Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.mrao.cam.ac.uk/~bn204/publications/2010/2010-10-14-Annual.pdf Äàòà èçìåíåíèÿ: Fri Aug 10 20:18:03 2012 Äàòà èíäåêñèðîâàíèÿ: Tue Oct 2 13:40:55 2012 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: http astrokuban.info astrokuban

EU FP6 ALMA Enhancement: Work Package 5 Prgress, First results at ALMA & Plans
B. Nikolic & J. S. Richer
Cavendish Laborator y/Kavli Institute University of Cambridge

14 October 2010 Annual ALMA Enhancement Meeting, Rutherford Appleton Lab

(University of Cambridge)

October 2010

1 / 34


ALMA Current AOS configuration
from http://wikis.alma.cl/bin/view/AIV/CommissioningStatusReview

Photo: N. Mizuno (University of Cambridge) October 2010 2 / 34


Introduction

Outline

1

Introduction Progress since last Annual Meeting First results at ALMA Plans Summar y

2

3

4

5

(University of Cambridge)

October 2010

3 / 34


0.0092

Introduction

Path fluctuations due to the atmosphere
0.009 750

500

250 L (µ m)

0 -250 -500 -750 6

6.2

6.4

6.6 t (hours UT)

6.8

7

7.2

(University of Cambridge)

October 2010

4 / 34


Introduction

Phase closure
Antenna 0 Vs 1
500

Antenna 0 Vs 2
1250 1000

0

L (µ m)

-1000

L (µ m)

-500

750

500

-1500

250

-2000

0
7.1 7.15 7.2 7.25 7.3 7.35 7.4 7.45

7.1

7.15

7.2

7.25

7.3

7.35

7.4

7.45

t (hours UT)

t (hours UT)

Antenna 1 Vs 2
-1000 1000 -1500 500 L (µ m) L (µ m) -2000

Closure phase

0

-2500

-500 -3000 -1000 -3500 7.1 7.15 7.2 7.25 7.3 7.35 7.4 7.45 7.1 7.15 7.2 7.25 7.3 7.35 7.4 7.45

t (hours UT)

t (hours UT)

(University of Cambridge)

October 2010

5 / 34


Introduction

Atmospheric Phase Fluctuations

(University of Cambridge)

October 2010

6 / 34


Introduction

Atmospheric Phase Fluctuations

(University of Cambridge)

October 2010

6 / 34


Introduction

Atmospheric Phase Fluctuations

(University of Cambridge)

October 2010

6 / 34


Introduction

Atmospheric Phase Fluctuations

(University of Cambridge)

October 2010

6 / 34


Introduction

WVR in the ALMA receiver cabin

(University of Cambridge)

October 2010

7 / 34


Introduction

Water Vapour cm/mm/sub-mm lines
1 mm precipitable water vapour
300 250 200 Tb (K) 150 100 50 0 200 400 ( GHz )
(University of Cambridge) October 2010 8 / 34

600

800

1000


Introduction

The 183 GHz Water Vapour Line
Blue rectangles are the production WVR filters

250

200

150 Tb (K) 100 50 0 175 177.5 180 182.5 (GHz)
(University of Cambridge) October 2010 9 / 34

185

187.5

190


Introduction

Sky brightness observed by WVRs
250 250 200 TB (K) TB (K) 6.25 6.5 6.75 7.25 200

150

150

100

100

50 6 250 7 t (hours UT)

50 6 6.25 6.5 6.75 7 7.25 t (hours UT)

200 TB (K)

150

100

50 6 6.25 6.5 6.75 7 7.25 t (hours UT)
(University of Cambridge)

Observed brightness temperatures of WVR on the three antennas involved in this test observation. The four colours in each panel are the four channels of the WVRs.
October 2010 10 / 34


Introduction

Correlation between WVR signals and path
8 1.5 1 0.5 TB,2 (K) 4 0 4 -0.5 -0.5 2 -1 0 -200 -100 0 L (µ m) 100 200 300 -200 -100 0 L (µ m) 8 1 1 8 100 200 300 -1 -1.5 0 2 6 8 1 6 0.5 TB,1 (K)

0

0.5 TB,3 (K)

0.5 6 TB,4 (K)

6

0 4 -0.5 2 -1 0 -200 -100 0 L (µ m) 100 200 300

0

4

-0.5

2

0 -200 -100 0 L (µ m) 100 200 300

(University of Cambridge)

October 2010

11 / 34


Introduction

Conver t WVR signals into path fluctuation estimates
This allows to correct the phase!

