Cn 2 profile reconstruction with ShackHartmann slope and

Cn 2 profile reconstruction with Shack-Hartmann slope and scintillation data: first on-sky results J. Voyez(1), C. Robert(1), J. -M. Conan(1), V. Michau(1), L. Mugnier(1), B. Fleury(1), E. Samain(2) in collaboration with Aziz Ziad (Nice University) (1) ONERA, The French Aerospace Lab (2) Observatoire de la Côte d’Azur

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Outline 2 • CO-SLIDAR motivation & principle • On-sky results on 1. 5 m telescope • ELT perspectives & conclusions

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Outline 3 • CO-SLIDAR motivation & principle • On-sky results on 1. 5 m telescope • ELT perspectives & conclusions

Why high resolution Cn 2 profiling ? extracted from ONERA ATLAS/MAORY AO analysis (E-ELT Phase A studies) 6 LGS WFAO on 4. 3 arcmin or 2 arcmin ring 4. 3 arcmin J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 2 arcmin 4 • Too coarse Cn 2 profile in turbulence & WFAO simulations may lead to very optimistic performance effect is increased when considering very large Fo. V systems • High resolution Cn 2 site characterization campaigns are needed sub-km resolution would be nice see also Elena Masciadri Poster-13542

CO-SLIDAR principle 1/2 : slope &! scintillation from S. -H. data Turbulent wavefront Microlenses J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Pupil plane 5 Detector Turbulent image Optical axis Focal plane SH data for a star at angular position α : • : slope in sub aperture m • : intensity in sub aperture m • : temporal average of Ø scintillation index: scintillation meas. implies few cm subapertures

CO-SLIDAR principle 2/2 : double star triangulation Altitude star 2 θ star 1 dmn: separation between subapertures Altitude layer h m n Ground layer dmn Correlation maps example: 2 layers {0, h} same strength One pixel gives auto/inter-correlation between all couples with given separation J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Ground layer 6 dmn = θh Altitude layer Autocorrelation of y-slopes Intercorrelation of y-slopes Altitude layer Autocorrelation of scintillation Intercorrelation of scintillation

Direct problem and inversion Correlations of slopes and scintillation Convergence noise Unknown J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 : matrix of weighting functions W 7 Correlations of detection noise The Cn 2 profile is estimated by minimization of J, the maximum likelihood criterion, under positivity constraint (+βJp) Inverse of the convergence noise covariance matrix Regularization term

Reconstructed Cn 2 profile: simulation data J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Voyez-SPIE 2012 Dtel = 1. 5 m 30 x 30 SH. 20’’ separation 8 CO-SLIDAR takes advantage of the complementarity slope vs scintillation leading to better precision / resolution here resolution is about 500 m in [0, 15 km] range

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Outline 9 • CO-SLIDAR motivation & principle • On-sky results on 1. 5 m telescope • ELT perspectives & conclusions

Pro. Me. O: profilometry with Me. O telescope Site : OCA (Calern, near Nice) Me. O telescope: Dtel = 1. 5 m, CO = 30% May 2012 observation campaign • 30 x 30 subapertures SH (dsub = 5 cm) • ANDOR 1 K x 1 K EMCCD • 2 nights of observations J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Me. O telescope 10 The bench

Shack-Hartmann image samples J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 STF 1744: Binary star with sep 14. 4’’, mag: 2. 23 and 3. 88 l = 517 nm ; Dl = 100 nm ; ~260 & 60 ph/subap/frame 11 14. 4’’ Temporal average of short exposures Short exposure: texp = 3 ms

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Preliminary data reduction: STF 1744 3 ms 12 15 radial orders reconstructed • r 0 is estimated on measurements excluding orders 1 and 2 • L 0 is difficult to estimate, of the order of 10 or 27 m, depending on the data set • We find good agreement with Kolmogorov turbulence, with outer scale effect

