The advisory team that assembled for a mid-project review of NatureServe's work under the NSF-funded grant "Developing a New Infrastructure for Dynamic Access to Multi-institutional Biodiversity Data" represented an excellent cross-section of the biodiversity and ecological informatics community. The wide range of experiences that these advisors brought to the meeting, along with their high level of engagement, made for a very successful meeting. Overall, the project team got a strong sense of validation from the assembled advisory team that our project approach is sound, and we gained some valuable insights and specific recommendations that will guide the remainder of the project in a more successful direction.
Summary of take-home messages:
Need to balance the short-term funding reality with the long-term vision, and focus the project on outcomes that are achievable within the project scope and budget
Focus on our primary audience and goals - we can't try to do everything for everybody
Take this opportunity to gain even more national prominence as the leader for observation-based biodiversity data
We have a good relationship with our network of 75 data providers, so we must remember to always give something back to them while we continue to encourage them to share most of their data openly
Validation for our Web services approach and for maintaining a federated, central data repository/cache
Specific recommendations:
Web services from other providers that we might want to consume:
GNIS/Gazetteer service from NBII for resolving place names - in order to respond to users that wish to define their search area by place name.
ITIS Web service - in order to cross-reference between the NatureServe taxonomy standard and ITIS.
Consider the taxonomic object service from the SEEK project for finding taxonomic concepts.
Consider a Web service from the specimen collections community (not sure if there is a specific service that exists yet for this) that would provide collection annotation updates for species of concern to NatureServe central scientists so we can keep our taxonomy up to date.
Web services that we might consider offering:
The group validated the basic menu of services already in development, which includes services that can respond to requests such as: Is a given species present in a defined area (yes/no response); Which species are found in a defined area (list of species reponse); Where is a given species known to occur (species distribution based on known species occurrence, at a spatial resolution appropriate to the user's security access, response).
Consider offering a service that allows contributors to submit observation data where there may be known gaps in the NatureServe species occurrences.
NBII is especially interested in invasive species issues, so any services that can contribute invasive species data would be welcome.
Overall, it was agreed that the emphasis for this project should be less about how NatureServe presents the data on the Internet and more about developing the Web services infrastructure to respond to user and application queries. However, several user interface recommendations were noted:
Consider a quick-search method to find species data (fewer clicks to return what majority of users are requesting).
Consider two main paths on the interface - search by species or search by place.
Given that data providers set access policies, present the user with a consistent spatial resolution for mapped species locations, but include metadata about the most precise level of data available from each data provider.
When presenting results to the user, highlight the most interesting components of the report up front, such as images and maps.
To address gaps between known species occurrences and the complete species distribution maps, consider displaying the species distribution range maps underneath the species occurrence layer so that users do not misinterpret the occurrence data as the only locations where these species exist (for example, some states do not track certain species, so these would appear as gaps in the NatureServe species occurrence data layer).
Consider handling more challenging and unique user stories through custom services, rather than trying to integrate these requirements into the public services.
Recommendations on the overall project development and management approach included:
Consider making all software developed under this project available to the broader community, under an open-source development framework or otherwise, so that others may expand on the results.
Consider the ability for our network of member programs to customize (or "skin") any applications developed under this project for their state.
Develop a suite of different user profiles (with different access control levels), develop a flow diagram for each profile, then test and fine-tune performance for each profile; suggest we focus on a small, finite number of standard roles initially and trace the user experience from end-to-end to ensure security and Web services infrastructure will integrate.
Some possible collaboration and follow-on opportunities were noted, including:
The NEON project will be looking to harvest biodiversity datasets.
There is some overlap between information that would need to be gathered from users to develop a trust relationship and customer relationship management (CRM) ways of managing data about people, although this may raise disclosure limitation concerns.
Look at approaches used in the medical community for disclosure and privacy protections related to sensitive data.
GBIF shares several problems in common with this project, including: the need for generalized portrayal of precise locational data for sensitive species; the need for a geospatial data standard that can handle non-point location data within an XML schema; ways to address data attribution across a network of providers; ways to design work flows for scientific data collection, and alignment of Web services with these work flows.
For Data Providers
Security