SESC currently monitors the Sun on a routine basis and issues alerts and
warnings based on one to three day predictions of solar flares and on
actual observations of the flares. Active regions are tracked as the Sun
rotates. On the basis of solar activity, they provide forecasts of future
geomagnetic activity as reflected in the 
and 
(the linear version
of 
) indices. Geophysical data are also distributed from a variety of
ground based and space based sources. Geosynchronous energetic particle,
X-ray and magnetometer data from the NOAA GOES satellite, energetic particle
data from the polar orbiting TIROS satellites, ground magnetometer data and
solar data are routinely available to SESC customers. AFSFC obtains a
variety of in situ data from the polar orbiting DMSP satellites, which also
will be shortly available in a post facto mode from the National
Geophysical Data Center. SESC and AFSFC share resources to serve their
respective customer bases. A synopsis of observing and forecasting
capabilities is given by Cliffswallow and Hirman [1993]. Verification
statistics show that geomagnetic forecasting is clearly a problem with only
a 44% accuracy for a one day lead time and a 40% accuracy for a five day
lead time. Forecasts over several days are not likely to significantly
improve until a better understanding of physical processes allows us to
accurately predict the development of events on the Sun and propagate those
effects to near-Earth space, or until an experimental capability to monitor
the progress of shock fronts through interplanetary space is developed.
A key to accurate, short--term (0 to 2 hr) forecasting of new disturbances is continuous, real-time solar wind data. Data taken at the libration point (240 Earth radii upstream), where the Earth's gravitational pull is balanced by that of the Sun, provide a 30 to 50 minute warning of when a shock or disturbance in the solar wind will encounter the magnetosphere [ Baker et al., 1983]. The precise time depends on the solar wind velocity. NASA launched the WIND spacecraft in November, 1994, to investigate and monitor the solar wind as part of the International Solar Terrestrial Physics Program. It is in a double lunar swing-by orbit, using synchronized close passages by the moon to constantly change the orbit in order to keep it in front of the Earth. It will spend a large fraction of its time at a geocentric distance greater than half of the libration point distance. The Geophysics Directorate of Phillips Laboratory in cooperation with the Air Force Space Test Program and NASA will be obtaining from WIND a limited amount of real-time solar wind data as part of an investigation called SWIM [ Heinemann et al., 1993]. Because of the modifications to the spacecraft for SWIM there is no spacecraft technical reason that limits the amount of real-time tracking. Options for increasing real time tracking are currently being explored within NASA and the Air Force. NOAA is providing resources to modify of the NASA Advanced Composition Explorer (ACE) to be launched in late 1997 to provide longer term real-time solar wind monitoring from the libration point.
The Air Force is making a significant investment in a technology transition program to develop specification and prediction models for various regions of near-Earth space to be used operationally in the AF Space Forecast Center [see Heinemann et al., 1993]. The Ionospheric Forecast Model (IFM) is a first principles model which uses the output of the Parameterized Real-time Ionospheric Specification Model (PRISM) as its initial conditions. It will provide a 12-hour forecast of when the ionosphere will recover from whatever disturbance is in progress. Forecasts of new disturbances are possible only when forecasts of the changes in the input driving functions are possible. The Magnetospheric Specification and Forecast Model (MSFM) also provides short--term forecasts of particle fluxes in space. In addition to being able to track particles in the inner magnetosphere, the MSFM also uses solar wind data in a neural network system to forecast key input parameters one hour ahead. This model is currently in its validation phase. The model successfully predicts most observed energetic particle flux enhancements and generally tracks the satellite data well. Upstream measurements of solar wind parameters must be translated to the magnetopause region for inputs into the MSFM and IFM. A Solar Wind Transport code (SWT) has been developed based on gas dynamics to translate in time and space the interplanetary magnetic field and solar wind density and velocity from the monitoring point to the bow of the magnetopause. Longer term (up to 2 days), but less accurate forecasts of the probability of major disturbances will be provided by the Interplanetary Shock Propagation Model (ISPM). It predicts shock arrival times using observations of solar radio bursts to characterize the solar flare determined energy release that drives the shock. Because the interplanetary magnetic field north-south component is not determined, it can not reliably predict magnetic disturbances. These models will be placed in operational use at AFSFC over the next two years and will establish the first short-term numerical forecast capability for regions of near-Earth space. An Integrated Space Environment Model (ISEM) executive system is also being designed to integrate the operation of these models and to quality control the results. It will build in the expertise of the model developers into the operations in a smart decision maker to lessen the burden of science related decisions on the forecaster.
An important improvement in accuracy of intermediate term forecasting on time scales of up to 48 hours will happen when it is possible to monitor the shock fronts of CMEs as they propagate toward the Earth. One such instrument, the Solar Mass Ejection Imager (SMEI) [ Jackson et al., 1991], which has been proposed by Air Force Phillips Laboratory and the University of California at San Diego to fly on a sun-synchronous polar orbiting satellite, will provide a complete map of the shock front every two hours. It is based on the measurement of Thompson scattered light from the increased density at the shock front. The concept was developed based on data analysis of variations in the Zodiacal light measurements by the HELIOS spacecraft. Measurements of the velocity and direction of the shock front could be used as more accurate boundary conditions in the ISPM model and would provide accurate arrival times of the impending shock front. The actual severity of the subsequent magnetic storm would not be predicted until the front had passed the libration point monitor and the interplanetary magnetic field orientation was known. The function of data from the SMEI instrument is analogous to that of the GOES images of global cloud cover which are on TV each night. The images track the frontal systems and provide data for the more accurate detailed models. Another possible method of tracking density and turbulence enhancements as they propagate out from the Sun has been through analysis of maps of interplanetary scintillation (IPS) in the 80 Mhz range. Ground-based antenna arrays determine maps of the ratio of the observed scintillation index to the average scintillation index. However, single station maps have been difficult to interpret during high solar activity, at least partly because of a changing ionosphere, and it has not been possible to define the rules necessary for an operational forecast technique. Maps from at least two widely separated stations will be needed if progress is to be made with this technique [ Leinbach and Ananthakrishnan, 1993].