misolfa solar monitor for the ground picard program
TRANSCRIPT
Astron. Nachr. / AN 331, No. 9-10 (2010) / P58
MISOLFA solar monitor for the ground PICARD program
T. Corbard1,?, A. Irbah2, P. Assus1, C. Dufour2, M. Fodil3, F. Morand1, C. Renaud1, and E. Simon1
1 Universite de Nice Sophia Antipolis, CNRS, Observatoire de la Cote d’Azur, BP 4229 06304 Nice Cedex 4, France2 Universite Versailles St-Quentin, CNRS/INSU, LATMOS-IPSL, Guyancourt, FRANCE3 CRAAG, Observatoire dAlger, BP 63 Bouzareah 16340 Alger, Algerie
Key words Atmospheric effects – methods: observational – site testing – Sun: general
Developed at the Observatoire de la Cote d’Azur (OCA) within the framework of the PICARD space mission (Thuillieret al., 2006) and with support from the french spatial agency (CNES), MISOLFA (Moniteur d’Images Solaires Franco-Algerien) is a new generation of daytime turbulence monitor. Its objective is to measure both the spatial and temporalturbulence parameters in order to quantify their effects on the solar diameter measurements that will be made from groundusing the qualification model of the SODISM (SOlar Diameter Imager and Surface Mapper) instrument onboard PICARD.The comparison of simultaneous images from ground and space should allow us, with the help of the solar monitor, tofind the best procedure possible to measure solar diameter variations from ground on the long term. MISOLFA is nowinstalled at the Calern facility of OCA and PICARD is scheduled to be launched in 2010. We present here the principles ofthe instrument and the first results obtained on the characteristics of the turbulence observed at Calern observatory usingthis monitor while waiting for the launch of the space mission.
c© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1 Introduction: why a new Solar monitor ?
Diameter measurement analysis revealed a dependence withthe seeing condition as represented by Fried’s parameterr0 (Irbah et al., 1994). Numerical simulations were devel-oped in order to better understand atmospheric effects ondiameter measurements. The error decreases with the see-ing but it is also strongly conditioned by turbulence coher-ence times (Fig.2, Lakhal et al., 1999). The error also showsweak dependence on the outer scale L0 for a small aperturetelescope. Existing solar monitors such as Solar DIfferen-tial Motion Monitors (S-DIMM) or arrays of scintillome-ters (SHABAR) are able to provide useful information onthe spatial scales of the turbulence and are commonly as-sociated for site testing (e.g. Beckers 2001). However ourgoal here is to obtain the accurate turbulence equivalent PSFneeded to properly interpret the ground based radius mea-surements. For this, estimates of the characteristic temporalscales are also needed and it was realized that a single in-strument could provide both the spatial and temporal turbu-lence scales by analysing Angle of Arrival (AA) fluctuationssimultaneously in its image and pupil ways.
2 Atmospheric parameters measured
An illustration of a multi-layer atmosphere model is givenon Fig. 3. MISOLFA will allow us to obtain information onall the following parameters using a Von Karman turbulencemodel:
– The Atmospheric structure constant of the air refractiveindex fluctuations C2
n(h).
? Corresponding author: e-mail: [email protected]
SODISM MountMISOLFA Focal Box
Fig. 1 Picture of MISOLFA monitor and the mount that will re-cieve the qualification model of the PICARD/SODISM telescopeat the Calern facility of OCA.
– Fried’s parameter r0 which is the diameter of the co-herence zone of the degraded wavefront. It correspondsalso to the image resolution obtained with the telescopeof diameter r0 placed outside the atmosphere.
– The spatial coherence outer scale L0 which defines themaximal size of wavefront perturbations remaining co-herent. It traduces the low frequency evolution of thewavefront.
– The isoplanatic patch θ0 which is the angle where AAor speckles remain correlated.
– The correlation time τ0 which is the time during whichthe atmosphere may be considered as frizzed for theconsidered structures (AA, speckles) i.e. the time dur-ing which they keep their coherence.
c© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
2 T. Corbard et al.: MISOLFA solar monitor for the ground PICARD program
Fig. 2 Estimated errors on diameter estimates as a function of the Fried’s parameter r0 for two values of Rτ (ratio between the exposuretime and the correlation time) (left) and as a function of Rτ for two values of r0 (right) (Lakhal et al., 1999)
Fried's parameter r0
Telescope pupil D
h1
hk
α 1 ( θ ) α 2 ( θ )
Extended object
Object image Telescope focal planeObserved fluctuations (d(θ ))
Angle of arrival
Characteristic time τ
Wave-front
Layer N
Layer k
Layer 1
Spatial coherence outer scale L0
C2n 1
C2n k
C2n N
θ
Fried's parameter r0
Telescope pupil D
h1
hk
α 1 ( θ ) α 2 ( θ )
Extended object
Object image Telescope focal planeObserved fluctuations (d(θ ))
Angle of arrival
Characteristic time τ
Wave-front
Layer N
Layer k
Layer 1
Spatial coherence outer scale L0
C2n 1
C2n k
C2n N
θ
Fig. 3 Illustration of a multi-layers atmosphere with VonKarman turbulence model (Irbah et al., 2003).
