What is the DDF area corresponding to the MAF QSONumberCountsMetric

Support
, ,
#1

Hi,

I would like to use the “QSONumberCountsMetric” from the metric analysis framework (MAF) and normalize this number by the sky area.

For example, for the COSMOS deep-drilling field (DDF). I find this metric from the website: baseline_v4.0_10yrs (baseline v4.0).

However, since the DDF pointings are dithered, I would like to understand whether this number should be normalized to the area of the telescope field-of-view (9.6 deg2) or to the area swept by the dither pattern (rough visual estimate yields ~18.6 deg2).

This would be trivial to estimate from the already plotted maps, but it seems that no information on the scale is given.

Thank you very much in advance.
Best,
-akke

#2

I see - you’re correct that no scale or total area is given. The number reported in the table below is the number integrated over the healpixel grid used in the calculation, and we haven’t reported the total area (we should think of a way to do this). I think this grid area may vary between simulation evaluations – in newer maf runs, we’re putting a tighter limit on the outer radius of the grid baseline_v5.0.0_10yrs

I’m not entirely certain what you gain from normalizing by the area though … are you trying to see if the estimated QSO number matches another luminosity function?
The LF used for the input to the simulation is derived here - GitHub - lsst-sims/create_NQSO_tables: Create tables used to calculate number densities of QSOs
(and I’m just realizing I may have failed to fold the latest update here back into the input data used in the MAF metric calculation, so we will update that as well! )

#3

Thanks, Lynne.

Yes, indeed. I am trying to use the normalized number counts to compare with the literature, and with another simulation.

Maybe the healpix maps are already available somewhere? Thanks again.

#4

@aetviita The team is actively working on a solution and will report back as soon as we have a clear path forward.

#5

Unfortunately, I have looked into what you’re asking for and we just don’t keep what you’d actually need – I think what you need is not just “the area in the DDF metric” but is the amount of area at a particular depth.
This is the healpix map of the depth, which you can then multiply by the luminosity function (which is what we do for this metric – see more details here).

But we don’t keep these maps, just due to the number of simulations we have.
I suppose another approach you could do is to use the average or median depth we report in a different metric (the ‘coaddm5 metric’ outputs), assume that the area is on the order of 10 deg, and multiply the LF we’re using for these metrics directly – the LF tables are in $RUBIN_SIM_DATA_DIR/maf/quasarNumberCounts.
(if you don’t have MAF installed, you could download just the data directory at https://s3df.slac.stanford.edu/data/rubin/sim-data/rubin_sim_data/maf_2024_06_13.tgz)
and then compare these values to what you calculate with your own LFs (or … just compare the LFs directly?).

Alternatively, if you wanted to plug your own LFs into the same metric, let me know and we can talk about how to get that working for you.

As an aside - the general point of the metric output here isn’t necessarily the absolute number, but how the number changes between simulations (i.e. if we have more r band visits than g band, does the number of expected QSOs change, or as we add more dithering, how does the number of QSOs change, or as we reduce the time spent in DDFs, how does the LF change as the depth changes).