14 January Observational Astronomy SPECTROSCOPIC data reduction Piskunov & Valenti 2002, A&A 385, 1095
14 January Echelle spectrum image looks strange … Spectral orders Spectral line
Data reduction concept Complex data layout requires in addition to the usual CCD calibrations: Spectral order localization Extracting 2D data into a sequence of 1D spectra Wavelength calibration Blaze function correction 14 January 20163
4 To performed data reduction we need calibration data Bias (shutter closed, 0-time exposures) Flat field (uniform source, short exposures) Dark current (shutter closed, different exposures) Order tracing (flat field with short slit) Wavelength map (emission line source, short exposures) Blaze calibration (flat field or a star with minimum sp. lines)
14 January CCD data reduction in one expression The intensity is given by: s – signal in science exposure b – bias level f – flat field signal g – gain (e - /ADU) d – dark current signal per unit time t – exposure times
14 January The problem is the errors: If f is close to b, the S/N is determined by the S/N of the flat field!!!
14 January Complete data reduction sequence 1. Create master bias and master flat 2. Trace spectral orders 3. Normalize master flat 4. Extract wavelength calibrations 5. Extract science spectra (optimal extraction) 6. Fit the continuum 7. Construct the wavelength solution
14 January Creating master bias and master flat The goal is to replace the actual calibration data with a model which is free of random noise but carries all the systematic signatures. S/N of a master must be much larger than the S/N in science frames!!! Main issue: getting rid of random errors, e.g. cosmic ray hits Method : filtering within a frame or across a stack of frames Cross-check between groups of calibration frames
14 January Example of constructing master flat
14 January Stack of individual flats (fragments)
14 January times larger vertical scale Individual flats with master flat subtracted
14 January Flat field Fragment of a master flat field
14 January Order tracing
14 January Order tracing (2)
14 January Wavelength calibration Pixel number Order number 2D polynomial Emission line spectrum (ThAr) with known laboratory wavelengths
14 January Wavelength solution and the PSF 1.Why the Gaussian misses the bulk of points? 2.Why do you see so much scatter? Line analysis for wavelength solution for ESO UVES spectrometer. Points show measured lines (scaled). Blue line shows best Gaussian fit.
14 January Slit function decomposition Idealized model : Image on the CCD is a sequence of monochromatic images of the entrance slit sampled with detector pixels Watershed pattern is due misalignment between spectral order and detector raws.
14 January Normalizing flat field Original flat Model flat “Spectrum” Normalized flat (data/model)
14 January Extracting science spectrum Extraction of science spectrum: -Simple summation -Gaussian fit to the slit function -Optimal extraction
14 January Continuum fit Blaze function is a good start:
14 January Fringing Accurate fringing removal requires slit illumination by the FF and by the science object to be identical Wide flat Narrow flat Science spectrum
14 January Exam Written exam on Thursday, March 14 th, starting 9am sharp Oral exam: book time with me