Extended Data Fig. 6: MESA+STELLA progenitor and degenerate light-curve models.

a, b, Ejecta mass M_ej and explosion energy \({E}_{\exp }\) inferred from Eq. (1) (Methods) as a function of progenitor radius R consistent with the bolometric light curve of SN 2018zd at the assumed luminosity distance of 9.6 ± 1.0 Mpc, along with the properties of the three degenerate explosion models. The blue and red shaded regions show explosion parameters expected for ECSNe^6,7,10 and typical of Fe CCSNe⁸⁶, respectively. c, d, Three degenerate MESA+STELLA explosion models providing good fits to the light curve and velocities inferred from the Fe ii λ5169 line during the plateau phase. Models are labelled by \({\rm{M}}[{M}_{{\rm{ej}},\odot }]\_{\rm{R}}[{R}_{\odot }]\_{\rm{E}}[{E}_{\exp ,51}]\). Error bars denote 1σ uncertainties. Note the observed early-time excess luminosity and suppressed velocity of SN 2018zd. This light-curve degeneracy highlights the inability to distinguish ECSNe from Fe CCSNe solely based on their light curves, suggesting that many ECSNe might have been overlooked owing to the lack of additional observations. e, f, Same as panels (c, d), but adding a dense wind profile (\({\dot{M}}_{{\rm{wind}}}=0.01\ {{\rm{M}}}_{\odot }\) yr⁻¹, v_wind = 20 km s⁻¹, and t_wind = 10 yr) to the three degenerate MESA models before handoff to STELLA. g, Comparison of the UV-colour models with the same wind CSM parameters as in panels (e, f). Error bars denote 1σ uncertainties. All three models with the same wind CSM parameters are able to reproduce the early-time luminosity excess and blueward UV-colour evolution almost identically, suggesting the insensitivity of a particular model choice. Despite a possible artificial velocity kink when the Fe line-forming region transitions from the CSM to the stellar ejecta, the velocity evolution with the early suppression is also reproduced.