U.S. JGOFS Data Policy
INTRODUCTION
From the inception of the program, the U.S. JGOFS Steering
Committee (SC) has recognized the necessity for a data management
strategy that addresses both program and individual PI needs. To
this end, U.S. JGOFS is committed to the timely submission and
sharing of all data collected in U.S. JGOFS field programs.
A Data Management System (DMS) has been developed with
financial support from NSF, by Glenn Flierl, Jim Bishop, and
David Glover. This system is being implemented in U.S. JGOFS
during 1994, and is now available to all participants in the
program, regardless of funding source. The approach taken by the
systems team was to design a data management scheme in which the
PI's current best versions of their data sets reside at the
originators' sites rather than being gathered into a central
archive. The data file format is likewise the PI's choice.
Users can access any of the individual data sets without regard
to storage method (format) or location, using the data system.
The data system is now being implemented, maintained and
operated by the U.S. JGOFS Data Management Office (DMO), headed
by Cyndy Chandler (formerly by Christine Hammond), at the Woods
Hole Oceanographic Institution. It is the mission of the DMO to
assure that all U.S. JGOFS data sets are readily accessible by
JGOFS PI's.
Ultimately, all data will be archived with the National
Oceanographic Data Center in response to national policy.
To achieve the goals set forth by national and program
policy, a strong commitment to data management is required of
each participating PI. In accepting NSF support within the U.S.
JGOFS program, each PI is obligated to meet the following suite
of data management requirements as an integral aspect of their
participation in the program.
NATIONAL POLICY
The national policy on the release of marine environmental
data to the public domain is very explicit regarding submission
to the National Oceanographic Data Center (NODC). To paraphrase
the national policy, recipients of federal funding supporting
collection of marine environmental data must release these data
to the NODC within two years of the date of collection. Thus,
U.S. JGOFS investigators have these conditions to meet. In
addition, the national policy requires that post cruise inventory
information, in the form of a ROSCOP form, be completed within 60
days of the end of the cruise, usually by the Chief Scientist.
The DMO will compile and submit to NODC the cruise inventory
information and ensure that U.S. JGOFS data sets are transferred
to NODC as specified.
Guidelines: U.S. JGOFS Chief Scientists will be
required to maintain a detailed event log for every
sampling operation for a given cruise/leg of a field
collection effort. This log will include the
following: operation number, station number, cast
number, date, time, position, sampling device, desired
comments (see attached sample).
U.S. JGOFS PI's are required to submit to the DMO
detailed sampling inventory information for a given
cruise/leg.
U.S. JGOFS POLICY
The U.S. JGOFS data management policy is based on openness
and data sharing for the mutual benefit of program PI's. The
policy sets guidelines for release of data from program defined
Core and non-Core Measurements, with the understanding that some
measurements require longer analytical procedures which prevent
speedy release after cruises.
Basic Core Measurements are defined as those needed to
depict general environmental conditions that existed during the
survey period. To obtain these fundamental measurements, NSF
funds specific PI's to make Core Measurements with the
stipulation that these data be readily available to the other
PI's on an unrestricted basis.
Although not comprehensive, the list of Basic Core
Measurements for any cruise may include: meteorological
observations; profiling of temperature, salinity, dissolved
oxygen, nutrients, carbonate system, chlorophyll/pigments,
particulate organic carbon and nitrogen, dissolved organic
carbon, fluorescence, beam attenuation, light transmission and
bio-optical profiling; primary and bacterial production,
phytoplankton, zooplankton and bacteria biomass, meso- and
microzooplankton grazing; biogenic, lithogenic, and grain size
components of sediments, sediment trap samples; and remotely
sensed observations in the form of processed sea surface
temperature and ocean color images to the level of chlorophyll
and attenuation coefficient.
The primary mechanism for data release in response to
program policy and for making data available to program PI's and
the DMO is via the U.S. JGOFS Data Management System. The
Steering Committee strongly urges all PI's with workstations to
install the DMS and use it as the method for data submission.
PI's who do not have access to workstations must submit data to
the DMO, where data sets will be loaded on workstations at
W.H.O.I.
Access to data sets on the DMS is not limited to
workstation technology. PI's with networked MAC or PC hardware
can readily access data on the DMS (see System documentation for
more detail).
The DMO will support the installation of the DMS on
workstation and PC/MAC equipment and provide systems training to
Program PI's.
