Mini Workshop on:

Atmosphere knowledge and Adaptive Optics for 8 to 100 m telescopes

European Southern Observatory, Garching; Germany

October 13th 2003, Auditorium

Mini -workshop initiative: N. Hubin, M. Sarazin

Mini-workshop objectives: The goal of this mini-workshop is to define the critical atmospheric parameters required for an optimum design of future AO-MCAO systems for 8 to 100m telescopes and to identify existing and potential instruments and facilities to measure in-situ these parameters.


Abstracts and pdf presentations

Executive summary


8:45 Welcome

9:00 Session 1: Possible AO concepts for 8 to 100m telescopes and relevant atmospheric parameters for an optimum design (Moderator: B. Ellerbroek)

First generation of AO systems for ELTs: a short review of possible systems: N. Hubin 20mn

Strategy for a better understanding of key atmospheric parameters: J.M. Conan 30mn

Impact of atmospheric parameters on AO system design: M. Le Louarn 20mn

10:10-10:25 Coffee break

10:25-11:25 Discussion: Summary of critical atmospheric parameters

11:25 Session 2: Existing instruments/ facilities for the measurement of atmospheric parameters, site testing (Moderator: J. Vernin)

Current knowledge of the atmospheric parameters: M. Sarazin, 30mn

12:00-13:00 Lunch at ESO

Atmospheric parameters part I: Present instruments able to measure the atmospheric parameters: J. Vernin, 30mn

MASS-DIMM: a turbulence monitor for Adaptive Optics: A. Tokovinin, 30mn

14:15-15:00 Discussion: Summary of available atmospheric parameters, accuracy/confidence and availability of database

15:00-15:15 Coffee break

15:15 Session 3: New instruments for the in-situ measurement of atmospheric parameters relevant for AO (Moderator: R. Ragazzoni)

Slodar turbulence profiling: R. Wilson, 20mn

Slodar with LGS and NGS MCAO WFS: B. Ellerbroek, 20mn

Development of a Single Star SCIDAR system for profiling atmospheric turbulence: D. Coburn, 20mn

Atmospheric parameters and AO design part II: New concept of Single Star Scidar: J. Vernin, 20mn

Direct 100-m scale wavefront sensing: R. Ragazzoni, 20mn

Measurements of atmospheric parameters with the VLTI: A. Glindemann, 20mn

Past and future measurements of atmospheric outer scale with interferometers: A. Quirrenbach, 10mn

Cn2 profiler with spatial scale larger than 10m: S. Esposito, 10mn

17:35-18:30 Discussion: list of recommended instruments, cost, timescale, collaborative efforts

18:30 End of workshop




Arne Ardeberg, Lund University

Derek Coburn, NUI Galway


Philippe Dierickx, ESO

Brent Ellerbroek, Gemini

Simone Esposito, Arcetri Firenze

Edward Graham, Uni. Fribourg

Denis Garnier, NUI Galway

Wolfgang Gaessler, MPIA Heidelberg

Andreas Glindemann, ESO

Norbert Hubin, ESO

Stephan Kellner, MPIA Heidelberg

Victor Kornilov,  Sternberg Moscow

Miska Lelouarn, ESO

Jerome Maire, LUAN Nice

Enrico Marchetti, ESO

Elena Mascriadri, MPIA Heidelberg

Richard Myers, Uni. Durham

Andreas Quirrenbach, Leiden Observatory

Roberto Ragazzoni, Arcetri Firenze

Tatyana Sadibekova, LUAN Nice

Marc Sarazin, ESO

Remko Stuik, Leiden Observatory

Andrei Tokovinin, CTIO La Serena

Jean Vernin, LUAN Nice

Richard Wilson, Uni. Durham




(Alphabetic order)

Development of a Single Star SCIDAR system for profiling atmospheric turbulence.

D.Coburn, D.Garnier, J.C.Dainty. National University of Ireland, Galway.

We are currently working on the development of a generalized SCIDAR system for characterizing atmospheric parameters using single star targets. The instrument, which is based on a commercially available 25cm diameter telescope, offers the potential for characterizing atmospheric parameters for wide areas of the sky. Here, we describe the system and outline the challenges in the data reduction and in solving the inverse problem needed to estimate the altitude dependence of refractive index structure constant. With the system development well underway we hope to test the instrument some time between November 2003 and March 2004.

Strategy for a better understanding of key atmospheric parameters

J.-M. Conan*, T. Fusco*, R. Conan**, G. Rousset, *Onera, Chatillon, France, ** LAOG, Grenoble, France.

We propose a list of key atmospheric parameters for the future AO developments. Their relevance for the different class of AO instruments and telescopes is justified.

