301.5. Euclid Deep Field South (EDFS)

301.5 Euclid Deep Field South¶

For the Rubin Science Platform at data.lsst.cloud.
Data Release: DP1
Container Size: Large
LSST Science Pipelines version: r29.2.0
Last verified to run: 2025-09-02
Repository: github.com/lsst/tutorial-notebooks

Learning objective: To explore the Euclid Deep South Field target field from commissioning with LSSTComCam, using Data Preview 1 (DP1) data.

LSST data products: visit_table, visit_detector_table, skymap.

Packages: lsst.daf.butler, lsst.afw.display, lsst.rsp, lsst.geom, astropy, numpy, matplotlib, scipy.

Credit: Originally developed by Andrés A. Plazas Malagón and the Rubin Community Science team. Please consider acknowledging them if this notebook is used for the preparation of journal articles, software releases, or other notebooks.

Get Support: Everyone is encouraged to ask questions or raise issues in the Support Category of the Rubin Community Forum. Rubin staff will respond to all questions posted there.

1. Introduction¶

This notebook examines the Data Preview 1 (DP1) data products in the LSSTComCam Euclid Deep Field South (EDFS) target field. It includes analyses of magnitude limits, visit distribution over time, data quality, and the distributions of stars and galaxies in color–magnitude and color–color diagrams.

The Euclid Deep Field South covers approximately 28.1 square degrees in the southern constellation of Horologium, named after the pendulum clock.

The LSSTComCam EDFS is one of the target fields selected to overlap with a key extragalactic deep field identified by the European Space Agency’s Euclid mission for intensive imaging and spectroscopic follow-up. This overlap enables valuable cross-survey comparisons and synergy.

The LSSTComCam EDFS field is centered at Right Ascension and Declination coordinates (RA, Dec) = (59.10°, –48.73°). Although its band coverage is more limited than that of the Extended Chandra Deep Field South (ECDFS), it remains highly relevant for extragalactic science and future multi-survey studies.

References Rubin-Euclid Derived Data Products: Initial Recommendations

Related tutorials: The 100-level tutorials demonstrate how to use the butler, the TAP service, and the Firefly image display. The 200-level tutorials introduce the types of image and catalog data. Other tutorials in the 300 series demonstrate additional LSSTComCam fields during commissioning.

1.1. Import packages¶

Import numpy, a fundamental package for scientific computing with arrays in Python (numpy.org), and matplotlib, a comprehensive library for data visualization (matplotlib.org; matplotlib gallery), including custom shapes (Polygon) and lines (mlines). itertools supports efficient iteration and combinatorics.

From the lsst package, import modules for accessing the Table Access Protocol (TAP) service, for retrieving datasets from the butler, and for image displaying from the LSST Science Pipelines (pipelines.lsst.io). Additional modules support spherical geometry (sphgeom), 2D geometry (geom), and standardized multiband plotting (lsst.utils.plotting) for LSST data analysis and visualization.

In [1]:

import numpy as np
import matplotlib.pyplot as plt
from matplotlib.patches import Polygon
import matplotlib.lines as mlines
import itertools

from lsst.rsp import get_tap_service
from lsst.daf.butler import Butler
import lsst.afw.display as afw_display
import lsst.sphgeom as sphgeom
import lsst.geom as geom
from lsst.utils.plotting import (get_multiband_plot_colors,
                                 get_multiband_plot_symbols,
                                 get_multiband_plot_linestyles)

1.2. Define parameters and functions¶

Create an instance of the TAP service, and assert that it exists.

In [2]:

rsp_tap = get_tap_service("tap")
assert rsp_tap is not None

Create an instance of the butler, and assert that it exists.

In [3]:

butler = Butler('dp1', collections='LSSTComCam/DP1', skymap='lsst_cells_v1')
assert butler is not None

Define the approximate central coordinates of the EDFS target field, in degrees, and a region within 1 degree of this center.

