Skip to main content

Protecting space assets from high energy particles by developing European dynamic modelling and forecasting capabilities

Final Report Summary - SPACECAST (Protecting space assets from high energy particles by developing European dynamic modelling and forecasting capabilities)

Executive Summary:
Space weather poses a serious threat to satellites on orbit. Radiation exposure to particles from the Sun and the Earth’s radiation belts have been linked to service outages and in exceptional cases total satellite loss. In 2003, ten percent of the entire satellite fleet malfunctioned during a major geomagnetic storm caused by space weather. Since then, the number of operational satellites has risen from 450 to over 1000. Better designs have been introduced, but service outages associated with space weather still occur.
Europe plays a major role in the design, construction and launch of satellites, and the provision of satellite services. The EU is deploying the Galileo radio-navigation system consisting of 30 satellites and has an extensive Earth observation programme called Copernicus. As Europe has a major investment in space it is essential that it develops the means to protect it. In this context, the SPACECAST project was set up to help protect spacecraft from high energy particles from the Sun and the Earth’s radiation belts by developing European models to analyse and forecast periods of high risk.
SPACECAST has delivered the first European system to forecast high energy charged particle radiation in the Earth’s outer radiation belt, and a “nowcast” of medium energy electrons and radiation dose. The forecasts are provided freely via the web (www.fp7-spacecast.eu) and are updated every hour automatically. The system has three unique features: First, it uses state-of-the-art models that include the physics of wave-particle interactions. Second, the forecasts cover the whole of the outer radiation belt and include satellites in medium and geosynchronous orbits. Third, it is a truly international collaboration using satellite and ground based data from the USA, Antarctica, Europe and Japan, and a network European modelling centres. To help users interpret the forecasts they are also presented in the form of a risk index for internal satellite charging, surface charging and radiation dose.
The forecasts are primarily for satellite operators, although designers and space insurance underwriters can also use the system. However, European citizens benefit from the results of SPACECAST via better forecasting of satellite risk of damage, an improved delivery of satellite services, a more competitive space industry from the knowledge gained, and new research which will lay the foundation for future applications.
SPACECAST has delivered major pieces of scientific research. The highlights include:
• Demonstrating that electron acceleration by wave-particle interactions is a key process that helps forms the outer radiation belt
• Developing better forecasting models as a result of including wave-particle interactions
• Modelling how electrons can be transported to geosynchronous orbit by time dependent electric field
• Showing how electrons loss from the radiation belts into the atmosphere is controlled by geomagnetic activity
• Developing an innovative idea to measure electron transport across the magnetic field which shows how transport increases with geomagnetic activity
• Developing a semi-empirical model for particle acceleration by shock waves near the Sun
• Conducting detailed modeling of solar energetic particle events for use in future prediction models
The SPACECAST project team has developed a network of stakeholders including satellite operators, designers, and insurance underwriters. Team members provided briefings for policy makers, NATO MPs, the UK House of Commons, and evidence for the UK Government on space weather. They also appeared on BBC2 TV “Stargazing Live - Back to Earth”, gave an interview for a Swedish TV documentary, a radio interview for the BBC World Service, participated in a podcast, issued three press releases, and the project results were covered by numerous newspaper and magazine articles. One press release in 2012 on the forecasting system led to 46 different press reports world-wide. Team members published 30 peer reviewed papers and gave numerous presentations at international conferences. They organised stakeholder discussion meetings with the European Space Agency each year at the European Space Weather Week.

