Last modified: 3 August 2023

URL: https://cxc.cfa.harvard.edu/csc/data_products/master/blocks3.html

Master Sources Bayesian Blocks File (blocks3.fits)

The Bayesian Blocks file is a single fits format file for each Master Source, that records the photometric and spectroscopic properties from blocks of observations within which the source may be considered to have a constant photon flux. The algorithm of Scargle et al. (2013, ApJ 764 167) is used to determine the blocks, using the marginalized posterior probability distributions for photon flux from individual observations as the input data cells. For ACIS observations, blocking is performed separated in the broad, soft, medium, and hard science bands, and the final blocks are determined from the intersection of the individual band blocks. Blocks for HRC observations are determined separately.

Bayesian Blocks

[Thumbnail image: Bayesian blocking analysis]

[Version: full-size]

[Print media version: Bayesian blocking analysis]

Bayesian Blocks

An example of the blocking analysis is shown in the accompanying figure, in which each panel displays the probability distribution for all observations of the source in, from top to bottom, the broad, soft, medium, and hard bands. Individual blocks are color-coded.

FITS HDUs (CXC Data Model blocks):

  • FLUXBLKS: photon fluxes from individual observations are sorted in increasing flux order for blocking analysis.
  • TIMEBLKS: photon fluxes from individual observations are sorted in increasing time order for blocking analysis.

Each HDU will contain one row for each block, containing all aperture photometry and spectral results for the combined data in the block, as well as a list of the obsids included in the block.

The header for each HDU will include a BEST_BLK keyword identifying the row containing the block with the largest total exposure.

The Bayesian Blocks FITS file is named: 〈m〉N〈v〉_blocks3.fits

where 〈m〉 is the Master Source name, formatted as "2CXOJhhmmsss{p|m}ddmmss[{<null>|A|B|C|...|X}]", and 〈v〉 is the data product version number.