In [4]:

ra_cen = 59.10
dec_cen = -48.73
radius = 1.0

region = sphgeom.Region.from_ivoa_pos(f"CIRCLE {ra_cen} {dec_cen} {radius}")

Define parameters to use colorblind-friendly colors with matplotlib.

In [5]:

plt.style.use('seaborn-v0_8-colorblind')
prop_cycle = plt.rcParams['axes.prop_cycle']
colors = prop_cycle.by_key()['color']

Define colors, symbols, and linestyles to represent the six LSST filters, $ugrizy$, as defined in RTN-045.

In [6]:

filter_colors = get_multiband_plot_colors()
filter_symbols = get_multiband_plot_symbols()
filter_linestyles = get_multiband_plot_linestyles()

Set afwDisplay to use Firefly to display images, and open Firefly frame 1.

In [7]:

afw_display.setDefaultBackend('firefly')
display = afw_display.Display(frame=1)

2. Coadd sky coverage and depth¶

The individual LSSTComCam images have been combined (stacked) into deep_coadd images, and sky maps of the accumulated exposure time and image depth have been created.

2.1. Display a deep coadd¶

Query the butler for all $r$-band deep_coadd images that overlap the defined region.

In [8]:

my_band = 'r'

In [9]:

coadd_datasetrefs = butler.query_datasets("deep_coadd",
                                          where="patch.region OVERLAPS"
                                          f" region AND band='{my_band}'",
                                          bind={"region": region},
                                          with_dimension_records=True,
                                          order_by=["patch.tract"])

In [10]:

print(f'There are {len(coadd_datasetrefs)} r-band deep_coadd images.')

There are 73 r-band deep_coadd images.

Retrieve the deep_coadd image associated with one of the dataset references returned by the butler query.

In [11]:

coadd = butler.get(coadd_datasetrefs[10])

Display the image, and set the mask plane to be fully transparent.

In [12]:

display.mtv(coadd)
display.setMaskTransparency(100)

2.2. Exposure time and depth¶

Retrieve the entire survey property map of the r-band magnitude limit as hspmap_rmaglim, and of the total r-band exposure time as hspmap_rexptime.

In [13]:

hspmap_rexptime = butler.get('deepCoadd_exposure_time_consolidated_map_sum',
                             skymap='lsst_cells_v1', band='r')
hspmap_rmaglim = butler.get('deepCoadd_psf_maglim_consolidated_map_weighted_mean',
                            skymap='lsst_cells_v1', band='r')

Extract the healsparse map values within the radius of the field center and divide the region into a 250x250 grid. Evaluate each survey property map at these grid points. Replace all negative values (no-data placeholders) with NaN to enable proper map display.

In [14]:

dec_size = radius
ra_size = dec_size / np.cos(np.radians(dec_cen))
ra_min, ra_max = ra_cen - ra_size, ra_cen + ra_size
dec_min, dec_max = dec_cen - dec_size, dec_cen + dec_size

ra = np.linspace(ra_min, ra_max, 250)
dec = np.linspace(dec_min, dec_max, 250)
x, y = np.meshgrid(ra, dec)

values_rmaglim = hspmap_rmaglim.get_values_pos(x, y)
values_rmaglim = np.where(values_rmaglim < 0.0, np.nan,
                          values_rmaglim)

values_rexptime = hspmap_rexptime.get_values_pos(x, y)
values_rexptime = np.where(values_rexptime < 0.0, np.nan,
                           values_rexptime)

Clean up.

In [15]:

del ra, dec

Print the mean, median, and min/max values for the $r$-band magnitude limit and exposure time.

