I've downloaded the 2MASS Photometric Redshift Survey data:

• 934,175 galaxies with a median redshift of z=0.07; photo-z's are estimated from the ANN
• Contains the following information (also listed here: http://ssa.roe.ac.uk//www/ssa_browser.html)
  • 2MASS_ID/WISEID/SCOSID
  • ra/dec/l/b
  • ebv, starDensity (log(density) of stars with K<14 at this position
  • j/h/k band mags and errors
  • w1/w2/b/r mags and errors
  • zSpec (if it exists) and zPhoto (from ANNz)

Main Takeaways so far

• There appears to be quite little overlap between the 2MPZ and WISExSCOS catalogs. At least so far, I've identified ~6000 overlapping galaxies between the two catalogs. They are preferentially at the bright end of WISExSCOS and the faint end of 2MPZ, which tracks with what the papers described (2MPZ targets brighter, nearby objects; WISExCOSMOS is for fainter, deeper galaxies).
• The 2MPZ catalog includes wise W1 & W2 AND the 2MASS J/H/K photometry, so there's a nice opportunity to compare the masses derived from Emily's estimator vs. Cluver+14's estimator for a large sample of galaxies. One complication is that Emily's K-band magntiudes are the k_m_ext (total, extinction corrrected magnitudes), whereas the J/H/K reported in the 2MPZ catalog are the "k_msig_k20fe" magntiudes (elliptical aperture at K=20mag/arcsec sq isophote), so they are slightly different (with Emily's likely being brighter in general). With that said, the Cluver+14 masses and Emily's masses are in VERY nice agreement. I expect that if I track down the k_m_ext magntiudes, the galaxies will be brighter --> more massive --> we'd get a similar ~0.6 dex offset between the mass estimates.
• GSMF/BHMF Takeaways:
  • The GSMFs with the combined catalog are in much better agreement (especially the low redshift bins). I'm still underestimating the GSMF that Emily quotes in her paper, but I'm quite happy with the general shape.
  • The low-redshift cutoff point only seems to affect the smallest galaxies in the GSMF, and has virtually no impact on the resulting BHMFs.
  • The shapes of the BHFs are quite nice, too, but I'm still underestimating the BHMF from Emily's paper, too.
• GSMF/BHMF with Mass Offset:
  • If we assume that the total masses are ~0.7 dex larger that our Cluver+14 estimates, we are nearly at the point of reproducing the GSMF and BHMF! We need to probably do a mass dependent correction which might improve things further, but we're definitely trending in the right direction.

• 5,923 sources in common between 2MPZ (~1 million total galaxies) and WISExSUPERCOSMOS (~18.5 million total galaxies). It's noteworthy that the overlapping sample is entirely confined to the brightest sources, which makes sense given the cuts to each catalog.
• I think it makes most sense to remove the overlapping sources not to count them twice.
• Here's a comparison of the W1/W2/B/R magntiudes for the different samples and their overlap, as well as plots of their redshift coverage:

• The Cluver+14 calculation is the same as before.

• Interestingly, the agreement between these are really great, in particular between log(M) = 9 to log(M) = 12 or so. The best fitting line is quite close to 1-to-1 as well, given that we have a long tail of stellar masses well below log(M) = 8.

• For simplicity, I'm using the photo-z's from the 2MPZ catalog (even though about 1/3 of those sources have spec-z's available). I can definitely swap out the proper redshifts, but for consisency/not over complicating things at this stage, I'm sticking with the photo-z's only.
• Throwing out the ~5900 overlapping wiseIDs from the 2MPZ catalog (sticking with the WISExSCOS data for those overlapping galaxies). Note that there are some more things to investigate for the overlapping galaxies (see above).
• I'm currently clipping the galaxy colors at W1-W2=[-0.2,0.6] following the recommendations from sevearl of the papers (Jarrett+23 and WISE2MBH, namely). But it's not clear we want to actually do this since we're using the Cluver+14 prescription (which DOESN'T suggest doing this).
• I'm not sure that the K-band magntiudes quoted in the 2MPZ catalog are the same K-band magnitudes that are used for the MASSIVE comparison we had before. Specifically, it seems like the 2MASS K-band magntiudes are the "k_msig_k20fe" magnitudes, not the "k_m_ext". I think in theory I could cross-match to get the correct magntiudes, but need to verify this a bit more closely.