2500

0 L (µ m)

-2500

-5000 0 1000 2000 Time (s) 3000 4000

(University of Cambridge)

October 2010

12 / 34


Progress since last Annual Meeting

Outline

1

Introduction Progress since last Annual Meeting First results at ALMA Plans Summar y

2

3

4

5

(University of Cambridge)

October 2010

13 / 34


Progress since last Annual Meeting

Highlights

1

2

Working interferometer at the AOS with good instrumental stability; WVRs installed in all antennas and showing good performance & reliability Completion and verification of a fully ALMA-compatible software framework End-user availability of initial phase-correction algorithms, testing on ALMA data & application to science data Initial version of User documentation, user training and feedback Documentation of current algorithms, implementation and testing framework, and effects of dispersion

3

4 5

(University of Cambridge)

October 2010

14 / 34


Progress since last Annual Meeting

wvrgcal ­ the

ALMA/CASA

phase correction framework

This is the user-facing application that does WVR phase correction Fully compatible with official ALMA software:
Reads ALMA/CASA Measurement Sets Writes CASA Gain Calibration ("T"-Jones) tables Callable and scriptable from CASA Inter-operates with CASA facilities like applycal, accum, plotcal & browsetable

Built on top of:
LibAIR, the phase correction librar y CASA-Core and CASA libraries for input/output

Available for public download, installed on computers at the JAO and at NRAO/Charolttesville A complete, working, end-to-end verified, fully compatible, software framework for phase correction
(University of Cambridge) October 2010 15 / 34


Progress since last Annual Meeting

The initial phase-correction algorithm
Developed in 2008, documented in ALMA memo 587 http://www.mrao.cam.ac.uk/~bn204/publications/ 2009/ALMAMemo587.pdf Based on:
The four absolute sky brightness measurements only Bayesian analysis, including priors and evidence calculation Extended to take into account bias due to clouds

Implementation in wvrgcal is documented in a note circulated to the project I will show later that this works reasonably well, but clearly significant room for improvement Tune-up and implementation of new algorithms now in progress (e.g., ALMA memo 588)
(University of Cambridge) October 2010 16 / 34


Progress since last Annual Meeting

User documentation at JAO
http://wikis.alma.cl/bin/view/AIV/Application

(University of Cambridge)

October 2010

17 / 34


Progress since last Annual Meeting

Testing framework
Goals: 1 Graphical summar y of
Test interferometric & WVR obser vations Atmospheric conditions Correlations between WVR and phase fluctuations
2

Regression testing:
Check all available data can be reduced without errors Check no regression in terms of phase-correction performance

3

Algorithm tuning and selection:
Which algorithms give best results ? Which options improve the performance?

4

Statistics:
What types of conditions/configurations have missing data? How well does phase correction work?

(University of Cambridge)

October 2010

18 / 34


Progress since last Annual Meeting

Testing framework
http://www.mrao.cam.ac.uk/~bn204/temp/wvrgcal- regress.html

List of all usable test observations and prevailing conditions
date 2010/01/02 2010/01/02 2010/01/14 2010/02/01 2010/02/15 2010/03/08 2010/03/11 2010/03/11 2010/03/11 ... time 05:32 05:32 06:42 16:41 23:39 22:55 03:32 05:45 06:56 ... X02 X02 X02 X02 X02 X02 X02 X02 X02 uid X68b9 X68b9 X1204a X25f10 X359cb X44600 X4860d X48f1f X49213 ... X1 X1 X1 X1 X1 X1 X1 X1 X1 nant 2 2 3 3 3 2 2 2 2 ... conditions None None None None None None None None None ... FlagA pwv 0.755 0.755 0.418 9.0 3.1 4.7 2.2 2.1 2.06 ... MaxBaseline 116.8 116.8 201.0 201.0 201.0 201.0 201.0 201.0 201.0 ... MaxWind 5.0 5.0 7.0 14.10 13.19 10.89 6.099 7.199 7.699 . ..

...

