SEPRAD - Workshop


Prediction of the Occurrence and Intensity of SEP Events with Energies >100 MeV and >500 MeV Using the UMASEP Scheme


Marlon Núñez
Universidad de Málaga, Málaga, Spain


Extreme solar events may accelerate solar particles to near the speed of light reaching the Earth in a few minutes. These particles may interact with the Earth’s atmosphere to produce penetrating neutrons known as Ground Level Enhancements (GLEs) which may irradiate passengers and flight crews in commercial aircraft flying at extreme polar latitudes [1].

This presentation summarizes the model and results of two tools that are able to predict the occurrence
and intensity of the first hours of Solar Energetic Proton (SEP) events with energies >100 and >500 MeV.

These tools use the UMASEP forecasting scheme, which infers a magnetic connection, along which energetic protons are arriving in the near-Earth environment, by estimating a lag-correlation between solar soft X-ray (SXR) flux and differential proton fluxes at near-Earth; if this correlation is high and the associated solar flare is also strong, then this scheme issues a SEP event prediction.

This forecasting scheme has been successfully tested for predicting >10 MeV SEP events (using the Low-Energy UMASEP model) in an operational level since 2010, when NASA’s integrated Space Weather Analysis system (iSWA) and the European Space Weather Portal started redistributing UMASEP forecasts. Since then, the developed tool (UMASEP-10) has predicted 83% of real SEP events, with a false alarm ratio of 25%.

The low-energy UMASEP model was the basis for the design of the high-energy UMASEP model, which was proposed for predicting >100 MeV events from 5-min GOES SXR and proton data. Recently, this model was also used to build the tool HESPERIA UMASEP-500 [4] for predicting >500 MeV SEP events (i.e. whose integral proton flux surpasses 0.8 pfu) from 1-min GOES SXR and proton data. The same predictions may be used for warning against GLE events before the detection by any neutron monitor station.

The performance of this tool with historical data for the period 2000-2016 may be summarized as follows: the Probability of Detection (POD) was 58% for the case of >500 MeV SEP events and 50% for the case of GLE events; the False Alarm Ratio (FAR) was 30% for both types of events; and, the average warning time was 16 min with respect to the GOES detection times of >500 MeV SEP events, and 15 mins with respect to the GLE Alert Plus [5] warning times.

Regarding the prediction of the intensity peak of >500 MeV SEP events, this tool obtained a mean absolute error of 0.73 in log10 of pfu units. In [4] authors speculate that UMASEP-500‘s high-energy predictions can possibly benefit from a >10 MeV electron contamination in high-energy proton detectors; for this reason, we conclude that an interesting research field is the correlation of SXR and relativistic electron data for predicting high-energy SEP events.


[1] Shea M.A. & Smart D.F. (2012), Space Science Reviews, 171.
[2] Núñez, M. (2015), Space Weather, 13, 11.
[3] Núñez, M. (2011), Space Weather, 9, S07003.
[4] Núñez, M., Reyes-Santiago P.J. & Malandraki O. (2017), Space Weather, 15.
[5] Souvatzoglou et al. (2014), Space Weather, 12, 11.


The development of the UMASEP-500 tool has received funding from the European Union’s Horizon 2020 research and innovation programme under agreement No 637324. The development of the UMASEP-100 tool has received funding from the Plan Propio de Investigación of the University of Malaga (Spain).