Notes from the Data Assimilation Workshop

1
Notes from the Data Assimilation Workshop
R.S. Weigel (CU/LASP)

• D. Baker (CU/LASP)
• D. Vassiliadis (NASA/GSFC)
• A. Klimas (NASA/GSFC)
• R. McPherron (UCLA)
• G. Siscoe (BU)
• N. Crooker (BU)
• H. Spence (BU)
• H. Singer (SEC/NOAA)
• T. Onsager (SEC/NOAA)
• N. Arge (SEC/NOAA)

Knowledge Transfer and Empirical Modeling Team
2
Data Assimilation Workshop Notes

• CU/LASP held a data assimilation workshop after
  Space Weather Week
• Copies of the talks are available at
  http//lasp.colorado.edu/cism/Data_Assimilation

• Why and What is Data Assimilation?
• What Data Assimilation is not
• Key Challenges in Data Assimilation
• Key Challenges with respect to magnetospheric DA
• How magnetospheric DA differs from meteorological
  DA

3
Lessons LearnedWhy and What is DA?

• Purpose of data assimilation is to combine
  measurements and models to produce best estimate
  of current and future conditions.
• Kalman filter is most often used as a method for
  data assimilation. It became popular because it
  is a recursive solution to the optimal estimator
  problem. (Only last time step of information
  needs to be stored.)
• Full implementation of Kalman is usually not
  possible. There is a growing field in the study
  alternatives.

• Data assimilation does not require a
  physics-based model.
• AD ? DA (The Assimilation of Data is not
  necessarily Data Assimilation)

4
Challenges in DA

• Analyzed field does not match a realizable model
  state
• Non-uniform and sparse measurements
• Observed variables do not match variables
  predicted by the model
• Observing systems are diverse and subject to
  error, sometimes poorly known.

5
Challenges For Magnetospheric DA

• Very sparse measurements
• Diverse set of both forward and inverse models
  that are highly specialized and are expert in
  different areas.

• How to combine forward models (MHD, particle
  pushing) with inverse models (empirical,
  stochastic).
• How to integrate data with these models

6
Differences between SW and Ordinary Weather

• SW is usually more concerned with unlikely events
• Magnetosphere is strongly forced