CKS-Gaia Catalogue

[bjfultn]

Hello Eric and/or Danley,

Would it be much trouble for you to create a jld planet catalogue using the new CKS-Gaia parameters from Fulton & Petigura (2018)?

Ideally, I'd like a planet catalogue filtered as we do in Fulton & Petigura (2018), or reasonably close (e.g. we don't necessarily need to merge in the Furlan catalogue) and a stellar catalogue to match the filters that apply (Teff cuts at 4700 and 6250 K is probably good enough). I'm set up to run the ABC calculations with the provided jld catalogues, but I've spent some time trying to figure out how to make my own catalogues. I'm new to Julia, so the learning curve is steep. I'm envisioning this being something that would take you guys a few min, but me >= a day or so. If it is significantly more work for you then please do say no and I'll continue to work on figuring it out.

Thank you, BJ

[bjfultn]

BTW, I forgot to mention that I'll provide the CKS-Gaia planet catalogue if you agree.

[dch216]

Hi BJ,

We'd happy to create catalogues for you. If you could produce a full list of the planet KOIs and a full list of the stellar target KepIDs, that would be immensely useful. I can follow most of the cuts you list in your paper, but I'd prefer to have an exact list so we can just match on IDs than to try to reproduce the filters.

Regards, Danley

[eford]

Hi BJ,

Glad to hear you're interested in applying our code. I agree with Danley that the easiest way to make sure that we're making the exact JLD that you want is for you to send us CSV files with Kepler IDs that you want included. That way there's no risk that our implementation of your cuts doesn't quite match yours.

I also wanted to mention that Danley and I just happened to be merging in several changes into the master branch today. Of course, we think all of those are improvements, but there's always a chance there's some unintended consequence. So if you're happy with what was being done for the old code, then you might want to either work from Danley's hsu_et_al_2018 branch or pin your local repositories of our packages at the master branch as of a day or two ago.

Danley's planning a follow-up paper that should describe the various improvements. However, we still have a few more feature upgrades that we're hoping to fold in from Darin and Keir before Danley's next paper. Realistically, I'd guess that we're at least a month away from having those in place and well tested (and knowing how research usually goes it could easily be longer). Hopefully, once you figure out how to use ABC and ExoplanetsSysSim, it'll be easy to just pull the most recent version and rerun things once we've merged all the upgrades into master branch. I'll try to remember to drop you a line once we think it's ready. But don't feel bad about asking for an update every month or so, as I could easily forget!

Cheers, Eric

On Wed, Aug 1, 2018 at 7:55 PM dch216 notifications@github.com wrote:

Hi BJ,

We'd happy to create catalogues for you. If you could produce a full list of the planet KOIs and a full list of the stellar target KepIDs, that would be immensely useful. I can follow most of the cuts you list in your paper, but I'd prefer to have an exact list so we can just match on IDs than to try to reproduce the filters.

Regards, Danley

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/eford/ExoplanetsSysSim.jl/issues/17#issuecomment-409761929, or mute the thread https://github.com/notifications/unsubscribe-auth/ABQywvB8Cu6ZCF2Y0XHqYXX0MO-gVVhiks5uMj_ggaJpZM4Vpe-q .

-- Eric Ford Professor of Astronomy & Astrophysics Center for Exoplanets & Habitable Worlds Center for Astrostatistics Institute for CyberScience Penn State Astrobiology Research Center Pennsylvania State University

[bjfultn]

Thanks so much guys!

Here are links to the catalogues:

https://www.dropbox.com/s/6pt00bd80miztxx/cksgaia-stars.csv?dl=0 https://www.dropbox.com/s/hbc8zvc2pwttr0j/cksgaia-planets.csv?dl=0

and a column definition file: column-definitions.txt

You'll want to use the gdir_prad, and gdir_srad columns for the planetary and stellar radii in the planet catalogue, and the gaia2_srad for the stellar radii in the parent stellar sample. Otherwise the KIC is found in id_kic and the KOI number is in in id_koi. Let me know if you have any questions about these files.

