Palmer Station LTER Information Management Site Byte - September 2010
Karen Baker

Information Management Committee Inquiry

Introduction
Site Bytes  are intended as a  general update on what  has happened at
your  site during the  last year.  Please highlight  new developments,
ideas, and issues.  Site bytes help us to stay  informed about what is
going on at  other sites. Please also be  forward thinking and include
your thoughts on the following subjects with regards to the developing
Network Information System.

We will discuss these questions further at the annual meeting:
Given limited resources, how do you prioritize data contributions to the NIS?	
1) Do you give a lot of attention to a few datasets or limited attention to a lot?
2) Do you deliver derived or raw data?
3) Do you communicate with your PIs regarding prioritization of data?

PAL Responses	
1) Do you give a lot of attention to a few datasets or limited attention to a lot?

Between these two extremes, we strike a balance. Our strategy has been
to  create  the technical  capability  to  rapidly  create and  update
datasets  together  with  a  management interface  for  submitting  to
community  collections.   Last  year  we  focused   on  improving  our
information system in ways that put  us in a position to work with the
NIS  development team. We  also improved  our metadata  approach while
working with  several local groups  with complex datasets who  had not
yet submitted data. In addition,  site work focused on streamlining of
weather data handling so as to have monthly updates for ClimDB as well
as summary plots  posted in DataZoo. Quality control  has been left to
the data  submitter to date but  we plan to  develop some capabilities
with  this in  the next  year as  part of  the  information management
services and hence to begin more routine submissions of more assuredly
high quality data to NIS.

2) Do you deliver derived or raw data?

The raw-derived question seems misleading  when stated as a binary. We
see multiple stages  as part of a processing  continuum that varies by
dataset  depending upon  sampling,  instrumentation, calibration,  and
analysis and so  we consider the decision as to  what dataset stage to
deliver as  dataset dependent. For many researchers,  the questions of
what  to  submit is  still  under  investigation  as they  may  submit
abundances  from  counts obtained  via  microscope  analysis and  then
submit carbon  converted counts  as they work  with the data.  We find
that over time and with  experience, our understanding of the data and
the data  audiences as well as  the shared projects  such as EcoTrends
influences our data delivery.

3) Do you communicate with your PIs regarding prioritization of data?

Yes, this is the strength of an LTER site to have IM guided in
practice by local data needs and by research scientists who collect,
analyze, submit, and use the data. Data inventories are created and
discussed periodically. The significant technical capacity improvement
this year coupled with our planned focus on metadata and data quality
reviews next year put us in a position to include submission of data
to Metacat as part of a more routine data process next year.u80

PAL Site Byte

1) Local Development
Focus this year was on technical development of DataZoo, a site-level
information system now used by four oceanographic sites (two LTER
sites (PAL, CCE) and two CalCOFI sites (SIO, SWFSC)), in an effort to
better meet local needs while developing web services as a
foundational element for upcoming site-network development. Redesign
of the study-project schema for DataZoo added flexibility in response
to metadata needs identified in practice and by the IMC working group
effort with project collections. Significant updates were made to the
DataZoo management interface that allows creation and editing of
organizations, studies, datasets, and their related metadata. New
approaches to web delivery through development of middleware and
standardized, asynchronous access to databases through a web service
layer successfully addressed issues relating to the increasing size
and diversity of data. Expanding DataZoo to a three-component system
accommodated the development of FileFinder, an interface to very-large
collections of hierarchically structured datasets. The prototype of
FileFinder was populated with oceanographic CTD data.u80

Additional work relating to DataZoo included adding the support for
taxonomic identifiers to the attribute qualifier system, use of the
Integrated Taxonomic Information System (ITIS) as a taxonomic
authority that prompted conversion of NODC taxonomic codes, and
redesign of query selection so users can select studies within a
designated time interval. Plot performance was improved by replacement
of the plot library and options added for contour and bubble
plots. Handling of external datasets like weather was streamlined to
facilitate monthly updates for ClimDB as well as summary plots posted
in DataZoo. Improvements were made to dataset documentation as
datasets were submitted for publication in EcoTrends.  An Information
Management media gallery was created in order to preserve posters as a
historical record of development. Finally, backups are now done via
the Computational Infrastructure Service at IOD in conjunction with
the San Diego Supercomputer.  In terms of physical infrastructure, a
new server with virtual machine capability was purchased (vSurf). u80