We discuss several ways of measuring these parameters, mainly through a statistical processing of data provided by existing or planned instruments (ASM, GSM, MASS, scidar, NAOS, MACAO, MAD, VLTI...). We then propose a multi-instrument campaign, for instance at Paranal, to improve our knowledge of the atmospheric models and parameters. The use of different instruments will limit misinterpretations, and will give simultaneous spatial and temporal information at various scales. The results of such a campaign would help defining a strategy for future site testing. They are also of great interest for the design of the new generation AO systems.

Measurements of atmospheric parameters with the VLTI

Di Folco E., Koehler B., Kervella P., Sarazin M., Schoeller M., Glindemann A.

Infrared interferometric observations on long baselines are strongly affected by the atmosphere. While the sensitivity of the fiber-fed near-IR instruments is limited by the atmospheric seeing, the precision of the visibility measurements is degraded by the differential piston. These effects can be compensated for with adaptive optics and fringe tracking facilities, but their performance still depends on the atmospheric characteristics (seeing, coherence time, outer scale length). We have used the commissioning instrument of the VLTI, VINCI, with two test-siderostats on various baselines (up to 140m in length) in order to estimate the interferometric coherence time from the temporal spectra of the observed differential piston. It proved to be correlated with the independent estimation provided by the DIMM monitor. In addition, we have obtained the first estimation of the outer scale measured at Cerro Paranal. We will report on the first measurement campaign at the VLTI and will describe plans for further measurements of the relevant atmospheric parameters with the VLTI.

SLODAR with LGS and NGS MCAO wavefront sensors

Brent Ellerbroek, AURA New Initatives Office

The spatial cross-correlations between the measurements from distinct wavefront sensors in an MCAO system may be used to determine weighted integrals of the atmospheric Cn2(h) profile. We derive the functional form of these weighting functions and present sample numerical values for the Gemini-South MCAO configuration, in which the guide stars are separated by 0.7 to 1.4 arc minutes and the subaperture size is 0.5 meters. Implementation issues (noise, closed-loop operation, data rates, and LGS tilt uncertainty) are briefly discussed.

Cn2 profiler with spatial scale larger than 10 m

S.Esposito, INAF-Arcetri

A device to measure the cn2 profile of the atmosphere is presented. We tried to design a relatively simple unit that can give information on the vertical distribution of the atmosphere. The proposed method could be used in some different configuration to measure the turbulence on spatial scale larger than 10m.

First generation of AO systems for ELTs: a short review of possible concepts

N.Hubin, ESO

We will present the status of the investigated AO-MCAO concepts and characteristics for ELTs. We will underline whenever possible the open questions related to the knowledge of the key atmospheric parameters.

Impact of atmospheric parameters on AO system design

M. Le Louarn, ESO

In this presentation, I will show some examples on how atmospheric parameters influence the design of AO systems. For example, the MUSE AO system, will aim at correcting the ground layer of turbulence in the visible to get an "improved seeing" over a 1 FOV. It will be shown how the Cn2 distribution is critical. Also the influences of tau0 and r0 on the design are demonstrated by the use of current Paranal statistics. I will also briefly show what MCAO systems might need in terms of information on the atmospheric parameters for the control (regularization) and operation (DM conjugation heights). Some simulations on the effects of AO performance as a function of telescope diameter will show the effect of the outer scale of turbulence.

Past and future outer scale measurements using interferometers

A. Quirrenbach, Leiden Observatory

Direct 100m-scale WaveFront sensing

R. Ragazzoni

I outline a couple of techniques to perform wavefront sensing measurement on specific layers or volumes of atmosphere that, with existing telescopic facilities, can achieve direct wavefront sensing on scales of the order of 100m.

These techniques are maybe, representative of a class of techniques and do not want to mean they are the only ways to achieve such a goal. Further to outline the used approaches and some initial calculation and simulations about their feasibility, I would like to summarize the outcome of such a measurement.

It is clear that 100m-scale WF sensing cannot be achieved routinely, on several sites, and hence is not strictly aimed to compete with any site-testing campaign. However it can provide unique information on parameters like the spectrum of the turbulence, the conformity of the atmosphere to the Taylor hypothesis, the maximum stroke required to DMs, the coherence of wind flow in such large footprints.

Current knowledge of the Outer Scale and other parameters

M. Sarazin

The AO relevant parameters routinely monitored at Paranal and La Silla observatories are reviewed and their statistics is updated.

Then follows a short non-exhaustive review of the to-date accumulated information about the outer scale in various sites and using various instruments and methods, including some information collected from VLT science operation data. To conclude, highlights of what remains to be answered about its vertical structure in particular.

