Fitting DESI spectrum with PyQSOFit

[astro-fermi]

Dear Sir,

Thank you for this nice software. I want to know if PyQSOFit can be used to fit the DESI-DR1 optical spectra. To do so, what changes are needed to the software? It would be nice to add a feature where PyQSOFit checks the spectral binning and automatically rebins the spectrum if necessary.

Thank you in advance.

[legolason]

Yes, you can use it. I believe no much difference expect the resolution.

[astro-fermi]

Dear Sir,

Thank you. I attempted to run PyQSOFit on one DESI spectrum and encountered some unexpected results. I first tried fitting only the continuum with the host + FeII + power law components. My input is as follows:

xx,kk,ll=np.loadtxt('39627677287583730.txt', unpack=True, usecols=[0,3,4])
q = QSOFit(xx, kk, ll, z=0.1969, ra=72.861939, dec=-4.522172169, path=path1)

q.Fit(name='39627677287583730_host', reject_badpix=False, deredden=True, decompose_host=True, host_prior=True, host_prior_scale=0.2, host_line_mask=True, decomp_na_mask=True, qso_type='CZBIN1', npca_qso=10, host_type='PCA', npca_gal=5, Fe_uv_op=True, linefit=False, rej_abs_line=False, save_result=True, wave_range=np.array([3800,5500]), rej_abs_conti=True, MC=True, save_fits_name='conti_param.fits', nsamp=200, param_file_name='qsopar.fits', plot_corner=False, kwargs_plot={'save_fig_path': '.'})

The fitting was successfully finished, generating the following plot, which shows that all worked well:

Next, I attempted to plot the fitted continuum model separately. My input was as follows:

plt.figure(figsize=(18,8))
plt.plot(q.wave, q.flux, color="k", label='Obs', lw=2)
plt.plot(q.wave, q.host, color="b", label=''host, lw=2)
plt.plot(q.wave, q.f_pl_model,'--',color="purple",label='Continuum', lw=2)
plt.plot(q.wave, q.f_fe_balmer_model,'-',color="orange",label='Fe II model', lw=2)
plt.plot(q.wave, q.f_conti_model,'-',color="r",label='FeII + PL', lw=2)
plt.xlabel('Rest Wavelength',fontsize=20)
plt.ylabel('Flux',fontsize=20)
plt.xlim(3800,5500)
plt.ylim(-2,175)
plt.tick_params(which='both', direction="in")
plt.yticks(fontsize=20)
plt.xticks(np.arange(3800, 5500, step=250),fontsize=20)
plt.legend(loc='upper left', prop={'size': 16}, frameon=False, ncol=2)
plt.savefig('plot.pdf', format='pdf', bbox_inches='tight')

Comparing these two diagnostic plots, we clearly see the difference in the fitted FeII model, which appears to be pretty smeared out in the second plot. Also, the power law model appears to be shifted to lower flux in the second plot. The sum of the power law and FeII model components also appears different from the first plot.

I am unable to understand why these two figures are so drastically different. The diagnostic plot generated by PyQSOFit shows a nicely fitted Fe II model, but the corresponding q.f_fe_balmer_model is totally different.

May I request that you kindly take a look at the problem and suggest a possible remedy? I have attached the spectrum text file and the qsopar.fits file below. Thank you.

39627677287583730.txt

qsopar.fits.txt

(Please rename the qsopar.fits.txt file as qsopar.fits file. I changed the name since GitHub doesn't allow uploading FITS files.)