In [16]:

print("r-band magnitude limit:")
print(f"  Mean: {np.nanmean(values_rmaglim): 5.2f}")
print(f"  Median: {np.nanmedian(values_rmaglim): 5.2f}")
print(f"  Min: {np.nanmin(values_rmaglim): 5.2f}")
print(f"  Max: {np.nanmax(values_rmaglim): 5.2f}\n")

print("r-band exposure time (s):")
print(f"  Mean: {np.nanmean(values_rexptime): 5.1f}")
print(f"  Median: {np.nanmedian(values_rexptime): 5.1f}")
print(f"  Min: {np.nanmin(values_rexptime): 5.1f}")
print(f"  Max: {np.nanmax(values_rexptime): 5.1f}")

r-band magnitude limit:
  Mean:  25.57
  Median:  25.81
  Min:  22.11
  Max:  26.51

r-band exposure time (s):
  Mean:  983.6
  Median:  720.0
  Min:  30.0
  Max:  2550.0

2.3. Coverage map¶

Define unique colors and linestyles for plotting each of the tracts in the dataset.

In [17]:

alltracts = [rec.dataId['tract'] for rec in coadd_datasetrefs]
unique_tracts = np.unique(alltracts)

tract_colors = plt.cm.tab10(np.linspace(0, 1, len(unique_tracts)))
linestyles = ['-', '--', '-.', ':'] * ((len(unique_tracts) // 4) + 1)

style_dict = {
    tract: {'color': tract_colors[i], 'linestyle': linestyles[i]}
    for i, tract in enumerate(unique_tracts)
}

The coadd_datasetrefs returned from the butler include region information for each dataset, stored as a ConvexPolygon3D with four vertices defining the sky footprint of each overlapping tract and patch.

Convert these vertices to 2D sky coordinates (RA, Dec) using geom.SpherePoint, and use matplotlib.patches.Polygon along with ax.add_patch() to draw each patch outline. Each tract is plotted with a distinct color and linestyle for visual clarity.

In [18]:

fig, ax = plt.subplots(figsize=(6, 5))

mesh = ax.pcolormesh(x, y, values_rmaglim, cmap='Greys_r', shading='auto')
fig.colorbar(mesh, ax=ax, label="r-band limiting magnitude (mag)")

for rec in coadd_datasetrefs:
    vertices = rec.dataId.patch.region.getVertices()
    vertices_deg = []
    for vertex in vertices:
        vertices_deg.append([geom.SpherePoint(vertex).getRa().asDegrees(),
                             geom.SpherePoint(vertex).getDec().asDegrees()])
    polygon = Polygon(vertices_deg, closed=True, facecolor='None',
                      edgecolor=style_dict[rec.dataId['tract']]['color'],
                      linestyle=style_dict[rec.dataId['tract']]['linestyle'],
                      linewidth=2)
    ax.add_patch(polygon)

ax.set_xlim(ra_max, ra_min)
ax.set_ylim(dec_min, dec_max)
ax.set_xlabel('RA (deg)')
ax.set_ylabel('Dec (deg)')

handles = [
    mlines.Line2D([], [], color=style['color'],
                  linestyle=style['linestyle'], label=f"Tract {tract}")
    for tract, style in style_dict.items()
]
ax.legend(handles=handles, loc='upper left', ncol=2)

ax.minorticks_on()

plt.tight_layout()
plt.show()

Figure 1: For the region of the Euclid Deep Fields South target field, the $r$-band magnitude limit survey property map is displayed in greyscale, ranging from $r$ magnitudes roughly 22 to 26.5. Patch boundaries are overplotted, colored by their tract number (as in the legend).

Clean up.

In [19]:

del coadd_datasetrefs
del hspmap_rmaglim, hspmap_rexptime
del x, y, values_rmaglim, values_rexptime
del style_dict

3. Visits (observations)¶

Retrieve all visits (individual observations of the sky) that fall within a circular region centered at (RA, Dec) = (ra_cen, dec_cen) with a radius of 1 degree. Return the visit ID, band, observation midpoint time in both MJD and calendar date.

