A Cyberinfrastructure Model for Ecological Applications

T. Little, B. Anderson, P. Cryan, M. Friedl, S. Gauthreaux, J. Kelly, T. Kunz, R. Larkin

January 25, 2006

Many fundamental questions in environmental sciences cannot be addressed without coordination between disciplines and access to multiple data sets. One such question concerns the nature of bird and bat movements at multiple spatial and temporal scales. Knowledge of how proximate factors influence these movements promises to inform understanding of complex relationships among ecological processes, weather patterns, land forms, and land cover. With greater understanding of these interactions, it will be possible to provide real-time observations, predictions, and forecasts of potential adverse affects on animal populations and health risks to humans. For example, knowing how extrinsic factors affect migratory events of birds and bats could be used to inform decision makers and developers of wind energy facilities on how best to mitigate fatalities, and knowledge of seasonal migrations would help inform the public health community regarding how to mitigate the effects of emerging diseases. A key challenge currently limiting the capacity scientists to develop such understanding is that available data sets are often incomplete, inaccessible, lack consistent semantics, or are otherwise difficult to integrate in a manner that can lead to useful scientific inquiry.

This proposal describes research to establish an environmental cyberinfrastructure (CI) prototype that integrates multiple environmental data sets and demonstrates new analytical capabilities to better understand how multiple extrinsic factors affect animal movements. The goal of this research is to create a rich collaborative environment among environmental scientists, domain specialists, and software developers that will lead to a successful use of modern CI. Driven by tangible scientific questions, the proposed research team will create a common lexicon and ontology appropriate for adopting CI components to better understand animal movements at different temporal and spatial scales. Software will be developed that exploits existing CI to expose multiple data sets through a common user and programmer interface that is amenable to both scientific inquiry and use of existing Grid computing middleware. The software infrastructure will provide standard web services interfaces suitable for (a) mature scientific analysis tools (e.g., R, IDL, ArcGIS, Matlab), (b) composite applications from other web services (e.g., integration with open source GIS tools and pop-software such as Google Earth), (c) exposure of data through dynamically-generated web portals, and (d) the development of stand alone applications that demonstrate ‘virtual observatories.’ The research will focus on data sets originating at multiple spatial and temporal scales including remotely sensed data, meteorological observations, NEXRAD Doppler radar, census, and ground-based animal sensor network systems. We describe several major use cases for this application, and propose user-centric methodologies for more extensive design of these use cases. The proposed work will involve the analysis, design, and adoption of CI software to meet the requirements of use cases established by a scientific community comprised of geographers, ecologists, biologists, and meteorologists. Significant efforts will be committed to identification and dissemination of this experience for influencing future use of CI in ecological networks.

Broader Impacts: The PIs will develop two new courses suitable for education and advancement of young scientists and engineers in the interdisciplinary domains of terrestrial, atmospheric, and information technology research. The first will focus on questions applied to the intersection of the targeted data sets in this proposal. The second course will be achieved by the creation of a software development lab in which both undergraduates and project participants will be exposed to collaborative design and development of robust web services based on open source CI software created by multiple generations of computer systems engineers. The enabling of complex queries and analysis of relevant data sets has the potential to empower researchers to investigate a broad set of environmental questions related to the data sets beyond North America. Thus, this broader impact will serve as a model for enabling analysis of other environmental questions in other geographic regions. Impacts from this research are also expected on public policy, due to the application domain, and in the education of future ecologists, biologist, and engineers. The evolution of IT education towards new CI tools and methods is relevant to national priorities of global competitiveness, and the education of a new generation of environmental scientists capable of addressing increasingly complex questions.