Project Context and Objectives:
Our modern society depends heavily on satellites and other space assets for a wide range of applications. These applications include communications such as mobile phones, television, and internet access. Satellites also provide accurate timing signals used for navigation in cars, ships and aircraft, as well as Earth observation which is used for agriculture, weather forecasting, security and defence, and many other areas. Over the last ten years satellite services have transformed business and have led to new innovations and economic growth [SIA, 2012]. For example, GPS timing signals are an essential part of high frequency trading on the London and New York stock markets, and are also used in agriculture to help farmers save fuel and fertiliser and hence reduce the chemical impact on the environment. The use of satellite services has also led to dependencies between different sectors of the economy in ways which were unforeseen.
Europe has always played a major role in the design, construction and launch of satellites, and the services that they provide. Two of the biggest satellite operators in the world (SES and Intelsat) are European; together they operate over 100 satellites. One of the major launch providers (Arianespace) is also European. Most of the financing of satellites for launch and in-orbit insurance is done through European financial centres. Europe is also deploying in the Galileo radio-navigation system consisting of 30 satellites and has an extensive Earth observation programme called Copernicus. To put this into context, the cost of a modern telecommunications satellite is around €200m to build and €80 – €100m to launch into geosynchronous orbit. With more than 380 operational satellites at geosynchronous orbit (more than 1,000 operational satellites on orbit altogether) and revenue of €65 billion a year from satellite TV alone [SIA, 2012] space is tremendously important. As Europe has a major share in this investment, it is essential that it develops the means to protect it.
Space weather poses a serious threat to satellites on orbit. Radiation exposure to particles from the Sun and the Earth’s radiation belts have been linked to service outages and in exceptional cases total satellite loss. In 2003 ten percent of the entire satellite fleet malfunctioned during a major geomagnetic storm caused by space weather. Despite advances in design service outages still occur. For example, in 2012 three satellites suffered a service outage for a few hours to a few days during a large geomagnetic storm and solar energetic particle event. The problems encountered included electrostatic charging, loss of solar array power and disruption to electronic memory circuits, all of which are classical signatures of space weather events. In July 2012 the Sun emitted one of the largest coronal mass ejections on record which was recorded by the STEREO spacecraft. Fortunately the CME did not come towards the Earth but if it had, it could have triggered an event as large as the 1859 Carrington event with major consequences. In the UK the Government takes the risk seriously and has put severe space weather on the National Risk Register of Civil Emergencies with the same relative impact score as heat waves and heavy snow.

Figure 1. Satellite orbits in relation to the Earth’s electron radiation belts. Satellites in geosynchronous, medium and elliptic orbits pass through the radiation belts. Those in low Earth orbit encounter the radiation belts at high latitudes.
Satellites are launched into four types of orbits (see Figure 1). Most telecommunications satellites are in geosynchronous orbit (GEO) which is about 36,000 km above the surface of the Earth. They maintain their position in longitude so that they are always in contact with a fixed point on the ground. Most Earth observation satellites, and the international space station, are in low Earth orbit (LEO) which lies between 400 and 1000 km. These satellites may take something like 90 minutes to orbit the Earth. Between these two orbits is medium Earth orbit (MEO) where most of the navigation and positioning satellites are located. The new Galileo radio-navigation satellites are in medium Earth orbit with an orbital period of approximately 14 hours. In addition to these orbits there are a few satellites in highly elliptic orbits (HEO) which are specially designed to maximize coverage over the high latitude and Polar Regions. Satellites in all these orbits (except low inclination LEO) pass through the Earth’s van Allen radiation belts where particle radiation levels are severe and can increase rapidly due to space weather. Similarly, solar eruptions can cause a rapid increase in energetic particles which penetrate the Earth’s magnetic field and cause damage to satellites in all orbits. These events are very difficult to predict and require more research to understand.
The goal of the SPACECAST project is to help protect space assets from high energy particles in the electron radiation belts and in solar energetic particle (SEP) events by developing European dynamic modelling and forecasting capabilities. The project addresses two of the most important radiation hazards for space vehicles and manned spacecraft. To achieve this goal the objectives of the SPACECAST project were:
• To forecast changes in the Earth’s electron radiation belts in near real time using a suite of physics based research models and near real time data
• To provide warnings and alerts to stakeholders based on dynamic radiation belt forecasts
• To improve understanding of the key processes responsible for dynamic variations in the radiation belts, and improve our ability to represent these processes in forecasting models
• To test model predictions of solar energetic particle events at 1AU against data, and improve SEP event modelling
• To develop understanding of how SEPs are accelerated at interplanetary shocks driven by coronal mass ejections and their storage in the plasma turbulent region ahead of the shock
• To provide a legacy of tools, skills and understanding that will last long after the completion of the SPACECAST project.

Project Results:
SCIENCE AND TECHNOLOGY RESULTS
The SPACECAST project began on the 1st March 2011 and ended on 28 Feb 2014. The project made major advances in three main areas
• Forecasting space weather
• Scientific research
• Dissemination and stakeholder engagement.

FORECASTING HIGH ENERGY ELECTRONS
For satellites on orbit one of the most important hazards is internal charging cause by high energy electrons at energies of a few hundred to several million electron Volts (eV). These electrons are trapped inside the Earth’s magnetic field in the Van Allen radiation belts (see Figure 1). During space weather events the flux of these electrons can change by five orders of magnitude causing internal satellite charging, electrostatic discharge and permanent damage to electronic components. Several types of satellite anomalies have been related to electrostatic discharges and in some exceptional cases they have caused total satellite loss.

