# 1. Summary

Data-derived magnetic field models of Earth's magnetosphere are generally developed using in-situ magnetic field measurements along with auxiliary parameters. These models predict the magnetic field at a position in the magnetosphere given values of auxiliary parameters that represent the state of the magnetosphere (i.e., disturbed or quiet). Early models required only a few parameters. For example, the Tsyganenko 1989 model required knowledge of Earth's dipole tilt and the level of geomagnetic disturbance (characterized by the Kp index). Modern magnetic field models are much more accurate but require many more auxiliary inputs. Qin et al. have developed a data set of input parameters for the Tsyganenko 2001, 2004, and 2005 magnetic field models. More information about this data set is available at Richard Denton's web page.

# 2. Acknowledgment

For data accessed through ViRBO, please cite

Weigel, Robert S., Daniel N. Baker, D. Aaron Roberts, and Todd King, Using Virtual Observatories for Heliophysics Research, Eos, Transactions American Geophysical Union 90, 44, 2009. [1]

in addition to the suggested citation below.

Appropriate acknowledgement would depend on the use. For minor use, you can add a sentence to the acknowledgements section such as "Magnetic field input parameters (derived from data on OMNIWeb) were provided by Richard Denton and Zhengui Qin." Please also acknowledge OMNIWeb as indicated on its website (see [2]). For a more substantial use (when the input parameters are crucial for the study) please also acknowledge a grant number that Richard Denton will give you when you contact him (see below). Then possible acknowledgement could look like "Magnetic field input parameters (derived from data on OMNIWeb) were provided by Richard Denton and Zhengui Qin, supported by so-and-so grant." If this data set is a major part of the paper (like a statistical study of the magnetic field for a large portion of the time period covered), we would appreciate an offer for us to collaborate as coauthors. Please feel free to make the decision yourself as to which form of acknowledgement is appropriate, but if possible let Richard Denton know when you publish a paper acknowledging the use of this database. You can contact Richard Denton at [3].

# 4. Update and Revision Notes

• Update 04/23/2009 - A correction in the code that affected the G2_status variable was applied. This correction only affects the hourly-averaged output data, which now has version number v1 in its file name.
• Update 05/15/2010 - The latest-available OMNI data were used to generated new data products for all resolutions.
• Update 10/20/2010 - Morley noted a potential problem with the high-res Dst output by code. See Talk:QinDenton
• Update 11/20/2010 - Source code was committed to svn http://virbo.org/svn/virbo/qindenton/.
• Update 02/06/2011 - All time resolutions were updated.
• Update 03/29/2012 - All time resolutions were updated.
• Update 03/05/2013 - All time resolutions were updated.
• Update 10/01/2013 - All time resolutions were updated.
• Update 10/01/2013 - Future updates will occur when [4] changes or when the source code changes. The files in the following subdirectory will always point to the most recent file in its parent directory.
• Update 04/27/2014 - D. Konrashov et al. [5] have made available a reconstruction of solar wind data in gaps and a README file and a file with reconstructed parameters (WGhourFS_72_13.txt) for 1972 to 2013. As described in the paper, this data will be superior to the data using the Qin et al. method for time periods before 1995 during which there are large gaps. After that, there shouldn't be much difference between the two methods.
• Update 08/18/2014 - The one-minute and 5-minute files are no longer being updated. The code used to compute the coefficients has a limitation (related to memory) on the time span over which the coefficients can be calculated that was discovered in early 2014. The issue has been reported to the author of the code used to calculate the coefficients.
• Update 08/18/2014 - Updates are now being done again on an ad-hoc basis.

# 5. Data: 1-hour

• Version 0: The baseline version.
• Version 1: A correction in the code that affected the G2_status variable was applied. This correction only affects the hourly-averaged output data, which is now has version number v1 in its file name.
• Version 2: The OMNI2 data set was updated to include new Dst data (see Newsletter announcement). This version uses the correction from Version 1 and the OMNI2 data downloaded on May 4th, 2009 (contains data through April 11th, 2009).
• Version 3-7: Version number was incremented because OMNI2 data changed.
• In the future, the version number will only increment if the source code used to process the input files changes.

Note that the ASCII output (WGhour-vVERSION.d.zip) and input (omni2_hour-vVERSION.dat.zip) files are available in the ftp directory [6].

# 6. Data: 5-minute

• Version 1: The OMNI data set has changes in many places and the version downloaded on May 15th, 2010 was used to create the version 1 files.
• Version 2: A processing error was corrected. See http://virbo.org/Talk:QinDenton
• Version 3-5: Version number was incremented because OMNIHR data changed.
• In the future, the version number will only increment if the source code used to process the input files changes.

Note that the ASCII input and output files are available at http://virbo.org/ftp/QinDenton/5min/merged

# 7. Data: 1-minute

• Version 1: The OMNI data set has changes in many places and the version downloaded on May 15th, 2010 was used to create the version 1 files.
• Version 2: A processing error was corrected. See http://virbo.org/Talk:QinDenton
• Version 3-5: Version number was incremented because OMNIHR data changed.
• In the future, the version number will only increment if the source code used to process the input files changes.

Note that the ASCII input and output files are available at http://virbo.org/ftp/QinDenton/1min/merged