Photometric zeropoints have been calculated from standard stars observations based on the Landolt catalogue (Landolt 1992, AJ 104, 340, used until 2008-03-31) and the Stetson catalogue (since 2008-04-01). You can find some information about conversion from the FORS2 filters here .
There are two kinds of zeropoints - per frame and instrumental.
The raw data have been:
de-biassed,
flat-fielded with a normalized twilight sky flat.
The twilight images of FORS2, which are used to create master sky flats, show evidence of a feature, which position depends on the rotator angle (see overview). This limits the achievable photometric accuracy to about 5%. For more information see Moehler et al. (2010, PASP 122, 93).
For the determination of a photometrically stable night we evaluate the variation of the extinction coefficient measured from standard stars frames. Since 2011 the FORS2 calibration plan foresees, for potentially photometric nights the initial acquisition of a low-high-airmass pair of standard fields, followed by further standard fields around midnight and in the morning if SCIENCE data requesting photometric
conditions are taken. A night is then defined as stable (S) if
the measurements cover an airmass range of more than 0.4 and
the variation of extinction coefficients is below 0.05 mag/airmass for v_HIGH and R_SPEC and below 0.06 mag/airmass for b_HIGH.
In addition we check for the time range covered by the measurements - if it is less than 1 hour the night is flagged as Sb (for 'stable begin'). Nights with higher extinction variations are marked as not-stable ('N'), while nights with an
airmass range of less than 0.4 or only one measurement are flagged as unknown ('U').
The frame zeropoints are based on a single STD_IMA exposure. They are calculated assuming a fixed extinction coefficient. A second result is calculated by the pipeline from the same input data, this time the extinction coefficient assuming a fixed zeropoint.
Disclaimer: Donot use these zeropoints for the reduction of science data. They have been processed assuming standard extinction coefficients and colour
terms, which may not be appropriate for an individual night. Instead we recommend to use the instrumetal zeropoints and extinction coefficients (available since 2012-06).
These coefficients were hardcoded in the old pipeline (for data observed until 2008-03-31) and are tabulated in configuration files for the new pipeline (for data observed since 2008-04-01). The
coefficients are frozen since 2011-11-28. We noticed a problem with some standard star field results, where the pipeline mis-identified standard stars. Usually such mis-identifications are caught during the QC certification process, but a few such cases (≤5%)
have slipped through. In such cases typically only few of the identified stars carry high weight. This problem has been fixed with pipeline version 4.9.2, which has been in use since 2011-05-25.
For the trending we select only zeropoints with errors of less than 0.1 mag and with more than 1 standard star found. In the same plot we trend the error of the frame zeropoints to check if outlying zeropoints have larger than usual errors.
Fig.1 Frame zeropoints (black circles) and instrumental zeropoints (red diamonds). Data for stable nights (flag 'S' or 'Sb') have filled symbols. The instrumental zeropoints are calculated with a delay of 28 days. They represent the true current instrument zeropoint, calculated with a measured extinction, while the frame zeropoint is calculated assuming a specific, fixed extinction.
Washing or re-coating of mirrors affects the zeropoints as do moves between telescopes.
General
For the old FORS2 CCD (in operation until 2002-03) the conversion factors were read from the FITS keywords "HIERARCH ESO DET OUTi CONAD" with i = 1...4 and averaged over the four CCD ports.
Until 2008-03-31
the old imaging pipeline was used, which determined the zeropoints as follows: SExtractor is run on the reduced images, to detect standard stars and to extract their fluxes. A fixed aperture of 10" radius is used. The fluxes of the unsaturated stars are then converted from ADU per sec into electrons per sec using the FITS keyword "HIERARCH ESO DET OUT1 CONAD" and corrected for atmospheric extinction and colour terms using hardcoded coefficients. For the zero points per frame the average of the zeropoints from the individual unsaturated stars is determined. Zeropoints fainter than this average by more than 0.3 mag are excluded and the average is re-determined; The rms error of the frame zeropoint (0 in general means that there was just one standard star in the field)
Until March 2006
the instrumental magnitudes were estimated with the MIDAS integrate/aperture command instead of using the SExtractor magnitudes from the pipeline, which were unrealiable due to an incorrect setting of the SExtractor MASK parameter. The instrumental magnitudes were corrected for extinction and colour using coefficients which were read from the QC1 Database and updated after every period.
