Measuring Seeing, The Differential Image Motion Monitor (DIMM)

1
Measuring Seeing, The Differential Image Motion
Monitor (DIMM)

• Marc Sarazin
• (European Southern Observatory)

2
List of ThemesHow to find the ideal site...and
keep it good?

• Optical Propagation through Turbulence
• Mechanical and Thermal
• Index of Refraction
• Signature on ground based observations
• Correction methods
• Integral Monitoring Techniques
• Seeing Monitoring
• Scintillation Monitoring
• Profiling Techniques
• Microthermal Sensors
• Scintillation Ranging
• Modelling Techniques

3
Why Differential Image Motion?

• The tracking errors are automatically subtracted
• The wind has no effect on the measurements
• The telescope optical quality is not important
  (nevertheless circular images are required, i.e.
  no coma allowed)
• Easy to implement with state of the art amateur
  astronomer detectors
• The DIMM gives two statistical estimates of the
  same variable

4
Optical PropagationThe Signature of
Atmospheric Turbulence
Seeing (radian, ??-0.2)
Fried parameter ( meter, ??6/5)
5
DIMM Principle

• Two images of the same star are created on a CCD,
  corresponding to light having traveled through
  two parallel columns in the atmosphere

6
DIMM Principle

• The variance of the image motion through a
  circular aperture of diameter D depends on the
  seeing as

• The variance of the differential image motion
  through circular apertures of diameter D,
  separated by d is

7
DIMM Principle
The final estimate of the seeing is the average
of both parallel and perpendicular motions
8
DIMM Principle

• Error Budget for a 10 accuracy goal

• The instrumental noise (sampling, centroiding) is
  measured in the lab on fixed sources. The
  constant part can be subtracted out, the noise is
  the remaining variance, about /- 0.002 pixel2,
  or 5 relative error at 0.2 seeing. The plate
  scale is calibrated on double stars of known
  separation
• The measurement noise might increase if the
  signal to noise ratio is too low images with
  low SNR due to scintillation have to be rejected.
• The statistical noise is inversely proportional
  to the square root of the number of samples in
  the time series. The relative error on the seeing
  is about 6 for 200 exposures.
• The temporal under sampling due to too long
  exposure time no way to correct for it because
  the velocity of the tilt is unknown. Interlacing
  two exposure times is the best way to control.
• The very bad seeing (gt2) is over estimated
  because the stellar image breaks into speckles

9
DIMM Precursor

• A visual DIMM was used in the 60s for site
  selection purposes in Chile and in Uzbekistan
  (photo Maidanak Observatory).
• See J. Stock and G. Keller, 1960, in Stars and
  Stellar System, Vol. 1, Chicago University Press

10
Portable DIMM Operation

• Preparing for nighttime measurements on the high
  chilean sites (5200m) in the vicinity of the ALMA
  project
• Source Cornell Atacama project
  http//astrosun.tn.cornell.edu/atacama

11
Portable DIMM Operation

• Alignment of C11 telescope mount on a high
  chilean site (5200m) in the vicinity of the ALMA
  project
• Pixel size0.7
• Pupil Diameter9cm
• Pupil Separation12cm
• Exposure Time10/20ms
• 50 frames/mn
• Photo credit P. Recabarren, Observatory of
  Cordoba, Argentina

12
Portable DIMM Operation

• 1m high platform and daytime protection of the
  portable DIMM on the high chilean sites (5200m)
  in the vicinity of the ALMA project
• Source Cornell Atacama project
  http//astrosun.tn.cornell.edu/atacama

13
Portable DIMM Operation

• 5m high tower and daytime protection of the
  portable DIMM at the observatory of Maidanak,
  Uzbekistan

The telescope stands in free air circulation to
prevent build-up of local thermal pockets
14
Automated DIMM Operation

• Daytime protection of the automated DIMM at the
  VLT Observatory
• The enclosure control is linked to the
  meteorological station (closes when windgt18m/s,
  Rhgt80)

15
Automated DIMM Operation

• 35cm Telescope for the automated DIMM at the VLT
  Observatory
• Pixel size0.7
• Pupil Diameter11cm
• Pupil Separation20cm
• Exposure Time5ms
• 600 frames/mn

16
Automated DIMM Operation

• The seeing is updated every minute for zenith
  observation at 0.5 micron wavelength
• The accuracy is better than 10 above 0.2
• The natural atmospheric noise is about 10 of the
  seeing