In [20]:

query = (
    f"""
    SELECT visit, band, expMidptMJD, expMidpt
    FROM dp1.Visit
    WHERE CONTAINS(
        POINT('ICRS', ra, dec), CIRCLE('ICRS', {ra_cen}, {dec_cen}, {radius})) = 1
    ORDER BY expMidptMJD
    """.strip())

In [21]:

job = rsp_tap.submit_job(query)
job.run()
job.wait(phases=['COMPLETED', 'ERROR'])
print('Job phase is', job.phase)
if job.phase == 'ERROR':
    job.raise_if_error()
assert job.phase == 'COMPLETED'

Job phase is COMPLETED

In [22]:

visits = job.fetch_result().to_table()
print(f"Total number of visits: {len(visits)}")

Total number of visits: 272

Option to display the table of results.

In [23]:

# visits

3.1. Filter distribution¶

Count and list the number of visits per filter.

In [24]:

all_bands = np.array([visit['band'] for visit in visits], dtype=str)
unique_bands, counts = np.unique(all_bands, return_counts=True)

for band, count in zip(unique_bands, counts):
    print(band, count)

g 61
i 42
r 87
u 20
y 20
z 42

In [25]:

field_filters = ['u', 'g', 'r', 'i', 'z', 'y']

Clean up.

In [26]:

del all_bands, unique_bands, counts

3.2. Visit dates¶

Plot two characteristics of the visit dates:

1. Plot the cumulative distribution of visit dates for each of the filters observed for this field.
2. Plot the distribution of the time separation between visits for each band and for all bands combined.

In [27]:

t_start = visits[0]['expMidptMJD'] - 1.0
t_end = visits[-1]['expMidptMJD'] + 1.0
time_bins = np.arange(t_end - t_start, dtype='int') + t_start
diff_bins = np.arange(0, np.max(time_bins) - np.min(time_bins) + 3, 1)
all_time_diffs = []

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(11, 4))

for band in field_filters:
    mjds = sorted([
        visit['expMidptMJD']
        for visit in visits
        if visit['band'] == band
    ])
    mjds = np.array(mjds, dtype=float)
    label = f"{band} ({len(mjds)})"

    counts, _, patches = ax1.hist(
        mjds, bins=time_bins, cumulative=True,
        histtype='step',
        linewidth=2, color=filter_colors[band],
        label=label
    )
    for patch in patches:
        patch.set_linestyle(filter_linestyles[band])

    tdiffs = []
    for comb in itertools.combinations(mjds, 2):
        tdiffs.append(comb[1] - comb[0])
    time_diffs = np.array(tdiffs)
    all_time_diffs.append(time_diffs)

    counts_diff, _, patches_diff = ax2.hist(
        time_diffs, bins=diff_bins, histtype='step',
        linewidth=2,
        color=filter_colors[band], label=band
    )
    for patch in patches_diff:
        patch.set_linestyle(filter_linestyles[band])

all_time_diffs = np.concatenate(all_time_diffs)
ax2.hist(
    all_time_diffs, bins=diff_bins, histtype='step',
    linewidth=2,
    color='black', label='all filters'
)

ax1.set_xlabel('Modified Julian Date')
ax1.set_ylabel('Cumulative number of visits')
ax1.minorticks_on()

ax2.set_xlabel('Separation between observations (d)')
ax2.set_ylabel('Number of observations')
ax2.set_yscale('log')
ax2.legend(loc='upper right', ncol=2)
ax2.minorticks_on()

plt.show()

Figure 2: The left panel shows the cumulative number of visits per filter as a function of time. The right panel shows histograms of visit-to-visit time separations for each filter individually (colors), and all together (black).

Clean up.

In [28]:

del time_bins, all_time_diffs

3.3. Visit image quality¶

Retrieve statistics about the image quality of the individual visit_images from the CcdVisit table.