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Preliminary data reduction: STF 1744 3 ms 13 The intensity distribution is very close to expected log-normal distribution • σ2χ << 0. 3 • We are in the weak perturbation regime

Correlation maps: STF 1744 3 ms J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Autocorrelation of x-slopes 14 Autocorrelation of y-slopes Autocorrelation of scintillation Turbulent signal ! Intercorrelation of x-slopes Intercorrelation of y-slopes Intercorrelation of scintillation

Cn 2 profile restoration • • • assumption : L 0 = 27 m (median L 0 at Calern, see R. Conan Ph. D thesis) Sequences of 1000 images at ~ 15 Hz, duration ~ 1 min Sub-aperture diameter : dsub = 5 cm Binary separation: θ = 14. 4’’ Zenith angle: ζ = 35° Altitude range and resolution from simple geometrical consideration: • Altitude resolution : J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 here: δh ~ 600 m 15 • Maximum altitude: here: Hmax ~ 17 km

CO-SLIDAR Cn 2 profiles & comparison with SLODAR / SCIDAR J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 discarding xentral autocorrelation point affected by of detection noise bias 16 • good agreement between CO-SLIDAR & SLODAR @ low altitude but overestimation of SLODAR @ medium altitude • SCIDAR reconstruction more questionable

CO-SLIDAR Cn 2 profiles & comparison with SLODAR / SCIDAR J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 with estimation and subtraction of the detection noise bias on scintillation variance 17 • better agreement between CO-SLIDAR & SCIDAR central point of the scintillation autocorrellation has its importance for global normalisation

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 CO-SLIDAR Cn 2 profiles: regularization 18 smoother profile (more realistic ? ) We plan a comparison with free atmosphere Cn 2 deduced from meteorological reanalysis [see Hach(. . . )Ziad et al. MNRAS 2011]

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Outline 19 • CO-SLIDAR motivation & principle • On-sky results on 1. 5 m telescope • ELT perspectives & conclusions

CO-SLIDAR in the Cn 2 profiler & ELT landscape • • CO-SLIDAR on meter class telescope provides high resolution Cn 2 profiles • site characterization to obtain relevant inputs for WFAO design and performance evaluation • data to help optical turbulence forecast [see E. Masciadri earlier talk ] Joint use of slopes & scintillation should lead to more robust profile restoration, with better resolution over whole altitude range compared to SLODAR or SCIDAR instruments • Of course inter-comparison campaigns are needed (SLODAR, gen. SCIDAR, J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 PBL. . . ) [see E. Masciadri Poster-13542, A. Ziad Poster on PBL] 20 • WFAO ELT instruments will include multi. WFS hence super-SLODAR capacities: should not need external measurements for control optimisation (also true for wind profiling? ) use of Cn 2 profile for LQG control WFAO : see Gaetano Sivo Friday talk first on-sky validation on Canary

Conclusions and perspectives • • A CO-SLIDAR profiler has been set up on a 1. 5 m telescope Cn 2 profiles have been restored from both slopes & scintillation correlations J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Other observation campaigns are needed in order to calibrate and compare CO-SLIDAR with other Cn 2 profilers 21 IR CO-SLIDAR ground-ground experiment conducted in fall 2012 (ONERA with French labs LTHE-INRA-CESBIO) 4µm ; 3 km range ; 21 days of SH data with SCIDAR and temperature probes data processing in progress

CO-SLIDAR principle 2/2 : double star triangulation Altitude θ double star assumed to be resolved by subaperture dmn = θh h Altitude of maximum sensitivity n m J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 dmn 22 Correlations of slopes: sensitivity to near ground layers • Correlations of scintillation indexes: better sensitivity to high altitude layers • We take advantage of both kind of correlations to retrieve the Cn 2 profile COupled SLope and sc. Intillation Detection And Ranging

J. Voyez -- Onera – AO 4 ELT 3 – May 29 th 2013 Cn 2 profiles: effect of outer scale L 0 23