IRAC Data Challenge 2015

To assess the repeatability of results, we are making available two exoplanet eclipse datasets on which to test your reduction techniques. This challenge focuses on the topic of whether or not different methods measure the same eclipse depths for a set of repeated observations. We encourage even those who will not be able to attend the IAU 2015 IRAC Workshop to try your hand at reducing the data, and submit your results. All submissions will be presented at the Workshop.
The first dataset contains 10 archival Spitzer observations of XO-3b during eclipse. The second dataset has been made with the IRAC data simulator to approximate 10 observations of XO-3b, with realistic correlated noise based on a model of spacecraft pointing, and with a known input light curve. Each of the ten observations represents a separate visit to XO-3 b and consists of 2 AORs, a 30-min pre-AOR (to accomodate short-term pointing drift) and a ~8.5 hour staring mode AOR with 2s exposure times, observed in sub-array mode with IRAC Channel 2 (4.5 microns). All real AORs have been pointed using PCRS Peak-Up to enhance sub-pixel pointing repeatability, using XO-3 as the target. (Simulated AORs have sub-pixel starting locations governed by PCRS Peak-Up statistics.)

Data Formats

The data for this challenge can be found in two forms: (1) tables of derived photometry and centroids, and (2) the usual FITS format Basic Calibrated Data (BCD) images. You are welcome to try your methods on both forms of data, but we request that teams entering the challenge at least analyze the tables so that we can separate the correction of instrument systematics from differences in photometric pipelines.
There are two CSV files for each AOR and provenance (real and simulated). Files include headers with column names. Columns include bmjd, flux at a number of different aperture sizes (2, 2.25, 2.5, 2.75, & 3.0 pixel radii), and flux uncertainties. Files labeled “XO3_r*_phot_ch2_pixvals.txt” contain the same basic information as the “XO3_r*_phot_ch2.txt” files, but also contain pixel values for the central 9x9 pixels of the Channel 2 subarray field. In the naming scheme of the headers pixel 0-0 corresponds to pixel [11,11] on the subarray (assuming that the bottom left corner has coordinates[0,0]).
An IDL routine for reading the CSV files can downloaded here: This can be executed using "a=READ_CSV_JIM('filename.csv')", and produces a structure "a" with tags equal to the column titles in the CSV file.

XO-3 b Datasets

Provenance CSV Tables Size
FITS Images Size
Real XO3_archival.tar.gz 67 MB XO3_ch2_eclipse (Spitzer Archive) 1.2 GB
Simulated XO3_simulation.tar.gz 69 MB XO-3_b_sim_ch2_150702.tgz 1.33 GB

"Truth" Light Curves for Simulated Data

Model light curves for XO-3 b that we used in the simulations are available at XO-3_b_sim_ch2_150702_predicted_lightcurves.tgz. The .tgz file unpacks to 20 ASCII whitespace-delimited text files, one corresponding to each AOR in the simulation.  Each file has a header explaining the details of the model, as well as two columns of data listing the time in days since the AOR start, and flux relative to the stellar flux.


To submit results to the data challenge please send ASCII text files with columns delimited by whitespace. Comment lines may be indicated by the pound sign “#” as the first character. For each Observation (pair of AORs), please submit the following:
  1. Corrected light curves in the form of a table with Time, Corrected flux relative to stellar flux, and Uncertainty in the relative flux.
  2. [OPTIONAL] A model exoplanet light curve. The text file should have columns for Time and Flux relative to stellar flux.
  3. Eclipse depth and its uncertainty for each AOR
  4. One slide describing your method and how you got your result, to be shown at the IAU workshop (pdf preferrable).

Uploading Results

If you want your results to be discussed during the IAU workshop, they are due Monday 27 July, 2015 (slides for the workshop can be submitted until August 6). Please upload your results to the Caltech Dropbox ( Under "Recipient Email Address", enter If your results are contained in multiple files, you will need to perform multiple uploads. Alternatively, you can combine multiple files into a compressed archive (zip or tar + gz format.)


Please email all questions or problems with the datasets to iracaas224 @ ipac . caltech . edu. We will try to respond within 24 hours.