Euclid Quick Data Release (Q1): First visual morphology catalogue

Image to be added after embargo lifts on 19th

Contents

Catalogue

List of ID columns and morphology columns

List of morphology questions and answers available

List of additional columns copied from MER (flux, ellipticity, area, etc.)

Images

Full documentation is available here. Below is a summary.

1. Catalogue

The morphology catalogue covers galaxies which are either bright or extended. Specifically, it includes galaxies matching one of the following criteria:

segmentation area > 700 pixels, or...

VIS < 20.5 AND segmentation area > 200 pixels

The measurements were made by Zoobot foundation models, finetuned on Euclid galaxies using the responses of Galaxy Zoo volunteers.

Our models were trained using galaxies from the selection cuts above but with the first option requiring 1200 pixels. Therefore, galaxies between 700 and 1200 pixels in area are may have less reliable measurements. 

The catalogue file is morphology_catalogue (.parquet or .csv, the contents are identical). It includes the following columns:

release_name
Always Q1_R1, for now

tile_index
Euclid tile index i.e. which MER tile hosts this galaxy

object_id
Euclid object id i.e. the MER catalogue identifier for this galaxy

segmentation_map_id
Alternative Euclid identifier. The first 9 digits are the tile index, the other digits match the internal segmentation id of the source.

right_ascension
in degrees, from the MER catalogue

declination
in degrees, from the MER catalogue

{question}_{answer}_fraction 
e.g. smooth-or-featured_smooth_fraction. The fraction of volunteers expected to give this answer to this morphology question. Probably the morphology columns you want. 

{question}_{answer}_dirichlet
e.g. smooth-or-featured_smooth_dirichlet. The concentration for a Dirichlet distribution (useful for uncertainties). See the paper.

warning_galaxy_fails_training_cuts
Marks galaxies between 700px and 1200px, where performance may be lower. See above.

cutout_width_arcsec
Width (and height) of cutout in arcseconds

The following questions and answers are available. 

Question
Answer
Notes

smooth-or-featured
smooth
May include face-on lenticulars, which are better identified with e.g. Sersic indices

how-rounded
round

how-rounded
in-between

how-rounded
cigar-shaped

smooth-or-featured
featured-or-disk
The question branch most commonly used by researchers

disk-edge-on
yes

edge-on-bulge
boxy

edge-on-bulge
none

edge-on-bulge
rounded

disk-edge-on
no

has-spiral-arms
yes

spiral-winding
tight

spiral-winding
medium

spiral-winding
loose

spiral-arm-count
1

spiral-arm-count
2

spiral-arm-count
3

spiral-arm-count
4

spiral-arm-count
more-than-4
Often overlaps with cant-tell

spiral-arm-count
cant-tell
Often overlaps with more-than-4

has-spiral-arms
no

bar
strong
Bar strength is a mix of length and width

bar
weak

bar
no

bulge-size
dominant

bulge-size
large

bulge-size
moderate

bulge-size
small

bulge-size
none

smooth-or-featured
problem

problem
star

problem
zoom
i.e. bad zoom, a cutout which is too wide

problem
artifact

artifact
satellite

artifact
scattered

artifact
diffraction

artifact
ray

artifact
saturation

artifact
other

artifact
ghost
Dichrotic ghosts

merging
none

merging
minor_disturbance

merging
major_disturbance
Primarily obvious tidal tails and similar features

merging
merger
Primarily "dramatic" ongoing mergers

clumps
yes
Not recommended; we are building clump-specific models

clumps
no
Not recommended; we are building clump-specific models

For convenience, we have also copied over some useful MER catalogue columns. The schema for the full MER catalogue is here. Additionally, Euclid also makes available many other tables with e.g. photometric redshifts, estimated masses, etc. These are documented here. All fluxes are in micro-janskies (uJy).

segmentation_area
Number of pixels included in SourceExtractor++ mask of galaxy (0.1 arcsec/pixel). 

flux_segmentation
Total VIS flux inside the segmentation mask above.

mag_segmentation
As above, converted to magnitude. ```mag = -2.5*log10(flux[muJy])+23.9```. Not technically in MER catalogue.

flux_detection_total
VIS flux measured within a Kron aperture in the detection image. FLUX_AUTO in SourceExtractor.

flux_vis_1fwhm_aper
VIS flux within an aperture of radius 1 FWHM. 

mumax_minus_mag
A star/galaxy diagnostic. The morphology catalogue uses the recommended filter MUMAX_MINUS_MAG>=-2.6 to reject stars.

mu_max
Peak surface brightness above the background in the detection band (directly from SExtractor)

ellipticity
A parametrization of how stretched an object is in the detection band (VIS, here), computed from the minor and major axes of the object itself (directly from SExtractor). [I assume this is the major/minor axis ratio]

kron_radius
Major semi-axis (in pixels) of the elliptical aperture used for total (Kron) aperture photometry on the detection image

2. Images

We are sharing the original cutout images as shown to Galaxy Zoo volunteers. The images are named like {tile_index}_{object_id}.jpg, where the negative sign ('-') in object id is replaced with 'NEG' to avoid path issues. You can construct the file paths from the morphology catalogue. For example:

df['file_loc'] = df['tile_index'].astype(str) + '_' + df['object_id'].astype(str).str.replace('-', 'NEG') + '_.jpg'

Each ZIP has images of every galaxy. There are three ZIP files, one for each image processing version. Volunteers were shown all three versions. The model predictions are made using the first version (the colour composite).

Fig 3 in the Q1 visual morphology paper shows an example galaxy in all three versions.

The VIS+Y images are composites with VIS in the blue channel and Y in the red channel (and the median of VIS and Y in the green channel, but this isn't visible). They use an arcinsh stretch, with the stretch designed to balance the contribution from each band.

The VIS only images are black-and-white, and also use an arcsinh stretch.

The VIS LSB images use a more complicated stretch to highlight LSB features.

Full details of the image processing are in the Q1 visual morphology paper.