README

TESS Transients

This webpage hosts TESS light curve data for transients reported to the Transient Name Server (wis-tns.org) that were also observed in TESS full frame images.

As of September 2022, TESS full frame images (FFIs) are released every seven days as a High Level Science Product (TICA) on the MAST. To capitalize on the rapid release of data, we are extracting light curves for transients reproted to TNS every 7 days and posting reduced data products here.

Only transients discovered in the TESS field of view while TESS was observing are presented. We plan to add prediscovery light curves in the future. A subset of confirmed supernovae per sector is also available.

How to get the data

Light curves are organized by sector. For a given sector, you can see all the light curves by navigating to that sector's webpage.

Above each figure is a link to a text file with the light curve data.

To find a specific transient, you need to know which sector it was observed in. You can use a tool like tess-point to find which sector a given transient was observed in based on its RA and Dec.

There is also an API to download individual light curves, for example wget https://tess.mit.edu/public/tesstransients/light_curves/lc_2022sfe_cleaned. In this case, you only need to know the IAU designation of the transient (replace "2022sfe" with the IAU designation). A GET request returns the data as a string.

For bulk downloads, you may prefer to visit the bulk downloads page. However, the data are also under version contol through this github repo, which you can use to access older versions of the data and/or keep local copies synced with the most up-to-date version.

Light Curve Files

Light curves are saved as text files, with 10 columns and two header lines. The header lines start with the # character.

Header

• The first line is labeld "reference_flux:" and gives the flux calibration offset measured from the reference image in units of counts per second (see Flux Calibration below). This number can be subtracted from the cts_per_s column to get the original differential light curve.

• The second line gives the labels for the columns. Many libraries can use this line to put names/labels/tags on the data structure. For example, with numpy you can do d = np.genfromtext('lc_2022sfe_cleaned',skip_header=1, names=True) and you will be able to access the timestamps and flux light curve with d['BTJD'], d['cts_per_s'], etc. (The skip_header=1 is necessary to ignore the reference_flux line.)

Columns

Image Processing and Light Curves

We use image subtraction and forced PSF photometry to produce the light curves. We are using a slightly customized version of the ISIS software package.

For image subtraction, we build a reference image from 20 images with low backgrounds. The reference image is essentially a clipped median. Starting in 2022 September, the 20 images are from the first 7 days of a TESS observing sector, because this time period coresponds to the first downlink of TESS data in each sector. Image subtraction is performed every 7 days, and begins as soon as the FFIs are publicly available at MAST.

We use the coordinates reported to TNS for forced photometry on each transient. We fit models of the pixel-response function (available on MAST) to the differential flux of the transient in each subtracted image. Uncertainties are estimated from photon statistics (source + background) in the original calibrated image. There is evidence that these uncertainties are underestimated by a factor of 1.2 to 2.0, depending on the brightness of the source and the surroudning scene.

More details are in Fausnaugh et al. 2021 and Fausnaugh et al., submitted.

Flux Calibration

To convert a differential flux light curve (delta-counts per second from subtracted images) to a flux-calibrated light curve (magnitudes), we estimate the flux of the source in the reference image. The flux in the reference image is added as a fixed offset to the differential light curve to obtain a flux light curve in units of counts per second. This flux light curve is given in the cts_per_s column in the light curve files. The reference image flux is given in the file headers. To convert to magnitudes, we use the formula in the TESS instrument handbook, -2.5log10(cts_per_s) + 20.44. The zeropoint has an uncertainty of 0.05 mag.

The advantage of this approach is that the magnitude units are easy to understand, and can be easily converted to physical units: (2416 Jy)*10^(-0.4*Tmag). This formula uses the Vega-mag zeropoint for Cousins I-band, which the TESS photometric system is defined to match.

There are two major disadvantages with this approach.

First, absolute photometry in TESS images is very uncertain. Uncertainty in the reference image flux is typically 20%, and can be greater than 100% for faint sources at 18th magnitude and fainter.

The main issue is that it is very difficult to estimate the correct background level due to the the large TESS pixels and high stellar density. There is evidence that background corrections estimated by traditional techniques tend to be overestimate, resulting in underestimated source fluxes. For example, the TESS mission pipeline applies an ad hoc correction for overestimated backgrounds (described in section 4.2 of this data release note document). Crowding from nearby stars that contaminate the photometric aperture also affects the flux calibration. We iteratively fit models of the PSF to the reference image in the vicinity of the transient to help mitigate crowding. However, this procedure is also highly uncertain and contributes to the large uncertainty in the reference image flux.

The second disadvantage of our flux calibration procedure is that it includes host galaxy flux in the light curve of extragalactic transients. For example, bright supernova may appear to peak at 13th or 14th TESS magnitude, but only because the integrated host galaxy light dominates the flux in the reference image.

For these cases, users should identify the light curve baseline and shift the light curve so that the baseline is forced to zero flux. We have chosen not to do this because identifying a suitable baseline is not always obvious, and may not be desirable for CVs, stellar flares, or unusual variability from other sources.

Users can also recover the original differential light curve by subtracting the reference_flux value in the light curve file headers from the cts_per_s column.

Lastly, users can calibrate the TESS cts_per_s light curve to match a light curve in physical units (such as from ASAS-SN, ATLAS, or ZTF) by fitting for a shift and scale factor. The TESS filter spans 600 nm to 1,000 nm, and so there may be non-negligble color terms when comparing to light cures in other filters.

Upper limits

When converting to magnitudes, we only use data points with signal-to-noise ratio = cts_per_s/e_cts_per_s > 3.0. Points with a signal-to-noise ratio below 3, including all negative fluxes, are converted to magnitude upper limits with the formula -2.5log10(3*e_cts_per_s) + 20.44. Negative fluxes are due to either an overestimated background or statistical noise that fluctuates below the background level. Upper limits are marked in the light curve files by setting the magnitude uncertainty to 99.9.

For low backgrounds, a typical 3-sigma limiting magnitude is about 19.6 in 30 minutes (secotrs 1--26), 19.0 in 10 minutes (sectors 27--55), and 18.4 in 200 seconds (sector 56--83).

Background Model

The main systematic error in these light curves are caused by time-variable scattered light from the Earth and Moon when they are above the TESS sunshade. In the worst cases, the sky background is brighter than 14th TESS magnitude per pixel, while the early time transient light curves are often fainter than 19–21st magnitude. When the background is this high, the background corrections must have a relative uncertainty less than 0.1% to isolate the transient signal. In practice, we have found that strong glints with high-frequency spatial features are the most difficult systematic error to remove.

We provide an estimate of residual background errors in the photometric aperture in the bkg_model column of the light curve files. Subtracting this columns from the cts_per_s light curve, in some cases, can improve the light curve. We show the background model as a purple line in the figures for each light curve, and the "background-model corrected" light curve (converted to magnitudes) in the bottom panel of each figure.

To produce the bkg_model column, we filter the subtracted images to remove point sources such as the transient and nearby variable stars. We used a median filter with a width of 100 pixels, and we applied the filter to the subtracted images first column-by-column and then row-by-row. We then produced the "background model" light curve by rerunning forced photometry on the filtered images at the position of the transient. The results of the forced photometry, and its uncertainty, are given in the bkg2 and e_bkg2 columns of the light curve files. Finally, we take a running median of the bkg2 column to reduce the statistical noise to produce the bkg_model columns.

Further details are in Fausnaugh et al., submitted.