Color transformations from DES, SDSS, PanSTARRS to LSST

I’m looking for an easy way to convert between DES, PanSTARRS and SDSS magnitudes and colors, and LSST values. Here’s a sample of what I’m looking for from DES, where it converts between PanSTARRS and SDSS values:

SDSS
Here, we matched stars from DES DR2 to stars in the Stripe 82 area of SDSS DR13, using the WAVG_PSF_MAG’s from DES and the PSF_MAG’s from SDSS. To reduce the effects of photon noise, we only considered stars with rms magnitude errors of ≤ 0.01 in DES and ≤ 0.02 in SDSS. For outlier rejection, we made use of iterated sigma-clipping, iterating over the fit 3 times and removing 3 σ outliers after each iteration. We found that the following to fall within reasonable bounds of accuracy, simplicity, and wide color coverage:

g{DES} = g{SDSS} - 0.061 (g - i){SDSS} + 0.008 ~(RMS: 0.017 mag) QA plot
r{DES} = r{SDSS} - 0.155 (r - i){SDSS} - 0.007 ~(RMS: 0.014 mag) QA plot
i{DES} = i{SDSS} - 0.166 (r - i){SDSS} + 0.032 ~(RMS: 0.018 mag) QA plot
z{DES} = z{SDSS} - 0.056 (r - i){SDSS} + 0.027 ~(RMS: 0.018 mag) QA plot

and

g{SDSS} = g{DES} + 0.060 (g - i){DES} - 0.005 ~(RMS: 0.018 mag) QA plot
r{SDSS} = r{DES} + 0.150 (r - i){DES} + 0.014 ~(RMS: 0.016 mag) QA plot
i{SDSS} = i{DES} + 0.167 (r - i){DES} - 0.027 ~(RMS: 0.015 mag) QA plot
z{SDSS} = z{DES} + 0.054 (r - i){DES} - 0.024 ~(RMS: 0.018 mag) QA plot

First-order polynomials based on a single color index (g-i for g and r-i for r,i,z). These equations are valid for stars with roughly -1.0 < g-i < 3.5 (g) and roughly -0.4 < r-i < 2.0 (r,i,z). (Note: the RMS listed after each transformation is RMS per star. The mean RMS for a collection of stars being transformed from one photometric system to the other – especially if that collection of stars covers a range of colors – should be correspondingly smaller.)
Pan-STARRS
As for the DES/SDSS transformation equations, here, we matched stars from DES DR2 to stars in the Stripe 82 area of PanSTARRS1 DR2, using the WAVG_PSF_MAG’s from DES and the MeanPSFMag’s from PanSTARRS1. Again, to reduce the effects of photon noise, we only considered stars with rms magnitude errors of ≤ 0.01 in DES and ≤ 0.02 in PanSTARRS1. Using the same type of outlier rejection and aiming for similar goals in accuracy, simplicity, and color coverage, we arrived at the following equations:

g{DES} = g{PS1} + 0.028 (g - i){PS1} + 0.020 ~(RMS: 0.017 mag) QA plot
r{DES} = r{PS1} - 0.142 (r - i){PS1} - 0.010 ~(RMS: 0.013 mag) QA plot
i{DES} = i{PS1} - 0.155 (r - i){PS1} + 0.015 ~(RMS: 0.012 mag) QA plot
z{DES} = z{PS1} - 0.114 (r - i){PS1} - 0.010 ~(RMS: 0.015 mag) QA plot
Y{DES} = y{PS1} - 0.031 (r - i){PS1} + 0.035 ~(RMS: 0.017 mag) QA plot

and

g{PS1} = g{DES} - 0.026 (g - i){DES} - 0.020 ~(RMS: 0.017 mag) QA plot
r{PS1} = r{DES} + 0.139 (r - i){DES} + 0.014 ~(RMS: 0.015 mag) QA plot
i{PS1} = i{DES} + 0.153 (r - i){DES} - 0.011 ~(RMS: 0.010 mag) QA plot
z{PS1} = z{DES} + 0.112 (r - i){DES} + 0.013 ~(RMS: 0.015 mag) QA plot
Y{PS1} = y{DES} + 0.031 (r - i){DES} - 0.034 ~(RMS: 0.017 mag) QA plot

– which are valid for stars with roughly -0.9 < g-i < 3.8 (g) and -0.4 < r-i < 2.4 (r,i,z,Y).

Dear Bob,
Yes, we have a couple of Rubin in-kind contributors who are working on just this – simple transformation relations that a typical astronomer can easily use. Since no on-sky data are available yet for LSSTCam, they are using synthetic photometry of the Pickles (1998) Stellar Atlas Spectral Energy Distributions (SEDs) multiplied with the current baseline for the LSSTcam filter bandpasses. These are matched with similar synthetic photometry for the Pickles atlas for other photometric systems (e.g., DES, PanSTARRS, Johnson-Cousins, Roman, EUCLID, etc.). Some initial results can be found in this iPoster from the the January 2025 American Astronomical Society meeting: iPoster.

It is still early, though, and only the LSST vs. DES has been done so far.

I will try to keep you updated!

Thanks!

Best regards,
Douglas

2 Likes

Thank you, Douglas, that’s helpful. I’ll keep reading, and that poster looks interesting!