Simulation of an NGS Speckle Observation with a 30 meter Telescope
Here is a movie showing simulations of wavefront propagation
through turbulence for a 30 meter telescope.
30 meter simulation using a single layer atmosphere
On the left side of the movie is the wavefront phase in the pupil
plane of a 30 meter aperture. On the right is the phasor amplitude of
the electric field in the image plane (i.e. the square root of the
intensity that would be detected by a camera). This simulation was
performed by creating a turbulent layer at the ground, placing an
aperture at the ground, applying these "optics" to a plane wavefront
to get the pupil plane phase, and then propagating the resulting
wavefront to the far field to get the image plane amplitudes. In this
movie you can see that the motion of individual speckles is correlated
with the wind velocity, which for this simulation was due north.
The parameters for this simulation were:
30 meter aperture
1 micron radiation
r0 of .3 meters at 1 micron
3 cm pixel scale in the wavefront
3 cm pixel scale in the turbulence layer
time step of duration 20 milliseconds
10 seconds (500 20 ms frames) of simulation time
This simulation was generated using the program simple_simulation, which is
provided as part of the arroyo distribution.
Here is another simulation in which a six layer atmosphere has been
used, and wavefronts have been propagated through this atmosphere
using diffractive rather than geometric propagation.
30 meter simulation using a 6 layer atmosphere
The simulation was run with parameters that differed slightly compared
to those above. First, the time step is 10 milliseconds rather than
20, and the duration of the simulation is 5 seconds instead of 10.
Second, the six layer atmospheric model that I used has an equivalent
r0 of .37 meters at 1 micron.
On the left appears the wavefront phase in the pupil plane of the 30
meter telescope. In the middle appears the wavefront amplitude in the
pupil plane of the thirty meter telescope. The wavefront amplitude
shows the effects of scintillation, which arises from diffractive
propagation effects. Briefly, phase errors introduced by high
altitude turbulence are mixed into amplitude fluctuations as the
wavefront propagates through free space. Finally on the right appears
the resulting PSF. I put the PSF into a separate movie to permit
closer inspection.
PSF only
There are a couple of interesting differences between the single layer
30 meter simulation above and this simulation. First, the multilayer
atmosphere permitted simulation of diffractive propagation, which
allows us to qualitatively inspect scintillation in the pupil plane.
The wavefront amplitude fluctuations from scintillation are of order
10 percent at 1 micron, and consequently the intensity fluctuations
would be of order 20 percent.
A second point of interest is that the apparent velocity of the pupil
plane phase is up and to the right, while that of the pupil plane
amplitude is up and to the left. This effect arises from the fact
that the velocities of the low altitude layers and the high altitude
layers were different in the simulation. The turbulence profile is
exponentially weighted towards the ground, so that low altitude layers
- which happen to move up and to the right in this particular
simulation - dominate the wavefront phase. Scintillation is dominated
by high altitude layers, which happened to be moving up and to the
left in this particular simulation.
A final point of interest is that the lifetime of speckles in the psf
appears much shorter in the six layer simulation than in the single
layer simulation. And while the speckles in the single layer
simulation appeared to translate in a direction correlated with the
wind vector of the layer, speckles in the six layer simulation appear
to wander around a short time before fading away.