apogee-maps

This repository contains the code for the papers "The stellar population structure of the Galactic disk" Bovy, Rix, Schlafly et al. (2015) and "On Galactic density modeling in the presence of dust extinction" Bovy, Rix, Green et al. (2015). This code is provided for illustration purposes only, but it should allow you to perform the same analysis as in these papers if you want. Much of this code runs for quite a long time and no guidance is given here as to how long each command takes (none of it takes ridiculously long though).

The code for "The age-metallicity structure of the Milky Way disk" Mackereth, Bovy, Schiavon et al. (2017), is also provided for completeness, as this paper made very similar analysis, and used an augmented version of the code for the above papers. The same disclaimers in terms of run-time should be adopted for this code, which is also provided for illustration only.

Contents

• Prerequisites
• The stellar population structure of the Galactic disk
  • Data
  • Fits to broad abundance-selected subsamples
  • Fits to mono-abundance populations (MAPs)
• On Galactic density modeling in the presence of dust extinction
• The age-metallicity structure of the Milky Way disk

Prerequisites

This code heavily depends on the apogee and mwdust general-use packages. You need to install these (including downloading all of the necessary dust data) first. This code also makes use of the numpy, scipy, matplotlib stack. Parts of the code require healpy. The code will complain if you do not have other required packages.

The stellar population structure of the Galactic disk

This code is mainly concerned with fitting spatial-density profiles to subsets of APOGEE data. Almost all of the commands below require the effective-selection function (see the papers above for a description of this), which you should first calculate as:

python calc_effsel.py ../essf/essf_green15.sav --dmap=green15 -m 32
python calc_effsel.py ../essf/essf_marshall06.sav --dmap=marshall06 -m 32
python calc_effsel.py ../essf/essf_sale14.sav --dmap=sale14 -m 32
python calc_effsel.py ../essf/essf_drimmel03.sav --dmap=drimmel03 -m 32
python calc_effsel.py ../essf/essf_zero.sav --dmap=zero -m 32

The first two of these are the most important, as they are used in most parts of the code. The option -m 32 calculates this in parallel (for different lines of sight) on 32 CPUs. There is some overhead in first computing and saving the overall APOGEE selection function (see apogee.select.apogeeSelect), so you want to run these in series (at least the first two).

The heart of the code are the densprofiles.py and fitDens.py files. These contain all of the different density profiles and the code to fit and MCMC-explore a given density profile to a subset of APOGEE data.

The notebook SpatialFitExample has general explorations of the fit provided by different types of density profiles. It formed the backbone of the exploratory phase of this project.

The following sections list the commands that generate the figures in the paper.

Data

Figure 1:

python plot_dustwapogee.py ./dustwapogee.ps

Figure 2:

python plot_vs_jkz.py -o ./rcmodel_mode_jkz_h_parsec.ps rcmodel_mode_jkz_h_parsec_newlogg.sav -b H -t mode --parsec

(this code is in jobovy/apogee-rc as it was part of an earlier paper).

Figure 3:

python plot_rcdistancecomparison.py ./rcdistcomp.ps

Figure 4:

python plot_afe_spectra.py afe_feh-0.3_residuals.sav ./spectra_ELEM.ps

Figure 5:

python plot_afefeh.py ./overall_afefeh.ps

Figure 6:

python plot_spatial_broad.py ./spatial_broad_SAMPLE.ps

Checking the abundance scatter of open clusters:

python check-abundanceScatter-clusters.py

Fits to broad abundance-selected subsamples

Figure 7:

python fitBroadSubsamples.py lowlow ../broadfits/lowlow.sav ./lowlow.txt ./lowlow_LOC.ps
python fitBroadSubsamples.py solar ../broadfits/solar.sav ./solar.txt ./solar_LOC.ps
python fitBroadSubsamples.py highfeh ../broadfits/highfeh.sav ./highfeh.txt ./highfeh_LOC.ps

Figure 8:

python fitBroadSubsamples.py highalpha ../broadfits/highalpha.sav ./highalpha.txt ./highalpha_LOC.ps

Figure 9:

python plot_broadsurfdens.py ./broadfits-radial.ps

Fits to mono-abundance populations (MAPs)

Fits are done as follows:

python fitMAPs.py tribrokenexp ../mapfits/tribrokenexp.sav
python fitMAPs.py tribrokenexpfixedflare ../mapfits/tribrokenexpfixedflare.sav
python fitMAPs.py tribrokenexpflare ../mapfits/tribrokenexpflare.sav
python fitMAPs.py tribrokenexpinvlinflare ../mapfits/tribrokenexpinvlinflare.sav
python fitMAPs.py tribrokenexplinflare ../mapfits/tribrokenexplinflare.sav
python fitMAPs.py tribrokentwoexp ../mapfits/tribrokentwoexp.sav

that is, fitMAPs.py just fits a given density profile to all MAPs. It also runs MCMC for each MAP and saves all of these results (in a rather large file!).

