Month: July 2016

Processing Models

So far processing the models have been pretty successful, but there seems to be a problem with the cross correlation function for calculating the required shift. So far all values get a zero shift EXCEPT when I do not normalize the function. Any help would be appreciated. In the meantime I will continue to develop the code under the assumption it is right.

Everything up to the cross correlation works as it should. But the output coming from the cross correlation function is showing a shift of zero no matter the shift applied to one of the model spectra’s star.

Below is all the spectra that’s used, and the code can be found here.

At first I thought it may have been my normalization function, but it appears to be working properly as seen below:

norm (line 119): The array to use to normalize the result of the cross correlation function.

The totalFlux was also verified to be working with the rest and shifted flux being added together as shown below:

totalFlux[0] (line 113): The combined models of a rest star and one shifted using martins rv tables to determine the velocity difference and added together.

Cspec was taken from aspcap so I can only assume it is correct.

cspec (line 116): The official continuum-normalized spectra of the binary from aspcapStar. Used to cross correlate against the totalFlux. Currently plotting locID=4587, apogeeID= 2M03250916+3126099.

vel (line 122): The result after cross correlating the totalFlux and cspec.

velNorm (line 123): The result after cross correlating the totalFlux and cspec and normalizing it.

Model Generation

Currently what I am attempting to do is to generate a spectrum from FERRE that represents a binary system and see how well it fits the observed spectrum.

So far I’ve generated the spectra of the two stars and shifting them by using the velocities generated by martins SB2 code.

The next step is now fitting the the model spectra of the individual stars with the observed spectra.

To do this I will create a “model temperature grid” holding a range that will be tested against the observed spectra and continue to narrow that range until we have a proper approximation.

Environment Setup Complete

The environment setup is now complete. The last of my problems in the setup of the apogee python package were:

was downloading a working gfortran compiler from GCC (the previous one did not work for some reason).
As well as the setup.py file needing to be updated to download the most recent version of FERRE as the previous version was no longer available. (setup.py: line 56, http://www.as.utexas.edu/~hebe/ferre/)

So far I’ve been playing with the package. Mainly the FERRE wrapper.

To do this, I generated gifs of changing parameter(s) values while holding the rest of the generated values static. Some examples to of this are below (click for larger view).

The purpose of these gifs were to get use to FERRE and just see what the data looks like. To get gifs of multiple parameters being dynamic, it’d take a little more work if we would like to go in that direction. The main question would be what order to display the plots when multiple parameters are changing.