[currently total of 41 records]

(University of Cambridge)

October 2010

19 / 34


Progress since last Annual Meeting

Testing framework
http://www.mrao.cam.ac.uk/~bn204/temp/wvrgcal- regress.html

List of tested versions of wvrgcal and options to it
ID Command line ­segfield ­nsol 5 ­toffset=-1 ­segfield ­disperse ­reverse ­disperse ­reverse ­cont 1 /home/bnikolic/d/p/wvrgcal-releases/0.17/bin/wvrgcal 2 /home/bnikolic/d/p/wvrgcal-releases/0.17/bin/wvrgcal 3 /home/bnikolic/d/p/wvrgcal-releases/0.17/bin/wvrgcal 4 /home/bnikolic/d/p/wvrgcal-releases/0.18/bin/wvrgcal 5 /home/bnikolic/d/p/wvrgcal-releases/0.19/bin/wvrgcal 6 /home/bnikolic/d/p/wvrgcal-releases/0.19/bin/wvrgcal 7 /home/bnikolic/d/p/wvrgcal-releases/0.17/bin/wvrgcal 8 /home/bnikolic/d/p/wvrgcal-releases/0.19/bin/wvrgcal 9 /home/bnikolic/d/p/wvrgcal-releases/0.19/bin/wvrgcal [currently total of 9 records]

(University of Cambridge)

October 2010

19 / 34


Progress since last Annual Meeting

Testing framework
http://www.mrao.cam.ac.uk/~bn204/temp/wvrgcal- regress.html

Matrix of results showing residual phase errors after application of WVR phase correction
ID 37 37 38 49 50 51 ... X02 X02 X02 X02 X02 X02 UID X68b9 X68b9 X68b9 X44600 X44600 X44600 ... X1 X1 X1 X1 X1 X1 Version 0.17 0.17 0.17 0.17 0.17 0.17 ... Env ID 1 1 2 1 2 3 ... Uncorr 3min 59.9 59.9 59.9 154.6 154.6 154.6 ... Corr 3min 21.9 21.9 22.0 47.1 46.8 47.8 ... Uncorr 1356.25 1356.25 1356.25 218 218 218 ... Corr 1337.09 1337.09 1351.54 101 100 102 ... PWV 0.755 0.755 0.755 4.7 4.7 4.7 ...

[currently 287 records]

(University of Cambridge)

October 2010

19 / 34


First results at ALMA

Outline

1

Introduction Progress since last Annual Meeting First results at ALMA Plans Summar y

2

3

4

5

(University of Cambridge)

October 2010

20 / 34


First results at ALMA

First application ­ February 2010
Cour tesy of Al Wooten

150

Phase of Corrected Data (deg)

100

50

0

02:41:16.8
(University of Cambridge)

02:52:48.0

Time

03:04:19.2

03:15:50.4

03:27:21.6
October 2010 21 / 34


First results at ALMA

Data-set A002 Xb9f5d X1: Shor t baseline
Red: uncorrected phase; Blue: corrected phase
50

25

0 Degrees

-25

-50

-75 0 500 1000 Time (s)
(University of Cambridge) October 2010 22 / 34

1500

2000


First results at ALMA

Data-set A002 Xb9f5d X1: Long baseline
Red: uncorrected phase; Blue: corrected phase
100

0 -100 Degrees -200 -300 -400 -500 0 500 1000 Time (s)
(University of Cambridge) October 2010 23 / 34

1500

2000


First results at ALMA

Data-set A002 Xb9fce X1
Similar conditions, shor tly after in time compared to the results in the previous slide
250

0

Degrees

-250

-500

-750 0 500 1000 Time (s)
(University of Cambridge) October 2010 24 / 34

1500

2000


First results at ALMA

Data-set A002 Xa0705 X1
Another example, only shor t baselines
20

0

Degrees

-20

-40

-60 0 250 500 750 Time (s)
(University of Cambridge) October 2010 25 / 34

1000

1250

1500


First results at ALMA

Data-set A002 Xba2ed X1
Shor t baselines, leak-through phase fluctuations (offset in coefficients due to time-constant cloud?)
10

5

0 Degrees -5 -10 -15 -20 0 500 1000 Time (s) 1500 2000

Very shor t baseline (A0-A1), essentially no phase fluctuations to correct
(University of Cambridge) October 2010 26 / 34


First results at ALMA

Data-set A002 Xba2ed X1
Shor t baselines, leak-through phase fluctuations (offset in coefficients due to time-constant cloud?)
10