3 Measurement principles
The MISOLFA principle is based on the statistical analy-sis of AA fluctuations, which are fluctuations at each pointof the normal of the perturbed wavefronts. The AA fluctu-ations can directly be observed in the image plane (case ofShack-Hartmann’s sensors used in adaptive optics) but alsoin the pupil plane if the observed source present an intensitydistribution with a strong discontinuity like the Solar limb.The intensity fluctuations observed in the pupil image are,at first order, proportional to AA fluctuations (Borgnino &Martin 1977, Borgnino 1978).
Fig. 4 MISOLFA optical design. See text for details.
Fig. 5 Illustration of the formation of several solar images by theentrance prismatic blade. The CCD camera is positioned accrossthe two first solar images.
c© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.an-journal.org
Astron. Nachr. / AN (2010) 3
Pixels
Pixe
lsImage acquise à '2009-07-01T07:09:09.484' - EXPOSURE = 2 ms
100 200 300 400 500 600
50
100
150
200
250
300
350
400
450
0 100 200 300 400 500 600 700100
150
200
250
300
350
400
PixelsInt
ensit
y (AD
U)
Coupe de l'image à mi-hauteur
α1(θ1)
α2(θ2)
Fig. 6 Image sample from the MISOLFA monitor (left) and hor-izontal cut (right). Arrows illustrate the motions used to computethe transverse covariance between to angular positions (see text).
4 MISOLFA Instrument design
Figure 4 illustrates the optical design of MISOLFA. It ismade of a Cassegranian telescope of 25 cm aperture equipedwith an alt-azimuth mount to which is associated a derotat-ing system. A flat glass blade with non parallels faces allowsthe formation of separate images of the sun (Fig. 5). In thefocal box, a first beam (image way) is imaging two oppositeparts of solar limb on a 480x640 CCD. A tipycal sequenceis 1.5 mn long with 32 images per second and a resolutionof 0.2 arcseconds per pixel. A second beam (pupil way) isimaging the pupil after passing through a spatial filter (slit)in the image of a part of solar limb. In the pupil image 4optical fibers (with diameters 2 mm, 1 mm and 2x0.5 mm)guide the light to 4 diodes. In addition, in the image of theslit a 2 mm fiber measure global flux passing through theslit.
5 Analysis Method
In the image way we estimate r0, L0 and θ0 from directmeasurement of AA fluctuations on the limb images:
– r0 is directly related to the variance σ2s of the mean AA
fluctuations by:
r0 = 8.25105 D−1/5λ6/5(σ2s)−3/5
where D is the pupil diameter and λ the wavelength.– The AA transverse covariance is computed from the ob-
served limb fluctuations by (Fig. 6):
Cα(θ) =< α1(θ1)α2(θ2) >
L0 and θ0 = r0/h are then obtained by adjusting theobserved covariance to the theoretical one given, in thecase of the one-layer model and Von Karman theory, by:
Cα(θ) = 0.0716 λ2 r− 5
30
∫ ∞0
f3
(f +
1L2
0
)− 116
. . .
. . . [J0(2πfθh) + J2(2πfθh)][
2J1(πDf)πDf
]2df
where h is the altitude of the equivalent impulse layergiving at ground level the same optical effects than thewhole turbulent terrestrial atmosphere.
– The turbulence profile C2n(h) is then obtained by invert-
ing (Bouzid et al., 2000) the structure function which, ina multi-layer model, is given by:Dα(θ) = 2 [Cα(0)− Cα(θ)]
= 2.4∫ ∞
0
C2n(h)K(θ, h,D,L0)dh
with:
K(θ, h,D,L0) =∫ ∞
0
f3
(f +
1L2
0
)− 116
. . .
. . . [1− J0(2πfθh)− J2(2πfθh)][2J1(πDf)πDf
]2df
In the pupil plane, we perform a spatio-temporal analysisof AA fluctuations (Berdja, 2007). The photodiode detec-tors used represent various collecting surfaces of the pupil-plane. They allow fast recordings (5KHz) and thus an accu-rate intensity signal time sampling. Such measurement timeseries will be used to study temporal properties of the di-urnal turbulence and continuously estimate the correlationtime τ0 of AA-fluctuations from the estimated temporal co-variance assuming the Von Karman model and Taylor’s hy-pothesis of frozen turbulence. Such as for other seeing mon-itors like the well-known S-DIMM or GSM, dual measure-ments from photodiode pairs on the same collecting sur-face will allow also estimate the Fried parameter r0. Theneeded transverse temporal covariance is obtained by doingθh = vτ in the previous formula, v denoting an equiva-lent speed (weighted average of the different layer speeds)and τ a temporal shift (Conan et al., 2000). The goal of us-ing photodiode detectors with different collecting surfacesis to estimate the spatial coherence outer scale L0 from AA-fluctuation statistics.
