nichollsh
commented
1 year ago Note that the code experiences some numerical issues with thick atmospheres when the instellation is updated. The solution to this is to set F_atm_bc=1 and implement an atmosphere model which solves for global radiative equilibrium. I did not include radiative time-stepping in this pull request because my toy model is not appropriate for the master branch.
timlichtenberg
Summary
This branch contains quite a few changes from the master branch, but I have confirmed that it runs PROTEUS fine. The main purpose of this branch was to include a method for running chemistry 'offline' from the main PROTEUS simulations; the script
tools/RunOfflineChemistry.pytakes the output from a PROTEUS run and provides it to VULCAN. This allows us to investigate the response of the upper atmosphere.Major features
Script to run VULCAN in offline-chemistry mode. This takes data from a completed PROTEUS run, and calculates the disequilibrium composition for several samples across the evolution of the planet.
Can now set a parameterised initial composition based on fO2, C/H, H2O inventory, and N2 ppm. This uses Dan's script for satisfying both equilibrium chemistry and melt-vapour partitioning functions. Can still manually choose the composition by disabling this in the config. Closes #42 .
Several more configuration options for existing features (e.g. to disable stellar heating, change tropopause type, choose spectral file).
Added skin-temperature as option for tropopause temperature.
Adaptive time-stepping for selecting new SPIDER step size. This can be turned off in favour of other time-stepping methods (e.g. const dt).
Introduction of 'eqm' loops, which run the SPIDER-AEOLUS iterations with dt=0 in order to adapt to changes in instellation. This is not effective for optically thick atmospheres and may be redundant soon.
Can limit the relative change in net upward flux each time-step, to prevent oscillations. This isn't physical in many circumstances.
Minor features
Can change the frequency of updating the instellation/spectrum, and for creating plots.
Limits on maximum and minimum temperature
Plotting scripts for visualising offline chemistry results and debugging the code.
Introduced 'input' directory containing stellar tracks, example cfg files, and stellar spectra.
Documentation updates, although much more is still needed. Includes changes to read-the-docs information.
Script for running PROTEUS using Screen. This is helpful for managing multiple simultaneous instances of PROTEUS.
For a given run 'x', files are stored in
output/x/. JSON and PKL data files are stored inoutput/x/data/.Changes to information printed to stdout. It was getting a bit unreadable with the new features, so I've just tidied this up.
Increased number of SPIDER levels from 100 to 150.
Split up some large py files into several smaller ones.
Different options for initialising offline chemistry. Can provide elemental abundances based on outgassed volatiles, or directly provide constant mixing ratios based on outgassed volatiles.
Calculated stellar spectra are passed to VULCAN during offline chemistry calculations.
Other scripts in
tools/.Can choose where the boundary condition
F_atmis calculated: either at the surface or at TOA. This will be necessary for future changes.More variables stored in the helpfile, which are needed for debugging the model.
Changes to which SPIDER variables are stored in
init_spider.optsand which are provided at runtime.New folder
examples/containing example outputs. Most notably, shows the result of running the current default configuration file.Fixes
Several bug fixes relating to stellar evolution calculations. These include scaling luminosity with the Baraffe model and also performing track interpolation.
Uses the new spectral file 'Mallard' by default, which can recover a runaway H2O greenhouse effect.
Stores updated stellar radius as the star ages. Previously was not being updated in the helpfile.
Much-improved method for choosing UV-PL boundary when using Mors evolution code. Previously it was fixed to be at the Balmer jump, but now is chosen (by bisection search) to be at a value that best reproduces the correct Lyman-alpha energy. This is helpful when running the model for various stellar classes.
Added Hamish Innes to contributors section.