0

Degrees

-10

-20

-30 0 500 1000 Time (s) 1500 2000

Also a ver y shor t baseline (A0-A2), some atmospheric-like phase fluctuation seen and corrected
(University of Cambridge) October 2010 26 / 34


First results at ALMA

Data-set A002 Xba2ed X1
Shor t baselines, leak-through phase fluctuations (offset in coefficients due to time-constant cloud?)
40

20

0 Degrees -20 -40 -60 -80 0 500 1000 Time (s) 1500 2000

Slightly longer baseline (A0-A3): atmospheric phase fluctuations clearly seen, corrected somewhat but clear "leak-through"
(University of Cambridge) October 2010 26 / 34


First results at ALMA

Data-set A002 Xba2ed X1
Shor t baselines, leak-through phase fluctuations (offset in coefficients due to time-constant cloud?)
40

20

0 Degrees -20 -40 -60 -80 0 500 1000 Time (s) 1500 2000

Slightly longer baseline (A1-A3): again WVR correction helps but "leak-through"
(University of Cambridge) October 2010 26 / 34


First results at ALMA

Dataset A002 X9c46d X1
Ver y good weather but dominated by instrumentals (BL A0-A1)
100

0

Degrees

-100

-200

-300 0 500 1000 Time (s) 1500 2000

(University of Cambridge)

October 2010

27 / 34


First results at ALMA

Dataset A002 X9c46d X1
Ver y good weather but dominated by instrumentals (BL A0-A2)
20

10

0 Degrees -10 -20 -30 -40 0 500 1000 Time (s) 1500 2000

(University of Cambridge)

October 2010

28 / 34


First results at ALMA

The variance of conditions at the site
These are measurements of total path fluctuation as derived from WVR data

A002 X9c15f X1
600
5000

A002 X9f64b X1

400 L (µ m)
2500 L (µ m)

200

0

0 -200 0 500 1000 Time (s) 1500 2000
-2500 0 1000 2000 Time (s) 3000 4000

Factor of 50 in the stability of the atmosphere

(University of Cambridge)

October 2010

29 / 34


Plans

Outline

1

Introduction Progress since last Annual Meeting First results at ALMA Plans Summar y

2

3

4

5

(University of Cambridge)

October 2010

30 / 34


Plans

Improving the algorithms
Implement the "empirical" algorithm (memo 588) Use more information:
Weather stations, ancillar y instruments to improve the priors used in the Bayesian analysis Use measured (rather than nominal) WVR filter frequencies and bandwidths significantly improve inference of water vapour column & the coefficients Use measured coupling to the sky & spill-over

Fur ther investigate methods to reduce effect of clouds Better handling of observations with many sources (i.e., how often the coefficients need to be recomputed?) Use test data as guidance for where the algorithms can be improved!

(University of Cambridge)

October 2010

31 / 34


Plans

Test observations
Longer baseline observations
Phase So far longer ALMA correction is much more challenging on long baselines only a little data around 600 m and no data on baselines than that will obser ve with 15 km baselines

Realistic fast-switching Higher frequencies
Check accuracy of dispersion correction Constrain the "dr y" component

More consistent, long-term, test observations
Current test obser vations are ad-hoc More consistent database needed for firm conclusions about how to improve the algorithms and which algorithms work best

(University of Cambridge)

October 2010

32 / 34


Plans

Software engineering, user suppor t & distribution

Improve feedback from the software to users
Highlight any problems with data Give an accurate estimate of how well phase correction is working

Pre-built packages for major data-reduction platforms
Currently only compilation from source-code is available

Suppor t the use of the software on early-science data!

(University of Cambridge)

October 2010

33 / 34


Summary

Outline

1

Introduction Progress since last Annual Meeting First results at ALMA Plans Summar y

2

3

4

5

(University of Cambridge)

October 2010

34 / 34


Summary

Summary
Working ALMA interferometer, working WVRs! Initial, but fully working, WVR phase correction software ­ used now to do phase correction at ALMA Impor tant to obtaining more test observations, especially:
On long baselines More consistent, long-term, measurements

Our work now focusing on:
Improving the existing algorithms Using more information (e.g., weather stations, ancillar y instruments) Implementing new algorithms (e.g., the `empirical') Rolling out to users for testing and feed back

Few remaining risks outside of our work package, good progress, but lots more work left to do!
(University of Cambridge) October 2010 35 / 34