Re: Transit Fitting plus referencing software in paper

Posted by karenacollins on Dec 17, 2014; 6:39am
URL: http://astroimagej.170.s1.nabble.com/Transit-Fitting-plus-referencing-software-in-paper-tp199p202.html

Yes, the AIJ fitting panel is designed to be useful for fitting partial
transits as well. As a disclaimer, it is sometimes difficult to extract
precise system parameters from partial transits. It helps to know the
expected transit ingress and egress times (i.e. Tc and the duration). If
these values (and the orbital period) are known, the prior center value
of a*/R* can be set to 'lock' and then the value  adjusted until the
transit duration matches the expected value.

For partial transits, locking u1 and u2 to theoretical values will help
a transit fit to converge to a reasonable model.

Reducing detrending variables can also help partial transit fits to
converge to a reasonable model. With too many detrending values, the
light curve may prefer a ~flat model instead of a transiting exoplanet
model. It might be best to start with no detrending or airmass only
detrending for partial transit observations.

The 'Prior Width' setting can be used to limit the range of a particular
parameter when the best fit is found (click 'Use' next to the box to
enable user input). For instance, if there is prior information to
suggest the the transit depth should be 1%, a reasonable approach might
be to set the prior center value of (Rp/R*)^2 to 0.010 (but not locked),
and then set the prior width to some value (say 0.003, which would limit
the depth range from 0.7% to 1.3%).

To set a specific prior center value for Baseline flux, (Rp/R*)^2, a/R*,
or Tc, either disable 'Auto Update Priors' which will enable all 4 boxes
for user entered setting for each of the parameters, or leave 'Auto
Update Priors' enabled and deselect one or more items in the 'Auto
Priors' menu to allow a user setting for individual parameters (i.e.
this disables the auto setting that is estimated from the light curve
data and allows the user to enter the value).

Karen