Database on Oceanography and Marine Ecosystems

1
Database on Oceanography and Marine Ecosystems
Janus Larsen, E-mail janus_at_ices.dk ICES, H. C.
Andersens Boulevard 44-46, DK-1553, Copenhagen V.
Tel [phone].
User requirements The primary advantage of an
integrated multidisciplinary database is
facilitation of the ecosystem approach. For
example, providing data for an investigation of
biological data of a specific area in relation to
the physical environment such as fish disease
data in oxygen depleted areas. Another major
advantage is easier access to data from many
disciplines through a single interface. Data
Centre requirements A driving force is lower
maintenance costs. Operating several diverse
database and coding systems is expensive. By
integrating our databases, ICES lowers operating
costs while gaining staff flexibility through
harmonization of databases as well as data
processing.
DOME is a relational database to store
oceanographic, environmental and fisheries data
at the ICES Data Centre, supporting a marine
ecosystem approach at the data level. Phase 1 of
the development is ongoing and it includes biota,
fish disease, sediment and water bottle data. In
phase 2 high-resolution oceanographic data such
as CTD data and biological community data will be
included. Some fisheries data will be considered
for a 3rd phase of DOMEs development.
Standard codes and handling of data from
different scientific communities is essential for
integration. Relevant multi-disciplinary codes
are implemented. However, specialized codes have
been incorporated as necessary. DOME draws on 2
standard code catalogues, the BODC data
dictionary and the Integrated Taxonomic
Information System (ITIS). Specific ICES codes
have been standardized in our RECO
database. Codes are stored in master databases
which serves all ICES databases including DOME.
For optimal performance, relevant codes are kept
up-to-date by automated replication to DOME.

• The data structure includes 3 common parts
• top level time, position, sampling platform,
  etc.
• bottom level definitions of codes, methods,
  access, etc.
• system level administrative tables.
• Measurements comprise DOMEs middle level. They
  are organised into separate structures
  corresponding to each of the major data
  categories. The separate structures are necessary
  due to the different natures of the data
  categories and their respective difference in
  types of data.

Reporting format and fixed field databases.
Reporting format
The example to the right illustrates the use of
codes and types/values to store data from a
measurement of mercury in a muscle sample.
Species information is stored higher up in the
hierarchy. The essential data, namely, the value,
the parameter code and the matrix code, are
stored in fixed fields. The rest of the data are
either linked to the measurement via a
many-to-many relation to a code table (blue
fields) or stored in a types/values table which
is linked to the measurement table (red
fields). It is easy to add new types of
information. If necessary, new definitions are
added to the respective common table, and the
measurement can then be associated with those
definitions no changes to the database
structure are needed.
DEPHU PERCE SIGND UNCRT DETLI LMQNT
35 NULL NULL 0.4 0.001 NULL
MATRX Matrix analysed MU
DEPHU Pressure/depth 35
PARAM Parameter code HG
MUNIT Measurement unit
BASIS Basis of determination Wet
AMLNK Analytical methods link 1
VFLAG Validity flag
QFLAG Qualifier flag lt
VALUE Value 12.3
PERCR Percentage recovery
SIGND Significant digits
UNCRT Uncertainty value .4
METCU Method of calculating uncertainty SD
DETLI Limit of detection value .001
LMQNT Limit of quantification
DOME
Types values
Codes
Type value
DEPHU 35
UNCRT 0.4
DETLI 0.001
The database is normalized by type/value
combinations. This makes the database very
flexible. New data types and meta-data can be
stored without any structural changes to the
database and without any adjustment of the
database interfaces. This approach also
eliminates most null values.
DOME is normalized with Types and Values
DOME will be equipped with 2 interfaces an
internal interface for handling data input and
output, and a web interface to allow searching
and downloading data from the ICES website.
Although the interfaces are separate, they use
the same search and export code as shown in the
figure to the right. This object-oriented design
allows easy maintenance of the system. Internal
interface The purpose of the internal interface
is to promote efficient data import and export
within the ICES secretariat. The internal
interface will feature facilities such as import
and update of data, assignment of access
rights. Web interface The web interface will
provide users with facilities to download data
directly from the ICES website. Various levels of
data restriction will be enforced some data will
be freely available, whereas others will be
password protected with various degrees of
restriction, including explicit permission of the
data owners. Eventually, the web interface will
be extended with a range of data products, see
the Gantt chart under When?.

• DOME is being developed in 3 phases
• Phase 1 (ongoing) develop database and
  interfaces for biota, sediment and water bottle
  data.
• Phase 2. Add high resolution oceanographic data
  and biological community data and modify the
  interfaces accordingly. Add data plot to web
  interface.
• Phase 3. Include selected fisheries data (under
  consideration).

The DOME database design succeeds in unifying
access to multi- disciplinary data, and will
facilitate integrated data products for use in
implementing an ecosystem approach. DOME
harmonizes structural aspects as well as many
common elements of the data such as code
definitions. However, the choice of coding
systems, the application of meta-data, and the
approach to data quality assurance are still data
category-dependent. These differences are, to a
large degree, due to different traditions in
different scientific communities.