Operational Space Weather Entering a New Era

Abstract

U.S. operational space weather maturing because of two competing factors. On one hand, directed agency funding at about $1 billion for model development over the past decade has brought modeling maturity to !ve broad Sun-toEarth domains, i.e., Sun, heliosphere, magnetosphere, ionosphere, and thermosphere. On the other hand, agency funding for transitioning these models into operations has been a small fraction of the level provided for model development. is situation has le# implementation of operational space weather largely unfunded and woefully undirected, with the exception of a few U.S. Air Force Weather Agency projects. A new vision needed so that operational space weather can help solve 21st-century challenges. e current paradigm for operational space weather began in the 1990s, when the idea was to build complete systems at single facilities. ese rapid prototyping centers were envisioned as a means for implementing new operational systems, but this path was unsuccessful. Although increased funding for operational space weather would have bene!ted the community, this did not happen, and in the lopsided funding pro!le, operational space weather was forced to develop an unusual but robust architecture. is architecture consists of distributed networks—automated systems of models, data streams, and algorithms at multiple, geographically dispersed facilities linked by operational servers. A central server manages the network and uses a database enabling remote nodes to deposit output data and access input data at their geophysical or measurement cadences and latencies. e emergence of distributed networks represents a paradigm shi# from 1995, when the National Space Weather Program (NSWP) strategic plan was !rst published. At that time, very few models existed that could be even considered for use in operations. Model development increased due to the funding surge, and by 2000 seventy-three models characterizing 15 space environment domains were listed in the NSWP implementation plan (2nd edition). During the past decade some of these models were implemented operationally and became important components of distributed networks. Early examples were the Magnetospheric Spec!cation Model (MSM; 2000), SOLAR2000 (S2K; 2001), and High Accuracy Satellite Drag Model (HASDM; 2004). More recently, systems of coupled models and data streams have been created, including Global Assimilation of Ionospheric Measurements (GAIM; 2006), HakamadaAkasofuFry (HAF; 2007), and Communication Alert and Prediction System (CAPS; 2008). Coupled model/ data systems that are currently being implemented include the JacchiaBowman 2008 (JB2008), the Nowcast of Atmospheric Ionizing Radiation for Aviation Safety (NAIRAS), U.S. Geological Survey Dst, ENLIL with Cone, and Ionosphere Operational System (IONS; i.e., the commercial version of GAIM at the Utah Science Technology and Research (USTAR) Center for Space Weather). Of added note that the Space Weather Modeling Framework (SWMF) and the Center for Integrated Space Weather Modeling (CISM) have developed multi domain systems to represent the entire Sunto-Earth environment, in which some of the models have been running for years. e need to test coupled systems existed even without funding, and the lack of funds drove small businesses, universities, and the multiagency Community Coordinated Modeling Center to create innovative, cost-e$ective solutions. Basically, these organizations had to !nd cheaper ways to accomplish testing and operational implementation. e result was the development of distributed networks to close the Technology Readiness Level (TRL) 7–9 gap. TRL the concept used by agencies and companies to assess the maturity of evolving technologies. TRL 7 models and data can operate in stand-alone mode within the context of an operational environment. TRL 8 signi!es the completion of a fully functional system prototype, and TRL 9 describes a fully delivered, operational system. e distributed network systems at TRL 7–9 are linked in a way that allows developers to maintain versioning and proprietary control over their models, permits data streams to be integrated at their own measurement cadence, results in testing %exibility, enables rapid