Astrometry 101 - Timing is the most critical parameter

The time associated with a precise astrometric position of a moving object in an image such as an asteroid or comet is the most critical observer controlled parameter. The position of the object can be derived very accurately from the relative positions of several dozen reference stars. But the time associated with that position is absolutely necessary to achieve small residuals less then 1 arc second. Timing errors of are probably the most serious type of astrometric error.

The faster a object is moving in the sky, the more critical this becomes.

Timing is of little importance for non-moving objects or objects whose image properties are not changing quickly with time, such as a supernova or a variable star or almost any other stellar object. Quasars would be an exception to that.

The accuracy required to obtain low residuals of comet or asteroid observations requires that the image capture program perform this time measurement automatically. The time measurement should be stored in the image header, so that it will not be altered by the image calibration process. The date-time-stamp given to an image file when it is saved by the computer operating system is not good enough for precise astrometry of moving objects.

SBIG cameras store the time of the start of image exposure and exposure duration in the header of either SBIG formatted images or fits images. It is intentionally set up so as to make modification of that difficult or tedious. Generally it should never require manual edit.

These are the lines containing keyword-value pairs in a SBIG image header which record date, time, and duration of exposure.


Exposure = 36000
Date = 12/28/09
Time = 11:23:44

Duration of exposure is in hundredths of a second. 36000 hundredths of a second.= 360.00 seconds
Note the precision.
The time of mid-exposure is then calculated by an astrometric software program.  In this case 180.00 seconds is added to the start time of 11:23:44 or 11:26:44 on 12/28/09 UT. The time of mid-exposure is the official time of an observation. It is this time along with the position that is submitted to the Minor Planet Center. The date and time is formatted as YYYY MM.DD.FFFFF where YYYY is 4 digit calendar year, MM is month o year, January is 01 etc., DD is day of month, and FFFFF is fraction of the day. eg: 2009 12 28.47690

There are many ways to synchronize a computer operating system with a time standard such as a NIST time server. For most astrometry of comets and asteroids, a program such as Dimension 4 set to update periodically will do the job. For very fast moving objects, the timing of the shutter on the camera can affect the astrometry. That would be a  more advanced topic.

For example, consider an object moving at 0.01 arc sec/second, a typical value for many bright comets and asteroids. A 5 second time error will then cause a  0.05 arc sec residual. For an object moving at 0.1 arc/second, a 5 second time error will then cause a  0.5 arc sec residual (too much). On the other extreme, for example Pluto, a slow mover, moving at 0.0005 arc/sec, a 5 second time error will then cause a  0.0025 arc sec residual which is probably negligible.

An observer should be very familiar with their image scale (or pixel resolution) and the time it takes for an object to travel one pixel in distance. One should understand quantitatively what the limits of your timing system and image acquisition are. Astrometry has no room for sloppiness. Precision time is essential.

Dimension4 software is freely available from the Thinkman software web site at:
 http://www.thinkman.com/

The synchronization history can tell you how far off your operating system gets off during times that the computer is off as well as how it varies with your computer usage. Performing CPU intensive tasks may be affecting time synchronization.

Astrometrica is a research grade shareware program astrometry:
http://www.astrometrica.at/

Guide to Minor Body Astrometry: http://www.cfa.harvard.edu/iau/info/Astrometry.html


Pixel scale in arcseconds = (206.265) * (pixel size in microns) / (focal length in mm)

Pixel time = pixel scale / target rate

If pixel scale is 2.5 arc sec per pixel, and an object is moving at 0.01 arc sec/sec, then it takes

2.5 /0.01 seconds or 250 seconds to move one pixel.

If pixel scale is 0.5 arc sec per pixel (very small), and an object is moving at 0.01 arc sec/sec, then it takes

0.5 /0.01 seconds or 50 seconds to move one pixel.

If pixel scale is 2.5 arc sec per pixel, and an object is moving at 0.10 arc sec/sec, then it takes

2.5 /0.1 seconds or 25 seconds to move one pixel.

If pixel scale is 0.5 arc sec per pixel (very small), and an object is moving at 0.1 arc sec/sec, then it takes

0.5 /0.01 seconds or 5 seconds to move one pixel.

From these examples, one can deduce that the smaller the pixel scale, the less time it takes to travel one pixel. This also means that the smaller the pixel scale, the less time you have for your CCD chip to integrate the light from the target to reach any given magnitude. The shorter your exposure limits, the more critical time becomes.

RESIDUALS STATISTICS FOR OBSERVATORY CODES: 

http://www.cfa.harvard.edu/iau/special/residuals.txt

The IAU MPC provides guidance on how to get an accurate time:
http://www.cfa.harvard.edu//iau/info/Astrometry.html#time