Refined HPF sky subtraction

The science and sky fibers in HPF have slightly different throughputs. In this notebook we show how muler can automatically apply a calibration correction based on twilight sky flats. We believe this change can reduce sky residuals by up to \(14\times\) over the naive subtraction of the sky fiber.

%config InlineBackend.figure_format='retina'
from muler.hpf import HPFSpectrum

example_file = '../../data/HPF/Goldilocks_20210517T054403_v1.0_0060.spectra.fits'
spectrum = HPFSpectrum(file=example_file, order=19)

spectrum = spectrum.normalize()

We see the conspicuous blaze pattern in each spectra order. Let’s apply a correction to one of those orders:

ax= spectrum.sky.plot(ylo=0, yhi=0.1);

These are the skylines we wish to remove. Since the sky emission is additive it should always be the first step before dealing with multiplicative terms, such as the blaze or telluric absorption.

ax = spectrum.plot(ylo=0.6, yhi=1.3, label='Observed')
spectrum.sky_subtract(method='scalar').plot(ax=ax, label='Sky Subtracted')
ax.legend();

You can see that that the sky subtraction successfully removes the emission lines. We can now proceed with deblazing:

spectrum.sky_subtract(method='scalar').deblaze().normalize().plot();

Awesome! Let’s see if we can spot the differences among the three sky subtraction methods:

ax = spectrum.deblaze().normalize().plot(label='observed');
for method in ['naive', 'scalar', 'vector']:
    spectrum.sky_subtract(method=method).deblaze().normalize().plot(ax=ax, label=method);

ax.set_xlim(10_830, 10_840)
ax.set_ylim(0.98, 1.09)
ax.legend();

Naive sky subtraction method is known to oversubtract the sky, see GitHub Issues.

We see that the naive method oversubtracts the sky line, as we expected. The scalar and vector methods appear to perform equally well for this source in this wavelength range.