6 Observations and first results
Sequences of images are recorded at Calern observatorysince june 2009. A typical sequence consists in a set ofabout 2500 images with an angular resolution of 0.2 arc-seconds per pixel recorded at a rate of 32 images per sec-ond. The exposure time of each image is 1 ms. The meanlimb position is evaluated for each image in order to cor-rect in real time the guiding errors. The limb profile of eachimage is first extracted using wavelet based denoising andinflexion point identification (Irbah et al. 1999) as illustratedFig. 7. Then, from the temporal evolution of all the points ofthe extracted limb profiles (Fig. 8), the variance of the meanAA fluctuations and the Fried’s parameter are estimated.When estimating the Fried’s parameter from increasing timeintervals, the value decreases while integrating clearly dif-ferent turbulence regimes (Fig. 9). Figure 10 shows Fried’sparameter r0 estimated on a continuous set of two secondsintervals which correspond to the exposure time of the fulldisk images that will be recorded by the SODISM instru-ment. The full time sequence is therefore cuted in 40 inter-vals of two seconds. On these time intervals r0 can reach up
www.an-journal.org c© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
4 T. Corbard et al.: MISOLFA solar monitor for the ground PICARD program
Fig. 7 The limb profile of each image is extracted using waveletbased denoising and inflexion point identification (Irbah et al.1999)
Fig. 8 Temporal evolution of all the points of the extracted limbprofiles.
0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0
2
3
4
5
6
7
8
T i m e ( s )
r 0 (cm)
D a t e : 2 0 0 9 - 0 7 - 0 1 0 7 : 0 9 - r 0 c a l c u l é s u r d e s p e r i o d e s d ' i n t e g r a t i o n c r o i s s a n t e ( r : p i x e l s a b e r r a n t s c o r r i g é s )
2
3
4
5
6
7
8
R 0 (cm)
Time (s)10 20 30 40 50 60 70 80
Fig. 9 Fried’s parameter r0 estimated from increasing time in-tervals.
to 14 cm while over the whole 75 s the value would be about2.6 cm (Fig. 9).
7 Conclusion
MISOLFA is a new type of solar monitor based on the anal-ysis of AA statistic that we record both in the image andpupil plane. It will allow us to get a full characterizationof the spatio-temporal parameters of the turbulence neededto identify the appropriate atmospheric model and build theturbulence + instrument equivalent PSF. The first sequencesof images obtained where analyzed to extract the Fried’s
0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0
2
4
6
8
1 0
1 2
1 4
i n d e x
r 0 (cm)
D a t e : 2 0 0 9 - 0 7 - 0 1 0 7 : 0 9 - r 0 c a l c u l é s u r d e s p e r i o d e s d ' i n t e g r a t i o n c o n s t a n t e ( r : s u p p r e s s i o n b a d p i x e l s )
2
4
6
8
10
12
14
R 0 (cm)
Time interval5 10 15 20 25 30 35 40
Fig. 10 Fried’s parameter r0 estimated on two seconds intervals.
parameter showing that good condition can eventually bereached at Calern observatory on some intervals for the 2 sexposure time of SODISM. The continuous record of the r0values will help us in the interpretation of ground SODISMimages while the PICARD satellite is operating. The pro-cedure to extract all the other spatial turbulence parametershave been tested and are now being implemented. The pupilway is still not operating properly and lots of effort are cur-rently made to increase the signal to noise ratio by improv-ing the electronic components.
Acknowledgements. This work has been performed with supportof the Observatoire d’Alger - C.R.A.A.G., the French Foreign Af-fair Ministry, the Observatoire de la Cote d’Azur and of the CentreNational d’Etudes Spatiales. We thank the European Helioseismol-ogy and Asteroseismology Network (HELAS) for supporting ourparticipation to the HELAS-4 conference. HELAS is a major in-ternational collaboration funded by the European Commission’sSixth Framework Programme.
References
Beckers, J.: 2001, ExA 12, 1Berdja, A.: 2007, Ph-D thesis, Nice Sophia-Antipolis UniversityBerdja, A., Irbah, A., Borgnino, J., Martin F.: 2004, SPIE 5237,
238Borgnino, J., Martin, F.: 1977, JOpt. 8, 319Borgnino, J.: 1978, Doctorat thesis, Nice UniversityBouzid, A., Irbah, A., Borgnino, J., Lanteri, H.: 2000, ASP Confer-
ence Proceedings, Eds : J.Vernin, Z. Benkhaldoun, C. Munoz-Tunon, Vol 266
Conan, R., Borgnino, J., Ziad, A., Martin, F.: 2000, JOpt Soc. Am.17, 1807
Irbah, A. Bouzid, A.-E., Lakhal, L. et al.: 2003, Lecture Notes inPhysics, Berlin Springer Verlag, Ed: J.-P. Rozelot, vol 599,159
Irbah, A., Bouzaria, M., Lakhal, L., Moussaoui, R., Borgnino, J.,Laclare, F., Delmas C.: 1999, SoPh 185, 255
Irbah, A., Laclare, F., Borgnino, J., Merlin, G.: 1994, SoPh 149,213
Lakhal, L., Irbah, A., Bouzaria, M., Borgnino, J., Laclare, F., Del-mas, C.: 1999, A&AS 138, 155
Thuillier, G., Dewitte, S., Schmutz, W., the PICARD team: 2006,AdSSpR 38, 1792
c© 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.an-journal.org