The GOES-R Sudden Impulse Detection Algorithm

The GOES-R Sudden Impulse (SI) Detection Algorithm will offer a powerful new tool to help forecasters and end users mitigate the effects of geomagnetic disturbances. Sudden Impulses often precede geomagnetic storms, which can cripple critical infrastructure such as the electrical grid. This new technique will therefore provide a way to help Space Weather forecasters prepare power companies, oil pipeline operators, and other affected parties with the opportunity to adapt their operations in such a way as to minimize impacts to the public.

The algorithm will work by combining measurements taken by the magnetometers aboard GOES satellites, ground magnetometers, and possibly measurements of the solar wind and magnetic field taken upstream of Earth at the L1 Lagrangian location by the Advanced Composition Explorer (ACE) or Deep Space Climate Observatory (DSCOVR). The algorithm then searches for a rapid change in these observations in a short time period. Two different methods are currently being employed to analyze the results for a relevant disturbance. The basic difference is that the first approach identifies time periods when an individual magnetometer is experiencing a rapid change, then counts how many magnetometers are affected within a certain time window to identify an SI. The other tries to develop a global picture of the change in the geomagnetic field first, then determines whether this global field proxy is changing rapidly to identify an SI. Identification of regional changes, for example a rapidly changing field in the magnetic longitudes spanned by the United States in the absence of a global sudden impulse, is also under consideration.

Each method has strengths and weaknesses which will be discussed in some detail. Each method also has a certain amount of scalability, which should mean that as the forecast center obtains access to additional magnetometers these data can be added to the algorithm, permitting results to improve throughout the life cycle of the algorithm. Ultimately, selection of the method for implementation will be based upon scoring the results of each algorithm versus a truth dataset.

Validation has been initiated on each algorithm using over a year's worth of high temporal resolution data provided by NOAA and the USGS. We plan to utilize data from a full solar cycle for the final validation and scoring. This extensive validation, combined with regular feedback from forecasters throughout the development cycle, should help to ensure that the end product substantively improves operators' abilities to protect the interests of the public.