Scientific Impact of Descopes

Rick Perley EVLA Advisory Committee MeetingMay 8-9, 2006

1

Scientific Impact of Descopes

Rick Perley

Version 3may06, 12PM


2

The Problem

• Contingency is getting short. • It may be necessary to consider descoping aspects

of the EVLA• We can consider descopes in one or both:

– Hardware

– Software

• Alternatively, there are some ‘delay’ options.• I review here the impact of these options on the

scientific productivity of the EVLA.


3

Hardware Descopes

We have considered:1. Removal of one or more bands

– X, Ku, S, Ka are the possibilities

2. Reduction of bandwidth from 16 to 8 GHz.

3. Reduce number of outfitted antennas.

4. Removal of special solar observing hardware.

– Items 1 and 2 were discussed last year.


4

Bandwidth Reduction

• Significant savings possible – but mostly in correlator. Not our money to save, and politically very tricky.

• Our savings come from DTS system. But, much of the needed components for all antennas already purchased.

• Scientific impact: 40% loss of continuum sensitivity per unit time at K, Ka and Q bands, less on Ku. None other bands.

• It is too late to obtain significant savings through this descope option.


5

Band Descopes

• Bands we could consider:

Total Parts Only1. S-Band $1.4M $400K

2. X-Band $1.0M $200K

3. Ku-Band $1.3M $200K

4. Ka-Band $1.2M $200K

• I consider the scientific implications of each band listed.


6

X-Band

• The VLA has a good X-band system – operating from 7.8 – 8.8 GHz.

• It has been argued that simply leaving this legacy system in place would be sufficient.

• This is a poor argument: – Continuum sensitivity will be less than half that of C-

band. – Only spectral observations of transitions lying between

8 and 8.8 GHz would be useful. • X-band would become both legacy and vestigial. • Nevertheless – this is the easiest band descope.


7

S-Band vs. U-Band?

• Removal of X-band might be considered an easy decision – but the decision of which band is next to go will not be easy.

• The debate of which is more scientifically productive will surely pit ‘non-thermal’ vs. ‘thermal’ science interests.

• Who is to say what is more valuable?

• Consider the following:


8

S-Band

• This band will become the band of choice for non-thermal imaging:– Likely better than twice the sensitivity of L-

band. • Twice the total bandwidth, • At least three times the available (RFI-free)

bandwidth• Nearly 1.5 times the efficiency.

– We anticipate superb performance from this band.


9

Examples of Key Lost Science

• TBD? (Do we need to)?


10

Ku-Band

• This is a key ‘thermal science’ band.– Transition zone (from optically thick to optically thin)

for the highest EM HII regions. – Dozens of molecular and atomic spectral transitions.– Lowest frequency where precise dust emission

observations can be made (lowest optical depth to dust extinction).

– Lowest frequency (= highest redshift) for key high frequency molecular studies (e.g. CO 1-0 at z – 5.8).

• Could be argued this is the highest ‘really good’ band for non-thermal science.


11

Examples of Key Lost Science


12

Ka-Band?

• Early in the project, this badn was identified for accelerated development – rich science mine!

• Relatively unexplored spectral region, esp. good for thermal emission processes.

• Hundreds of molecular spectral transitions.• Low sky emission – between H2O and O2

emission. Decent efficiency, excellent Tsys.• Testing of the designed horn/polarizer is about to

begin.• We could review this decision, if necessary.


13

Ka-Band Science

• Thermal Science! • HII regions, dusty obscured disks, star

forming regions.• Hundreds of spectral transitions.• Hi-z lines. • Excellent continuum sensitivity – far better

than Q-band, even for optically thick thermal emission.


14

My Opinion

• Full Frequency Coverage a Primary Goal of this project.

• Any retreat is an admission of failure by NRAO (and by AUI!)

• The science impact of band descopes is enormous. At least 10 years would be needed to recover a descoped band.

• If we must cut back, then an acceptable compromise is to simply purchase the parts, and build the new systems in post-construction years.

• This, in essence, shifts the implementation costs to operations.


15

Solar Mode Descopes

• Not a lot of money (~$200K)

• Scientific impact limited to the ‘solar community’ and solar science.

• Solar community use of VLA is now very light (nearly non-existent).

• Only two bands (L-band and one other, currently not identified) budgeted for special solar mode systems.


16

Implementation Descope?

• A suggestion has been made to reduce the number of EVLA-outfitted antennas.

• Savings limited – in many areas we have already bought ‘in bulk’ for the whole project.

• Very strong science impact: – Point-source sensitivity reduced by linear fraction. – Imaging capability (particularly for complex fields)

reduced by a greater fractions – baselines rise as N2. • This descope would affect all bands, all observers,

all programs, all science. • I think this idea is a ‘non-starter’. It should be

dismissed at the highest level, now.


17

Software Descopes

• Software in three major areas:– M&C

– E2E

– Post-Processing.

• We consider M&C as sacrosanct. The group is of just-sufficient size.

• Post-processing group already minimal – sub-critical. No fat here!

• We look at ‘e2e’ for descope opportunities.


18

Descopes in e2e?

• Level 2 and Level 3 software e2e deliverables already descoped.

• Budget is currently unable to provide all ‘Level 1’ deliverables.– Three more staff needed to achieve Level 1 deliverables. – This does not include extra staff for imaging algorithm R&D.

• Approximately $1M of contingency already utilized in addressing shortfalls in `e2e’.

• Could review ‘Level 1’, for further descopes. • Probable course of action is to delay some deliverables.• Effect would be to move some to operations.


19

E2E Descope Options

• (A list of Level 1 deliverables which would be deferred/removed, should we have to go this route …)

• (Perhaps a reminder that e2e, and the One Observatory, is an NRAO/AUI obligation, and that the failure of the 2000 DMD model is not the fault of the EVLA Project).

• (Hence, more resources from NRAO/CV, and AUI, are needed).


20

Deferment Options

• Although not formally a ‘descope’, perhaps the best options are to defer planned capability.

• Hardware: Already mentioned that costs could be offset into future operations by purchasing early, with implementation deferred until post 2012.

• Not a great option, but doable, deferring $xxxK to future generations.

• Risk: Much better receivers might be available later? (I doubt this, given the existing capabilities).


21

Deferment Options

• Software. • Clearly an area where deferment is possible – even likely:

– Real-time scheduling: Can be done by a person indefinitely, until good software can be written.

– Automatic image generation: Can be done by trained individuals until more automated systems can cover harder and harder configuration/frequency/weather combinations.

– Post-processing software. Although development of improved algorithms is non-deferrable, we can match scheduling of tough experiments (full-field, noise-limited, high-resolution polarimetric imaging at L-band, for example) until the software is ready.

– Nothing new in this – done by the VLA. – Other examples…


22

Summary

• Advertised scientific productivity of EVLA requires all hardware and software deliverables to be met.

• It is unwise policy to remove frequency bands. History shows recovery will be at least a decade away.

• All software deliverables in e2e are key for full scientific usage – especially by those not trained to be ‘radio astronomers’.

• The preferred solution is an increase in the budget sufficient to keep us on track and on time. Estimate of this: $5M. (… only 6% of total EVLA1 budget).

• The acceptable, but not preferable, means is deferment of promised capability into the operational years.

Scientific Impact of Descopes

Documents

Transcript of Scientific Impact of Descopes