Figure 2. An example of satellite risk taken from the SPACECAST web site (www.fp7-spacecast.eu).
One of the major achievements of the SPACECAST project is to provide the first European system to forecast high energy charged particle radiation in the Earth’s outer radiation belt [Horne, 2012; Horne et al., 2013]. The system uses a forecast of geomagnetic activity to drive two independent computer models, one in France and one in the UK. The models compute the electron flux and the 24 hour electron fluence at selected energies and present a forecast of up to 3 hours ahead. The forecasts are updated every hour automatically. To help users interpret the forecasts they are also presented in the form of a risk index (red, amber, and green – see Figure 2) for internal satellite charging and surface charging, for different orbit types. The forecasts are primarily for satellite operators, although satellite designers and space insurance underwriters are also known to use the system, and more technical information is also provided for scientists and engineers. The system was released after the first year of the project, in March 2012, and has been operating ever since. A major upgrade of the forecasting models was made in 2014 after input from the research studies described below.
The SPACECAST forecasting system has three unique features. First, it uses physics based models to make the forecasts which include the physics of wave-particle interactions (described below). Second, it makes a forecast of radiation levels for the whole of the outer radiation belt which includes satellites in medium and geosynchronous orbits, and third it is a truly international collaboration using satellite and ground based data from the USA, Europe and Japan, and a network European modelling centres.
One of the new innovations is to use a forecast geomagnetic activity derived from the ACE satellite to drive the forecasting models. ACE is located between the Sun and the Earth at the L1 position and measures the polarity of the interplanetary magnetic field which is essential for forecasting the severity of magnetic storms and the intensity of the radiation belts. Since it takes 30 to 60 minutes for the solar wind to reach the Earth from the ACE spacecraft this enables us to forecast ahead. In fact the time for the radiation belts to respond to the solar wind can be much longer than this, which is why the forecasting could be extended up to 3 hours ahead and possibly longer. The use of ACE data has helped to reduce false alarms. When ACE data is not available the system uses a forecast of geomagnetic activity derived from ground based magnetic observatories.
In the forecasting system satellite and ground based data are collected from sites in Sweden, Germany, Japan, UK and the USA by a system managed by a small company in Belgium. The data are used to drive the forecasting models at distributed sites and the results exported back to Belgium where they are displayed on the SPACECAST web site (see www.fp7-spacecast-eu). Supporting data from satellites in the solar wind and magnetosphere and ground based data from the Antarctic and are also displayed on the web site to help verify the forecasts and aid scientific interpretation. The feedback from one satellite engineer – unrelated to the project - was “the web site is an order of magnitude (maybe several) more sophisticated than the NOAA SEC one”.
The forecasting models used in the SPACECAST system are state-of-the-art physics based models. They are analogous to general circulation models used in climate research. They solve a diffusion equation that includes electron transport across the magnetic field, electron acceleration due to wave-particle interactions, losses to the atmosphere and losses to the outer boundary of the geomagnetic field.
The SPACECAST forecasting system has proved to be very robust. In March 2012 there was a large space weather event consisting of a geomagnetic storm and solar energetic particle event. During this event data from the ACE and GOES satellites, which are two of the primary satellites used to monitor space weather, became unreliable. GOES in particular was showing an exceptionally high electron flux which was contaminated by solar energetic particles. In contrast, the SPACECAST system continued to provide electron forecasts throughout the period [Horne et al., 2013]. The resilience of the SPACECAST system is due to its ability to switch between different sources of data to drive the forecasting models, and the distributed nature of the modelling centres so that the service can continue even if one of the links goes down.

NOWCAST OF MEDIUM ENERGY ELECTRONS
While high energy electrons at energies of a few hundred to several MeV cause internal satellite charging medium energy electrons, typically 1- 100 keV, are known to cause surface charging. Surface charging can also cause an electrostatic discharge, loss of solar array power, uncontrolled phantom commands and many other types of unwanted behavior. Surface charging is a particular hazard when there is a large increase in medium energy electrons which often occurs near local midnight and early morning at geosynchronous orbit. It is a particular hazard when satellites go into eclipse in March and September each year as this reduces photo-emission and the satellite can charge positive to very high levels, as high as several kV.
It is not possible to provide a forecast of medium energy electrons that cause surface charging since the basic science governing the timing and transport of these electrons is not understood and since there are a very limited number of measurements. Instead of a forecast, the SPACECAST project provides a “nowcast” of the electron flux in the energy range 40 – 150 keV based on a sophisticated electron transport model. A nowcast is very valuable since the flux of electrons can vary spatially and well as temporally, so that at a given time the electron flux along the geostationary orbit can vary by orders of magnitude depending on position.
The model that produces the nowcast is called IMPTAM [Ganushkina et al., 2013] and uses data on the solar wind to define the source population of medium energy electrons and to drive the model [Amariutei and Ganushkina, 2012]. The source population resides beyond geosynchronous orbit on the night side of the Earth. The model takes into account the solar wind interaction with the geomagnetic field to transport electrons from the source inwards towards the Earth. The model calculates the differential electron flux at midnight for a range of orbits between approximately 2 and 7 Earth radii from the planet.
Comparisons between the IMPTAM model and the few satellites that do measure low energy electrons showed a correlation between satellite charging and periods of high electron flux, and that these events in space are related to magnetic field disturbances measured on the ground in the Antarctic. Again, the nowcast model works automatically and is updated every hour. Although there is a reasonable agreement between the model and data there is still more work to do to capture all the variability for very disturbed conditions. These studies offer very exciting new opportunities for satellite operators and scientists to work together.

