Figure 1: A typical X-class solar flare observed by TRACE in September 2005.
Solar flares (SFs) are sudden, violent and very energetic explosions occurring in active regions (ARs) around sunspots. They are powered by sudden large changes of the local magnetic field topology through reconnection processes that lead to huge releases of magnetic energy. During SFs, magnetic energy of 1028–1034 erg is released in the solar chromosphere and corona over a few minutes[Benz, 2008]. This energy release in turn leads to plasma heating, particle acceleration and mass transport and produces electromagnetic radiation across the electromagnetic spectrum at all wavelengths from long wave radio to the shortest wavelength gamma rays [Anastasiadis, 2002]. During a large solar flare, the X-ray flux increases by many orders of magnitude compared to the pre-flare X-ray levels.
Coronal Mass Ejections (CMEs)
Figure 2: The massive CME of 23 July 2012 as recorded by STEREO-B (from http://svs.gsfc.nasa.gov/cgi-bin/details.cgi?aid=4178).
Coronal mass ejections (CMEs) are large-scale solar eruptions during which mass and energy are being released from the Sun to the interplanetary (IP) medium. CMEs are being regularly monitored by white-light coronagraphs at 1 AU onboard the SOHO [Domingo, Fleck, and Poland, 1995] and the Solar TErrestrial RElations Observatory (STEREO) missions [Kaiser et al., 2008]. Most of the ejected material comes from the low corona, although cooler, denser material probably of chromospheric or photospheric origin is also sometimes involved. The CME plasma is entrained on an expanding magnetic field, which commonly has the form of helical field lines with changing pitch angles, i.e., a flux rope [Webb and Howard, 2012].
Solar Energetic Particle (SEP) Events
Solar Energetic Particles (SEPs) are observed as flux increases above a background level, and their energies range from ~ 10keV to 10GeV/nuclei, while their duration can last from a few hours to several days. SEPs consist of electrons, protons, alpha particles and heavier ions up to Fe [Reames, 2012]. The classical paradigm is to divide SEP events into two categories: the impulsive and the gradual ones based on their parent solar events [Reames, 1999]. Impulsive SEP events are considered to be associated to SFs, while the gradual ones are considered to be accelerated by CME-driven shocks.
Figure 3: An illustration of the classical SEP classification. Intensity-time profiles of electrons and protons in ‘pure’ (a) impulsive and (b) gradual SEP events. Panels on the left presents a series of impulsive SEP events and their generator: the solar flare, while panels on the right present a series of the gradual SEP events and their causative, CMEs (taken by Reames, 2012).
SEP events are produced in the solar atmosphere by particle acceleration processes in association with powerful M and X class flares and in interplanetary shocks created by the interaction of CMEs with the solar wind. Accelerated particles then propagate through the interplanetary medium, spiralling along the interplanetary magnetic field (IMF). Figure 4 presents cartoons of the two possible acceleration mechanisms. In the left panel, the flare acceleration scenario is presented. It implies that SEPs are accelerated together with the particles radiating γ rays (γR), Hard X-rays (HXR), radio and Soft X-rays (SXRs), escaping to open flux tubes towards the IP space. In the right panel, a CME, its interplanetary counterpart (ICME) and the CME-driven shock are presented. SEP events are thought to be accelerated at the CME-driven shock of relatively fast CMEs.
Figure 4: Acceleration processes for SEP events. The 'flare acceleration' is presented in the left panel, whereas the 'CME acceleration' in the right panel (top row). Actual solar storms (SEP events) generated by SFs and CMEs, respectively (bottom row) [Images courtesy of the NASA/ESA SOHO and the NASA/SDO missions].
