Skip to Main Content
Species distribution data are becoming increasingly available over the past few years and the availability is likely to increase significantly in the near future due to technological advances. While traditionally GIS are used to visualize the distributions of a limited number of species and to generate biodiversity indices in predefined regions in an offline mode, it is desirable to manage such data in a spatial database environment and allow customer applications to efficiently query the database with arbitrary dynamically defined regions. In this study, we have developed a Variable-Fanout Space Partition (VF-SP) tree structure to represent species distribution maps by extending the classic quad-tree data structures to accommodate user-defined raster tessellations. Subsequently we have developed an approach to import multiple VF-SP trees representing a large number of species distribution maps into a spatial database for efficient query processing. Experimental results using NatureServe 4000+ bird species distribution data demonstrate that the proposed approach can be 30-300 times faster than the baseline approach that manages the same data as polygons in the same spatial database with respect to the average query response time using a query window size of 0.1 degree to 1 degree at a global scale. The average response times for such queries are less than 1 second when querying more than 15 million boxes in a PostgreSQL database. The results are encouraging with respect to using state-of-the-art spatial database technologies to manage large-scale species distribution data and answer dynamic queries in generating indices that are important to biodiversity research.