LSST crowded field static science discussion

crowded-fields
milkyway
Tags: #<Tag:0x00007fd892551708> #<Tag:0x00007fd8925515a0>

(federica bianco, NYU) #1

Discussion about science drivers for observations on Bulge and Crowded fields and development of static sky (very) crowded fields photometry pipeline
On February @ivezic addressed the SMWLV Science Collaboration (contact @jgizis for more info on this SC) and a few other experts on MW science asking to generate a list of well-motivated science drivers for observations in the bulge, plane, and Magellanic Clouds where the extreme crowdedness would not be well handled by the current LSST pipeline. LSST project is looking for science cases to appropriately direct resources to crowded field photometry pipeline construction. Note that the current tests indicate that the LSSt pipeline would perform well even on extremely crowded fields (200k start/deg^2) for transient science (i.e. the difference imaging pipeline adequately handles these fields) but the static sky photometry would not perform at this crowdedness limit.

This post is designed to continue the discussion, including asking questions on the content and format of the science-case documents that the Project has requested.


(John Gizis) #2

Is there a convenient standard map of the stellar source density for LSST on the sky?

(I feel like it’s a part of MAF but I’m not finding it right now. I’m not worried if the maps are based on an older model that could be off by factors of ~2.)


(Lynne Jones) #3

The maps in MAF (which are actually downloaded as part of the data from sims_maps, but read/used in MAF) are based on stellar density estimates derived from other simulations catalogs inputs, which in turn are derived from GalFast.

The notebook here - https://github.com/lsst/sims_maps/blob/master/notebooks/Star_Map_Examples.ipynb - may provide some useful information on how the stellar density maps were created and their general range of values, and there is more information on the underlying simulated stellar catalog here - https://confluence.lsstcorp.org/display/SIM/Database+Contents+--+Catalog+Simulations (under Galactic Structure Catalogs).


(John Gizis) #4

Thanks! That’s perfect.


(Leo Girardi) #5

Our team has been preparing simulations alternative to GALFAST, which are partially uploaded to NOAO Data Lab (https://datalab.noao.edu/query.php?name=lsst_sim.simdr1) , but they are not yeat a good replacement to the maps in MAF mentioned above (especially because the DataLab files have big gaps in the Plane). Anyway, here are new estimates of the crowding limits in the r band, for seeing = 0.4, 0.5, 0.6, that is: these are the magnitudes at which photometric errors due to fainter sources will get larger that 0.1 mag. They agree with the DECAPS-based observation that the outer disk should not be crowded for LSST. Some areas in the inner Plane also appear as uncrowded, but simply because they have a huge extinction. Please contact me if you want to see some other related map – e.g. luminosity functions, distances we can reach for RC stars, etc.


(Zeljko Ivezic) #6

@lgirardi these are great maps! Since you offered, could you please add:

  • expression you used for computing photometric errors due to fainter sources
  • similar maps that would give cumulative source counts per sq. deg. for sources
    brighter than the crowding magnitude limit
  • for extra credit, integrated above counts within the Galactic plane diamond
    (|b| < 10*(90-l)/90 for l<90 and symmetric for 270<l<360) would be super
    useful, too.
    Thanks!

(Michael Wood-Vasey) #7

Are there cases or (l,b) where it’s okay to assume that all flux comes from point sources?

Conversely, are there cases where identifying nebulosity is of particular importance?


(Leo Girardi) #8

Hi Zeljko

  • the expression comes from Knut’s paper http://adsabs.harvard.edu/abs/2003AJ…126…452O , and there’s a python version here: https://github.com/lsst/sims_maf/blob/master/python/lsst/sims/maf/metrics/crowdingMetric.py
    In a few words one computes the LF and at each magnitude bin i, evaluates the noise from fainter stars as
    0.5sqrt(pi/area_in_sqrsec) * seeing * sqrt( sum_i (n_iluminosity_i^2) ) / luminosity_i
    It depends on the LF hence differs from halo to bulge.

  • the density maps are below: that’s the log(Nstars/deg2) observable either above r=27.5, or above the crowding limit. There’s a weird effect that as the real stellar density increases towards the bulge, reaching the crowding limit suddenly causes the observable stellar density to decrease. Across the bulge and inner disk, for seeing=0.5 arcsec, we have a plateau with ~200000 stasr/deg2, similar to the limit you were mentioning for the current photometric pipeline! (if I got it right).

  • number of stars in the diamond area: (that’s a lower limit because the modeled area is somewhat smaller):
    2.2e9 for seeing=0.6 arcsec
    3.8e9 for seeing=0.5 arcsec
    7.0e9 for seeing=0.4 arcsec

  • just a correction, in these maps, the crowding limit is considered to be reached at sigma=0.05 mag, not at 0.1 mag as I wrote in the previous post.

Cheers