RBW
From ViRBO
ViRBO is hosting documents and notes for this focus group. For information about other meetings, see Resources/Science.
Contents |
1. Overview
The Radiation Belt and Wave (RBW) focus group is a part of the NSF Geospace Environment Modeling (GEM) program. The RBW focus group will concern itself with identifying and quantifying the contributions and effects of various sources of heating, transport, and loss of radiation belt ions and electrons, and developing global and local models of the radiation belts and the wave environment of the inner magnetosphere.
The RBW focus group is chaired by Yuri Shprits, Jacob Bortnik, Scot Elkington, and Craig Kletzing. The RBW focus group was incorporated into the GEM program as part of the Inner Magnetosphere and Storms Research Area in 2010, and is slated to continue its activities through 2014. A full description of the RBW focus group and its objectives can be found here.
2. GEM-RBW meetings, workshops, and resources
- 2011_CEDAR-GEM_Joint_Workshop 2011
- RBW FG Agenda pdf
- GEM Summer Workshop 2010
- 2010 Fall AGU GEM Mini-Workshop
- RBW Notes and Resources
3. 2010 RBW Challenge
At the 2010 GEM Summer Workshop RBW Planning Session it was decided to organize a Global Radiation Belt Modeling Challenge, with the intent of better understanding the relative strengths of available physical and analytical models in capturing global radiation belt dynamics, defining necessary data inputs and model requirements, and working towards defining appropriate comparative metrics in evaluating the various models.
The challenge defines an interval of time comprising several geomagnetic storms which were associated with changes in the trapped space radiation environment. The challenge provides for a "tune-up" or training interval of time, independent of the challenge interval, for adjusting model inputs and settings. The goal of the challenge is to accurately simulate variations in the trapped radiation environment for either the entire challenge interval, or for select storms within the interval.
The purpose of the challenge is not to declare a “winner” among the models participating in the challenge, but to learn what physics and numerics are important for reproducing real observations.
3.1. Challenge interval and validation
The RBW group selected the following periods and storms for the challenge and training intervals
- Training interval: August 15, 1990 to October 15, 1990
- August 26-30, 1990 geomagnetic storm
- September 11-16, 1990 geomagnetic activity
- October 9-15, 1990 geomagnetic storm
- Challenge interval: February 1, 1991 to July 31, 1991
- February 1-5, 1991 geomagnetic storm
- February 24, 1991 geomagnetic activity
- March 24-31, 1991 geomagnetic 'superstorm'
- May 17-21, 1991 geomagnetic storm
- June 5-17, 1991 geomagnetic storm series
- July 9-11, 1991 geomagnetic storm
Several of these storms are discussed in more detail in Iles et al., 2006. The March 24-31, 1991 event was studied in Chu et al., 2010.
Model results will be evaluated against data from the CRRES mission. We will compare CRRES observations of energetic electron phase space densities as a function of L* at the following first and second invariants. L* and K calculations are available with the other challenge input data.
- 1860 MeV/G (~1 MeV at geosynchronous), K=0 (equatorial) and K=0.1 (off-equatorial)
- 750 MeV/G, K=0 and K=0.1
- 550 MeV/G, K=0 and K=0.1
More details about the Global Challenge interval, storms, and validation points are available here; an overview of CRRES observations for the challenge interval are available here and listing of hourly Dst values for the challenge interval is available here. Please see Challenge data resources for a full list of input data sources available for the challenge.
3.2. Challenge rules and guidelines
- Two pools for evaluation: Data assimilators and physically-based models.
- Simulate the entire interval if you can, select specific storms if you must.
- A “Tune-Up” or training interval has been defined, during which simulators can experiment with input parameters to select the best settings.
- The "Challenge" interval should be simulated using a single set of model parameters, determined from the training interval.
- Input and validation data:
- Solar wind input conditions, provided by UCLA.
- Validated against CRRES phase space density (PSD) vs M, K, L*.
- Allowed data sources:
- Anything else (including other in situ data, like Akebono, GPS, GEO).
- Must be made available for use by all challenge participants.