The FORecasting Solar Energetic Particles and Flares (FORSPEF) tool provides forecasting of solar eruptive events, such as solar flares (SF) with a projection to coronal mass ejections (CMEs) (occurrence and velocity) and the likelihood of occurrence of a solar energetic proton (SEP) event. It also provides nowcasting of SEP events based on actual SF and CME near real-time alerts, as well as SEP characteristics (peak flux, fluence, rise time, duration) per parent solar event. A more detailed describtion on FORSPEF Tool can be found in Papaioannou et al. 2015.
FORSPEF Forecasting mode
Within FORSPEF's forecasting mode, first the SF forecasting takes place, utilizing LOS SDO/HMI full-disk magnetograms, providing 24-hour conditional flare probabilities and eruptive flare likelihoods that can be refreshed at intervals of 3 hours. To achieve this, it calculates values of a prediction parameter of choice, namely Beff, “effective connected magnetic field strength” [Georgoulis and Rust, 2007]. Besides eruptive flare likelihoods it independently provides also projected near-Sun CME speeds and assessments of proximity to Coronal Holes (CHs). All this information is provided per Active-Region Identification Algorithm (ARIA)-identified AR. Next, the detection of SEP events, in terms of probability of occurrence, is performed based on the derived probability for solar flare occurrence (forecasting pre-flare scheme) for a given AR, made available in the previous step. In the forecasting pre-flare emergence scheme, for a given location of an AR, the information stored in a newly constructed SEP database, hereafter FORSPEF database [Papaioannou et al., 2016] is used to derive the Distribution Functions (DFs) of the flare magnitudes associated and/or not associated to SEPs. This corresponds to the local statistical model that is being used for the construction of the DFs. These latter DFs are folded with the solar forecast leading to a weighted forecast [Sandberg et al., 2014]. Using this scheme, statistical outputs (e.g. percentiles) for the SEP occurrence are derived. In what follows, an actual example of the FORSPEF's forecasting scheme is presented. For a given AR (in this case AR12157), the probabilities of flare and CME occurrence are first derived, (right top panel of Figure 5). Results stemming from the FORSPEF database are presented in the bottom left panel of Figure 5. The abscissa corresponds to the flare peak photon flux and the ordinate corresponds to the derived probabilities of SEP occurrence. The probabilities derived by the Solar Flare Prediction are folded with the probabilities derived by the FORSPEF database. This is illustrated at the bottom right panel of Figure 5. FORSPEF provides the maximum probability of SEP occurrence from that bottom right panel of Figure 5.
Figure 5: An actual illustration of FORSPEF's forecasting work scheme (taken by Papaioannou et al., 2015)
FORSPEF Nowcasting mode
Detection of SEP events, in terms of probability of occurrence, is performed based on the actual near real-time solar flare data (i.e. maximum solar flare magnitude and position) downloaded in near-real time mode from Solar Demon (http://solardemon.oma.be/flares.php?) a service operated by the Royal Observatory of Belgium (ROB). A typical nowcast (post-flare) warning window is of the order of ~15-20 minutes, while nowcasts are renewed with the identification of any new solar flare event. The SEP nowcast provides the probability of SEP occurrence for a given SF as well as the expected SEP characteristics, calculated based on sub-samples of the FORSPEF database that occur under the pre-requisite of the nearest (in longitude) 4000 historical solar flare events (in case of only SF information made available) and/or 1000 historical flare events accompanied by CME ((in case that the CME occurrence information is made available). The output of the SEP nowcasting (post-flare) scheme is presented in Figure 6. Left panel presents the probability of SEP occurrence using only the SF information, while the right panel presents the same probability in the case that a CME has occurred [Anastasiadis et al., 2015]. Furthermore, nowcasting is also performed based on actual near real-time CME data (i.e. velocity and width) from CACTus, another service from ROB. For each, identified CME, proper fits per CME width (e.g. halo, partial halo, non halo) and velocity, established based on the FORSPEF database are being employed.
Figure 6: Derived probabilities of SEP occurrence from FORSPEF's nowcasting work scheme under the presence (or absence) of a CME (taken by Anastasiadis et al., 2015).