FORS1 IMAGE QUALITY

QC1 PLOTS IMAGE QUALITY: image quality

QC1 ASCII DATA

CURRENT PERIOD
	gif	ps
imagequality:
imagequality-DIMM:

(updated regularly)

QC1 database (expert mode)

General

Starting with Period 65 (April 2000), all reduced FORS1 images taken in Service Mode are analyzed in order to estimate their Image Quality (hereafter IQ).

This is done in a fully automatic way using a stand-alone procedure based on the Sextractor object detection alghoritm. Stars are identified based on Sextractor parameters (Stellarity Index > 0.5) and their FWHM is measured by the program. Then the mode of the IQ distribution is found and the RMS deviation is estimated using the left wing of the distribution only. The estimated IQ is then logged into a database together with other parameters, like alt-az field coordinates, DIMM seeing (hereafter DS),wind speed, wind direction and so on. The data is undergoing an extensive analysis in order to asses whether there are any relations between the measured IQ and other physical parameters.

IQ vs DIMM

In this page we present the current results (updated at the end of Period 71) obtained for FORS1 comparing the DIMM seeing and the measured IQ.

In order to allow for a meaningful comparison, the measured FORS1 IQ has been corrected to zenith and to the DIMM central wavelength (0.5 mu). An additional correction due to the different telescope diameters has also been included,following the prescriptions by Marc Sarazin. A quick description of the adopted correction is given here.
To reduce the uncertainty on the measured IQ, only values obtained from more than 10 stars have been considered. The DIMM seeing values are the average values during the FORS1 exposure. Note that DIMM measurements are obtained once a minute. In the following plots, only images for which at least one DIMM measurement is available during the exposure have been included. No selection has been performed on the basis of the exposure times, which range from 1 minute up to 30 minutes.

In the next figure, we have plotted the deviations IQ-DS (in arcsec) vs. the DIMM seeing (upper panel) and the UBVRI FORS1 IQ vs. the DIMM seeing (lower panel).
The black dotted line represents the locus where DIMM seeing and FORS1 IQ are identical, while the red dotted line is a fit to the data. The scatter is quite large, but as a general rule, FORS1 IQ is on average better than DIMM seeing when DS >0".7, while it tends to be larger than DIMM seeing for DS<0".7. Again, the scatter appears to be quite large, indicating that the measured IQ depends not only on the actual seeing, but also on other parameters, like wind speed and wind direction with respect to the telescope pointing.

IMPORTANT NOTE: The Standard Resolution (SR) collimator produces a detector scale of 0".2/pixel (1x1 binning). When the seeing is less than 0".4-0".5 this leads to some significative undersampling of the Point Spread Function (PSF). Even though Sextractor applies a correction for undersampling effects down to a FWHM=1 pixel (assuming an underlying gaussian PSF), a residual effect is expected in the region of excellent seeing. Of course the effect should be much less pronounced when the High Resolution (HR) collimator is used, with a detector scale of 0".1/pixel (1x1 binning). This is in fact visible in the above figure, where the measurements obtained with the HR collimator have been plotted in green. They clearly distribute at the lower edge, even though this is not totally due to the sampling effect. In fact, the HR collimator is used when seeing conditions are very good and this naturally tends to bias the correspondent distribution towards the lower seeing values.

In the next figure we show the distribution of IQ-DS deviations (upper panel) and the distribution of DIMM seeing and corrected FORS1 IQ (lower left and right panels). In each panel the dashed line marks the median of the distribution.

The lower figure shows the cumulative function of the FORS1 IQ-DIMM deviations. This shows that in ~73% of the cases (light shaded region) the measured IQ is within 0".1 from the DIMM seeing, while in ~52% of the cases (dark shaded region), it is equal or better than DIMM seeing. The dashed line marks the median of the distribution.

Postscript versions of the above figures can be found here:

QC home

QC and trending