Wednesday a.m. 2

SDO Knowledge Base

Leader: Karel Schrijver

• 10:30 CCMC Capabilities and SDO, Peter MacNiece

• 10:45 SDO Knowledge Base - Introduction, Karel Schrijver

By about the end of 2008, SDO will give us over 2TB of data per day. That is too much for almost all of us to retrieve, store, or review on a routine basis. Finding data of interest within the vast archive of SDO data therefore requires a largely automated system to identify events, and an efficient system to find and retrieve the metadata on these events and, ultimately, link to the observations themselves.

As you hopefully know, we are developing a meta-data knowledge base (http://www.lmsal.com/helio-informatics/hpkb/) that contains information that describes what happened on the Sun. During the SDO teams meeting, we will demonstrate the developing system, and discuss with you, the user community, how the system should function, how it should present its findings in response to your queries, and what kinds of 'events' it should contain. And then, of course, we shall have to discuss which of the aspects that we would like to have are actually affordable and implementable within the coming year, and how we make optimal use with work done within the project and in the community.

Specifically:

• In the morning of Wednesday, March 26, we will present our thoughts on the system (including how our initial system focused on the Sun can grow into a broad-reaching Heliophysics system and how it complements the Virtual Observatories), demonstrate its first incarnation, discuss how features and events may be found by the analysis of the data flowing into the archive, and start a discussion on the fundamentals of the system.

Later Knowledge Base Sessions

Wednesday Afternoon

Then four sessions follow in which we hope you will actively participate:

• Part 1 (starting at 2:30 on Wednesday)

  • 1) One session on user and software interfaces with the system (led by Neal Hurlburt) will be followed by three parallel sessions on features and feature finding:

• Part 2 (starting at 4:15 on Wednesday)

  • 2a) Photosphere-to-heliosphere events (such as active regions from AIA and HMI, sunspots, coronal holes, flares from AIA images and from EVE lightcurves, filaments and their eruptions, atmospheric wave phenomena, etc.); this session is chaired by Karel Schrijver;
  • 2b) Events found from modeling efforts (open-field perimeters, active-region field energies and helicity, field topology, helioseismic events related to fields and flows, etc.); chaired by Mark Cheung, and
  • 2c) Helioseismic events (far-side measurements, subsurface events including emerging regions, flare-induced oscillations, etc.) chaired by Sasha Kosovichev.

After the teams meeting, we need to have a prioritized list of which kinds of features we should plan the knowledge base to contain, and how these are to be parametrized by attributes. We also need to have assessments of 1) what it will take to 'find' all features (how much research work and funding, how much CPU time, what kind of data, 2) how much ancillary data is needed, 3) what software exists, 4) which methods need to be developed, etc.) And we need a plan to move forward, including concepts to validate the system, validate the feature-finding software, develop the software, meeting opportunities for working groups, etc. We hope you will give all this thought in preparation of the meeting.

Friday Afternoon

The final session on the knowledgebase (on Friday afternoon, starting at 1:30) will be very practical: what keywords do we need, how do we standardize them, what schema, which keys, what kind of IDs, etc. If you are interested in providing events, working with events through the knowledgebase, or if you have catalogs and event finders that can be incorporated into the system, this is a session for you!

• Back to [:SDOTeamsMeeting:SDO Teams Meeting Agenda ]