An Introduction to Alert Follow-up Networks Bryan Miller 1 LSST - - PowerPoint PPT Presentation

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

An Introduction to Alert Follow-up Networks

Bryan Miller

1

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Why all the fuss? In operation LSST will generate ~107 transient alerts per night!

2 CANDELS (http://www.spacetelescope.org/images/heic1306d/) 104 alerts every ~2 min

Science teams must select interesting targets. Observatories must prepare for Order(10-100) triggers/night/tel

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Such high event rates will likely swamp current ToO handling procedures

At Gemini, ToOs make up about 20-25%

• f the time in Band 1
• Max rates are 1-2/night
• Many more will overwhelm the manual

scheduling process and observers

3

At CTIO/SOAR, ToOs interrupt visitor nights

• Many more will be

disruptive to standard programs

A new solution is needed!

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

The proposed solution is a system that dynamically turns alerts into requested data.

4

Talk by Soraisam on ANTARES. Talk by Bianco ZTF example: talk by Ridgway

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

TOMs request observations from the appropriate network resource, then a scheduler takes over

Scheduling needs to be automatic, update frequently or as new

• bservations arrive.

Targets assigned to nodes based on site conditions, target visibility, requested resource, etc.

5

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Las Cumbres Observatory (LCOgt) is an example

• f an operating telescope network

21 robotic telescopes: 0.4m, 1m, 2m Imaging and spectroscopy (on 1m and 2m) Both hemispheres and nearly complete longitudinal coverage

6

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Rachel Street (LCOgt, PS) Bryan Miller (Gemini, PS) Stephen Ridgway (NOAO, PS) Cesar Briceno (NOAO/SOAR) Andy Adamson (Gemini) John Blakeslee (Gemini) Adam Bolton (NOAO) Bob Blum (NOAO/LSST) Todd Boroson (LCOgt) Jay Elias (SOAR) Steve Heathcote (NOAO) Catherine Merrill (NOAO) Joanna Thomas-Osip (Gemini)

7

NOAO/SOAR/Las Cumbres/Gemini are working to dynamically schedule SOAR, Gemini, Blanco...

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Astronomical Event Observatory Network (AEON):

1. Develop interfaces (APIs) a. Receive requests from TOMs b. Send schedule to telescopes c. Receive status 2. Incorporate SOAR into LCOgt network, execute observations on dedicated nights 3. Coordinate data pipelining and archiving efforts 4. Incorporate Gemini, modify scheduler to handle queue 5. Be ready to incorporate other facilities (Blanco, etc)

8

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Las Cumbres and SOAR are working on APIs and automation

Automating Goodman spectrograph

• perations via scripts

Modifying LCOgt scheduler Will develop interfaces to accept a schedule and create scripts for taking observations. Tests in late 2019

9

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

We close the loop via data archives and reduction tools

Gemini has developed a pure python image reduction package (2019) SOAR and Gemini are collaborating on spectroscopy reduction in python starting in 2018 New Gemini instruments come with reduction tools that work in our pipeline environment (DRAGONS). Automated processing is a goal for the LSST era

10

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Gemini’s observation planning/scheduling tools must be improved to handle a higher ToO rate and any new follow-up modes.

Current weaknesses:

• Queue observers switch between static

plans as conditions change.

• Rapid ToOs disrupt the plan
• Managing timing windows
• It is difficult to support complicated

cadences (eg. observe every N days, logarithmic cadences)

11

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Improvements are being made under the OCS Upgrades Program

Rethink system from first principles. New UIs for ease of use. AEON APIs for search, status, and triggering. Automated scheduling built on LCOgt adapts to ToOs and changing conditions. Part of AEON in ~2020+

12

See Oct 2017 Gemini Focus, pg. 20

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

The community should consider various issues related to increased follow-up.

Allocation options:

• PI projects through existing TACs (as now)
• Community campaigns?
• Increase the fraction of time in Fast Turnaround?
• Single, system-wide AEON TAC (e.g. LLP)?

▪ Allocates time on all network resources ▪ Avoids multiple jeopardy from different TACs

Should there be new modes?

• “follow-LSST”?
• Other?

13

?

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Summary

NOAO/SOAR/Gemini/LCOgt are developing an instance of an alert follow-up network that includes:

• Broker (alert filter)
• TOM (target/resource matching)
• Scheduling/execution of network nodes
• Data reduction pipelines

There are many implications for how we will use or the network and facilities. We want your feedback so that we can develop a system that works for everyone and delivers the best science.

14

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

Gemini will support AEON with the ongoing OCS Upgrades Program

Goals:

• Rethink the purpose and UI

from first principles

• Make Phase 2 preparation

much easier

• Include new features that are

not possible in the current code

• Make the code maintainable

and scalable

15

See Oct 2017 Gemini Focus, pg. 20

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17 16

Task-specific but consistent UIs ⇒ lower learning curve Calibration steps and observations automatically added, configs linked Physical conditions on target (FWHM) instead of conditions bins

LSST Project and Community Workshop 2018 • Tucson • August 13 - 17

The Phase 2 tool should be easier for new users, more capable for advanced users

Include tools for visibilities, planning, etc

17

Programmatic access (eg. APIs for scripts) for search, full control, and access to services (ITCs, AGS).