Feasibility of Using Existing Web-Services for On-Demand Data Access within Distributed Environmental Decision Support Systems

Environmental decision support systems (DSSs) are becoming more data-intensive as data availability increases. Significant effort has been invested in making these data accessible through web services, which allow for on-demand data access through standardized, machine-accessible interfaces. Although the typical approach is to use web services to gather and store data as a separate, off-line step, there are advantages to being able to build distributed environmental DSSs where data are accessed on-demand from web services for direct use within the DSS. The objective of this study is to assess the current feasibility of this approach by testing the response times of existing, national-scale web services in the United States that would be commonly used within a distributed environmental DSS. A stormwater management application was used to define typical data needs for an environmental DSS, these data needs were mapped to available national-scale web services, and then the service response times were measured under different scenarios. Results of the analysis showed that service requests, which included soil data access, river network tracing, watershed delineation, and streamflow statistics, took from less than a second to more than one minute to respond. Some services scaled well to increasing study area sizes, while others did not. All services showed significant variability depending on the geographic location of the client. These results highlight specific areas where future research and development are needed to improve service response times in order to achieve the vision of distributed environmental DSSs.