REAL TIME RADIATION DOSE RATE
A third important step forward has been the real time calculation of the radiation dose rate at geostationary orbit due to solar energetic particles. Usually the dose is accumulated gradually but during a solar energetic particle event the dose rate can rise significantly for a period of a few days or more causing an increase in the number of single event upsets in electronic components. Radiation dose is one of the key factors that limit the operational life of satellites on orbit and must be carefully assessed at design stage. For a satellite with a design life of 15 years or more one of the major uncertainties is planning the number of solar energetic particle events that might take place and monitoring the accumulated dose.
During the project a new tool called Dosetool was developed to compute the radiation dose at different depths in various materials and for different thicknesses of shielding. The incident particles are assumed to be protons and are assumed to travel in a straight line in any material. The tool computes the dose [rad] and dose rate [rad/s] for 74 different materials for two shielding configurations; planar and spherical. The tool also enables the analysis of synthetic SEP events which are available from the SOLPENCO tool for more detailed event analysis.
A near-real time calculation of radiation dose rate was implemented on the SPACECAST web site. By using proton differential intensities measured by the GOES spacecraft at GEO orbit the dose rate is calculated by assuming a spherical shielding of 2mm of Aluminium for a silicon target, which is a typical value for geosynchronous orbit. The results are displayed on the SPACECAST web site and provide an immediate awareness of the hazard for satellite operators.

SCIENTIFIC HIGHLIGHTS

ACCELERATION AND LOSS BY WAVE-PARTICLE INTERACTIONS
Two of the major scientific problems in radiation belt physics are to understand how electrons are accelerated to very high energies to form the radiation belts, and what controls the variability of the radiation belts. Over the last few years it has been established that various types of very low frequency plasma waves can cause electron acceleration to very high energies. One of the major advances in the SPACECAST project was to include wave-particle interactions into a global model of the radiation belts and demonstrate that a particular type of electromagnetic waves, known as chorus waves, can accelerate electrons up to several MeV in energy. The results show that acceleration occurs mainly inside geostationary orbit.

Figure 3. Comparison between electron data from the CRRES satellite (top panel) and three simulations. The best results are when wave-particle interactions are included (panel d). Geomagnetic activity is shown at the bottom.
Figure 3 shows a set of comparisons between satellite data and the computer model. The top panel (a) shows electron data from the CRRES satellite at an energy of 0.976 MeV for a period of 100 days or so in 1990. When the model was run using only radial transport the second panel (b) shows that there is very poor agreement between the model and the data and the model flux is too high. By including various types of plasma waves the third panel (c) shows that there is a much better agreement with the data but the peaks in the electron flux (red) are not well reproduced. The model showed that these waves contribute to electron loss into the atmosphere. When wave-particle interactions due to chorus waves are included the fourth panel (d) shows that there is a much better agreement between the model and data. The results show that chorus waves cause electron acceleration on a global scale. More generally, the variability of the radiation belts is better reproduced when wave-particle interactions are included [Glauert et al., 2014]. The results of this study were used to improve the forecasting models substantially in 2014.
Another major development in the research models was to study the rapid drop in the electron flux that is often observed during geomagnetic storms. By using data on the solar wind dynamic pressure and polarity of the interplanetary magnetic field, changes in the outer boundary of the Earth’s magnetic field were incorporated into the research models. These studies showed that when the outer boundary moved inwards towards the Earth there was an increased outward transport of electrons and a rapid reduction in the radiation belt flux. In effect, the losses in the radiation belts were taking place at the outer boundary of the magnetic field. Again, this effect has been included into the forecasting models in 2014.

