One of the topics discussed at the TMT Science Forum was the need for a searchable archive that would allow PIs to access their data and, when data were public, would provide data access for a broader scientific community. The Keck Observatory was considered a model for the TMT operations concept, and I think that the Keck Observatory Archive (KOA) is also a good model for a TMT archive. KOA is operational, hosts over 20 TB of data and is ingesting data from all active instruments. I presented a poster on this topic at the Forum, shown below.
First, though, let’s take a look at the specifications for the TMT archive, taken from Section 7.7 of the Operations Requirement Document (TMT.OPS.MGT.07.002.CCR) (September 23 2012):
KOA is a model for the Keck Archive because it meets many of these specifications already. It archives and curates the raw science and calibration files from multiple instruments and the associated metadata – including metadata on the program, telescope and instrument status etc – and performs QA on these data during the ingestion process. It implements a data access policy, and has procedures in place for managing exceptions to it. It has developed algorithms for automated associated calibration files with the science files. It has begun to work with the VO to develop VO-compatible services. Finally, it produces value-added browse products for all the instruments.