In [29]:

query = f"""
    SELECT visitId, ra, dec, band, magLim, seeing
    FROM dp1.CcdVisit
    WHERE CONTAINS(POINT('ICRS', ra, dec), CIRCLE('ICRS', {ra_cen}, {dec_cen}, {radius})) = 1
    ORDER BY visitId
    """

In [30]:

job = rsp_tap.submit_job(query)
job.run()
job.wait(phases=['COMPLETED', 'ERROR'])
print('Job phase is', job.phase)
if job.phase == 'ERROR':
    job.raise_if_error()
assert job.phase == 'COMPLETED'

Job phase is COMPLETED

In [31]:

ccd_visits = job.fetch_result().to_table()

Plot the distribution of seeing, which is the mean full width at half maximum of the PSF, for all visits and detector combinations.

In [32]:

min_seeing = np.min(ccd_visits['seeing'])
max_seeing = np.max(ccd_visits['seeing'])
seeing_bins = np.arange(min_seeing, max_seeing, 0.05)

fig = plt.figure(figsize=(7, 5))
for band in field_filters:
    print(f"Median in {band}-band (arcsec): \
            {np.ma.median(ccd_visits[ccd_visits['band'] == band]['seeing']): .2f}")
    n, bins, patches = plt.hist(ccd_visits[ccd_visits['band'] == band]['seeing'],
                                seeing_bins, histtype='step',
                                linewidth=2,
                                color=filter_colors[band],
                                label=band)
    for patch in patches:
        patch.set_linestyle(filter_linestyles[band])

plt.legend(loc='upper right')
plt.xlabel('PSF FWHM (arcsec)')
plt.ylabel('Number of visits')
plt.minorticks_on()
plt.show()

Median in u-band (arcsec):              1.88

Median in g-band (arcsec):              1.25
Median in r-band (arcsec):              1.20
Median in i-band (arcsec):              1.10
Median in z-band (arcsec):              1.18
Median in y-band (arcsec):              0.99

Figure 3: Histogram of delivered image quality of detector images, separated by the six $ugrizy$ bands. The majority of observations have delivered PSF size between 0.7 to 1.5 arcseconds.

Clean up.

In [33]:

del seeing_bins

Plot the distribution of the magnitude limit, magLim.

In [34]:

min_mag = np.min(ccd_visits['magLim'])
max_mag = np.max(ccd_visits['magLim'])
maglim_bins = np.arange(min_mag, max_mag, 0.1)

fig = plt.figure(figsize=(7, 5))

for band in field_filters:
    print(f"Median in {band}-band: \
            {np.ma.median(ccd_visits[ccd_visits['band'] == band]['magLim']): .2f}")
    n, bins, patches = plt.hist(ccd_visits[ccd_visits['band'] == band]['magLim'],
                                maglim_bins, histtype='step',
                                linewidth=2,
                                color=filter_colors[band],
                                label=band)
    for patch in patches:
        patch.set_linestyle(filter_linestyles[band])

plt.legend(loc='upper left')
plt.xlabel('limiting magnitude')
plt.ylabel('Number of visits')
plt.minorticks_on()
plt.show()

Median in u-band:              23.30
Median in g-band:              24.50
Median in r-band:              24.31
Median in i-band:              23.92
Median in z-band:              23.20
Median in y-band:              22.16

Figure 4: Histogram of limiting magnitude at signal-to-noise=5 on detector images, separated by the six $ugrizy$ bands. The deepest images are in the $r$- and $g$-bands, reaching typical limiting magnitudes of roughly 24.5 mag.

Clean up.

In [35]:

del maglim_bins

4. Object (detections)¶

The Object table contains forced measurements on the deep coadded images at the locations of all objects detected with signal-to-noise ratio > 5 in a deep_coadd of any filter.

Query the Object table for objects in the field, and return their PSF and cModel AB magnitudes.