TRANSPORT OF MEDIUM ENERGY ELECTRONS
Time dependent changes in the Earth’s magnetic field and corresponding induced electric fields are very important factors controlling the transport of medium energy (1 – 100 keV) electrons inside the magnetosphere. These time dependent fields are triggered suddenly by the solar wind interaction with the magnetosphere (a substorm), but how this is done, and the spatial and temporal scales over which it occurs are major scientific questions.
A series of studies were conducted using the IMPTAM model to assess the importance of the electric and magnetic fields. Using data on the interplanetary magnetic field and solar wind to drive the model [Amariutei and Ganushkina, 2012], the model could transport electrons for the outer region of the magnetosphere at 10 Re on the night side of the Earth to geostationary orbit. However, the flux was found to be substantially lower than that observed suggesting that improved boundary conditions are required. It was also found that the results were more sensitive to the choice of magnetic field model than the large scale electric field models [Ganushkina et al., 2013]. A series of electromagnetic pulses were also applied to the model to represent the effects of a substorm. These pulses resulting in a rapid change in the electron flux near geosynchronous orbit, but it was found that more work is required to match the timing and the magnitude of the flux observed by spacecraft [Ganushkina et al., 2014]. As a result of these studies a scaling of the source population was implemented and used to improve the nowcast model in 2014.

WAVE-PARTICLE INTERACTIONS
Wave particle interactions play a key role in radiation belt dynamics. They lead to particle heating and acceleration in the radiation belts and also particle scattering in direction (pitch angle) leading to loss from the radiation belts into the atmosphere. Dynamic global models of the radiation belts are diffusion models and require diffusion rates that depend on wave properties. Since wave properties can vary significantly with spatial location and geomagnetic activity the state of the art is to develop global models of the wave power for the various relevant wave modes using data from several satellites. Several types of plasma waves were studied in the SPACECAST project as described below.
WAVE DATABASE
In order to analyse wave data for research studies and use it in the forecasting models a comprehensive database of plasma waves in the inner magnetosphere was developed. Data from seven scientific satellites were obtained and used to extend the coverage and improve the statistics of existing models. Data from each of the satellites was combined and weighting by the number of samples to produce a database of the wave power as a function of frequency band, L*, magnetic local time (MLT), magnetic latitude ?m, geomagnetic activity as measured by the AE and Kp indices, and location with respect to the plasmapause. Details of the instrumentation, data analysis techniques and binning method are described in detail in Meredith et al. [2012].
CHORUS WAVES
Chorus waves play a major role in the formation of the outer radiation belt by accelerating electrons to MeV energies. Under certain conditions they deplete the radiation belts by contributing to electron loss into the atmosphere. They are a special type of circularly polarised waves at frequencies of typically a few kHz, below the electron cyclotron frequency. The global distribution of these waves and how they change with space weather are essential for better forecasting of the radiation belts.
The wave database was used to determine the global distribution of chorus as a function of geomagnetic activity, magnetic local time, latitude and radial extent. The waves were split into an upper and lower frequency band as these waves had very different distributions. Figure 4 (middle panel) shows that in the equatorial region lower band chorus was most intense between 23:00 and 12:00 magnetic local time (MLT). In contrast, upper band chorus (right panel) is weaker and less extensive and is found between 00:00 and 11:00 MLT. The distributions at higher latitudes were somewhat different. At mid-latitudes lower-band chorus is restricted to the dayside and no significant upper band chorus wave power is observed at mid to high latitudes [Meredith et al., 2012]. While the magnetic local time distribution of the waves is reasonably well understood it is still not clear why there is such a difference between waves observed near the equator and higher latitudes, and what controls the radial distribution of the waves.


