Strategies to identify strongly lensed type Ia supernovae in Rubin LSST
Prajakta Mane
Final year, Integrated BS-MS,
Indian Institute of Science Education and Research (IISER) Mohali, India
Joint SLSC/DESC-SL Telcon
Anupreeta More, Surhud More
Inter-University Centre for Astronomy and Astrophysics (IUCAA), India
April 17, 2024
Time delay cosmography:
accurate time delays, surface mass distribution and redshift of lens
Hubble constant with strong lensing time delays
well-constrained Hubble Constant
Why are SNe Ia a better source for TDC?
-
well-studied light curves -> more accurate time delay measurements
-
transient source -> improved constraints on the lens galaxy model
-
standard candle nature -> help overcome degeneracies
How to find lensed SNe Ia?
DIA: run LSST Stack's difference imaging pipeline to identify multiply imaged SNe
CMA: study outliers in the colour-magnitude space for SNe to identify lensed SNe Ia
- Raw exposures from Subaru Hyper-Suprime Cam (HSC) -> Public Data Release 1
- UltraDeep Survey's COSMOS region, Tract number 9813 (~1.7 deg wide)
- 149 visits in grizy from 10 nights in 2014 and 2015: 16,688 CCD images
- LSST Science Pipeline w_2022_12
- Strong lensing positions, magnifications, and time delays simulated using a code adapted from More & More, 2022
- SNe light curves: SALT2 (Guy et al., 2007, 2010) template built in SNCosmo (Barbary et al., 2017) manually incorporating strong lensing observables
SNe light curve simulations
Sky data for DIA
Difference imaging analysis
Template Image
Science Image
Difference Image
Difference imaging analysis
Preliminary results
Template Image
Science Image
Difference Image
Yellow circles of radius 1": injected lensed SNe
Red circles of radius 2": diaSources
Difference imaging analysis
- Total recovery fraction ~70%
- Slightly higher for quads
- Decreases significantly for the y-band
| Injected | Detected | % recovery | |
|---|---|---|---|
| Total | 74k | 51.5k | 70 |
| Doubles | 70k | 48.3k | 70 |
| Quads | 4.1k | 3.2k | 78 |
| g | 310 | 260 | 83 |
| r | 5.4k | 4.7k | 86 |
| i | 14k | 11k | 80 |
| z | 21k | 16k | 76 |
| y | 34k | 20k | 58 |
Lens system level analysis:
Is there a diaSource within the radius of 3" from each lens center in each epoch?
Difference imaging analysis
Preliminary results
Recovery fraction variation with injected unresolved magnitude for each band
- Recovery fraction decreases for fainter systems -> cutoff brightness.
- Cutoff brightness varies across bands.
Lens system level analysis
Difference imaging analysis
Preliminary results
Individual image level analysis:
How many unique diaSources were found corresponding to each lens system?
A single diaSource for entire system: Pipeline could not resolve the system (DUR): ~49k
Two or more diaSources for a system: Pipeline could resolve the system (DR): 1285
Doubles
Quads
vertical dashed line: median seeing
Difference imaging analysis
Preliminary results
Color-magnitude analysis
Quimby et al. 2014
: Unlensed SNIa
: Unlensed core collapse
: Lensed SNIa
: Lensed core collapse
The redder supernovae for given i-band magnitude are more likely to be lensed SNeIa when studied for unresolved photometry on rising phase of light curve.
The black bold curve separates the lensed SNeIa from unlensed both on the rising and the falling edge of light curve for z < 2.4
The criteria hold well for the observed unlensed and lensed SNeIa.
Contamination from unlensed SNe is neglegible.
Contamination from CC SNe
Color-magnitude analysis
Contamination from unlensed SNe is neglegible.
Contamination from CC SNe
Color-magnitude analysis
Primary contaminants are lensed SNe Ib and Ic.
Summary so far...
- Employed difference imaging pipeline recovers ~70% of the injected data of lensed SNe Ia. The recovery fraction decreases for fainter injected systems. It is slightly higher for quads and decreases significantly for the HSC y-band. The resolved fraction is weakly correlated with the angular separation of system for the doubles.
- Colour-magnitude effectively selects lensed SNe Ia for simulated and observed (un-)lensed SNe Ia till redshift 2.4, both on rising and falling edge. Contamination from unlensed CC SNe is low. Primary contaminants are lensed CC SNe Ib and Ic.
Ongoing work
- Combine detections across epochs and study trends.
- Study if extendedness can be used as a marker for a lensed system.
- Next DI run: Incorporate LSST cadence.
- Run it on DC2 data.