In [36]:

query = """SELECT objectId, coord_ra, coord_dec, ebv,
        {}_psfMag, {}_cModelMag, {}_psfMag, {}_cModelMag,
        {}_psfMag, {}_cModelMag, {}_psfMag, {}_cModelMag,
        {}_psfMag, {}_cModelMag, {}_psfMag, {}_cModelMag,
        refExtendedness
        FROM dp1.Object
        WHERE CONTAINS(POINT('ICRS', coord_ra, coord_dec),
              CIRCLE('ICRS', {}, {}, {})) = 1
        ORDER BY objectId ASC
        """.format(field_filters[0], field_filters[0], field_filters[1], field_filters[1],
                   field_filters[2], field_filters[2], field_filters[3], field_filters[3],
                   field_filters[4], field_filters[4], field_filters[5], field_filters[5],
                   ra_cen, dec_cen, radius)
print(query)

SELECT objectId, coord_ra, coord_dec, ebv,
        u_psfMag, u_cModelMag, g_psfMag, g_cModelMag,
        r_psfMag, r_cModelMag, i_psfMag, i_cModelMag,
        z_psfMag, z_cModelMag, y_psfMag, y_cModelMag,
        refExtendedness
        FROM dp1.Object
        WHERE CONTAINS(POINT('ICRS', coord_ra, coord_dec),
              CIRCLE('ICRS', 59.1, -48.73, 1.0)) = 1
        ORDER BY objectId ASC

In [37]:

job = rsp_tap.submit_job(query)
job.run()
job.wait(phases=['COMPLETED', 'ERROR'])
print('Job phase is', job.phase)
if job.phase == 'ERROR':
    job.raise_if_error()
assert job.phase == 'COMPLETED'

Job phase is COMPLETED

In [38]:

objtab = job.fetch_result().to_table()

4.1. Magnitude distributions¶

Plot histograms of object magnitudes, separating the samples into stars and galaxies.

Use the refExtendedness flag to distinguish them: treat objects with refExtendedness == 1 as likely galaxies (extended) and those with refExtendedness == 0 as likely stars (point sources).

Use cmodelmag mags for galaxies and psfmag for stars.

In [39]:

ptsource = (objtab['refExtendedness'] == 0)

In [40]:

mag_bins = np.arange(15.4, 28.6, 0.2)

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(9, 4), sharey=True)
plt.subplots_adjust(hspace=0, wspace=0)

for band in field_filters:
    mag_psf = objtab[f"{band}_psfMag"]
    mag_cmodel = objtab[f"{band}_cModelMag"]

    valid_mags = (
        (15 < mag_psf) & (mag_psf < 30) & (15 < mag_cmodel) & (mag_cmodel < 30)
    )
    print(f"Number of objects in {band}-band: {np.sum(valid_mags)}")

    n_psf, bins_psf, patches_psf = ax1.hist(
        mag_psf[valid_mags & ptsource],
        bins=mag_bins,
        histtype='step',
        linewidth=2,
        color=filter_colors[band],
        label=band,
    )
    for patch in patches_psf:
        patch.set_linestyle(filter_linestyles[band])

    n_cmodel, bins_cmodel, patches_cmodel = ax2.hist(
        mag_cmodel[valid_mags & ~ptsource],
        bins=mag_bins,
        histtype='step',
        linewidth=2,
        color=filter_colors[band],
        label=band,
    )
    for patch in patches_cmodel:
        patch.set_linestyle(filter_linestyles[band])

ax1.set_xlim(mag_bins.min(), mag_bins.max())
ax1.set_yscale('log')
ax1.set_xlabel('Magnitude')
ax1.set_ylabel('Number of objects')
ax1.set_title('Point sources (PSF mags)')
ax1.legend(loc='upper left', ncols=2)
ax1.minorticks_on()

ax2.set_xlim(mag_bins.min(), mag_bins.max())
ax2.set_xlabel('Magnitude')
ax2.set_title('Extended sources (cModel mags)')
ax2.minorticks_on()

plt.show()

Number of objects in u-band: 229871
Number of objects in g-band: 325230
Number of objects in r-band: 322792
Number of objects in i-band: 317241
Number of objects in z-band: 314406
Number of objects in y-band: 278285

Figure 5: Histogram of object magnitudes separated by morphological type. The left panel shows likely stars (refExtendedness = 0) using PSF magnitudes, while the right panel shows likely galaxies cModel magnitudes.