- Post small data files directly to wiki. Post large files to ftp://virbo.org/tmp and link to them from this wiki page. See Uploading challenge results for an example.
- Metrics: Root mean squared error (RMSE) in log PSD at CRRES for 5 reference M, K
- Pass-by-pass RMSE
- Storm-by-storm RMSE
- Participants are encouraged to submit entire 4-D (M,K,L;t) data cube to challenge data repository
- Also, evaluate PSD at 2 RBSP s/c
- Challenge coordinators will procure RBSP sample ephemeris
- Prepare us to start thinking about how to analyzer RBSP data
3.3. Challenge data resources
3.3.1. Input and validation data
- Reconstructed solar wind data (time, By, Bz, P) and geophysical indices (Kp, Dst) for the challenge and training intervals, courtesy Kondrashov and Shprits. Latest version (2013-08-02): [1] and original version [2].
- CRRES/MEA flux data, courtesy Bob Johnston (file list).
- Abstract: A new cross-calibrated data set of energetic electron observations from the Medium Electron Sensor A (MEA) and High Energy Electron Fluxmeter (HEEF) instruments on the CRRES satellite has been completed. This includes cleaning for proton contaminated MEA data and MEA/HEEF data with incomplete pitch angle data, spectral correction of MEA data, and adjustment of HEEF data at high flux levels. Description of the data set contents is provided.
- CRRES magnetic field data (ASCII) (file list and data description).
- CRRES electric field data (ASCII): (file list and data description).
- Reanalysis obtained with VERB 1D data assimilative code using Akebono, CRRES, GPS, and GEO measurements, courtesy Y. Shprits (file list or zipfile containing all results). Note: These values are provisional. L* was computed with ONERA-DESP and results have not yet been validated.
3.3.2. Other data resources
- Akebono data during the CRRES interval, courtesy T. Nagai (file list).
- Simulated RBSP locations for the challenge interval, courtesy Paul O'Brien (rbspa.cdf and rbspb.cdf).
- L* and K calculations using IRBEM for Akebono, CRRES, GEO 1989, and GPS ns18, courtesy Y. Shprits (file list or zipfile containing all results).
- http://virbo.org has lots of merged and versioned data files that may be useful. (These versioned data files were created in part to support model validation/comparison studies, which require fixed inputs, even if the inputs are out-of-date.)
3.4. Uploading challenge results
To upload a file, login (if you do not want to create an account, use username=RBWM and password=MWBR) and edit the page and enter text similar to this Description of My File. After saving, you will see a red link as in Media:My File Name.ppt. When you select the red link, you will be taken to an upload page.
An example data file posted to the wiki: Media:data.txt. Same data file uploaded to ftp://virbo.org/tmp/weigel/data.txt (create a subdirectory with your last name). For security purposes, you will not be able to see any subdirectory or files that you create until after it is moved. Send an email to virbo@virbo.org when you have completed the upload.
3.5. FAQ
- Question: What is the relationship between the challenge events and the events listed at RBSPb? Answer: The RBSPb event list is more exploratory and is intended for a loose collaboration for attendees of the Radiation Belt Saint Petersburg meeting. The RBW event list is a part of a formal challenge as a part of the GEM Workshop.
- Question: What is the difference between the Training intervals and the Challenge intervals? Answer: The set of training intervals are intended to be used for tuning models and the challenge intervals for testing model predictions. Model parameters should not be selected based on any analysis performed using the test set intervals. (Ideally the test intervals would not be known in advance.)
- Question: Is everyone required to use the same inputs? Answer: Ideally yes, but many models use different inputs in part because they are based on very different physical assumptions and approximations, so this is not mandatory. The inputs that are common to all models (solar wind velocity, for example) should be the same (and are made available on this web site) to simplify the comparison.
4. RBW Notes and Resources
- Radiation Belt, Saint Petersburg (RBSPb) wiki page and RBSPb challenge event info. See also #FAQ for explanation about relationship between RBW and RBSPb challenge.
- Teleconference and meeting notes: RBW/Telecons