The CELT/GSMT site testing program (talk cancelled)

Matthias Schoeck, CELT, Caltech, Pasadena, CA, USA

Last year, after a previous period of informal collaboration, the California Extremely Large Telescope (CELT) and AURA's Giant Segmented Mirror Telescope (GSMT) started a formal collaboration with respect to site testing for their 30-m telescope projects. I briefly describe the current state of and the plans for the site characterization effort. I will concentrate on the on-site measurements, which are centered around a DIMM/MASS system, but will also include other instruments such as SODARs and microthermal probes.

MASS-DIMM: a turbulence monitor for adaptive optics.

A. Tokovinin

A short introduction to the principle of multi-aperture scintillation sensor (MASS) and its combination with DIMM is given. This instrument measures seeing, free-atmosphere seeing, isoplanatic angle, AO time constant, and low-resolution turbulence profile. Characteristics of MASS and results of its inter-comparisons with other instruments are reviewed. Properties of turbulence relevant to AO operation that were revealed by MASS are listed, and the use of MASS data for predicting the performance of ground-layer AO is presented. Plans for use of MASS-DIMM in site testing and monitoring are outlined. Mass database is:

Atmospheric parameters and Adaptive Optics design: key parameters, measurement equipment and impact on AO design

J. Vernin, LUAN, Nice University

We will present a definition of the outer scale of optical turbulence which explains the large discrepancy between our definition (From Tatarski Lo ~ 1m) and the coherence outer scale (From von Karman Lo ~ 20m). We will summarize the instruments able to measure the above mentioned atmospheric parameters. We will present the new concept of Single Star Scidar and recall an old paper (Coulman, Vernin 1991) in which the phase structure function is expected to rise again, after a saturation region.

SLODAR Turbulence Profiling

Richard Wilson

The SLODAR method for profiling of the atmospheric optical turbulence strength with altitude and velocity will be reviewed. Example and statistical results from recent observations at the WHT and Mercator telescopes will be presented, as well as details of the instrumentation deployed. The design and predicted performance of a proposed portable SLODAR instrument for site characterization will be discussed.


