O3A: Data Assimilation (DA)


Moderator:
Tom Berger (CU Boulder/SWx TREC)

Advisors: (proposed)
Marcin Pilinski (CU Boulder/LASP)
Nick Pedatella (NCAR/HAO)

This overarching activity focuses on the research, development, and operational implementation of data assimilation systems for space weather forecasting. The activity will primarily focus on developing DA systems for the Ionosphere-Thermosphere-Mesosphere (ITM) domain due to the prevalence of physics-based forecasting models for this domain as well as the relatively large amounts of assimilative data available compared to the magnetosphere or solar domains. While the magnetosphere and solar domains have sparse data availability for development of assimilation systems, research into these domains will also be supported, particularly considering upcoming missions prioritized in the Heliophysics Decadal Survey and/or international operational space weather observatories in the L1, L5 and L4 Lagrangian orbits.  In addition to research into observation operators for both NASA satellite and ground-based ITM system observations, the cluster will coordinate the development of novel DA methodologies such as hybrid variational/ensemble methods. Both the observation operator and DA system development efforts may take advantage of novel DA frameworks such as the NASA/NOAA/DOD JEDI system. Physics-based ITM models to be utilized include the NCAR TIE-GCM and WACCM-X models, the NASA/Goddard GEOS upper atmospheric model, the CGS DRIVE center MAGE model, and the NOAA FV3/WAM-IPE model. Empirical models include the DRAGSTER, JB08, and MSIS models.  The cluster will also support Observation System Simulation Experiments (OSSEs) to investigate the efficacy of including upcoming sources of ITM, magnetospheric, or solar observations, as well as Observation System Experiments (OSEs) to test the efficacy of inclusion of existing data sources in increasing the predictive skill of various models. There will be a synergy of this cluster with the Machine Learning cluster (O3B) as new methods of purely data-driven nonlinear system prediction and hybrid ML/DA systems are developed for operational forecasting of space weather. 

Activities:

  • Development of observation operators for NASA satellite missions such as TIMED/SABER and GOLD, NOAA missions such as COSMIC-II, SWFO-L1, and GOES, commercial LEO satellite GNSS data, and ground-based systems managed by NSF, NASA or international coalitions such as meteor radars, Fabry-Perot Interferometers, etc. 
  • Development and test of JEDI model interfaces, observation operators, and DA methodologies for application across agencies. 
  • Performance of OSSEs and OSEs with various observational data sources.
  • Model performance evaluation including model validation (cotemporal) and forecast verification (predictive skill) with and without DA. 
  • Designing and evaluating new observational systems for ITM, magnetospheric, and solar data assimilation including instrumentation, data formatting, data latency, and formatting requirements.
  • Support the integration of LEO orbital determination systems with ITM models and DA systems for improved satellite navigation, collision avoidance, and re-entry planning.  


Action topics:

  • Improve space weather modeling and prediction capabilities through the implementation of ensemble and variational data assimilation methodologies for the ITM regime where data availability is highest.
  • Investigate particle filter ensemble techniques for sparse data regimes such as magnetosphere and solar space weather prediction. 
  • Explore the synthesis of DA and ML methodologies including purely data-driven forecast models (Cluster O3B activity) with novel real-time data ingest and inference methodologies.
  • Develop large time-span reanalysis datasets for use in training ML learning models for space weather prediction (Cluster O3B activity).
  • Understand the impact of various current and future space weather observations on model performance through OSSEs and OSEs.
  • Interface with the operational space weather forecasting enterprise to understand forecaster and end-user requirements, particularly in the field of LEO satellite navigation and collision avoidance. 
  • Facilitate the development of new instrumentation concepts for low-latency data suitable for data assimilation into existing and future forecasting models. 

 

Action Teams: 

O3A-1: SWORD – Space Weather Operational Readiness Development Center (Lead (proposed): Mark Miesch, CU/CIRES)

O3A-2: Operational forecasting for geospace satellite navigation (Lead (proposed) Shayla Mutschler, SET)