Optimising tomographic reconstructor parameters for adaptive optics using real turbulence profiles

The performance of tomographic adaptive optics systems will depend on the vertical profile of turbulence in the atmosphere. Since the profile is changing over time, to maintain optimal correction the tomographic reconstructor must be continually updated with new profile information. Several reconstructor parameters must then be chosen to optimise performance given the constraints of real-time computing resources: the number of reconstructed layers, reoptimisation period and averaging time. We analyse the effect of changing these parameters by coupling fast Fourier-domain AO simulation with a large database of over 10,000 high resolution turbulence profiles measured by the Stereo-SCIDAR at Paranal.