Does Lightkurve smooth-out or minimize intrinsic variability? #1252
Unanswered
Brad-Schaefer
asked this question in
Q&A
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
-
Does Lightkurve smooth-out, minimize, or remove instrinsic variability in the light curve of a target star in TESS light curves? I am asking because my Lightkurve extracted light curves for four TESS sectors all have the target showing only small amplitude chaotic variability with a time scale of one to ten days with full peak-to-peak amplitudes of 0.004 mag. This is a trouble because I have four other unimpeachable light curve sources that show the same target to have chaotic variations on the 1-10 day timescale -- but with full amplitudes of 0.11--0.20 mag. It is very unlikely that the target knew to turn off most of its variability just for the four TESS sectors and then switch back on for my other data sources. Two of my alternative data sources are contemporaneous with the TESS sectors, and they have similar spectral sensitivity, so we have no explanation from the amplitude due to long-term change in the variability or to differences due to color. With four independent data sources giving 0.11--0.20 mag amplitude variability, yet the LightKurve light curve giving only 0.004 mag amplitude, I have to be concerned that something is wrong with the Lightkurve photometry. Further, the SPOC light curve for sector 52 shows variability at the 0.2 mag level, while the Lightkurve product shows only Poisson noise below the 0.004 mag level for the exact same input data.
OK, let me give some details that people can use for consideration. My target is IRAS 00500+6713, which is the remnant left behind by the supernova seen by Chinese and Japanese observers back in the year 1181 AD. My science question is to find a coherent periodic modulation from ~40 seconds up to ~10 days, or to place an upper limit on any such periodic modulation. (It turns out that a severe limit on any periodic modulation from an orbital period is the last and biggest constraint that eliminates one of the two competing models, and this solves the case. Or is there really an orbital period in the TESS data that has been minimized by Lightkurve? So my final science conclusion hangs on whether there is a strict limit on any orbital modulation as seen by TESS.) This target is a roughly 15th mag isolated star. With MAST, the target position is about two arc-seconds off of the central star's coordinates. The J2000 coordinates of the target star is 00:53:11.21 +67:30:02.4. The target was well measured with TESS sectors 18, 24, 25, and 52. The first three sectors only have FFI data at 1800s time resolution. Sector 52 has 20 TPFs, while MAST contains SPOC light curves with 20-second and 120-second time resolutions. I have run Lightkurve for the FFI data, for the PLD, PLD2, and RC correctors. Here is the plot of sector #18 from Lightkurve:
The point of this plot is to see that the variability returned is ~0.1% for this sector. To see what my photometry aperture is, here is the plot:
I have also run the same code with a shifted 3x3 box to a background area, and the average PDCSAP flux is 1101. The PDCSAP flux from the target with the RC corrector is 2660 with no background subtraction.
Again, we see that Lightkurve gives only a very small variability (consistent with the Poisson noise) of up to 0.4%. (Again, the question in hand is whether this variability in Lightkurve has somehow removed much of the intrinsic variability?) The SPOC light curve for sector 52 shows a large amount of noise with 20-seconds time resolution, but when binned up to 1800 seconds, the light curve shows variations of 0.2 mag. This directly contradicts the amplitude of variability from Lightkurve for the exact same data. So here we have a strong pointer that Lightkurve is somehow taking out intrinsic variability in the light curves it produces.
Beta Was this translation helpful? Give feedback.
All reactions