Since April 2008
the new imaging pipeline has been in use. It uses the Stetson catalogue, which does not contain U band data. Therefore U_SPEC zeropoints are not available anymore.
Since April 2009
the standard filter set contains R_BESS, I_BESS, b_HIGH, and v_HIGH.
Since Nov 2011
the following extinction coefficients and zeropoints are used.
Since Jun 2012
the V_2010-09 version of the Stetson catalogue is used. Also the SourceExtractor Parameters have been changed to: DETECT_THRESH 1.5; PHOT_APERTURES 40, BACK_PHOTOTYPE LOCAL, BACKPHOTO_THICK 30
November 2014
change of the Longitudinal Atmospheric Dispersion Corrector (LADC); zero points increased
2016-09-02
after M1 re-coating the zero points increased by about 0.2 mag in all filters
Ocober 2018
cleaning of LADC and annual maintenance, zero points increased
The SExtractor software (Bertin & Arnouts, 1996,
A&A 117, 393) is run on the reduced images, to detect standard stars and to extract
their fluxes. A fixed aperture of 10" radius is used and flux from neighbouring
stars within this aperture is corrected. The
fluxes of the unsaturated stars are then converted from ADU per sec into electrons
per sec using the inverse gain stored as FITS keyword "HIERARCH ESO DET OUT1
CONAD". They are then corrected for atmospheric extinction and colour terms
using tabulated
coefficients. For the zero points per frame the optimally weighted average
of the zeropoints from the individual unsaturated stars is determined. The optimal
weights are determined in such a way, that the error on the frame zeropoint is minimised.
Individual zeropoints deviating more than either 0.3 mag or 5 σ from the mean
zeropoint are rejected as outliers, and the mean is re-determined.
The sig_zp_frame parameter corresponds to combination of the rms error of the individual zeropoints and the accumulated errors from the frames (raw frame, master bias, master twilight flat).
Fig.2 Frame extinction (black circles) and night extinction (red diamonds). Data for stable nights (flag 'S' or 'Sb') have filled symbols. The night extinction for a specific date is calculated with a delay of 28 days. The night extinction is displayed for stable nights only, since it is defined only for those nights. It is the true extinction coefficient for the considered night, applicable together with the true instrumental zeropoint (Fig. 1).
Fig.3 Airmass distribution for the nights with FORS2 photometry. All data points from the night (per filter and chip) are grouped together and connected by line. This plot has the colour coding for 'S' or 'Sb' nights (green, stable), 'N' (not stable, red), and 'U' (unknown, black). See text for the criteria.
after M1 re-coating the zero points increased, which affected calculation of extiction; as a result, the extinction coefficients were systematically underestimated; it went back to normal within a month of resumed operations
2020-12-16
PHOT_TABLE adjusted with more appropriate zeropoints for the current situation; It resulted in lower extinction values derived from single frames and in more nights classified as stable
Assuming a value for the zeropoint an extinction coefficient is derived. The values assumed for the zeropoints can be found here. If the assumed zeropoints are over/under-estimated, the derived extinction coefficients are correspondingly shifted. The final values per stable night are calculated as below
The error of the extinction coefficient is the rms value.
With a time delay of one month, we calculate in a second step the instrumental zeropoints and extinction coefficients per stable night.
For the instrumental zeropoints and extinction coefficients per night, we collect data from at least 7 stable nights around the night of interest. The combinations of high airmass coverage and stable extinction for these nights allows us to determine a common zeropoint and nightly extinction coefficients with the pipeline recipe fors_photometry. The zeropoint and the extinction coefficient derived for the night of interest are stored in the database table fors2_photometry. We take care that data are not combined across mirror coatings.
The extinction coefficient for the night more accurately refers to the times of the night when standard star images were taken. If, for instance, only two standard star images were taken at the beginning of the night, the extinction coefficient describes only the conditions for that short time interval.
For trending we select only data from stable nights, since the fors_photometry recipe is very sensitive to variations of the extinction coefficient during the night, which may affect the determination of the instrumental zeropoint.
The instrumental zeropoint is zeropoint derived by the procedure described above. Since the recipe can determine the exticntion coefficient from the data this zeropoint should be free from atmospheric effects and reflects only changes in the instrument and/or telescope.
The error is the rms error of the instrumental zeropoint.
The extinction coefficient is derived for the night. It is defined only for stable nights.
The error is the rms error of the extinction coefficient for the night.