17
Automated DIMM Operation

• The system automatically switches to another star
  in case of clouds
• The seeing is independent of cloudiness (although
  sometimes pretty good with high cirrus clouds)
• Aperture photometry alows to monitor the sky
  variability

18
Automated DIMM Operation

• Aperture photometry on ca 5000 DIMM short
  exposures allows to monitor the flux variability,
  equivalent to the extinction variability (June
  2000 statistics below)

The threshold for photometric sky is between 1
and 2 relative flux rms
19
DIMM Seeing vs. VLT Image Quality
DIMM converts image motion into large telescope
seeing with the assumption of an infinite outer
scale of the turbulence. UT images turned out
about 10 better than predicted by DIMM,
confirming the finite character of the outer
scale.

• Comparison of DIMM seeing (Y axis), with FORS
  Science Verification (SV) Image Quality (X axis)
  as processed by the SV team, corrected for zenith
  and 500nm.

20
Corrected DIMM Seeing vs. VLT Image Quality
DIMM converts image motion into large telescope
seeing with the assumption of an infinite outer
scale of the turbulence. UT images turned out
about 10 better than predicted by DIMM,
confirming the finite character of the outer
scale. Correcting for that effect is possible by
removing from the DIMM the share of the tilt of
an 8m aperture.

• Comparison of DIMM seeing (Y axis) after
  correction for outer scale, with FORS Science
  Verification (SV) Image Quality (X axis) as
  processed by the SV team, corrected for zenith
  and 500nm.

21
Corrected DIMM Seeing vs. VLT Image Quality
DIMM converts image motion into large telescope
seeing with the assumption of an infinite outer
scale of the turbulence. UT images turned out
about 10 better than predicted by DIMM,
confirming the finite character of the outer
scale. Correcting for that effect is possible by
removing from the DIMM the share of the tilt of
an 8m aperture.

• Comparison of DIMM seeing (Y axis) after
  correction for outer scale, with UT1 Science
  Verification (SV) Image Quality (X axis) as
  processed by the SV team from Test Camera long
  exposures, corrected for zenith and at 500nm.

22
DIMM Seeing vs. Large Telescope Image Quality
DIMM converts image motion into large telescope
seeing with the assumption of an infinite outer
scale of the turbulence. Assuming that the outer
scale larger than the telescope aperture, a first
order correction is obtained by removing the one
axis image jitter (Gradient tilt) variance from
the long exposure FWHM

• Outer scale correction coefficient to apply to
  the DIMM estimates of the image quality of a 8m
  telescope limited by the atmosphere, for 0 and 60
  degree zenith angle, as a function of the
  observing wavelength (the following central
  wavelength of the bands U, B, V, R, I, J, H, K,
  L, M, N corresponding to 0.36, 0.44, 0.55,
  0.64, 0.79, 1.25, 1.65, 2.2, 3.4, 5.0, 10 in
  mm).

23
Monitoring Turbulence Height with the DIMM
Scintillation through DIMM apertures of 10-12cm
diameter can be related to the isoplanatic angle
(Loos Hogge, Appl. Opt. 18, 15 1979) and then
to the normalized 5/3rd moment of the turbulence
height (Hbar).

• The atmospheric seeing (black lower curve, in
  arcsec) is the cumulative effect of several
  turbulent layers at various altitudes monitoring
  the characteristic altitude of the turbulence
  (red upper curve, in km) is necessary for
  planning adaptive optics instrumentation. In this
  example, the bad seeing is located at low
  altitude while good conditions are produced by a
  few layers at high altitude.

24
Local Seeing Ground Layer Turbulence at Paranal

• Measurement of the microthermal activity and
  Seeing at Paranal (GSM Campaign, Nice University)
  during a night presenting variable conditions
  (F. Martin, R. Conan, A. Tokovinin, A. Ziad, H.
  Trinquet, J. Borgnino, A. Agabi and M. Sarazin
  Optical parameter relevant for high angular
  resolution at Paranal from GSM instrument and
  surface layer contribution Astron. Astrophys.
  Supplement, v.144, p.39-44 June 2000).

25
Local Seeing Seeing Impact of Ground Layer

• Measurement of the microthermal activity and
  Seeing at Paranal (GSM Campaign, Nice
  University) The contribution of the layer 7-21m
  above ground is marginal both during good and bad
  seeing conditions .

26
Conclusion
Intercalibration of the site monitoring
instruments is recommended