Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.mrao.cam.ac.uk/~bn204/almasim08/papers/0809-alma-simu-5.pdf Дата изменения: Tue Dec 2 15:00:05 2008 Дата индексирования: Tue Oct 2 13:06:50 2012 Кодировка: Поисковые слова: arp 220

Developing On-the-fly observations for ALMA

Nemesio RODRIGUEZ FERNANDEZ Jerome PETY, Frederic GUETH IRAM


Wide field imaging
M51 CO(2-1) M51 CO(2-1) M51 CO(2-1)

Schuster et al. (2007)

ALMA field of view at 230 GHz


Mosaics
· Stop-and-go

· On-the-fly (OTF)
­ Advantages:
· Faster, time efficient, better quality data · Larger fields · Open new ways of data editing and processing ... although this is non-trivial, never done before






OTF simulator
An OTF simulator based on the GILDAS ALMA simulator


OTF simulator

PdB Conf D


Imaging as a classical mosaic


Deconvolution


OTF simulator

Without short-spacings


OTF-specific algorithm


Ekers & Rots (1979)

; "super-visibility function"

for constant u

Direct measurement of the true visibility of the source in the surroundings of the point u


ER79: intuitive idea

d

v (m)

D

D-d

D+d

u (m)


ER79: intuitive idea
Analyzing how the visibilities change from one position of the source to another, it is possible to recover the information around a (u,v) point

v (m )

u (m)






From Nfields uv planes to a global "wide-field" uv-plane

v (m)

Nfie

lds

uv-planes with "averaged"

1 uv-plane for the whole mosaic

u (m)

visibilities


How to develop an ER79-based algorithm
One uv-plane per pointing (lp,mp)
V i s i b i l i ty m a p V(ui,vj) (lp,mp)

?

(ui,vj)

mp lp

FT 1/FT(B15)

~Super-visibilities P(up,vp)/FT[B(up,vp)]

......

vp up

......

FT[I](ui+up,vj+vp)

......

- Weight estimation - Weighted linear combination of the functions

to get the global uv-plane - Global imaging and deconvolution


How to develop an ER79-based algorithm
V(lp,mp,u,v)
S a m pl in g ? G r idi n g ?

?
Vg(lp,mp,u,v) ......
(ui,vj)

Vg(lp,mp,u,v) ...... ...... ...... ......

...... ...... ......

Visibility map for (ui,vj)

...... mp

......

Cell size?
G r idi n g ? S a m pl in g ?

P(up,vp)/FT[B(up,vp)]

mp lp

FT 1/FT(B15)

vp up

lp

- Weight estimation? - Weigthed linear combination of the functions to get the global uv-plane - Global imaging and deconvolution


Griding and sampling in 4D (lp,mp,u,v)

Equation similar to the classical relation between the gridded dirty image, gridded dirty beam, and beam-apodized sky brightness distribution Problems: - if we continue working with Pg(up,vp,u,v) it is not possible to correct for the griding convolution - otherwise a deconvolution (4D) is already needed to obtain P(up,vp,u,v)


Griding and sampling in 2D (u,v)
Limiting cell size Gridding in (lp,mp) by observational procedure

FT(Vg)


Griding correction and final imaging and deconvolution


Map size and scanning velocity
time Vscan Mosaic size

?
Shift in the uv plane from first field to last field => u, v (l, m)~ (1/u, 1/v) Mosaic size

PdBI =30є


Dumping time

PdBI =30є

min tdump=1 sec max vscan=10''/sec


Summary
· We have now an On-the-fly (OTF) observations simulator to develop OTF-specific imaging techniques. · We have done feasibility studies of OTF-specific imaging algorithms. ­ The general case is far from trivial... ­ A workaround is to impose some constrains on the observation procedures (Cartesian grids, mosaic size) PS: In addition, we have done analytical and numerical calculations of

the error transfer in the pseudovisibilities computations to optimize the addition of short-spacing data to wide-field interferometric observations


What next ?

· Publications of reports:
­ Short spacings addition: pseudovisibilities weights ­ Feasibility studies of OTF-specific imaging algorithms

· Implementation of the restricted algorithm · Exploration of the limits of the algorithm using the OTF simulator as input data



Weights and global uv plane
2


On the griding in the uv plane
We can be sure that the uv points corresponding to the different fields are not shifted to a distance larger than the cell size ... ... but this does not mean that they will lie exactly in the same cell, they can be in contiguous cells ... ...... ... and to be close to the theory we should have a measurement of the visibility at a the same (u,v) point for all the mosaic fields ......

...... ...... ......

...... ...... ......


Dumping time & effective beam

(Vscan tdump ) < 1/3 FWHM of the primary beam