Scintillation - aperture dependence

   Main publication

  D. Dravins,  L.Lindegren,  E.Mezey  &  A.T.Young

III.  Effects for Different Telescope Apertures
PASP  110, 610-633 (1998);  erratum: PASP  110, 1118 (1998)

ABSTRACT.  Stellar intensity scintillation in the optical was extensively studied at the astronomical observatory on La Palma (Canary Islands).  Atmospheric turbulence causes "flying shadows" on the ground, and intensity fluctuations occur both because this pattern is carried by winds, and is intrinsically changing.  Temporal statistics and time changes were treated in Paper I, and the dependence on optical wavelength in Paper II.  This Paper III discusses the structure of these flying shadows, and analyzes the scintillation signals recorded in telescopes of  different size and with different (secondary-mirror)  obscurations.  Using scintillation theory, a sequence of power-spectra measured for smaller apertures is extrapolated up to very large (8-m) telescopes.  Apodized apertures (with a gradual transmission falloff near the edges) are experimentally tested, and modeled for suppressing the most rapid scintillation components.  Double apertures determine the speed and direction of the flying shadows.  Challenging photometry tasks (e.g., stellar microvariability) require methods for decreasing the scintillation "noise".  The true source intensity I (lambda) may be segregated from the scintillation component  delta­I (t, lambda, x, y) in post-detection computation, using physical modeling of the temporal, chromatic, and spatial properties of scintillation, rather than treating it as mere "noise".  Such a scheme ideally requires multi-color high-speed (less than about 10 ms) photometry on the flying shadows over the spatially resolved (less than about 10 cm) telescope entrance pupil.  Adaptive correction in real-time of the two-dimensional intensity excursions across the telescope pupil also appears feasible, but would probably not offer photometric precision.  However, such "second-order" adaptive optics, correcting not only the wavefront phase but also scintillation effects, is required for other critical tasks such as the direct imaging of extrasolar planets with large ground-based  telescopes.

Fig. 1.  Power spectral density of scintillation in telescopes of different size.  The symbols are values measured on La Palma for a sequence of small apertures.  Their fit to a sequence of synthetic spectra predicts the scintillation also in very large telescopes up to 8 m diameter.  Bold curves are for fully open apertures.  A central obscuration (secondary mirror) increases the scintillation power, while apodization decreases it for high temporal frequencies. .

Fig. 2.  Power spectral content of scintillation in different apertures, i.e. the amount of integrated power, as function of frequency.  Observations and simulations are as in Figure 1.  This illustrates where in the spectrum the power is located.  For smaller apertures, the power distinctly shifts towards higher frequencies.  This trend continues until aperture diameters around 5 cm, where the structures in the "flying shadows" begin to get resolved. 

Fig. 3.  Scintillation autocovariances, showing dependence on wind direction (position on the sky) and on central obscuration (secondary mirror).  For a circular and open 20 cm aperture, the function is shown for the zenith, and for two wind-azimuth angles at zenith distance Z = 35 deg.  The scintillation in a 2.5 m telescope is much less but shows somewhat complex time structure, caused by its 90 cm secondary mirror.  The plot shows autocovariance + epsilon;  = 0.0002 for the smaller and = 0.00003 for the larger aperture. The true zero levels (= epsilon) are marked.  Although this figure contains synthetic data only, both amplitudes and timescales were fitted to empirically determined values for summer conditions on La Palma. 

Fig. 4.  Transmission profiles of apodization masks used on La Palma.  These "unsharp" telescope apertures were made from airbrush-painted MylarTM films.  The radial dependence of optical transmission in two of these is shown, as measured on a microphotometer.  The amplitude scale (left) is relevant for computing diffractive effects of light, while the scale at right  is the ordinary light intensity.

Fig. 5.  Measured and synthetic autocovariances for sharp and apodized apertures.   The "sharp" one is a clear Mylar window of similar material as the "strongly apodized" one (Fig. 4).  The increased autocovariance (power) for the apodized aperture is caused by its smaller effective diameter (due to its apodized edges), mimicking a smaller sharp aperture.  The differences at the highest frequencies are seen in Fig. 6. 

Fig. 6.  Measured and synthetic scintillation power spectra for sharp and apodized apertures.  As in Fig. 5, the increased power at most frequencies for the apodized aperture originates from it being effectively smaller.  However, for the highest frequencies, there is a tendency for the spectrum to fall off steeply enough, that the power seen with the apodized aperture becomes less than  with the clear one. 

Fig. 7.  Synthetic power spectra for apertures with and without cental obscuration; with and without apodization.  The scintillation power at 10-100 Hz may differ by an order of magnitude between telescopes that otherwise would appear to be nearly equivalent.  For investigations that are limited by atmospheric effects, this shows the potential for improving sensitivity by optimizing the geometry of the telescope's entrance pupil.  These synthetic data were normalized to observed summer conditions on La Palma, both in power and frequency.

Fig.8.  Scintillation measured through masks with two apertures, at different separations and position angles.  If the same flying-shadow pattern crosses both apertures, a secondary peak appears in the autocorrelation, revealing the flying-shadow speed and direction.  The autocorrelation changes significantly with position angle: the secondary peak is reproducible only within a rather narrow range (about 30 degrees).  For apertures separated by 30 cm, typical delays of 20 ms indicate a flying-shadow speed of about 15 m/s.

Fig. 9.  Double and single apertures, and different colors.  Autocorrelations were measured through a mask with two 10 cm apertures, separated by 45 cm.  The position angle was adjusted to show a secondary peak due to flying shadows crossing at a speed of about 15 m/s.  This secondary peak remains essentially unchanged between 400 and 700 nm, but the function is strongly different from that seen in a single 10 cm aperture.

Fig. 10.  Telescope concepts for reducing scintillation "noise" in stellar observations.  The passive system (left) incorporates a photometer that rapidly (< 10 ms) and with good spatial resolution (< 10 cm) measures the two-dimensional brightness distribution over the entrance pupil, thus resolving the spatial, temporal and chromatic signatures of scintillation.  The active system (right) incorporates second-order adaptive optics which measures the pupil illumination in real time, and corrects the two-dimensional intensity excursions across it (e.g., by imaging it through an  adaptive two-dimensional neutral-density filter).

Comments are welcome to

        Top of DD research pages

   Lund Observatory

Updated JD 2,455,775