2) Partner Activities and Leadership
Our team (Mason Kortz, Lynn Yarmey, James Conners) is leading a
network-level LTER unit registry and unit best practices effort that
received support from a LTER post All Scientist Meeting proposal
award. Two cross-site visits (KBS, JRN) and a visit to the Network
Office were carried out as part of this project. This work has
informed redesign of DataZoo architecture to a web-services
orientation. The new unit registry together with its management
interface replaces a static, isolated unit dictionary thus enacting
one type of data comparability as well as demonstrating a new type of
site-network model that enables site participation in NIS development
through web services. Our team (Mason Kortz, James Conners) is also
leading a joint IMC-LNO web services working group and serving as a
NIS tiger team member (James Conners). Karen Baker is co-leading the
IMC Governance Working Group (GWG). The GWG was created in order to
facilitate the conduct of LTER business in an arena of growing
complexity and responsibility and to support the development of Terms
of Reference or by-laws for committees and working groups. u80

3) Professional Development and Publication
Professional development of  information managers was provided through
exposure to  site-level and network-level  development activities with
the Unit  Registry project. In  addition, a weekly  summer informatics
reading group  stimulated discussion  and joint learning.  A sociology
graduate student  studying the Design Studio  stimulated reflection on
our  work  as well  as  on  our work  arena.  Lynn  Yarmey received  a
fellowship  to  pursue two-year  masters  in  Library and  Information
Science with  a concentration in Data Curation  through the University
of Illinois, one of the national iSchools. u80

At the LTER Information Management Committee (IMC) annual meeting
CCE/PAL members contributed as co-chairs of the IM Governance Working
Group and Unit Working Groups. Lynn Yarmey reported on a Unit
Dictionary Best Practices Guideline. Karen Baker was elected a member
of the Network Information System Advisory Committee (NISAC). Mason
Kortz and James Conners prototyped the LTER Unit Registry and
contributed to formation of a new Web Services Working Group. At the
LTER All Scientist Meeting, contributions included a series of
posters, participation on a panel about Network level efforts, and a
workshop focusing on curriculum for information managers.u80

Collaborating on two chapters about the EcoTrends Project contributed
to description and understanding of network level information
management efforts as well as development of a set of recommendations
for future work (Laney et al, in press). Another book chapter explores
the topic of digital infrastructure growth (GCBowker, KSBaker,
FMillerand, and DRibes, 2010. Towards Information Infrastructure
Studies: Ways of Knowing in a Networked Environment. In Int. Handbook
of Internet Research). Collaboration with Science Studies partners led
to publications in peer-reviewed journals. An interdisciplinary
investigation into work with data over time spanning decades provides
insight into ÔBig DataÕ efforts and the LTER network model (Aronova,
Baker, and Oreskes, 2010. From the International Geophysical Year to
the International Biological Program: Big Science and Big Data in
Biology, 1957-present. Historical Studies in the Natural Sciences
40(2):183-224). A study of design provides insight into the complexity
of the development processes at an LTER site (Millerand and Baker,
2010. Who are the users? From a single users to a web of users in the
design of a working standard. Information Systems Journal
20:137-161).u80

4) In Summary
We plan to focus next year on metadata population, quality control,
and data delivery as well as on documentation of the DataZoo
system. Technical work with the unit registry and web services will
continue at a network level. Technical issues that remain to be
addressed include integration of the personnel module with DataZoo, on
geolocation based on the earlier gazetteer effort, and on management
of dataset use logging. Our website content will be updated but
migration to drupal/wordpress/etc is not yet scheduled. The
application module MediaZoo was redesigned as an API. The move from a
photo gallery to a plot gallery and eventually to the media gallery
module has reached an appropriate level of abstraction that it is time
to make uniform the five instances of installation currently in
use.u80

Despite significant  experience with information  management practices
and  issues  at sites,  innovation  at sites  is  limited  by lack  of
resources and  incentives. The 2010  information management supplement
stabilizes our work on DataZoo  and collaboration on the LTER NIS this
year but  highlights the  need either to  continue such support  or to
collaborate locally  with additional projects  in order to  maintain a
team with technical expertise at the site-level over the long-term.