The Space Weather Reanalysis

SWR Methodology

Figure 4

[12] In order to drive the three main models used in the SWR effort it was necessary to create/enhance three separate databases. In that effort NGDC added to the SPIDR data management system the Hemispheric Power Index (NOAA/DMSP) and the 1.0 minute Interplanetary Magnetic Field data. In addition the Geomagnetic Holdings at NGDC were expanded by including stations not normally part of the World Data Center system. This expanded the number of stations for the 1998 period by over 50% (see Figure 4).

Figure 5

[13] To understand the significance of the newly added stations compare the results in Figure 5 which displays the AMIE model run with minimal data (middle) and run with full data (left) as well as the net difference. Clearly in trying to create a realistic representation of the environment the more robust the data set, the better.

[14] Next it was necessary to develop a uniform Application Programming Interface (API) for the SPIDR system so that individual models could request the supporting observations they needed directly, without resorting to file based interactions. The API has been fully integrated and tested with both SIMM and AMIE is is able to provide data via the network to support the automated runs of these models.

[15] When trying to run models over an extended period, several factors come in to play which may not be a consideration in normal operation. Things such as instrument baseline shift, undetected spikes, or calibration drift can lead to a skewed result for the model. As such part of the SWR effort was to develop and extend automated algorithms to handle the long term data archives necessary to run the 1999-2004 time period. This effort has yielded vastly improved algorithms for detecting spikes and other quality issues but continues to be an area of focus for the SWR team. Currently the data is being reviewed by the automated quality control at each step.