EXECUTIVE SUMMARY (edited from notes taken by A. Tokovinin)
The goal was to define critical atmospheric parameters required to design future AO-MCAO systems and to identify the instruments and techniques to measure those parameters.
Adaptive optics needs and the instruments
SEEING DIMM is the answer, although reliable seeing is now extracted from AO WFS data (NAOS) and otherwise. MS presented the comparison of DIMM with IR images, showing the effect of outer scale: 30% better FWHM in K-band compared to the standard theory. The strict sense of seeing (i.e. r_0) is defined as the intensity of the high-frequency part of turbulence spectrum (NOT resolution, as originally set by Fried, because of outer-scale effects).
Cn2 PROFILE needed for many things including seeing. The low resolution (2-3 layers) is useful for AO performance optimization in real time, higher resolution (several layers per DM) is needed for simulations. For seeing improvement (ground-layer compensation), a resolution of 100m is claimed to be necessary (NH). In fact, the profile resolution for GLAO is (actuator pitch/field size), depends on system parameters. MASS resolution (0.5km in the first km) is sufficient for 3-arcmin. field and 0.4m actuators, but for the MUSE project a 200m change of layer altitude already matters. Instruments discussed below.
AO TIME CONSTANT ESO DIMM now delivers TAU_0 in real time by using model wind velocities in the troposphere. Seasonal trends revealed (bad in June, July), correlation with seeing (good seeing -> high and fast turbulence). Prediction of TAU_0 is available. Comparison with TAU_0 from MASS is encouraging.
ISOPLANATIC ANGLE THETA_0: Again, ESO DIMM now derives THETA_0 from scintillation, statistics is available. It was stressed that THETA_0 is only indicative of AO isoplanatic field, which depends on system parameters and is usually wider. B. Ellerbroek mentioned that small corrected field on Altair is a concern (bad DM conjugation?).
OUTER SCALE L_0: Important parameter that influences DM stroke (especially at ELTs), interferometer operation, astrometric accuracy, uncorrected resolution in the IR, etc. Now the L_0(h) profile is needed for MCAO but there is no instrument to measure it. J. Vernin presented a typical profile L_0(h) derived from balloons, reconciled his previous estimates (metric L_0 from temperature gradients) with decametric L_0 parameter in von-Karman spectrum by noting that "V-K L_0" is ~6 times larger than "gradient L_0". The instrument for L_0 measurements is GSM (Generalized Seeing Monitor), showing decametric L_0 and its large variability. However, on many sites GSM results are very similar, with median L_0 around 20-25m: no more use to measure other sites? GSM results are being confirmed now by analysis of WFS data in AO systems and by direct measurements of path-length fluctuations in interferometers (MarkIII, SUSI, PTI, GI2T, VLTI). However, temporal analysis of interferometer data is wrong, only wave-front spatial structure leads to correct L_0 (JV). Nobody believes that V-K model is valid beyond ~10m, so L_0 remains model-dependent. AT suggested that PSF in M or L band is a way to get model-independent L_0, but WFS data on VLT can probably do the same.
SCINTILLATION may be an issue for extreme AO, but this is covered by the knowledge of profile or THETA_0.
Turbulence profilers
Several instruments/ideas were presented. B. Ellerbroek noted that a special "profilers" meeting (or calibration campaign?) would be desirable to compare, collaborate, verify, etc. A.Quirrenbach suggested Mount Wilson as test ground (high solar tower, 300m radio mast).
GENERALIZED SCIDAR No new developments. Remains off-line, for existing observatories only.
MASS Mentioned frequently, becomes an established technique for monitoring. Ideas on planetary scintillation: can have 3-4 points in the first km, to be pursued further.
SLODAR (R.Wilson, Durham): Now a working technique, an instrument for existing telescopes costs ~20k. Portable SLODAR (40-cm Meade, 3L CCD) under development, will be more expensive. Lim. magnitude V=7-8, profile is robust when normalized with simultaneous r_0 data. Wind velocity is measurable as well. Profile is derived by cross-correlation, thus noisedoes not produce any bias. A great advantage is double resolution: ~2km over whole atmosphere using close binary star, ~150m in the first km using wide binary star. Thus, good for the first km. B. Ellerbroek presented the theory of SLODAR tailored to profile extraction from MCAO WFS data.
Single-Star SCIDAR (J. Vernin): Uses cross-correlation with 2 exposure times to separate contributions of different layers by their velocity difference. Wind dispersion in layers is important, also included in the model. An instrument with 25-cm telescope is under development, but diffraction on small pupil in generalized mode remains an unsolved issue.
Single-Star SCIDAR (D. Coburn). A different approach: measure the autocorrelation in the pupil plane (also in generalized mode, with several shifts) and restore the profile by inverse-problem solution. A 25-cm telescope is planned, no instrument or results yet.
LIDAR (S. Esposito and R. Avila): Just an idea to launch a low-power (100mW) laser and to measure differential image motion with a very fast detector, following the laser pulse upwards propagation. No details or simulations.
KITES: J. Vernin advocated kites with micro-thermal probes to study the first km, experiments at Mauna Kea in Dec. 2002 successful.
Outstanding issues
OUTER SCALE and, more generally, power spectrum of turbulence at large spatial scales, remain uncertain. Profile of L_0(h) is needed. If von Karman were indeed valid for ELTs, there would be huge consequences (diffraction-limited cores in uncorrected images in the IR, no need for tip-tilt compensation with LGS, etc.).
TAYLOR HYPOTHESIS validity range remains to be established. If valid on 100ms scale, it gives tools to predict wavefronts and increase guide-star magnitude limits (RR). J. Vernin noted that decay of wave-front correlation with time is mostly due to wind velocity dispersion, not deviation from Taylor, and that Taylor hypothesis does not work for inside-dome turbulence. B. Ellerbroek noted that AO systems often compensate vibrations more than turbulence, so Taylor validity may be not so important after all.
BOUNDARY LAYER (first kilometer): need profiles with high vertical resolution and a solid statistics. Curiously, SODAR was not mentioned at all, some people do not even know what it means. New crazy ideas welcome.
NON-STATIONARITY: So far, turbulence variations in time and space are poorly characterized, people keep using stationary theory. We need to quantify the variations. Seeing variations across a 100-m ELT aperture is one example: should we split the aperture dynamically and direct its "best" portions to the most demanding instruments? Shall we suspend exposures when seeing suddenly degrades?
Consequences for AO design
Given a large range of variations of all atmospheric parameters, a poor knowledge of some, and a need to operate AO under worse-than-average seeing (especially on ELTs that are AO-dependent), there was an opinion that robust (bullet-proof) design of future AO is needed (B. Ellerbroek). Using some representative profile for AO/MCAO design should be replaced by using distributions of relevant quantities.
Consequences for site testing/selection
Little was said on this. A. Tokovinin insisted that understanding turbulence can be more important than measuring yet another mountain. Some progress on this is achieved. M. Sarazin said that altitude of turbulent layers identified at Paranal by MASS often correlate with instability regions in atmospheric profiles recuperated from on-line meteorological forecast models (GFS). Correlation with jet stream and predictions of good/bad conditions are encouraging.