Figure 4. Average wave intensities for chorus waves around the Earth. The Earth is in the centre of each plot and midday is at the top. Low frequency waves (left) are most intense near noon whereas lower band waves (middle) are most intense from midnight through dawn to noon. The inset panels show the number of samples.
CHORUS DIFFUSION RATES
The results of the wave study were used to calculate new chorus diffusion matrix for use in the global radiation belt models. In general the frequency of maximum wave power was lower than that used in previous models and as a result energy diffusion was found to extend to several MeV at large pitch angles. The results indicated losses at low energies and net electron acceleration at high energies and large pitch angles. It was also found that diffusion due to upper band chorus is restricted to L* < 6 whereas that due to lower band chorus was significant even at L* = 8. The diffusion matrix includes more levels of geomagnetic activity, magnetic local time resolution and radial coverage than previous models and has enabled a much better evaluation of electron acceleration and loss rates [Horne et al., 2013a]. The matrix was incorporated into the forecasting model in 2014.
LOW FREQUENCY CHORUS
During the analysis of the wave database it was noticed that there is substantial wave power at much lower frequencies which has not received much attention. Lower frequency waves are potentially important since they could affect the highest energy electrons. Therefore the wave database was used to construct a new model of low frequency chorus in the frequency range between the lower hybrid resonance frequency and one tenth of the electron gyrofrequency [Meredith et al., 2014a]. These waves were strongest at mid-latitudes in the pre-noon sector (Figure 4, left panel). These waves are not currently included in radiation belt models, but the observed wave power suggests that they could be important for electron acceleration and loss and should be taken into account in radiation belt models.
PLASMASPHERIC HISS AND LIGHTNING-GENERATED WHISTLERS
Other types of plasma waves contribute to electron scattering and loss from the radiation belts into the atmosphere within about 25,000 km of the Earth. These waves are known as plasmaspheric hiss and lightning generated whistlers. Lightning generated whistlers are radio waves originating from lightning which leak out of the atmosphere and into space and in a frequency range from a few hundred Hz to a few kHz. Again the global distribution of these waves and how they change with space weather are essential for better forecasting of the radiation belts.
The wave database was used to determine the global distribution of plasmaspheric hiss and lightning generated whistlers as a function of geomagnetic activity. Figure 5 shows the average wave intensity for plasmaspheric hiss (bottom panels) and lightning-generated whistlers (top panels) for, from left to right, quiet, moderate and active conditions. Plasmaspheric hiss peaks largely on the day-side and afternoon (bottom panels) in the region 2 < L* < 4. This suggests that electron loss from the radiation belts into the atmosphere should be most important in this region.
Waves in the frequency band 2 < f < 4 kHz, (top panels) are attributed to lightning generated whistlers. On the dayside the intensity increases with increasing geomagnetic activity (left to right). However, at night the intensity is almost independent of geomagnetic activity. At night absorption by the ionosphere is weaker and so these waves are associated with signals originating from lightning. However, waves on the dayside as associated with geomagnetic activity which suggests that they are related to substorms and the injection of medium energy electrons.

Figure 5. Average wave intensities for plasmaspheric hiss (bottom) and lightning generated whistlers (top). The Earth is at the centre and noon is at the top. Quiet, moderate and active conditions are shown left to right. The inset panels show the number of samples.
CHORUS AS A SOURCE OF PLASMASPHERIC HISS
Recently it has been proposed that chorus waves which are observed further out in the magnetosphere could be the source of plasmaspheric hiss that is observed much closer to the Earth. The idea is that chorus waves propagate along the magnetic field and then gain entry to the lower region at relatively high latitudes and then become trapped in the inner region. Using the wave database we were able to show that chorus does extend along the magnetic field to high latitudes in the pre-noon sector, and, that in the equatorial region, there is a clear gap of 1-2 Earth radii between plasmaspheric hiss at L* < 4 and chorus further out. The results of this analysis are consistent with other studies looking at the propagation of these waves. The observations confirmed two of the key predictions of the new theory and provided the first statistical evidence for chorus as the embryonic source of plasmaspheric hiss [Meredith et al., 2013].
EMIC WAVES
Theory suggests that electromagnetic ion cyclotron (EMIC) could cause major losses of radiation belt electrons during space weather events. However, there is very little data on these waves as only a few satellites carry the instruments needed to detect them. They are usually observed at frequencies of a few Hz, in space and on the ground. To test the theory a new EMIC wave database was constructed using data from the CRRES satellite. The data are very sparse, but it was found that the wave power generally increased with geomagnetic activity in the afternoon sector with an average percentage occurrence of 2.6% [Meredith et al., 2014b]. The average spectral properties of the waves were used to determine the effects on scattering high energy electrons. It was found that the waves can cause losses to the atmosphere of electrons at very high energies, typically greater than 2 MeV, but not below. The work showed there should be a special type of signature in the distribution of the electrons left behind in space, namely that the distribution should become more peaked towards 90o pitch angles as energy increases from 6 to 10 MeV. In addition, it was found that these waves could not be responsible for the complete reduction or ‘drop-out’ that is sometimes observed in the radiation belts since they cannot remove electrons at very large pitch angles.

