5 Conclusion

In this paper we have described an implementation of generalised SCIDAR that provides estimates of the profiles of the refractive index structure function constant C_n²(h) and velocity profile V(h). The technique has been verified using Monte Carlo propagation code and sample results from three locations presented. These show a fairly similar form, and as a general rule there are apparently just one or two dominant layers in addition to the inevitable layer in the region of the telescope itself.

The SCIDAR technique gives an excellent picture of the relative vertical distribution of optically significant turbulence; however, for the finite sampling levels here, there appears to be a tendency to obtain estimates of C_n²(h) that are too small, producing (for example) values of the Fried parameter r₀ that are too large. This could be due to spatial or temporal averaging effects, or to non-Rytov or non-Kolmogorov effects, since the theory assumes both Kolmogorov turbulence and the Rytov approximation. It would be preferable to simultaneously make a measurement (of, for example, the Fried parameter) that would allow a calibration of C_n²(h) to be made.

SCIDAR requires a medium-sized telescope, typically on the order of 1 m, and suitable double stars of combined magnitude brighter than $m_v\approx$5. Both of these facts are rather restrictive: for example, it would be desirable to have a C_n²(h) measurement device attached to a large telescope equipped with an adaptive optics (AO) feed to its instruments, so that the AO system could continuously operate in an optimised fashion such as being conjugated to a mean layer height. As an alternative to SCIDAR, single star techniques based on inverting the correlation of the scintillation may provide crude but adequate estimates of both C_n²(h) and V(h).

Acknowledgements

We wish to thank Jean Vernin and Michel Tallon for the help and encouragement they have given us during our development of the SCIDAR technique and particularly in introducing us to the generalised SCIDAR method, and we also acknowledge Jean Vernin's contribution to this field over a number of years. We are also indebted to Nielson Roberty for his great contribution to our development of algorithms for the C_n²(h) inversion problem.

We wish to thank all of those observatory staff who helped in the preparation of our observations at the JKT at La Palma, the 40 inch at Siding Spring and the 2.2 m at Calar Alto. In

particular we wish to thank, Richard Wilson at RGO, John O'Byrne at Sydney University and Steve Beckwith at the Calar Alto Observatory for their continued support of, and practical help with, this project.

The observatory site at La Palma (and its sister on Tenerife, the Observatorio del Teide) is operated by the Instituto de Astrof¡sica de Canarias. The Isaac Newton Group (of which the JKT is a part) is operated by the Royal Greenwich Observatory on behalf of the UK Particle Physics & Astronomy Research Council (PPARC) and the Nederlandse organisatie voor wetenschappelijk onderzoek (NWO).

The Visiting Astronomer, German-Spanish Astronomical Centre, Calar Alto, is operated by the Max-Planck-Institute for Astronomy, Heidelberg, jointly with the Spanish National Commission for Astronomy.

This work is supported by the Particle Physics & Astronomy Council, UK.