OPENCoastS builds on-demand hydrodynamic and water quality forecast systems for user-selected sections worldwide and maintains them running operationally everyday.
Coastal forecast systems are nowadays essential tools to support sea and harbour management, civil protection emergencies, navigation procedures, military operations and recreation activities. They not only anticipate natural disasters but also the impact of human-made coastal disasters (e.g. wastewater spills, oil and chemical spills) and support in search and rescue operations. These systems are thus very useful for port and coastal managers, civil protection agencies, companies operating in coastal regions as well as researchers and the local communities.
Setting up and maintaining such systems requires a team of experts on coastal dynamics and computer science as well as the availability of computing resources to guarantee the timely delivery of predictions.
Scientists Anabela Oliveira, Alberto Azevedo, André Fortunato, Marta Rodrigues and their teams at the National Laboratory for Civil Engineering in Lisbon, Portugal, have tackled this challenge by taking advantage of a strong coastal dynamics team and IT development, in collaboration with LIP, the University of La Rochelle (CNRS) and the University of Cantabria.
On top of that, they made use of EGI’s computational core services. Indeed, the robustness and quality of EGI computational resources along with the tailored support offered to their users was key to the selection of the infrastructure to provide the computation and storage for this real time, daily service with strict availability time restrictions.
The European team developed OPENCoastS: an open-source platform to develop on-demand hydrodynamic and water quality forecast systems. With OPENCoastS, a user can simulate hydrodynamics forced by tides, winds and river flows, and include the interactions with waves and stratification effects. Through the EGI-ACE project, OPENCoastS evolved to OPENCoastS+ and it now also predicts water quality.
The service’s Web interface is relatively straightforward to set up a new forecast system at the user-selected coastal region and provides a forecast manager, and a visualisation and download interface. Furthermore, EMODNET’s in-situ data and processed Sentinel satellite data comparisons are now available to validate the forecasts, aiming at providing confidence to the user on the quality of the results.
Considerable computational resources are necessary to guarantee the timely delivery of multiple, simultaneous simulations with dozens or a hundred thousand model grid points. OPENCoastS is operated using EGI resource providers and takes advantage of EGI services for authentication, Cloud Compute and High-Throughput Compute, simulation dispatch management as well as agile storage.
On average, each of the simplest OPENCoastS forecast simulation requires daily 60h of CPU and 2GB of storage to produce a period of 48h of forecast results. Considering that OPENCoastS has approximately 30 active recurring users, and each user can deploy up to 5 forecast systems, the service has to deal with nearly 9000 computing hours and 300GB of storage daily. This intensive computational usage can only be possible with the support of the EGI core services and the infrastructure provided by the European Open Science Cloud.
The EGI-ACE project, led by the EGI Foundation, expanded OPENCoastS to water quality forecasts and climate change scenario simulation. The computational and storage capacity requirements are even higher for the expanded service.