Euclid

We have investigate one of the first steps necessary to analyse the incredible galaxy catalogs will be available with Euclid: the selection of quiescent and star-forming galaxies using observed colors. 

To perform this analysis we derived three different mock catalogs i) the first interpolates the multi-wavelength observations of galaxies in the COSMOS2015 catalog; ii) the second uses the best theoretical template describing the multi-wavelength observations of galaxies in the COSMOS2015 catalog; iii) the third takes galaxy parameters from the Euclid Flagship mock galaxy catalog.

The main results are:

• By selecting galaxy types in the commonly accepted UVJ plane derived from these mock observations, we only recover ∼ 20% of the original quiescent galaxy population up to redshifts z = 3.

• the most effective way to separate quiescent from star-forming galaxies with observed colors is the combination of (u-VIS) and (VIS -J) colors. The u-band is not among the filters on board of Euclid, but it will be available thanks to the Euclid-specific follow-up ancillary ground-based observations.

• Among the Euclid-only color combinations, the (VIS - Y ) and (J - H) colors are the most efficient for isolating quiescent galaxies. However, due to the absence of strong spectral features inside these filters at z < 0.75, quiescent and star-forming galaxies have similar colors.

If you are interested on knowing more about this work, you can have a look at this paper. 

Machine learning methods are increasingly becoming the most efficient tools to handle this enormous amount of data, because they are often faster and more accurate than traditional methods.

We investigate how well redshifts, stellar masses, and star-formation rates (SFR) can be measured with deep learning algorithms for observed galaxies within data mimicking the Euclid and Rubin/LSST surveys. We find that Deep Learning Neural Networks and Convolutional Neutral Networks (CNN), which are dependent on the parameter space of the training sample, perform well in measuring the properties of these galaxies and have a better accuracy than methods based on spectral energy distribution fitting. 

CNNs allow the processing of multi-band magnitudes together with H-band images. We find that the estimates of stellar masses improve with the use of an image, but those of redshift and SFR do not. Our best results are deriving 

• the redshift within a normalised error of less than 0.15 for 99.9% of the galaxies with S/N>3 in the H-band; 

• the stellar mass within a factor of two (~0.3  dex) for 99.5% of the considered galaxies;  

• the SFR within a factor of two (~0.3 dex) for ~70% of the sample. 

If you are interested on knowing more about this work, you can have a look at this paper.