Guidelines: Depending on shipboard processing
capabilities, it is envisioned that some Core
Measurements (i.e., CTD and some bottle cast
measurements) will be available in preliminary form at
the end of each cruise. Distribution is limited to
cruise participants with the data labeled preliminary.
Final versions of most Core Measurements should be
placed on the DMS within 6-12 months of the sampling
date. It is also understood that some Core
Measurements will not be available within the time
frame set by this guideline, due to the nature of the
analytical process. However, all PI's making these
measurements must strive to meet a timely release date.
Within one year of each cruise all Core and non-Core
Measurements must be placed on the DMS. To further
U.S. JGOFS science objectives, PI's responsible for non-
Core Measurements are encouraged to share data with
their colleagues prior to the release deadlines.
DOCUMENTATION
Because many of the analytical procedures being used in U.S.
JGOFS are "state of the art," and thus subject to change,
documentation on sampling and analytical methods is essential.
The DMS allows for submission of documentation text files. They
are identified in the DMS Dictionary as supporting documentation
files.
Guidelines: All data placed on the DMS must be
accompanied by a text file of documentation on sampling
and analytical protocols. This documentation should
include descriptions of: shipboard sampling procedures,
sample preparations, analytical procedures, equipment
calibrations, data reduction techniques, computation
algorithms, citations and anything else deemed
necessary.
DATA FORMATS
To maintain a true data base management system each data set
must contain some critical control parameters. Each observation
must contain the unique operation/event number. If sampling is
from a bottle cast, the cast number and bottle number must be
identified.
The DMS was designed to be format independent and those
PI's placing their data on local workstations have few, if any
concrete format requirements. These PI's have several options
available to them: (1) The systems team has written a number of
data translation programs (methods) for a number of the more
basic data formats. With only minor reformatting, a PI may be
able to adapt his or her data to one of these structures
eliminating the need to write a method. (2) The PI can prepare
an interpreting program that fits his or her internal data
format. In either case, contact the DMO for more detailed
information.
Those individuals placing their data with the DMO for
installation on the DMS have minimal data formatting
requirements. The formatted data must contain the above listed
critical control parameters. The data formats described in the
attachment are acceptable structures that the DMS can utilize
with a minimal amount of system preparation on the part of the
DMO. The data must be in ASCII code and can be transmitted to
the DMO via FTP or mailed on magnetic media. Those individuals
having data sets that cannot be formatted to one of these forms
should contact the DMO for guidelines.
Guideline: All data sets must contain the following
mandatory elements; Each observation must contain the
event/operation number. If sampling from a bottle
cast, each samples data record must contain the event,
cast and bottle number.
Guideline: Those PI's submitting their data to the DMO
for installation on the Data System, at minimum, must
comply with one of the formats described in the
attachment.
QUALITY ASSESSMENT/ASSURANCE
The DMO provides routine quality assessment of all data
placed on the Data System. This effort is accomplished either by
the DMO directly or through assistance provided by experts
working with the DMO. All data will be reviewed and categorized
according to an established set of standards, where applicable.
Investigators will be notified of data quality issues, and asked
for comment and/or corrections.
Guideline: PI's are responsible for maintaining the
quality and correctness of the data placed on the DMS.
Questions raised by the DMO should be addressed by the
PI as quickly as possible.
PROPRIETARY RIGHTS
While recognizing the legitimate rights of data originators
to the first use of the data they collect, U.S. JGOFS also
encourages the oceanographic community to use data collected by
the program, and in particular, believes that data availability
should be restricted only in exceptional cases. Data normally
becomes publicly available for use without restriction two years
after origination. It is expected that all users will properly
acknowledge the source of the data.
Guideline: Basic Core Measurements will be available
to U.S. JGOFS investigators without restriction
following submission to the DMO. However, it is
expected that published works will cite the originator
and will be sent to him or her for prior review.
Guideline: All non-Core Measurements will be available
to U.S. JGOFS investigators with restrictions requested
by the PI, such as: can be viewed by all JGOFS PI's,
view and use only with consent of originator, use only
in collaboration with originator, etc.
Guideline: Data will be released to the public domain
two years after the date of collection. Before then,
all requests from investigators outside the JGOFS
program will be referred to the PI unless the DMO has
been given explicit permission to distribute the data
freely.