Check SEDs

I have assigned SEDs to all the galaxies, as outlined in my previous post.

In this post I am doing some checks to see if these SEDs are reasonable.

Distribution of Free Parameters

First, I simply plot the distribution of the different parameters used to calculate the SEDs. These parameters are age $a$, star-formation time $\tau$, the SFR $\psi$, metallicity $Z$, dust attenuation $\bar{\tau}_V$ and gas ionization parameter $U_S$.

These values make sense with what was imposed.

Star Formation Rates

The SFRs were imposed from relations in Shreiber et al. 2017.

These look like they were assigned correctly

$Z$-$U_S$ Relation

This was imposed (Eq 18 in W18).

This looks pretty good. It is a bit above the relation for high Z, I’m guessing because of the truncation I placed in the distribution (if I didn’t place a truncation, I would occasionlly end up with galaxies with incredibly unrealistic $Z$ and/or $U_S$ values).

UV properties

Unlike W18 I did NOT impose these relations. Therefore I compare the recovered $M$–$M_{\rm UV}$ and $M_{\rm UV}$–$\beta$ relations (as calculated here) to those used in W18 (dotted lines). Note that this only includes SF galaxies.

The $M$–$M_{\rm UV}$ looks fairly similar to the relation from W18 and has the proper dependance on redshift, though our objects seem to be slightly brighter, and the slopes slightly steeper.

The $M_{\rm UV}$–$\beta$ relation does not behave as expected. While the $\beta$ values look reasonable, the trend with $M_{\rm UV}$ is backwards. I am hoping that this is a problem with how I am calculating $\beta$—I plan to double check this later.

UVJ diagram

I also got FSPS to return the magnitudes in the Johnson U and V bands (calculated from the SEDs). This is so I can plot the galaxies in a UVJ diagram, and check that the SF and Q classifications are consistent with the colours.

Taking galaxies with redshifts $z<1$, this is what the UVJ diagram looks like

This doesn’t look great, but I think it is mostly because I didn’t take the restframe U,V and J values. When I restrict it to galaxies with $z<0.2$ it is a bit better:

However, there do seem to be more quiescent galaxies in the lower left corner than expected.

I should actually get the restframe filters instead to check this.

Summary

This mostly looks reasonable, the only worrisome thing is the $M_{\rm UV}$–$\beta$ relation. I need to try and make sense of this, and figure out how to make the results look more reasonable.

I also want to plot observational data on top of these trends. Other things I can check against data are the star formation rate density (e.g. Fig 18 in W18), and the sSFR (Fig 19 in W18).

---

« SED Methods

Mock Catalog Detection Predictions »