RADIAL DIFFUSION
One of the most important processes that contribute to the increase or decrease of particles in the radiation belts is the transport of particles across the magnetic field. Usually high energy charged particles gyrate around the magnetic field and it is very difficult to transport them across the magnetic field. However, fluctuations in the magnetic and electric fields can cause a diffusion of particles across the magnetic field, and this process is known as radial diffusion. The diffusion is not driven by collisions between particles as these are very rare, but by the fluctuating fields.
Radial diffusion is an essential process that is included in the forecasting models. They key to improving the models is to measure the diffusion coefficients but this is exceptionally difficult. For example, the magnetic field can be measured on the ground but these measurements have to be mapped into space leading to large uncertainties. On the other hand satellites can only make point measurements in a vast area of space. In the spacecraft project a new and innovative approach was adopted. Theoretical work shows that if the Earth’s magnetic field is separated into a part that is symmetric about the Earth and an asymmetric part then only the asymmetric part can drive the diffusion [Lejosne et al., 2012]. By using seven years of satellite measurements from the GOES satellites at geosynchronous orbit the symmetric and asymmetric components were identified and new diffusion coefficients were calculated. These calculations were performed for different levels of geomagnetic activity and for different particle energies. They show that the diffusion rates at energies less of a few hundred keV are higher than that used in current models, and at energies of several MeV they are lower [Lejosne et al., 2013]. These results are some of the best measurements to date and have been incorporated into the forecasting models.
INTERPLANETARY SHOCK ACCELERATION
Solar energetic particle events cause some of the highest proton radiation levels for satellites in Earth orbit and interplanetary missions. Some of the most intense events are caused when coronal mass ejections create a shock wave as they travel out into space. At the shock there are turbulent electromagnetic fields which accelerate protons to very high energies allowing them reach the Earth in tens of minutes. By modeling these events and understanding the acceleration and transport processes it is hoped to provide some measure of forecasting the intensity and duration of solar energetic particle events in future.
Current models of interplanetary shock waves are limited. In one class of models, known as MHD models, turbulent electromagnetic fields cannot be resolved but have to be represented by an empirical term leading to a lot of uncertainty. On the other hand kinetic plasma simulations can include the turbulence self-consistently but they require such large amounts of computer time they cannot be used for forecasting. In the SPACECAST project we have developed a semi-analytical model that describes particle scattering in the foreshock region, which has the potential of being applied in real time [Vainio et al., 2014].
The model calculates the energetic proton spectrum at various distances ahead of the travelling shock wave and the spectrum of turbulent fluctuations in the foreshock region ahead of the shock wave [Vainio et al., 2014]. The model is based on the theory of diffusive shock acceleration whereby the particles gain energy from the turbulent fields as they are scattered across the shock many times, but modified to take into account time dependence and a diverging magnetic geometry in space. The model was tested against fully self consistent simulations to calibrate the results and showed that theoretical models over-estimate the upper energy limit at the shock by up to two orders of magnitude. The model also shows that the cut-off energy in the proton spectrum observed at the Earth can be used to obtain information on the resulting energetic storm event even when the shock is very close to the Sun and inaccessible to in-situ observations. The results can be used for SW modeling for future spacecraft missions such as Solar Orbiter and Solar Probe Plus as well as developing acceleration models for SEP events.
MODELLING LARGE SOLAR PROTON EVENTS
One of the large uncertainties in modelling solar energetic particle events is to understand how the shock wave evolves as it travels out into space and distorts the interplanetary magnetic field. The evolution affects how the Earth is connected to the shock via the interplanetary magnetic field which in turn affects the particle flux. To model these transport effects a two-dimensional solar wind model was developed using magnetohydrodynamic (MHD) theory. Figure 3 shows an example where the shock has travelled outwards from the Sun and shows how different locations at Earth orbit are connected to different locations on the shock front (the cob point) via the magnetic field (grey lines).

Figure 4. A shock wave travelling away from the Sun. Different parts of the shock front are connected via the magnetic field (grey lines) to different parts of the orbit of the Earth.
In contrast to previous models, the acceleration of the solar wind was included by assuming a source of heating in the solar corona. A new automated method to identify the cob point in the simulations was developed in order to extract plasma conditions at the cob point as the shock moves. Two simulations were performed for solar energetic particle events on 4th April 2000 and 13the December 2006. The simulations showed that the jump in the speed across the shock (VR) was very similar in these two events, and similar to other previously modeled events. By using a separate transport model to fit the particle intensity observed at the Earth as a function of time for the two solar proton events it was possible to obtain the jump in the shock speed and the particle injection rate at the shock. The key result is that a simple relation between the shock speed ratio VR and the particle injection rate Q could be established. The significance of this result is that it enables software tools to be developed to predict the peak intensity, fluence and proton-intensity time profiles of SEP events that cause damage to spacecraft.

