Data-intensive environmental research: re-envisioning science, cyberinfrastructure, and institutions

Transcription

1 Data-intensive environmental research: re-envisioning science, cyberinfrastructure, and institutions Patricia Cruse John Kunze California Digital Library University of California

2

3 Environmental research and global change Global change presents a complex scientific and societal challenge Society needs good data in order to build good science inform wise policy-making enable sustainable resource management decisions Good data and data-intensive research need solid technical infrastructure sound organization community engagement (you)

4 Data curation is hard Data sets encompass everything, including regular object types Documents, images, audio, video, etc. Tension between establishing standards and fostering innovation Data is like software, but even more specialized Heavy processing requirements imply tricky long-term migration/ emulation of custom data/software Heavy provenance and snapshot coherence requirements Instability: value of some preserved data depends on ongoing change, in particular, on researcher annotation

5 The complexities of global change Smith, Knapp, Collins. In press.

6 Data challenge 1: dispersed sources ( finding the needle in the haystack ) Data are widely distributed Ecological field stations and research centers (100 s) Natural history museums and bio-collection facilities (100 s) Agency data collections (100 s to 1000 s) Individual scientists (1000 s to 10,000s to 100,000s)

7 Data challenge 2: diversity the flood of increasingly heterogeneous data Data are heterogeneous Syntax (format) Schema (model) Semantics (meaning) Jones et al. 2007

8 Data challenge 3: poor practice data entropy Time of publication Specific details Information Content Accident General details Retirement or career change Death Time (Michener et al. 1997)

9 CDL Data challenge 4: loss Natural disaster Facilities infrastructure failure Storage failure Server hardware/software failure Application software failure External dependencies (e.g. PKI failure) Format obsolescence Legal encumbrance Human error Malicious attack by human or automated agents Loss of staffing competencies Loss of institutional commitment Loss of financial stability Changes in user expectations and requirements Source: S. Abrams, CDL

10 Data challenge 4: more loss 1,000,000 Petabytes Worldwide 900, , , , , , , , ,000 Information Available Storage Transient information or unfilled demand for storage Source: John Gantz, IDC Corporation: The Expanding Digital Universe

11 Cumulative impact on data longevity Study Resource Type Resource Half-life Rumsey (2002) Legal Citations 1.4 years Harter and Kim (1996) Scholarly Article Citations 1.5 years Koehler (1999 and 2002) Random Web Pages 2.0 years Spinellis (2003) Markwell and Brooks (2002) Computer Science Citations Biological Science Education Resources 4.0 years 4.6 years Nelson and Allen (2002) Digital Library Objects 24.5 years Koehler, W. (2004) Information Research 9(2): 174.

12 Data Observation Network for Earth The goal of DataONE is to enable new science through universal access to data about life on earth by: engaging the scientist in the data preservation process supporting the full data life cycle, encouraging data stewardship and sharing promoting best practices engaging citizens DataONE and Data Conservancy (JHU) are two DataNet awardees recommended for funding by the US National Science Foundation (NSF), and between which collaboration is expected

13 CDL Initial data types Biological genes to biomes Environmental Atmospheric Ecological Hydrological Oceanographic

14 Existing biological data archives ESA s Ecological Archive Distributed Active Archive Center National Biological Information Infrastructure Fire Research & Management Exchange System Long Term Ecological Research Network Knowledge Network for Biocomplexity

15 Existing cyberinfrastructure: tools 15

16 New distributed framework Coordinating Nodes Member Nodes retain complete metadata diverse institutions catalog subset of all data serve local community perform basic indexing provide network-wide resources for managing services their data ensure data availability (preservation) provide replication services Flexible, scalable, sustainable network

17 CDL/UC3 roles in DataONE Member of overall DataONE leadership team Chairing Governance and Sustainability working group Chairing Preservation working group

18 DataONE management and partners William Michener, University of New Mexico Suzie Allard University of Tennessee Bob Cook Oak Ridge National Laboratory DAAC Patricia Cruse California Digital Library Mike Frame USGS, National Biological Info. Infrastructure Matt Jones University of California Santa Barbara Steve Kelling Cornell Lab of Ornithology DataONE Partners plus Kepler- CORE and SEEK/KNB Teams We welcome your involvement!

19 Building global communities of practice and long-lived cyberinfrastructure Community engagement Involve library and science educators Build on existing programs Involvement of cultural memory organizations brings centuries of preservation experience to datasets

20 Other data researcher support projects DataCITE initiative to encourage data publishing through global data citation support, including citation standards and persistent reference to datasets in regional archives Curation and hosting for existing museums and archives (e.g., UC Berkeley s Media Vault Project) Support for publication of extended data description appendices, typically omitted from journals, but which drops crucial details support research conclusions and credit for data producers (reducing incentive for data sharing)

21 Summary Libraries and cultural memory organizations play a vital role in meeting the global change challenge through dataintensive research Getting started includes projects such as DataONE to work directly with scientists DataCITE for data citation support Joining forces with museums and archives