The Atmospheric Emission Signal as seen with SHARC-II

1
The Atmospheric Emission Signal as seen with
SHARC-II

• Alexander van Engelen
• University of British Columbia

2
Who am I?

• Working with
• Douglas Scott (UBC)
• Andy Gibb (UBC)
• Tim Jenness (JAC, Hilo HI)
• Dennis Kelly (UK ATC, Edinburgh)
• on the data reduction pipeline software for
• SCUBA-2
• Scanmap research

3
Introduction

• In sub-mm experiments, atmospheric emission due
  to water vapor is the strongest component of the
  data
• Bright
• Varies on long spatial and temporal scales
• For data reduction studies it is important to
  model this in a useful and accurate way

4
Current Model

• Use a fluid dynamic model (Kolmogorov) to
  describe the characteristics of the signal
• In SCUBA-2 simulations the emission from various
  altitudes is approximated by a single, constant
  screen of emission
• Gaussian 2-D random field
• Fixed at an altitude of 800 m, and blows past
  the observatory at the local wind speed

5
Atmospheric Emission Image

• Features here are very large
• This constant screen blows past the observatory
  at 15 m/s (5000 arcsec/sec)

6
SCUBA-2 scanmap basics

• Simple raster scan
• To fill in the under-sampled 450µm array, scan at
  an angle of arctan(1/2)26.5 to array axes

(courtesy D, Kelly)
7
SCUBA-2 scanmap simulation

• Simulated scan over a regular 2-d array of point
  sources
• Just a simple reprojection of the time series
  onto a map nothing fancy here!
• Note streaks due to atmospheric emission signal

8
Issues

• Is this truly a Gaussian random field?
• Power spectrum?
• Constant wind vector?
• Component on scales smaller than the array?
• How stable are the properties of the screen?

9

• In order to learn more about the properties of
  this signal Colin Borys kindly gave us some
  SHARC-II data
• Lissajous scan of MS0451
• considered to be faint enough that source flux
  can be neglected here

10

• The data is overwhelmingly common-mode across the
  array

11
Sample array-mean signals
12
Sample spectra
Model predicts (in the timestream)
13
Animation

• From a data reduction perspective we are
  interested in whether the atmosphere is
  completely described by a common-mode signal.

14
Residual after a common-mode signal is subtracted
away
15
Residual after common-mode subtraction
Still some correlated structures remaining
16
Zero-timelag cross-correlations in residuals
-Difficult to understand
17
Direct detection of Kolmogorov model? (I)

• Compare derivative of array mean with slope of
  fit plane across the array positive correlation
  indicates a comfirmation
• It turns out that for the SHARC-II data the
  gradient is overwhelmed by instumental effects.

18
Direct detection of Kolmogorov model? (II)

• There should be a shift in the signal (of a
  fraction of a sample) between detectors if the
  wind speed is reasonable
• Unfortunately it is difficult to measure this
  explicitly.

19
Conclusions

• Since the data is so common-mode, subtracting a
  simple mean is not ruled out as a way of dealing
  with the atmospheric emission signal. However,
  there seem to be some small-amplitude correlated
  structures that remain.
• No direct detection of the wind-blown screen
  model was made in SHARC-II data.

20
(Fin)