DISSEMINATION AND STAKEHOLDERS ENGAGEMENT
Members of the SPACECAST team disseminated results widely via a number of channels. Team members provided briefings for policy makers, a presentation at the UK House of Commons for NATO MPs, and provided oral and written evidence for the UK Government on risk. They also appeared on the BBC2 program “Stargazing Live - Back to Earth” in 2014, gave an interview for a Swedish TV documentary in 2013, a radio interview on the BBC World Service in 2012, a podcast, 3 press releases, and numerous newspaper and magazine articles. One press release in 2012 on the forecasting system led to 46 different press reports world-wide. Team members published 30 peer reviewed papers and gave numerous presentations at international conferences.
The SPACECAST project made a particular effort to engage with stakeholders in the satellite industry. Each year (2012-2014) the SPACECAST project held a special plenary session at the European Space Weather Week (ESWW) where commercial operators and agency representatives were invited to speak. A specially held business lunch-time discussion was held each year and organized jointly with the European Space Agency. Approximately 30 stakeholders from satellite design and construction, operators, space insurance and from Government agencies attended as well as scientists and members of the ESA. The lunch-time discussion was led alternately by SPACECAST and by ESA each year and covered all aspects of space weather as it affects satellites. Reports of these meetings were written jointly with the ESA and circulated to all stakeholders. Following feedback from these meetings the SPACECAST web site was re-organised to better meet the user’s needs.
A close-out meeting was held in Cambridge UK in February 2014 specifically for stakeholders. This also included representatives from the USAF and the World Meteorological Organisation. One of the stakeholders subsequently wrote “The web site is an order of magnitude (maybe several) more sophisticated than the NOAA SEC one”. All the stakeholders agreed that they wanted to continue meeting each year with the SPACECAST team as they found the work and the dialogue very useful. As SPACECAST has come to and end it was agreed to take this forward and meet each year at ESWW as part of the new EU SPACESTORM project which runs from March 2014 – March 2017 and where the most important elements of SPACECAST will continue and will be expanded to include work on extreme space weather events.

Potential Impact:
The SPACECAST project came to an end on the 28th February 2014. The project will have significant impact in four distinct areas:
1. Significantly contribute to the European capacity to prevent damage/protect space assets from space weather events
The space weather forecasting models developed in the SPACECAST project now provide Europe with a state-of-the-art forecasting capability of impending, disruptive, space weather events in the electron radiation belts, and have significantly improved engineering tools for SEP event modelling. The project now delivers a near real time forecasting and nowcasting capability for Europe and has significantly contributed to European capacity by developing the expertise and models of the European partners. Early warnings and alerts are now provided for stakeholders as a result of these forecasts. These warnings enable operators to take action to mitigate the effects of disruptive events to aerospace vehicles, satellites, and other vulnerable technologies in space and on the ground.
Forecasts from the SPACECAST models on the electron radiation belts are now accessed by the UK Met Office for their space weather forecasting.
The DOSETOOL provides the space weather user community with a new tool to calculate radiation dose and dose rates during SEP events, as demonstrated in the on-line plots already available in the Spacecast web server. The hourly dose rate estimate has been added to the spacecraft risk quantities on the web site. Used in conjunction with the synthetic SEP events data of the SOLPENCO tool, DOSETOOL can provide radiation doses and dose rates for different SEP events simultaneously detected at 0.4 AU and at 1.0 AU; i.e. it gives the possibility to easily obtain radiation doses in the inner heliosphere, including < 5 MeV proton population not considered in SOLPENCO2.
The Q(VR) relation permits the development of software tools, like SOLPENCO and SOLPENCO2, that are necessary for the prediction of peak intensities, fluences and proton intensity-time profiles of SEP events for energies that are relevant to space weather; that is, energies at which protons are capable of traversing typical shielding conditions, E > 10 MeV.
2. Significantly contribute to both improving forecasts and predictions of disruptive space weather events
The SPACECAST project has delivered basic research in scientific peer reviewed journals that has increased our understanding of the basic physical processes that control radiation belt dynamics and solar energetic particle events. These results have been used to update the forecasting models and improve forecasts of disruptive space weather events.
3. Identify best practices to limit the impacts on space- and ground-based infrastructures and their data provision.
Members of the SPACECAST project have worked closely with stakeholders by holding Stakeholder Discussion meetings at the annual European Space Weather Week Workshops, attending other user related meetings, and by including a stakeholder on the Steering Committee. This has enabled the consortium participants to better understand the problems encountered with satellites and other space assets. Two way discussions between the SPACECAST project team and Stakeholders have enabled greater awareness on the part of European companies and Government to assess impact of space weather events. For example, information from SPACECAST contributed directly to an assessment by the UK Government which led to Space Weather being included on the UK national risk register.
4. Facilitate and promote an open exchange of information on incidents potentially caused by space weather events between different European (and international) actors in the field.
Space weather forecasts provided by the models developed in the SPACECAST project are now available via the SPACECAST web site and provided to the ESA Space Weather Environment Network (SWENET) web portal for wider dissemination and the UK Meteorological Office. A historical record of space weather forecasts together with the relevant satellite data are now archived on the SPACECAST web site and are openly available. These data will be available to help identify the cause of incidents potentially caused by space weather events and hence facilitate the open exchange of information. Some of the SPACECAST models are now being coupled to other space weather models as part of the work of the ESA Virtual Space Weather Modelling Centre.

List of Websites:
http://fp7-spacecast.eu/