Clean up.

In [41]:

del (mag_bins, mag_psf, mag_cmodel,
     valid_mags, n_psf, bins_psf,
     patches_psf, n_cmodel,
     bins_cmodel, patches_cmodel)

5. Analysis¶

A short demonstration of some stanard scientific analyses.

5.1. CMDs and CCDs¶

Plot $r$ vs. $(g - r)$ Color-Magnitude diagrams (CMDs) and $(r - i)$ vs. $(g - r)$ Color-Color diagrams (CCDs.) Create separate plots for stars and galaxies using PSF magnitudes for stars and cModel magnitudes for galaxies.

In [42]:

fig, ((ax1, ax2), (ax3, ax4)) = plt.subplots(
    2, 2, figsize=(7, 6), height_ratios=[1, 2.5]
)
plt.subplots_adjust(hspace=0, wspace=0)

grmin, grmax = -0.9, 2.3
rimin, rimax = -1.3, 2.8
magmin, magmax = 15.8, 26.8

ax1.hexbin(
    objtab[ptsource]['g_psfMag'] - objtab[ptsource]['r_psfMag'],
    objtab[ptsource]['r_psfMag'] - objtab[ptsource]['i_psfMag'],
    gridsize=(150, 150),
    extent=(grmin, grmax, rimin, rimax),
    bins='log',
    cmap='Grays',
)
ax1.set_title('Point sources (PSF mags)')
ax1.set_xlim(grmin, grmax)
ax1.set_ylim(rimin, rimax)
ax1.set_ylabel(r'$(r-i)$')
ax1.set_xticklabels([])
ax1.minorticks_on()

ax2.hexbin(
    objtab[~ptsource]['g_cModelMag'] - objtab[~ptsource]['r_cModelMag'],
    objtab[~ptsource]['r_cModelMag'] - objtab[~ptsource]['i_cModelMag'],
    gridsize=(150, 150),
    extent=(grmin, grmax, rimin, rimax),
    bins='log',
    cmap='Grays',
)
ax2.set_title('Extended sources (cModel mags)')
ax2.set_xlim(grmin, grmax)
ax2.set_ylim(rimin, rimax)
ax2.set_xticklabels([])
ax2.set_yticklabels([])
ax2.minorticks_on()

ax3.hexbin(
    objtab[ptsource]['g_psfMag'] - objtab[ptsource]['r_psfMag'],
    objtab[ptsource]['r_psfMag'],
    gridsize=(150, 200),
    extent=(grmin, grmax, magmin, magmax),
    bins='log',
    cmap='Grays',
)

ax3.set_xlim(grmin, grmax)
ax3.set_ylim(magmax, magmin)
ax3.set_xlabel(r'$(g-r)$')
ax3.set_ylabel(r'$r$ magnitude')
ax3.minorticks_on()

ax4.hexbin(
    objtab[~ptsource]['g_cModelMag'] - objtab[~ptsource]['r_cModelMag'],
    objtab[~ptsource]['r_cModelMag'],
    gridsize=(150, 200),
    extent=(grmin, grmax, magmin, magmax),
    bins='log',
    cmap='Grays',
)
ax4.set_xlim(grmin, grmax)
ax4.set_ylim(magmax, magmin)
ax4.set_xlabel(r'$(g-r)$')
ax4.set_yticklabels([])
ax4.minorticks_on()

plt.show()

Figure 6: CMDs and CCDs for objects classified by morphology. The left column shows likely stars (refExtendedness = 0) using PSF magnitudes, and the right column shows likely galaxies using cModel magnitudes. Top panels display $(r - i)$ vs. $(g - r)$ CCDs; bottom panels show $r$ vs. $(g - r)$ CMDs.

In [ ]: