APSViz: Creating, managing, and visualizing data products from the ADCIRC Prediction System

Recent tropical cyclone impacts on the eastern US and Gulf of Mexico coasts have demonstrated that storm surge and wave predictions from the ADCIRC Prediction System (APS) have substantial demand and value to end-users, such as emergency managers, the US Coast Guard, NOAA, DHS/FEMA, state governments, and private companies. To meet this demand, the core APS operations community relies on high performance computing (HPC) resources to compute and deliver real-time storm surge guidance information to end-users, in conjunction with the APS’s ecFlow-based Floodwater model management system. The distributed nature of APS activities (people and computers) has presented a significant challenge for coordinating these activities and ensuring that critical information reaches the intended end users and decision makers. This includes knowing what prediction simulations are being computed and where, the status and progress of simulations, and (particularly) when updated results can be expected.

To address this information dissemination challenge, the DHS Coastal Resilience Center has developed the ADCIRC Prediction System (APS) visualization engine (APSViz), a new open-source approach to creating, managing and disseminating products from APS, focusing on rapid, efficient, and scalable cloud-ready components. A significant component is the web mapping visualization system that displays real-time results from APS. The complex software system manages the acquisition, processing, and dissemination of real-time predictions from APS, developed using open-source tools and software and leveraging software containers, Kubernetes for workflow management, and community resources such as Unidata’s THREDDS. This presentation will describe the software development and architecture of APSViz, how products are generated, its hosting in Amazon Web Services, future extensions to the system, and examples of its use during recent tropical cyclone events.