panel shows (data − model)/σ; dotted lines mark ±3σ. Error bars are those used in the likelihood. Host subtraction changes the displayed fluxes but leaves residuals unchanged.
These are conditional results for archived Case A. The fit does not uniquely determine the dust geometry. Full parameter uncertainties require the corresponding profile analysis.
Source physics, dust geometry, grain properties, equilibrium, energy conservation and the domain of validity.
Read the physical guide ↗02Eight steps with exact input files, copyable commands and the expected outputs. Begin with one verified model.
Open the guided tutorial ↗03Understand the objective, numerical convergence, grid search and conditional parameter intervals.
Explore the inference workflow ↗Computed: time-, radius- and grain-size-dependent temperatures, secondary heating, geometric delays, observer attenuation and absolute synthetic photometry.
Assumptions: an axisymmetric effective medium, graphite grains, a fixed source spectral shape and rapid thermal equilibrium. Scattering and steady background heating are omitted.
Static opacity: the solver checks temperatures against a sublimation threshold. It does not evolve irreversible grain destruction. Rendered points sample the effective medium.
Inputs and object-specific settings ↗