1. From the 2MPZ catalog, grab the wiseIDs, W1, W2, and ZPHOTO measurements.
2. From WISExSCOS, grab those same columns.
3. Cross-match 2MPZ and WISExSCOS on wiseID, throwing out the rows from the 2MPZ data and retaining only the WISExSCOS data.
   • The result here is roughly ~19.5 million galaxies in my data.
4. Compute the stellar mass estimate using the Cluver+14 relation with the Jarrett+23 k-corrections:
   • This K-corrects the W1 flux and W1-W2 colors
   • We also clip the W1-W2 colors to be between [-0.2,0.6]
   • Distances are computed from photo-z's, assuming H0=70 km/s/Mpc and Omega_m=0.3 to go from apparent to absolute magnitudes
5. For computing the GSMF:
   • Compute maximum luminosity distance each galaxy can be placed at and still be observed, assuming a limiting W1 magntiude of W1=17.0
   • Convert this luminosity distance to a redshift, again for every galaxy. This is then the maximum redshift we could put this galaxy and still observe it.
   • We then find the volume to this maximum redshift, and we call it Vmax. Technically we multiply this volume by 68% since this is the fraction of the sky covered by our survey (roughly).
   • In each mass bin for the GSMF, we add up 1/Vmax for all galaxies in that bin, and divide by the bin width. That is to say, each galaxy contributes 1/Vmax of a galaxy to their mass bin (rather than just 1).
6. For computing the BHMF:
   • Once I computed the stellar masses from Step 4 above, I "paint-in" black holes by drawing a black hole mass for each galaxy given by the relation in MM13. Specifically I draw from a Gaussian distribution centered at log(Mbh) = alpha + beta*(M*-11) [alpha and beta given in MM13], with a width of sigma=0.34 dex.
   • The BHMF is then computed in the same manner as the GSMF, but I consider the black hole masses rather than the stellar masses.

• We might want to consider using the k_m_ext magntiudes for consistency with Emily, but just wanted to bring this up here. Here's a histogram of the differences:

Hoping to address some of the questions Chung-Pei had this morning about the K-band magnitudes. As a quick reminder, here are the various datasets I'm playing around with:

• 2MASS XSC
• 2MPZ
• WISExSCOS

I started by matching the WISE XSC with the 2MPZ catalog so we can compare the resulting k_m_ext vs. k_m_k20fe for the 2MPZ galaxies alone:

I then found the galaxies which are in both 2MPZ and WISExSCOS. There are 5923 galaxies in common between the two catalogs.

• Note that both catalogs quote extinction corrected magntiudes, but the two catalogs use DIFFERENT EXTINCTION CORRECTIONS! So we shouldn't expect an exact replication of the magntiudes, but they're quite close. More specifically, here is what each catalog says about the extinction correction:
  • 2MPZ: All of the magnitudes were corrected for Galactic extinction using the Schlegel et al. 1998, ApJ, 500, 525 dust maps. In the case of SuperCOSMOS, the A_λ/E(B-V) coefficients were hemisphere-dependent: ANB=4.165 and ANR=2.773 in the North (δ1950>2.5), ASB=4.011 and ANR=2.778 in the South (δ1950<2.5), and AI=2.016 for the entire sky. 2MASS J/H/K_m_k20fe magnitudes use A_λ/E(B-V) from Cardelli et al. (1989), whereas for WISE w?mpro magnitudes the coefficients are AW1/EBV=0.231 and AW2/EBV=0.194 (Flaherty et al. 2007).
  • WISExSUPERCOSMOS: 2) All of the magnitudes were corrected for Galactic extinction using the Schlegel et al. 1998, ApJ, 500, 525 dust maps, with revised coefficients derived from Schlafly & Finkbeiner 2011. In the case of SuperCOSMOS, the Aλ/E(B-V) coefficients were AB/EBV=3.44 and AR/EBV=2.23, whereas for WISE w?mpro magnitudes they were AW1/EBV=0.169 and AW2/EBV=0.130 (Indebetouw et al. 2005). Note that these coefficients were substantially revised with respect to those used for 2MPZ.

With that caveat, here's a comparison of the W1 and W2 measurements in the two catalogs, which appear to be in excellent agreement.

• Start with the 2MPZ catalog, which contains W1/W2 magnitudes and the K-band (though it has the wrong K-band to start).
• Match the XSC to the 2MPZ catalog to swap out the K-band reported in 2MPZ (k_m_k20fe) with the K-band reported in the XSC (k_m_ext).
• The result is that I have the original 2MPZ W1/W2 measurements with the correct K-band (k_m_ext) taken from the XSC.
• I then compute the stellar masses in the same way:
  • For the W1/W2, I use the Cluver+14 prescription (using photo-z's to get D_L for the absolute magntiudes).
  • For the K-band, I use the photo-z's given in 2MPZ to get a D_L, convert the apparent K-band to absolulte K-band, and then use Emily's relation.

• Technically the 2MPZ catalog has a different extinction correction, so we might have to be careful when using the W1/W2 from 2MPZ vs. W1/W2 from WISExSCOS, as noted above, instead of just combining the arrays and computing. Need to check but this extinction might affect the K-band measurements, too. Just need to verify.