Figure 10:

python plot_maprmax.py ../mapfits/tribrokenexpflare.sav ./mapfits-rpeak.png

Figure 11:

python plot_mapsurfdens.py ./mapfits-radial.ps
python plot_mapsurfdens_highalpha.py ./mapfits-radial-highalpha.ps

The resulting output profiles for the low- and high-alpha MAPs displayed in this figure can be found under out/mapsurfdens.csv and out/mapsurfdens_highalpha.csv.

Figure 12:

python plot_mapflarepdf.py flare_lowalpha.sav ./mapfits-flare-lowalpha.ps
python plot_mapflarepdf.py flare_highalpha.sav ./mapfits-flare-highalpha.ps

Figure 13:

python plot_mapflare.py ./mapfits-radialflare.ps
python plot_mapflare_highalpha.py ./mapfits-radialflare-highalpha.ps

The resulting output profiles for the low- and high-alpha MAPs displayed in this figure can be found under out/mapflare.csv and out/mapflare_highalpha.csv.

Figure 14:

python plot_maphz.py ./mapfits-hz.png

Figure 15:

python plot_maptwohz.py ./mapfits-twohz.ps

On Galactic density modeling in the presence of dust extinction

Many fewer figures in this paper (phew!), but here we go. Figure 1:

python plot_dust_gaia.py 5.0 ./dust_5.0kpc.ps

Figure 2:

python plot_gaia_rcmag.py ./gaia_mg.ps

(see gaia-rc.py for some code to get the RC's properties in the Gaia passband). Figure 3:

python plot_powspec.py 5.0 ../savs/PowspecDensAndDustAndESSF ./powspec_dens_dust_essf_D5.0.ps
python plot_powspec.py 6.3 ../savs/PowspecDensAndDustAndESSF ./powspec_dens_dust_essf_D6.3.ps

Figure 4:

python plot_distanceintegral.py ../savs/distInt.sav /dev/null

(and similar for subsets of the sky, see options in plot_distanceintegral.py).

Figure 5:

python plot_ah_location.py 4240 ./ah_4240.png

and similar for other locations (like 4240). Figure 6:

python plot_effsel_location.py 4240 ./effsel_4240.ps

also similar for other locations.

The age-metallicity structure of the Milky Way disk

The following section concerns the adaptations made to the code above for the Mackereth et al. (2017) paper. Most of the additions were made in order to allow the use of the code with the full APOGEE red giant branch (RGB) sample, which has a larger sample with measured ages, and for which the selection function is more easily applied to the stellar evolution models (for calculating the surface-mass density contributions of populations).

There are quite a few extra requirements necessary to reproduce the results which make up this paper, the main ones being the extra data tables (with DR12 RGB distances, and the ages from Martig et al. 2016), and the PARSEC isochrones - implemented via the isodist python package.

To get hold of the DR12 distances, which have not yet been made publicly available, you need to have access to the SDSS SAS (SDSS collaborators only) and the file at this link. You should place this file in ../catalogues . If you dont have access to that catalogue, you could download the (publicly available) DR14 distance VAC here, which can be cross-matched with the DR12 catalogue to get distance estimates. This would, however, require some tweaks to the existing code where the distance catalogue is used.

The age catalogues can be obtained from vizier via ftp by running:

python get_agetables.py ../catalogues

which will download and convert the tables into the required format.

You will then need to run the code which extracts and calculates the weights for the PARSEC isochrones (first installing the isodist package), by simply running:

python make_isochrone_grids.py

This can take quite some time due to the large number of nodes in the grid...

Yet another prerequisite for these results is to re-calculate the effective selection function for the RGB sample. This requires a sampling of the $M_{mathrm{H}}$ distribution, which is implemented using the isochrone grid(s) calculated above.

This is run by calling:

python calc_effsel_monoage.py --dmap=marshall06
python calc_effsel_monoage.py --dmap=green15

These scripts will also calculate the raw APOGEE selection function if this file does not exist (i.e. if you havent run the code for the previous papers). These can take quite some time to run depending on your system. You have now calculated all the required files to start performing the calculations which make up the bulk of results in the paper using these scripts:

python fit_monoage.py
python mass_script.py

These will perform the density fits to the mono-age mono-$mathrm{[Fe/H]}$ populations (including MCMC explorations), and then calculate their surface-mass density contributions (again using the precalculated isochrone grids). Again, these scripts are pretty time consuming depending on your system, but not prohibitively so. Results are saved into files in the ../out folder.

The plots for the paper can then be produced by running the code in the apogee-monoage iPy notebook.