back
 

BAS Radiation Belt Model Forecast

Radiation belt forecasts

The BAS radiation belt model (BAS-RBM) calculates the electron flux across the entire outer radiation belt for the last 7 days and provides a forecast up to 1 day ahead.

Benefits

  • The forecasts cover the entire outer radiation belt
  • The forecasts are available for a range of different energies
  • The forecasts include wave-particle interactions that improve accuracy
  • The forecasts provide a time resolution of 1 hour, not a 24 hour average
  • The forecasts are updated every hour, 24/7

The BAS-RBM

The BAS-RBM is a research model that solves the Fokker Planck equation to find how the differential electron flux varies with time. The model includes electron transport across the magnetic field by radial diffusion, pitch angle and energy diffusion by wave-particle interactions, electron loss into the atmosphere via wave-particle interactions and collisions with atmospheric gasses, and losses due to magnetopause compression. For more details see Glauert et al. [2014].

How we make our forecasts

We use a near real-time forecast of the Kp index, combined with data from the DSCOVR satellite in the solar wind and electron data from the GOES 16 satellite at geostationary orbit as input to the model. The model is then run to re-construct the radiation belts for the past 7 days and to provide a forecast for the next 24 hours. The results cover the entire outer radiation belt. The input data are updated every hour and the model is re-run to update the radiation belt forecast every hour. The results are then displayed on the public web site.

For more background information on the forecasting method see Horne et al. [2013]

Forecasting timescale

Since it takes approximately 30-60 minutes for the solar wind to flow from the DSCOVR spacecraft to the boundary of the Earth’s magnetic field one may ask how we can forecast up to 24 hours ahead reliably. However, research shows that it takes approximately 4 hours for information about changes in the solar wind to be transferred to electrons at geostationary orbit [Borovsky et al., 1998]. This additional time delay helps to make the forecast up to about 4 hours ahead more reliable than one might think

Beyond 4 hours the forecasting relies on the accuracy of the Kp index, and becomes much more uncertain. Here the Kp index is a forecast based on the 27 day rotation of the Sun. In this case the forecast depends on the accuracy of the Kp forecast and the time history of the radiation belts.

The method implicitly assumes that the level of geomagnetic activity will not change much over the next 24 hours. If it changes slowly and the solar wind becomes less geo-effective then activity will generally drop and our forecast will tend to be a little high. However, this will be corrected within one hour. If the solar wind becomes more geo-effective, then the forecast will tend to be a little low.

If the solar wind changes rapidly, on a timescale less than an hour or so, then the forecasts will become less reliable. For example, a large coronal mass ejection (CME) can push the outer boundary of the Earth’s magnetic field inside geostationary orbit on a timescale of minutes, totally changing the radiation environment. However, when the solar wind changes rapidly one of the most common features of the outer electron radiation belt is that the electron flux drops rapidly on a timescales of minutes to hours. This is known as a flux dropout and is the subject of much research. We cannot reproduce these events reliably at present. Since satellite operators are more interested in high flux, which causes internal charging problems, the omission of these events from our model should not pose too much of a problem, at least for the outer radiation belt beyond about 4 Earth Radii (about 19,000 km altitude). On the other hand, a large coronal mass ejection can cause a major increase in the electron flux below 19,000 km altitude. In this case there could be a substantial risk of satellite charging.

Skill scores

The forecasting skill has been evaluated and the resulting paper is currently under review. The skill for predicting the electron flux in medium Earth orbit against the average flux ranges from 0.6 to 0.8. Here a skill score of 1.0 corresponds to a perfect prediction and 0.0 is no better than taking the long term average. Thus a skill score greater than 0.5 suggests the model is doing very well.

References