Difference between revisions of "How do you find variables?"
From CoolWiki
Jump to navigationJump to searchm |
m |
||
Line 1: | Line 1: | ||
− | ''Based on materials developed by Peter Plavchan, NExScI/IPAC'' | + | ''Based on materials developed by Peter Plavchan, NExScI/IPAC/Caltech'' |
=Getting a light curve= | =Getting a light curve= |
Revision as of 05:40, 12 August 2010
Based on materials developed by Peter Plavchan, NExScI/IPAC/Caltech
Getting a light curve
There are many steps from an astronomical image to a science-quality light curve.
- Reducing the image
- Artifact corrections, biases, flats, etc.
- Aperture or differential photometry
- Generating image catalogs of photometry
- Cross-correlating / rectifying images/catalogs
- Extracting the light curve for each source
- Ensemble Analysis
- Zero point corrections
- Detrending
- Red noise filtering (e.g., removing long slow variations due to, for example, the source rising and the consequent changing airmass)
How do you find variables?
In a word, STATISTICS.
NStED provides a series of standardized variability statistics for all light curves it serves, including:
- Mean, median
- Standard Deviation (aka RMS) w/r/t the median
- Chi-Squared
- Fraction & # of data > 5-sigma deviant from median
- Median Uncertainty
- Median Absolute Deviation Statistic (or MADS)
This is all important, because if you find a characteristic time scale for variations in your light curve, you have a clue to the physical mechanism producing the variability.