DREAM: An R2O Success Story?

The Dynamic Radiation Environment Assimilation Model (DREAM) is a data-assimilative model of the Earth’s radiation belts that has, until recently, been used primarily as a research tool to understand radiation belt dynamics and to develop of Kalman filter techniques for application to magnetospheric modeling. More recently, the emphasis of the DREAM program has shifted toward implementation of an operational prototype for testing and validation at the Air Force Research Laboratory’s (AFRL) Space Weather Forecast Laboratory (SWFL). The transition has required significant effort, funding, and shifting of priorities that serve as a recent example of the opportunities and challenges of transitioning a model from research to operations (R2O). Two recent technical developments will be discussed: (1) implementation of real-time data-assimilation in DREAM and the associated user interfaces and data products and (2) adapting DREAM for service oriented architectures that enable distributed computation and “easy” integration of services that couple codes without those services even having to be aware of each other. (Google Earth is one example of a highly-developed service oriented architecture.) We will briefly illustrate some of these capabilities but we will also discuss requirements for 24/7 operation, self restarting, direct application of outputs, etc. Of equal importance, we will discuss some of the logistical aspects of our experience with R2O. Those include: How do you fund it? How do you identify a feasible subset of user needs? How do you (or should you) support activities at multiple R2O facilities such as SWFL, SWPT, and CCMC? What inter-agency coordination is required? The transition of DREAM from research to operations has been ongoing for just over one year and is not yet complete. So, it can not yet be called a success - but, it hasn’t failed and it continues to provide an interesting, contemporary test case as it crosses the R2O “valley of death”.