Maintenance and Development of the Subseasonal to Seasonal Prediction Project (S2S) Database and Tools in IRI Data Library

DESCRIPTION: The proposed work will maintain and further develop the existing implementation of the WWRP/WCRP Sub­seasonal to Seasonal Prediction Project (S2S) database and tools in the IRI Data Library (IRIDL), for the duration of Phase II of the S2S Prediction Project (2019–2023). This will include adding the S2S ocean data from the primary S2S archive at ECMWF, together with any extensions to it made at ECMWF such as new variables and models. In so doing, the IRI Data Library (IRIDL) will serve as a third archiving centre for S2S data, along with ECMWF and the China Meteorological Administration (CMA). Beyond an S2S database copy in the US, IRIDL will provide important new resources compared to those available through ECMWF and CMA, accelerating S2S research and development of S2S applications. These will include pre­computed lead­time dependent model climatologies, with large labor savings for users. A suite of online tools with a Jupyter Notebook Python interface will enable users to access the S2S database via OpenDAP without the need to download data, a significant advantage especially in bandwidth ­limited environments and for for S2S training workshops. These Python tools will enable diverse users to easily construct weekly forecast anomalies, skill scores, and other derived variables such as weather regime indices and vertically­ integrated moisture fluxes. In addition, having the S2S database archived at the same place as the NOAA SubX and NMME databases at IRI is a big advantage for researchers, enabling a direct comparison of the American models with S2S models.

The proposed work addresses NOAA’s 2nd Long­term mission goal Weather­Ready Nation, where Society is prepared for and responds to weather ­related events. The S2S timescale between 2 weeks and a season, bridging the gap between existing weather forecasts and seasonal climate outlooks is recognized to be a critical one for early warning of weather hazards and one in which significant advances are expected due to recent advances in models and MJO prediction skill in particular. Serving the S2S database in IRIDL will significantly contribute to achieving this goal by facilitating access by a broad spectrum of researchers and applications­ developers, eliminating the need for users to laboriously download their own “dark copy” of the database and write their own analysis codes. The increased uptake will translate into development of improved S2S forecast products and user ­applications, and will pave the way for applications of cloud computing to multi-model forecast datasets in the future.