WCS Time Coordinates as Applied to Solar Physics

There is a new draft paper out which discusses how to incorporate time into the FITS World Coordinate System. Here is the message sent to FITSBITS announcing the latest draft:

Version 0.90 of the FITS WCS Paper V on Time is now available for review:


This draft is in a near-final form and is put out for community comments before we start formal reviewing. It is available in PS and PDF format.

We request that comments be directed to the FITSWCS discussion list:

Thank you,

  - Arnold Rots

As one of the co-authors of this paper, I wanted to bring this to the attention of the solar physics community, and share some of my own thoughts as to how this applies to our field.

First of all, like all of the other WCS papers so far, it can seem a little intimidating because it's intended to serve a wide range of communities with differing needs. It has a lot of bells and whistles, but not all situations require all of the available features. Basically, it's very simple. Time axes are expressed as an offset from a scalar time reference. Typically, the time reference is given in ISO-8601 format, e.g.

DATE-OBS= '2011-10-14T19:03:05.540'
which has been in common use in the solar physics community since the launch of SOHO. When one of the axes is time, then the CRPIXi, CRVALi, and CDELTi values are defined to return an array containing the number of seconds (CUNITi='s') since the reference time, which is also common practice. The draft WCS paper simply codifies and refines this common practice.

There are a number of keywords that are commonly used in our community to express times related to the observation time. For example, EXPTIME is typically used to express the exposure time of the observation. Unfortunately, this keyword has not been consistently used across all missions. For example, the STEREO/SECCHI Heliospheric Imager (HI) telescope observations are built up from a series of short exposures which are then summed onboard into one long effective exposure before sending down to the ground. This is done to remove cosmic rays. The EXPTIME values for HI reflect the total summed exposure time of all the individual exposures. However, some other (non-solar) missions which do the same kind of image summing have used EXPTIME to give the exposure time of an individual exposure. Thus, it was decided to define a completely new keyword XPOSURE to express the effective exposure time for the data, corrected for dead time and lost time.

It should be emphasized that just because there's a new standard for encoding time (which is currently only in draft form), this does not mean that current FITS files are non-compliant, or that existing observatories need to start using the new keywords. Instead, this system should be considered as a tool, and one should ask oneself, does this tool meet all of my needs? Some parts of the solar physics community may have more stringent requirements than others. For example, analysis of flare timing may require knowledge of the exact position of the observatory, or one may wish to express times as heliocentric rather than topocentric. Is the methodology in this paper sufficient to address these and other time-related issues?

Please address any comments to the FITSWCS discussion list rather than to myself.

Author: William Thompson
Responsible Official:
Last revised: 14-Oct-2011