eResearch

You are here

Prototype: Semantic WildNET

Scientists and biodiversity experts need to be able to integrate species databases, observational data, maps, remote sensing images, climatology data, vegetation distributions and ecological data in order to identify endangered species or fragile, threatened ecosystems and to prepare and implement environmental impact analyses and conservation strategies. Semantic WildNet is a case study that applies Semantic Web/Grid technologies to integrate distributed ecological databases for the purposes of biodiversity monitoring.

Datasets

Our current datasets concern South East Queensland:
  • Over 10 000 Bird and Snake Sightings, provided by the EPA,
  • Bird and Snake Taxonomies, provided by the Museum of Australia,
  • Habitat data from the National Herbarium
  • Climate Sensor readings from the Bureau of Meteorology,
  • Topographical maps from Geoscience Australia.

Features

The Semantic WildNet prototype adds an ontology-based semantic search layer over distinct ecological datasets, representing wildlife listings, taxonomies and sightings, as well as climatological and physical environment information.

Wildlife sighting data is kept by organisations including museums, government environmental protection agencies, and non-profit environmental groups. While these records are often in a simple format, they can have a great amount of intepretation in terms of species taxonomies and exact definitions for time and point location. Additionally, merging these databases can result in duplication due to multiple records, or the same sightings being reported by different observers, and spelling mistakes can result in equivalent sightings passing unchecked.

We have developed ontologies to allow a common understanding of these datasets to be reached. Equivalence between records can be declared and automatically considered when searching.

Semantic WildNET Architecture

Implementation

Server:

Client:

  • Web based application that maps search results onto Google Maps
  • Translations into KML allow viewing in 3D environments such as Google Earth

The following screenshots show the SemanticWildNet client interface.


Semantic WildNet SPARQL query interface

Semantic WildNet map interface

Semantic WildNet Satellite interface

Semantic WildNet hybrid interface

Semantic WildNet in Google Earth

Semantic WildNet ontologies


Semantic WildNet ontology classes and properties

The SemanticWildNET ontologies are available from metadata.net:

Demo

Live Demo System (UQ internal only, requires username and password, Firefox 2.0 or Safari browsers)