VERTICAL SCALES OF TURBULENCE AT THE MOUNT-WILSON-OBSERVATORY

The vertical scales of turbulence at the Mount Wilson Observatory are inferred from data from the University of California at Berkeley Infra red Spatial Interferometer (ISI), by modeling path length fluctuations observed in the interferometric paths to celestial objects and those in instrumental ground-based paths. The correlations between the stell ar and ground-based path length fluctuations and the temporal statisti cs of those fluctuations are modeled on various timescales to constrai n the vertical scales. A Kolmogorov-Taylor turbulence model with a fin ite outer scale was used to simulate ISI data. The simulation also inc luded the white instrumental noise of the interferometer, aperture-fil tering effects, and the data analysis algorithms. The simulations sugg est that the path delay fluctuations observed in the 1992-1993 ISI dat a are largely consistent with being generated by refractivity fluctuat ions at two characteristic vertical scales: one extending to a height of 45 m above the ground, with a wind speed of about 1 m s(-1), and an other at a much higher altitude, with a wind speed of about 10 m s(-1) . The height of the lower layer is of the order of the dimensions of t rees and other structures near the interferometer, which suggests that these objects, including elements of the interferometer, may play a r ole in generating the lower layer of turbulence. The modeling indicate s that the high-altitude component contributes primarily to short-peri od (less than 10 s) fluctuations, while the lower component dominates the long-period (up to a few minutes) fluctuations. The lower componen t turbulent height, along with outer scales of the order of 10 m, sugg est that the baseline dependence of long-term interferometric, atmosph eric fluctuations should weaken for baselines greater than a few tens of meters. Simulations further show that there is the potential for im proving the seeing or astrometric accuracy by about 30%-50% on average , if the path length fluctuations in the lower component are directly calibrated. Statistical and systematic effects induce an error of abou t 15 m in the estimate of the lower component turbulent altitude.