LSST Science Advisory Council Face-to-Face meeting Friday, August 15, 2014 Phoenix, Arizona SAC members attending: Chris Hirata, Bhuvnesh Jain, Jason Kalirai, Mansi Kasliwal, David Kirkby, Rachel Mandelbaum, Renu Malhotra, Michael Strauss (chair), Lucianne Walkowicz, Beth Willman, Michael Wood-Vasey Also attending: Chuck Claver, Mark Giampapa, Steve Kahn, Nigel Sharp This meeting of the SAC occurred at the heels of a full LSST team meeting during the week of August 11. An important component of that meeting was a workshop discussing the LSST cadence, and ways to quantify how well any given realization of the cadence meets the needs of any specific science goals. A summary of this workshop is being prepared, and will be distributed publicly soon. At the beginning of the LSST team meeting, Steve Kahn and Victor Krabbendam presented excellent summaries of the status of the LSST project. You can find them as links from https://project.lsst.org/?q=content/lsst-2014-workshop-project-status *******Commissioning Plans Chuck Claver led a discussion on the commissioning plans for the telescope. The schedule may be found at http://lsst.org/lsst/science/timeline . Engineering first light will be in mid-2020. After an initial Shack-Hartman instrument for alignment, we will use a commissioning camera ("ComCam") which consists of a single raft (9 CCDs) with identical electronics to the full LSST camera. This will be used for the first phase of commissioning, and will be a powerful scientific instrument in its own right, with an etendue about half of that of the Dark Energy Camera on the Blanco 4-meter at Cerro Tololo. It will have a filter exchanger with 3 filters available at a given time. Science verification with the full camera will start in mid-2021, and full science operations are to start in late 2022. There will be opportunities for community involvement in commissioning activities at at least two levels: -Helping to plan the commissioning activities themselves. Commissioning needs to work through a checklist of system-level requirements flowed down from the Science Requirements Document. Once one gets beyond the purely engineering requirements, many of the commissioning tasks require exposures on the sky. While these tasks require fairly specific sets of observations (e.g., specific dither patterns to test scattered light), they are agnostic to where in the sky to point. Input from the community to make these commissioning data scientifically useful will be solicited (although the project will start by writing down a first set of commissioning observations). Chuck is planning an external review, including members of the community, of the plans for comissioning. This review would include a library of possible observing plans, ideas for which can and should come from the community. It will also be useful to talk to folks from DES and HSC, to learn about their commissioning experiences and to hear what tests they found to be particularly valuable. -Helping to analyze the commissioning data themselves. While the LSST Project has the formal responsibility to use the commissioning data confirm that the system-level requirements are met, it will be tremendously valuable to have scientists work with these data in order to search for more subtle problems. We imagine a process whereby scientists or groups of scientists formally apply for access to commissioning data to carry out specific commissioning tests. Note that the folks who contribute ideas for commissioning plans need not be the same folks who analyze the data. Members of the science collaborations will be particularly well-positioned to write compelling proposals for taking part in the commissioning activities, but applications will be solicited from the community as a whole. The SAC discussed the difficulties of managing community expectations for all this. In particular, we have to make clear from the outset that commissioning data do not belong to those folks who contribute ideas for specific commissioning observations. When white papers were solicited for Kepler 2.0, this was not made clear, causing some confusion and bad feelings in the community. Note that the commissioning data (at least that produced with a mostly working system, and thus likely to be scientifically valuable), will be released to the public, presumably as part of the first operational data release. We will have a challenge managing the information flow during the very busy time of commissioning. Teams of scientists will be finding subtle problems, and will have many questions about the nature of the data and how to access it (remember that the pipeline software, the databases, and the user interface are also being commissioned in parallel), and the people who can answer these questions will be extremely busy with their own commissioning responsibilities. Open web forums where people can air their concerns may be the best mechanism for communication. Given how scientifically powerful the commissioning data are likely to be, this led to a discussion of publishing rules. How concerned should we be that people publish junk with commissioning data? Some of the SAC members said that this should not be an over-riding concern; a larger danger follows from trying to place too much control over the publication of early LSST papers. In this context, it is worth mentioning that the current LSST publication policy is going to need an overhaul: it will be refocussed on LSST Project technical papers, with individual science collaborations responsible for their own rules. We talked through the different phases of commissioning. There will be purely engineering phases, where, e.g., interfaces between telescope and camera are exercised and tested, wavefront sensing is refined, and the telescope pointing model is developed. There will also be phases (with both ComCam and the full camera) that will be in "science verification phase": the telescope will be working well, and we will get scientifically valuable data. The commissioning plan includes a period of downtime before the survey proper starts, giving time to go through a punchlist of remaining things to be fixed before the survey proper begins. The current commissioning plan is 5.5 months shorter than had been originally planned; some of the final refinements in the Data Management systems will take place in the first months of full operations. *******Science Collaborations We discussed the status of the various science collaborations. The LSST Project is controlling their activities, including their membership, much less than it did before. Several science collaborations are developing roadmaps and white papers for the work ahead. There is some concern about the lack of activity in the Solar System science collaboration. The range of topics covered by the collaboration is very broad, and the communication between those interested in KBOs and NEOs, say, is relatively small. (Of course, other science collaborations also cover a broad range of topics; this is not an issue limited to Solar System!). A possibility is to change the name of the science collaboration: essentially all the science it will do is small bodies (LSST is not likely to teach us much new about Mars or the Sun), and perhaps the name of the science collaboration should reflect that. The visibility of LSST within the solar system community is poor. We plan a substantial presence at the upcoming DPS (Division of Planetary Sciences) meeting (November in Tucson) which should help a lot. We should also have an LSST presence at an asteroid-specific meeting. We also need to communicate with the Small Bodies Assessment Group within NASA: http://www.lpi.usra.edu/sbag/committee/ To increase activity within the science collaboration, we need to reach out to individuals or groups in specific areas. For example, there is a very energetic NEO team at JPL. We have approached NASA multiple times over the years to see if they are interested in LSST; the time may be ripe to do so again. This led to a discussion of the understanding, or lack thereof, of what DM will provide for orbits for main belt asteroids and Trans-Neptunian Obits. One of the high-level goals of LSST is to "inventory the Solar System." While the LSST pipeline processing will attempt to fit orbits to all detected asteroids, the community (and indeed the DM team) does not yet have a clear understanding of the limitations of the MOPS code, and the extent to which the project will provide science-ready data for Solar System science. In any case, there will be plenty of work to be done as part of a Level 3 analysis! For example, understanding the completeness of the asteroid sample as a function of orbital parameters can be determined by injecting false sources into the data stream; this is beyond the standard pipeline processing. We also discussed the cadence. There is a broad range of opinions in the Solar System community whether the current uniform cadence is adequate for orbit determination, although some of the skepticism is based on experience with inferior cameras, in which multiple repeat observations are needed to identify moving objects cleanly and link them together for orbit determination. What is needed are metrics that quantitatively assess the ability of a given realization of the LSST cadence to measure asteroid orbits; generating those will be real work. The problem of the optimal cadence for KBO orbit determination is very different from that for NEOs! ********Cadence in general The LSST Project developed what they called the Uniform Cadence (as described, e.g., in the LSST Science Book). It was a demonstration that we could satisfy the core LSST science goals in the 10 years of the project; demonstrating this was key to garnering approval for funding the project. Now that the project has formally started construction, we can explore alternate cadences that enhance the science further. For example, there was a lot of interest in "rolling cadence", in which the area of sky imaged in any given month, say, is limited to a relatively narrow range of LST, rather than trying to maximize the sky covered per night. Exploring these options was the theme of the NOAO/LSST-sponsored cadence workshop. In particular, the workshop focused on development of "metrics", i.e., pieces of code that take as input a specific 10-year realization of LSST's pointing history as output by the Operations Simulator (OpSim), and output a quantitative measure of how well this cadence satisfies specific science goals. Participants in the workshop (and indeed, the community as a whole) were/are encouraged to develop metrics in every area of science that LSST will enable, from orbits of near-Earth asteroids (see the discussion above) to weak lensing systematics. These metrics will certainly not all be optimized for the same set of choices for the Operations Simulator, and so difficult choices will need to be made weighting different science areas in choosing the actual cadence. The Director will seek the advice of the Science Advisory Committee in making these choices. Making these decisions will also require understanding the derivatives of any given science metric with respect to small changes in cadence parameters; perhaps a tweak in one parameter could greatly improve some metrics while only having a small deleterious effect on others. We recognize that this is truly a difficult problem: the parameter space of possible cadence models is huge, and we won't be able to truly optimize in the mathematical sense. We discussed some of the things we want to balance when we think about cadence. For example, we will want to consider the ability to do science early (say, after the first year) and want to maximize the potential overlap and synergy with other facilities of finite lifetime (JWST, WFIRST, etc). We also briefly discussed the deep drilling fields. LSST went through an exercise some years ago of calling for white papers for suggestions on the placement and cadence of those fields, and a nominal set of 10 fields was defined at that time. Four of those have been announced publicly, to stimulate observing with other facilities. So we definitely should do those! But the others are worth rethinking. This was not a topic that was discussed in any detail at the Cadence workshop. Deciding on the right cadence is more than simply developing the appropriate set of metrics that span the space of LSST science projects. The OpSim runs, and the parameters that run them, also need to span some reasonable space of possibilities. This is ultimately a project responsibility (which will develop a survey scheduler based on what is learned from OpSim). There is a real need to understand better the algorithms that currently drive OpSim, and to suggest alternatives in a variety of directions. The SAC remains unclear to what extent it will be realistic for the community to get access to the OpSim code itself and carry out some of these explorations themselves, but we heard at the cadence workshop that the code should be robust, fast, and user-friendly enough to make this possible on the timescale of a year or so. The OpSim operates via a series of "proposals", i.e., distinct algorithms for deciding on field choices for different aspects of the survey. At the moment, the main wide-fast-deep survey proposal has some specific choices, such as a sharp airmass=1.4 limit, as well as sharp limits in Galactic latitude, and we wish to explore variants on these choices. While LSST will be largely focused on carrying out a uniform survey, and thus for the most part we will want to keep a fixed cadence for the full ten years, unanticipated scientific opportunities may arise through the survey. Thus Steve suggests that during operations, we set up a Cadence Optimization Committee, who will advise the then-director on changes to the cadence as the survey progresses, as well as targets of opportunity from such things as LIGO gravitational events (which can have an error circle similar in size to the LSST field of view). In any case, to make progress on the cadence decisions, we need to get broad community involvement. The 100+ people who showed up to the cadence workshop shows that there is broad interest indeed. The organizers of that workshop are drafting a report, to be posted publicly (to arXiv?) to solicit further input.