Geospatial web services within a scientific workflow: Predicting marine mammal habitats in a dynamic environment

摘要: Abstract Our ability to inform conservation and management of species is fundamentally limited by the availability relevant biogeographic data, use statistically robust predictive models, presentation results decision makers. Despite ubiquity presence-only where available, survey effort should be included in modeling process limit spatial bias. The archive therefore able store serve related information such as lines or polygons study area, best accomplished through geospatial web services Open Geospatial Consortium (OGC) Web Feature Service (WFS). Ideally data could then easily fetched modelers into a scientific workflow, providing visually intuitive, modular, reusable canvas for linking analytical processes without need code. Species distribution model accessible makers, web-based support system (SDSS). With these principles mind, we describe our progress date serving marine animal from OBIS-SEAMAP ( http://seamap.env.duke.edu ), consuming environmental cetaceans. Using automate workflow process, mammal observations are used sample date-synchronous remotely sensed satellite building multivariate habitat models using variety statistical techniques (GLM, GAM, CART). We developed custom workflows ESRI Model Builder, ArcGIS geoprocessor, R package, Python scripting language, PostGIS geodatabase, UMN MapServer. These outputs passed an SDSS with summary capability. Custom products will open-source freely available. In future, hope integrate technologies OGC WCS, OPeNDAP, Kepler. lessons described here can broadly applied modeling, within ecological informatics community.