Getting started#

Contents#

Opening example spectra#

If you want to do this with your own data, make sure it is reduced by following the steps in IRAF

Example ARCES spectra are available online for you to work with. You can download them via Python like this… this will download temporary copies of the famous Kepler target KIC 8462852, also known as Boyajian’s Star, and spectroscopic standard O star BD+28 4211. We also create an EchelleSpectrum object for each star:

>>> from astropy.utils.data import download_file

>>> target_url = 'https://stsci.box.com/shared/static/mu4fa1fmq1lw8boem12e2umyi99skbdl.fits'
>>> spectroscopic_standard_url = 'https://stsci.box.com/shared/static/18fa008byy2500yrfhuooyxs5d6pwc6e.fits'

>>> target_path = download_file(target_url, show_progress=False)
>>> standard_path = download_file(spectroscopic_standard_url, show_progress=False)

>>> from aesop import EchelleSpectrum

>>> target_spectrum = EchelleSpectrum.from_fits(target_path)
>>> standard_spectrum = EchelleSpectrum.from_fits(standard_path)

You can check basic metadata for an EchelleSpectrum object by printing it:

>>> print(target_spectrum)
<EchelleSpectrum: 107 orders, 3506.8-10612.4 Angstrom>

The EchelleSpectrum object behaves a bit like a Python list – it supports indexing, where the index counts the order number, starting with index zero for the order with the shortest wavelengths. Elements of the EchelleSpectrum are Spectrum1D objects. Suppose you want to make a quick plot of the 73rd order of the target’s echelle spectrum:

order73 = target_spectrum[73]
order73.plot() # doctest: +SKIP

(Source code, png, hires.png, pdf, svg)

_images/getting_started-1.png

You can see the H-alpha absorption in this O star at 6562 Angstroms.

Normalizing your spectra#

You can continuum-normalize your echelle spectra with two methods:

  • continuum_normalize_from_standard will remove the blaze function from each echelle order by fitting polynomials to the continuum of each order of a spectroscopic standard star, and then remove those polynomials from each order of the target star

  • continuum_normalize_lstsq will attempt to remove the blaze function from each echelle order by using a robust least-squares fit to the continuum in each order of the target spectrum – no spectroscopic standard observation is required by this method.

Let’s first use continuum_normalize_from_standard on our previous example to see order containing the H-alpha line after the blaze function has been mostly removed:

>>> target_spectrum.continuum_normalize_from_standard(standard_spectrum,
...                                                   polynomial_order=8)
>>> target_spectrum[73].plot()

(Source code, png, hires.png, pdf, svg)

_images/getting_started-2.png

As you can see in this example, the standard star normalization will approximately flatten the continuum, but not normalize it to unity. We can now flatten the continuum and normalize it to unity with the other continuum normalization method, continuum_normalize_lstsq:

>>> target_spectrum.continuum_normalize_lstsq(polynomial_order=2)
>>> target_spectrum[73].plot()

(Source code, png, hires.png, pdf, svg)

_images/getting_started-3.png

Merge all orders into a 1D spectrum#

Now that you have a great normalized echelle spectrum, let’s collapse the echelle spectrum down to one, big 1D spectrum, using to_Spectrum1D, which will give us a Spectrum1D object:

>>> spec1d = target_spectrum.to_Spectrum1D()
>>> print(spec1d)
<Spectrum1D: 3562.4-10380.9 Angstrom>

>>> spec1d.plot()

Of course, this plot is going to look a bit bonkers because there is a lot of noise in the extreme red and blue, cosmic rays here and there, and whopping telluric absorption. Here’s what it looks like:

(Source code, png, hires.png, pdf, svg)

_images/getting_started-4.png