Skip to content

BPZ and the COSMOS field

Jump to bottom
Adam Wright edited this page Dec 12, 2017 · 7 revisions

For how to run/modify the scripts that are meant specifically for running BPZ on the COSMOS fields, this is the right place.

For how to run BPZ in wtgpipeline in general, see Photometric Redshifts with BPZ.

For how to utilize COSMOS data to get a Mass Bias catalog, see Quantifying Mass Bias with COSMOS.

NOTE: We'll need to look at ugrizY, ugriz, BVRIZ, and BVRI. For one of our clusters we've got only 4 bands (BVRI) and we'd like to get a p(z) mass out of it if possible.

Cosmos catalog details

Note that keys of the name 'rp', 'ip', and 'zp' denote the Subaru filters r, i, and z. The 'p' indicates "plus", because they are the 'W-S-R+', 'W-S-I+', and 'W-S-Z+' filters.

  • 'zp' denotes data taken on Subaru with the 'W-S-Z+' filter in 2008 or earlier (when it had blue-sensitive ccds)

  • 'zpp' denotes data taken on Subaru with the 'W-S-Z+' filter after 2008 when the red-sensitive ccds were put on the camera

Status as of 10/24/2017

script that runs bpz on COSMOS is up and running! See run_bpz_on_cosmos.py

see output here: /u/ki/awright/COSMOS_2017/run_bpz_on_COSMOS/B-V-rp-ip-zpp_output

Here is bpz's feedback:

input params:  NMAX  =  None
input params:  FC  =  None
input params:  ODDS  =  0.95
input params:  DELTA_M_0  =  0.0
input params:  MIN_RMS  =  0.05
input params:  PLOTS  =  yes
input params:  INTERACTIVE  =  no
input params:  ZP_OFFSETS  =  0.0
input params:  MIN_MAGERR  =  0.001
input params:  SED_DIR  =  /nfs/slac/kipac/fs1/u/awright/InstallingSoftware_extension/bpz-1.99.3//SED/
input params:  ADD_SPEC_PROB  =  None
input params:  ONLY_TYPE  =  no
input params:  CONVOLVE_P  =  yes
input params:  PROBS2  =  no
input params:  COLUMNS  =  /u/ki/awright/COSMOS_2017/run_bpz_on_COSMOS//B-V-rp-ip-zpp_output/cosmos_ultravista.B-V-rp-ip-zpp.CWWSB_capak.list.rp.columns
input params:  NTYPES  =  None
input params:  CHECK  =  /u/ki/awright/COSMOS_2017/run_bpz_on_COSMOS//B-V-rp-ip-zpp_output/cosmos_ultravista.B-V-rp-ip-zpp.CWWSB_capak.list.rp.flux_comparison
input params:  MAG  =  yes
input params:  OUTPUT  =  /u/ki/awright/COSMOS_2017/run_bpz_on_COSMOS//B-V-rp-ip-zpp_output/cosmos_ultravista.B-V-rp-ip-zpp.CWWSB_capak.list.rp.bpz
input params:  PROBS  =  no
input params:  ZMAX  =  4.0
input params:  GET_Z  =  yes
input params:  ZMIN  =  0.01
input params:  INTERP  =  8
input params:  DZ  =  0.01
input params:  Z_THR  =  0
input params:  MERGE_PEAKS  =  no
input params:  COLOR  =  no
input params:  N_PEAKS  =  1
input params:  EXCLUDE  =  none
input params:  AB_DIR  =  /nfs/slac/kipac/fs1/u/awright/InstallingSoftware_extension/bpz-1.99.3//AB/
input params:  PROBS_LITE  =  /u/ki/awright/COSMOS_2017/run_bpz_on_COSMOS//B-V-rp-ip-zpp_output/cosmos_ultravista.B-V-rp-ip-zpp.CWWSB_capak.list.rp.probs
input params:  PHOTO_ERRORS  =  no
input params:  SPECTRA  =  CWWSB_capak.list
input params:  ZC  =  None
input params:  P_MIN  =  0.01
input params:  PRIOR  =  hdfn_SB
input params:  NEW_AB  =  no
input params:  INPUT  =  /u/ki/awright/COSMOS_2017/run_bpz_on_COSMOS//B-V-rp-ip-zpp_output/cosmos_ultravista.B-V-rp-ip-zpp.CWWSB_capak.list.rp.cat
input params:  VERBOSE  =  no
input params:  FILTER_DIR  =  /nfs/slac/kipac/fs1/u/awright/InstallingSoftware_extension/bpz-1.99.3//FILTER/
input params:  ADD_CONTINUOUS_PROB  =  None
input params:  MADAU  =  yes
 plots= matplotlib  interactive= 0
checkSED= True
0.95 1.95993
ML * prior -- NOT QUITE BAYESIAN

PHOTOMETRIC CALIBRATION TESTS
If the dmag are large, add them to the .columns file (zp_offset), then re-run BPZ.
(For better results, first re-run with -ONLY_TYPE yes to fit SEDs to known spec-z.)

  fo/ft    dmag   nobj   filter
  1.038    0.041 200175   B
  1.004    0.005 201251   V
  0.975   -0.028 201118   rp
  1.007    0.008 190689   ip
  1.009    0.010 201689   zpp
fo/ft = Average f_obs/f_model weighted by f_obs/ef_obs for objects with ODDS >= 0.95
dmag = magnitude offset which should be applied (added) to the photometry (zp_offset)
nobj = # of galaxies considered in that filter (detected and high ODDS >= 0.95)

To be continued...

Still some things must be figured out. Mainly with the filters and with making this script more flexible to run different configurations, do it on the batch/etc, include the spec-z comparison, etc. As well as figuring out which final COSMOS catalog to use.

Have to figure out how to match up remaining filters:

  • SUBARU W-C-RC and W-C-IC aren't in the COSMOS catalog (I think), but that's fine, just use W-S-R+ and W-S-I+ instead ('rp' and 'ip')
  • There is an HSC Y-band, but is it in the matched.cosmos.cat file we're currently using? The ultravista Y-band probably isn't comparable to the DECam Y-band, but it's worth looking into anyway. This relationship can be explored by once Lucie processes the DECam-imaged COSMOS field.
  • The CFHT u-band is comparable to the DECam u-band.
  • is 'i_c' the CFHT i-band? (maybe it's the W-C-IC tophat filter?)
Clone this wiki locally