@eford I'll be sure to work on the hsu_et_al branch for now, assuming there aren't operational differences between that branch and master and I'll keep a lookout for the paper on the future upgrades! On the timescale of a few months I'd like to understand this well enough to apply/adapt it to RV surveys.

[dch216]

@bjfultn I've noticed that gdir_srad and gaia2_srad are different values for the same star. Since SysSim only uses one set of stellar parameters in its calculations, I'll need to know which of the two columns should be used. If you want gaia2_srad, then I can work with what's given. In contrast if you want gdir_srad, then I'll need the values for all the stars in the parent stellar sample (since the planet catalogue only contains a subset).

Also, are you fine with DR25 values for stellar masses and densities?

As an extra note, in SysSim the planetary radius is calculated from the stellar radius and the observed planet-star radius ratio, so the planetary radius column doesn't get referenced.

[bjfultn]

gaia2_srad will work for both. We get gdir_srad using the parallax from Gaia + Teff from spectroscopy, so this isn't feasible for the parent sample (although it will be when we have Gaia Teff's).

D25 values are good for anything else you need.

This issue of calculating the planetary radii within the ABC model may have had a part in tripping me up when I was trying to make these catalogues. I couldn't find where the planetary radii were being used in the catalogue generation. We can use the DR25 radii ratios and the giaia2_srad column for the stellar radii. This won't be a perfect comparison to what I'm doing, but I think its close enough to get an idea of the biases incurred on the relative occurrence rates with my IDEM method in the regions we care about.

[dch216]

So I slightly misspoke regarding how planetary radii are calculated. We take the sqrt of the observed transit depth and multiply it by the stellar radius to get the planetary radius. If you'd like to look over the code where this is done, it's in the calculation of the summary statistics of the observed catalogs when planets get sorted into the appropriate period-radius bin:

https://github.com/dch216/ExoplanetsSysSim.jl/blob/hsu_etal_2018-v1.0/examples/hsu_etal_2018/christiansen_func.jl (Line 357)

I'll get onto creating the stellar jld catalog for you. The planet catalog csv actually works fine as an input to SysSim if one of the KIC columns gets its header name changed to kepid (SysSim uses this column to join the stellar and planet catalogs on target star). Let me know if you'd like me to do this for you or if you are fine just creating a copy of the csv with one header column name changed on your end.

[bjfultn]

OK, great! I can make that change to the planet catalogue on my end.

[dch216]

A few more changes you need to make to the planet catalogue:

1) The column id_koicand needs to be renamed kepoi_name

2) Either add an entire new column named koi_pdisposition and set all values to the string "CANDIDATE" or edit the SysSim code to not use the is_cand check here: https://github.com/dch216/ExoplanetsSysSim.jl/blob/hsu_etal_2018-v1.0/src/kepler_catalog.jl (Lines 149 & 153)

Hopefully that should be it in terms of changes to the planet catalogue csv to make it work with SysSim. Sorry for not catching these in the previous response!

[dch216]

Alright, here's the stellar catalog jld file for the CKS-Gaia catalogue. As a heads-up, there are 89 target stars missing from the original 24981 stars because the DR25 catalogue (that we are using) doesn't have CDPP values for those stars.

https://psu.box.com/s/53n5klhf5g2ey4woslx84ugmq3mxdnfd

I also realized that there may be a potential issue with the stellar catalogue I've created for you. As of a few months ago, the DataFrames package of Julia updated how it handles N/A values in csv files (i.e. they now implement a Missings type). The process of creating the JLD file should have excluded all N/A values so you won't run into compatibility issues, but let me know if you run into any warning/error messages regarding DataFrames.

[bjfultn]

Awesome! Thanks again Danley. I'll let you know if I run into any issues while trying to run it.

I'll close this issue for now, and re-open if I run into problems.