search

California-Planet-Search / grandsol

Grand Solution - Doppler forward modeling
MIT License
0 stars 0 forks source link

Normalize lines on MAD by iter plots #29

Closed bjfultn closed 8 years ago

bjfultn commented 8 years ago

Need to normalize all lines on the MAD by iter plots to see how the residuals are changing as a function of iGrand iteration.

bjfultn commented 8 years ago

This is complete. Here is an example. The iGrand_sun_resMAD_byiter.pdf plots from the recent solar simulation are updated in our shared Box folder.

iGrand_sun_resMAD_byiter.pdf

JeffValenti commented 8 years ago

Looking at the plot, i see that the MAD of the residuals does not change significantly after iGrand iteration 4. Some orders get slightly better, others get slightly worse. This suggests that the increasing spread in velocity errors seen in later iGrand iterations is not caused by the solver finding a better spectrum fit. Instead it seems that there is a degeneracy that allows equivalent fits with a range of parameter value. We should perhaps consider an additional constraint, for example an LSF centering constraint. First we need to see if Jay's work on the wavelength solution fixes the diverging velocity errors without adding a constraint.