The database holds a variety of information, and together with a GIS system,
can provide a wide range of functionality to the user. The features and
structure of this system are described here in greater detail.
Database Design Details
Sources of Data
Compliance with Data Standards
Capabilites- See Output
section for examples
For a given species (or higher taxon)
of interest, OBIS can:
For a given geographic area of interest,
Produce a table of locations where the taxa have been found
(suitable for export and further analysis)
Produce a map of these locations, either globally or for
a user-defined region.
Overlay maps of environmental factors (such as sea surface
temperature, or primary productivity) onto the range map, to investigate
possible correlations between environmental factors and distribution patterns.
Detail who first described the species, give references for
relevant published work, web sites, or graphics, and indicate where specimens
of each taxon are held and who to contact regarding the specimens.
list all species found in that region
produce a map of the region with locations of different species
overlay environmental data onto the regional map
produce tables of counts of species richness, for export
into statistical analysis software
link the above capabilities to a web interface which is freely
accessible to the public
produce range maps from point location data using the GARP
model (developed by David Stockwell). This model, which is currently used
by the Australian ERIN
system and the Mexican
Bird Project, uses available physical data to produce estimates of
ranges, including confidence estimates.
All data sets for OBIS will be provided by systematists specializing in
the given taxa. The role of the systematist is to be aware of current state
of the taxonomy and to reexamine and confirm identification of specimens.
For many groups the taxonomy has changed greatly over time, and determinations
attached to specimens in museums represent a variety of schemes. Having
a single expert directly examine specimens and resolve taxonomic conflicts
is a prerequisite for a uniform dataset. It will also ensure that as later
revisions are made, the data set can be updated and conflicts resolved.
For every record of a species found at a certain location, there should
exist a preserved specimen. This ensures that the original specimen can
be reexamined should later taxonomic revisions occur, or questions about
that record arise. Since the taxonomy of many marine groups is understudied
and likely to change, this is the only way to ensure that the database
remains current and conflicts can be resolved.
We do acknowledge that, for some groups, the taxonomy is relatively stable
and identification is straightforward, and restricting the data base to
specimen-based records may limit the geographic coverage available. We
have therefore designed OBIS so that it can also hold information on samples
that are not specimen based, and will make determinations on a case-by-case
basis about requirements for specimens.
Initially, OBIS is collecting data sets into a central data bank at the
Institute of Marine and Coastal Sciences at Rutgers University. The OBIS
designers have worked with David Vieglais of the Z39.50
Biological Implementers Group, who is currently developing a client-server
application to allow distributed networks of biological collections information.
OBIS will be compatible with the application, allowing it to take part
in the distributed system.
Advantages of a distributed system are:
as the data set grows, it can be stored in multiple locations, avoiding
the hardware requirements of having a single, very large, data set. Computer
requirements for the application are modest: just a Windows NT server.
All software is free and self-maintaining (has automated maintenance features).
systematists providing data sets can retain control of the data by setting
up their own data server. They can continually update their own data sets,
instead of OBIS having to coordinate periodic data transfers and updates,
and work with multiple versions of the data set. The systematists can also
screen portions of the data set, to protect sensitive information. Systematists
who do not wish to set up a server can still transfer their data set to
the central archive for storage.
there is no perceptible change to the user of the OBIS system. Users may
create queries and maps as if they were working off a single system. The
distributed system is "hidden" behind the application.
the OBIS user can also access museum collection data for those institutions
on the system
Sources of Data Sets
All data sets within OBIS have been freely contributed by the researchers
who created them. These researchers are interested in seeing their data
used by a larger audience, and saved in a managed archive. As OBIS grows,
it will rely on the desire of experts to share their information. If you
have or know of data sets that are appropriate
for the OBIS database please contact Jen Gregg at firstname.lastname@example.org.
Data currently in the system, and future data sources, are listed on
the input data page.
1. Content Standards
OBIS should be fully compliant with the
Standard for Digital Geospatial Metadata, developed by the
Federal Geographic Data Committee, which is the current national standard
for spatial data.
2. Transfer Standards
Data Transfer Standard is the national standard. It describes protocols
for the exchange of information, not the structure of information while
it is stored. OBIS has been designed to hold the information necessary
to meet this transfer standard.
Core Profile for Biological Collections and Taxonomic Data/ANSI
Z39.50 Information Retrieval Protocol
ANSI Z39.50 is the reigning standard for information retrieval.
Currently, the Z39.50 Biological Implementers Group (ZBIG) are developing
a profile to apply Z39.50 to collections and taxonomic data, and are creating
client-server applications which will allow data sets using the profile
to be linked.
The Association of Systematics Collections, Computerized Networking
Committee developed a model in 1992 describing the information categories
and relationships that should appear in a database for biological collections.
It has since been updated by Stan Blum at the University of Kansas. Because
it was developed for specimen collections, it requires some information
not relevant to spatial diversity and range information, and excludes other
information necessary for our purposes. Therefore, our database model diverges
somewhat from their reference model. However, meetings with Stan Blum at
the University of Kansas have assured us that our general design is in-line
with the design of the reference model, and that there is enough compatibility
to allow cross-transfer of information between data sets in OBIS and data
sets based on the ASC reference model. We have also relied heavily on the
ASC model for parts of our database design, notably the Agent/Person/Organization