getnewenv.f

GETNEWENV moves the outer fitting point of the interior model outward, by turning part of the envelope into new interior points, so that the envelope again holds the fraction of the mass it held before the step's mass change ($\log(m_\mathrm{fit}/M)$ = SENVOLD). Its header (MHP 07/02): "move the outer fitting point when the total mass is being decreased; add new points by dropping a sinkline from the surface down and including new points up to the desired envelope mass." It integrates the envelope once from the surface, then copies the points between the old and the new fitting point into the model.

When it runs. The main program calls it when MASSLOSS has set LNEWFIT, which it does whenever its rate DMDT is positive (massloss.f line 131), whichever rate applies: the constant DMDT0 or the Reimers rate CREIM$\,L/(gR)$. MDOT adds DELM = DMDT$\cdot\Delta t$ to the total mass, so a positive DMDT raises $M$ while the interior points keep their masses, and the envelope above the last point grows; main.f's comment reads "restore prior fitting point if mass accretion is being included." The defaults of both rates are negative (DMDT0 = $-10^{-14}$, CREIM = $-4\times10^{-13}$, parmin.f), so with them the routine is not called; the header's "mass is being decreased" disagrees with this trigger.

• Who calls it: only the main program (main.f), when LNEWFIT is set (above); it then calls FPFT with its own OMEGA.
• Inputs: SENVOLD (the desired fitting point, below); the model arrays (HCOMP, HD, HL, HP, HR, HS, HT, LC, M), the surface BL ($\log L/L_\odot$) and TEFFL, the rotation arrays, and CMIXL, the mixing-length parameter, passed on to ENVINT.
• What it calls: ENVINT (the envelope integration) and, for rotating models, GETROT.

• Outputs: the model with M increased and the new points filled; HS1, HS2 (unlogged masses and shell masses) for the new points; for rotating models, HJM, HI, ETA2, QIW, R0, HKEROT (not OMEGA, which is local: suspected bug 1) and the totals SJTOT, SKEROT; LNEW = true; SENV in COMP_MOD. It prints the new points to standard output.

• Modules: ENVSTRUCT_MOD (the envelope solution ENVP, ENVT, ENVS, ENVD, ENVR, ENVX, ENVZ, LCENV, NUMENV, written by ENVINT); NMLPAR_MOD (the envelope step sizes ENVMAX, ENVMIN, ENVBEG, temporarily set to HPTTOL(8) and restored; the starspot parameters SPOTF, SPOTX; LROT, WALPCZ; LEXCOM); COMP_MOD (STOTAL = $\log M$, SENV); DEBHU_MOD (the Debye–Hückel composition, set here from the surface shell); CONSTANTS_MOD; DIMENSIONS_MOD (JSON).

1. Surface quantities. From $L = 4\pi R^2\sigma T_\mathrm{eff}^4$ and $g = GM/R^2$: $$\log R = \tfrac12\left(\log\frac{L}{L_\odot} + \log L_\odot - 4\log T_\mathrm{eff} - \log 4\pi - \log\sigma\right), \qquad \log g = \log G + \log M - 2\log R.$$ With spots (fraction $f$ = SPOTF, temperature ratio $x$ = SPOTX) and a convective surface, the ambient temperature is $\log T_\mathrm{amb} = \log T_\mathrm{eff} - \tfrac14\log(f x^4 + 1 - f)$.

2. Envelope integration from the surface down to the current fitting point with ENVINT, at the spacing HPTTOL(8), without rotational distortion ($f_P = f_T = 1$).

3. Matching. The envelope's first point is at the mass of the last interior point; the differences there in $\log P$, $\log\rho$, $\log T$, $\log r$ are added to all envelope points, which shift down by one (the convective flag LCENV with them), and points closer than $10^{-10}$ in $\log m$ are dropped.

4. New points. Each envelope point below the new fitting point becomes an interior point, with $L$ held at its value at the old fitting point and the abundances other than $X$ and $Z$ copied from it (lines 184–201↗). The copy matches the envelope: ENVINT integrates it with the surface $X$ and $Z$ of shell M at every point (ENVX, ENVZ: lines 112–113↗, envint.f, lines 633–634↗, envint.f, lines 760–761↗), so the whole envelope has the old fitting point's composition (settled from the code, Nov 9, 2026). The last point is interpolated linearly in $\log m$ to the new fitting point.

5. Shell masses. $m = 10^{\log m}$, $\Delta m_j = \tfrac12(m_{j+1} - m_{j-1})$, and the last shell takes the rest of the star.

6. Rotation. The new points get, as a first guess, constant $j$ if $\alpha$ = WALPCZ $\le -2$, solid body if $\alpha \ge 0$, and otherwise $\Omega \propto r^{\alpha}$, the law WCZIMP imposes (PARMIN limits $\alpha$ to $[-2, 0]$); with $I = \tfrac23 r^2$ and $j = \Omega I$. GETROT then finds the rotation law, and the totals $J = \sum j\,\Delta m$ and $E_\mathrm{rot} = \sum\tfrac12\Omega j\,\Delta m$ are recomputed.

SENV is the fitting point's $\log(m/M)$, as can be read from HS = ENVS + STOTAL; it is limited to at most $-10^{-12}$.