 |
| Titel |
A Constrained and Versioned Data Model for TEAM Data |
| VerfasserIn |
S. Andelman, C. Baru, S. Chandra, E. Fegraus, K. Lin |
| Konferenz |
EGU General Assembly 2009
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 11 (2009) |
| Datensatznummer |
250029053
|
|
|
|
|
|
| Zusammenfassung |
| The objective of the Tropical Ecology Assessment and Monitoring Network
(www.teamnetwork.org) is “To generate real time data for monitoring long-term trends in
tropical biodiversity through a global network of TEAM sites (i.e. field stations in tropical
forests), providing an early warning system on the status of biodiversity to effectively guide
conservation action”. To achieve this, the TEAM Network operates by collecting
data via standardized protocols at TEAM Sites. The standardized TEAM protocols
include the Climate, Vegetation and Terrestrial Vertebrate Protocols. Some sites
also implement additional protocols. There are currently 7 TEAM Sites with plans
to grow the network to 15 by June 30, 2009 and 50 TEAM Sites by the end of
2010.
At each TEAM Site, data is gathered as defined by the protocols and according to a
predefined sampling schedule. The TEAM data is organized and stored in a database based
on the TEAM spatio-temporal data model. This data model is at the core of the TEAM
Information System - it consumes and executes spatio-temporal queries, and analytical
functions that are performed on TEAM data, and defines the object data types, relationships
and operations that maintain database integrity.
The TEAM data model contains object types including types for observation objects (e.g.
bird, butterfly and trees), sampling unit, person, role, protocol, site and the relationship of
these object types. Each observation data record is a set of attribute values of an observation
object and is always associated with a sampling unit, an observation timestamp or time
interval, a versioned protocol and data collectors.
The operations on the TEAM data model can be classified as read operations, insert
operations and update operations. Following are some typical operations:
The operation get(site, protocol, [sampling unit block, sampling unit,] start time,
end time) returns all data records using the specified protocol and collected at the
specified site, block, sampling unit and time range.
The operation insertSamplingUnit(sampling unit, site, protocol) saves a new
sampling unit into the data model and links it with the site and protocol.
The operation updateSampligUnit(sampling_unit_id, attribute, value) changes
the attribute (e.g. latitude or longitude) of the sampling unit to the specified value.
The operation insertData(observation record, site, protocol, sampling unit,
timestamps, data collectors) saves a new observation record into the database
and associates it with specified objects.
The operation updateData(protocol, data_id, attribute, value) modifies the
attribute of an existing observation record to the specified value.
All the insert or update operations require: 1) authorization to ensure the user has
necessary privileges to perform the operation; 2) timestamp validation to ensure the
observation timestamps are in the designated time range specified in the sampling schedule;
3) data validation to check that the data records use correct taxonomy terms and data values.
No authorization is performed for get operations, but under some specific condition, a
username may be required for the purpose of authentication.
Along with the validations above, the TEAM data model also supports human based data
validation on observed data through the Data Review subsystem to ensure data quality. The
data review is implemented by adding two attributes review_tag and review_comment to each
observation data record. The attribute review_tag is used by a reviewer to specify the
quality of data, and the attribute review_comment is for reviewers to give more
information when a problem is identified. The review_tag attribute can be populated by
either the system conducting QA/QC tests or by pre-specified scientific experts. The
following is the review operation, which is actually a special case of the operation
updateData:
The operation updateReview(protocol, data_id, judgment, comment) sets the
attribute review_tag and review_comment to the specified values.
By systematically tracking every step, The TEAM data model can roll back to any
previous state. This is achieved by introducing a historical data container for each editable
object type. When the operation updateData is applied to an object to modify its attribute, the
object will be tagged with the current timestamp and the name of the user who
conducts the operation, the tagged object will then be moved into the historical data
container, and finally a new object will be created with the new value for the specified
attribute.
The diagram illustrates the architecture of the TEAM data management system.
A data collector can use the Data Ingestion subsystem to load new data records into the
TEAM data model. The system establishes a first level of review (i.e. meets minimum data
standards via QA/QC tests). Further review is done via experts and they can verify and
provide their comments on data records through the Data Review subsystem. The data editor
can then address data records based on the reviewer’s comments. Users can use the Data
Query and Download application to find data by sites, protocols and time ranges. The
Data Query and Download system packages selected data with the data license and
important metadata information into a single package and delivers it to the user. |
|
|
|
|
|
|