Research Challenges in Using Distributed GIS Services to Support Biodiversity Visualization and Analysis
Robert Guralnick 1, David Neufeld 2. 1Asst. Professor and Curator Dept. of EEB and CU Museum University of Colorado Boulder Boulder CO 80309-0265. 2CU Museum University of Colorado Boulder Boulder CO 80309-0265.During development of an NSF supported online mapping application for visualizing and analyzing biodiversity information, we confronted two major research challenges. The first challenge was determining how to best access distributed point-based information about species occurrences across the landscape given network bandwidth limitations. The second challenge was building an application that could access heterogeneous distributed datasets of collections information and spatial environmental reference layers. In order to determine the most effective means to transport potentially large datasets of geospatial point data, we calculated file sizes returned from remote servers for 100, 1000 and 10,000 records using three different methods: XML, ArcSDE, and PNG images. The best solution in all cases was to develop a remote image mapping service to generate PNG images. This solution scales particularly well as the number of records returned increases. Using image fusion had the added advantage of allowing access to heterogeneous data sources as long as the returned data is a spatially referenced image. Our architecture is built on a Java Server Pages Model-View-Controller2 (MVC2) design pattern that allows for access, in our case, to remote ArcIMS servers and Web Map Services (WMS) servers. The solutions presented here are not specific to biodiversity information and should be of general use for all developers of web-based mapping services that access distributed data sources.