All-Sky Automated Survey for Supernovae

How to Properly Use the ASAS-SN Public Light Curve Interface

This effort is primarily funded by grant GBMF5490


UPDATE (September 2021): we are now providing three photometry methods: aperture photometry (as before), image subtraction photometry with reference flux added (new) and image subtraction photometry without adding reference fluxes (new), see below for details.

UPDATE (September 2018): we are now serving both V-band (as before) and g-band (new) light curves, see below for details.

This website serves aperture photometry performed on request on ASAS-SN images. For details on how the photmetry is done, see Kochanek et al. (2017, PASP).

When using ASAS-SN light curves in publications, cite both Shappee et al. (2014) and Kochanek et al. (2017).

WARNING: V-band (cameras ba-bh) and g-band (Cameras bi-bt) magnitudes obtained via our interface are calculated upon query using aperture photometry, with zero-points calibrated using APASS catalog. ASAS-SN images have 8" pixels (~15" FWHM PSF,) so blending/crowding might be very significant in some cases. Proceed with caution before publishing---see "Examples" below, and also in the Kochanek et al. paper.


Make sure you press the "I'm not a robot" button.
DO NOT attempt to use this website via an automated script!

Note that your search will produce a permanent http link with the light curve plot and the data, which you can share, so you do not need to run the script again to get the same light curve.

________________________________________________________________________________________________

Example "-3": Now also serving image subtraction data (September 2021)

Retrieved on September 7th, 2021

* Source: ASASSN-21kl (Transient, SN Ia)
* Enter right ascension: 56.125667
* Enter declination: -44.63250
* Enter number of days to go back: 160

Light curve (aperture photometry):
(click to enlarge).

Light curve (image subtraction with reference flux added):
(click to enlarge).

Light curve (image subtraction without reference flux added):
(click to enlarge).

Notes:
Aperture photometry: This does aperture photometry directly on the individual images. This is the original Sky Patrol photometry method.
Image Subtraction (reference flux added): This does aperture photometry on the coadded, image subtracted data for each epoch and then adds in the flux of the source on the reference image.
Image Subtraction (No reference flux added): This does aperture photometry on the coadded image subtracted data for each epoch but does not add the flux of the source on the reference image to the light curve.

Because the image subtracted light curves use the coadded data for each epoch (usually 3 images) while the aperture photometry light curves used the individual images, there will be fewer points but with deeper flux limits. When the reference flux is not added, the light curves will have both positive and negative fluxes because it is the change in flux relative to the reference image. Because it can be both positive and negative, these light curves can only be defined by flux and not magnitudes. ________________________________________________________________________________________________

Example "-2": Now also serving image subtraction data (September 2021)

Retrieved on September 7th, 2021

* Source: ASASSN-V J160939.56-350200.7 (Variable star, Mira)
* Enter right ascension: 242.414550
* Enter declination: -35.033580
* Enter number of days to go back: 2000

Light curve (aperture photometry):
(click to enlarge).

Light curve (image subtraction with reference flux added):
(click to enlarge).

Light curve (image subtraction without reference flux added):
(click to enlarge).

Notes:
Aperture photometry: This does aperture photometry directly on the individual images. This is the original Sky Patrol photometry method.
Image Subtraction (reference flux added): This does aperture photometry on the coadded, image subtracted data for each epoch and then adds in the flux of the source on the reference image.
Image Subtraction (No reference flux added): This does aperture photometry on the coadded image subtracted data for each epoch but does not add the flux of the source on the reference image to the light curve.

Because the image subtracted light curves use the coadded data for each epoch (usually 3 images) while the aperture photometry light curves used the individual images, there will be fewer points but with deeper flux limits. When the reference flux is not added, the light curves will have both positive and negative fluxes because it is the change in flux relative to the reference image. Because it can be both positive and negative, these light curves can only be defined by flux and not magnitudes. ________________________________________________________________________________________________

Example "-1": Now also serving g-band data (September 2018)

Retrieved on September 24th, 2018

* Enter right ascension: 299.66840
* Enter declination: 8.37033
* Enter number of days to go back: 30

Light curve: (click to enlarge).

Significant number of measurements in both V-band (top) and g-band (bottom). Return page at this address gives camera names, V-band is cameras "ba" to "bh", and g-band is "bi" to "bt". ________________________________________________________________________________________________

Example 0: Random spot on the sky, no data taken in the last 20 days ("behind the Sun")

Retrieved on June 19th, 2017

* Enter right ascension: 04 00 00.0
* Enter declination: +30 00 00.0
* Enter number of days to go back: 20

Output: "No data was found for this coordinate and time window."

________________________________________________________

Example 1: Random star, not known to be variable

Retrieved on June 19th, 2017

* Enter right ascension: 05 44 17.056
* Enter declination: 75 13 31.44
* Enter number of days to go back: 100

Light curve: (click to enlarge).

Significant number of measurement (filled points with errorbars), with some upper limits. Data from one CCD only. Note that the last data point is from April 2017, as this position is now "behind the Sun".

_______________________________________

Example 2: Random spot on the sky

Retrieved on June 12th, 2017

* Enter right ascension: 17:00:00
* Enter declination: -15:00:00
* Enter number of days to go back: 100

Light curve: (click to enlarge).

As expected, mostly upper limits, with one (spurious) data point. One can see the influence of the Moon phase on the depth of the photometry (but one can go deeper by co-adding 3 consecutive 90-sec exposures).

________________________________________________________

Example 3: Known Long Period Variable (V814 Sco)

Retrieved on June 12th, 2017

* Enter right ascension: 244.07975
* Enter declination: -12.78665
* Enter number of days to go back: 100

Light curve: (click to enlarge).

This variable star is present in two different CCD chips ("ba" in Hawaii and "be" in Chile), and they complement each other well. Below we show the complete ASAS-SN light curve for the same variable (it takes several minutes to compute the full light curve, so you should plan accordingly). Note that the full light curve comes from 4 different CCD chips, but the majority of data are still from "ba" and "be".

Full light curve: (click to enlarge).

_____________________________________________

Example 4: Galactic Nova ASASSN-16ma

Retrieved on June 19th, 2017

* Enter right ascension: 18:20:52.12
* Enter declination: -28:22:13.5
* Enter number of days to go back: 1000

Light curve: (click to enlarge).

We started observing the entire Galactic plane in 2016, hence shorter coverage here. You are first seeing a number of upper limits (less deep in the Galactic plane), followed by the nova caught still on the rise. After that there is the seasonal break (field too close to the Sun), with the nova still present in our data right now (June 2017).

____________________________________________________________________________

Example 5: Bright Galactic Cepheid VY Car, saturated in ASAS-SN data

Retrieved on June 19th, 2017

* Enter right ascension: 161.136213
* Enter declination: -57.565366
* Enter number of days to go back: 300

Light curve: (click to enlarge).

This is an example discussed in the Kochanek et al. (2017) paper: nominally our bright saturation limit is between 10th and 11th magnitude (V-band, TBD for future g-band data). However, we employ a "saturation correction" procedure, so even for object significantly brighter than our saturation limit useful variability information can be obtained. Indeed, one can see a clear Cepheid-like variation in the light curve above. However, for a number of images the correction did not work/was not applied, resulting in spurious measurements roughly at V~13. So again, proceed with caution.


See below our sky coverage plot for the last 365 days - we are observing the entire sky!


If you have questions/comments/suggestions, e-mail Josh Shields (shields.217@osu.edu) and Kris Stanek (stanek.32@osu.edu) (but study the examples above first).

Updated Mon Sep 24 16:30:37 EDT 2018