Community Research and Development Information Service - CORDIS


FLARECAST Report Summary

Project ID: 640216
Funded under: H2020-EU.2.1.6.

Periodic Reporting for period 1 - FLARECAST (Flare Likelihood and Region Eruption Forecasting)

Reporting period: 2015-01-01 to 2015-12-31

Summary of the context and overall objectives of the project

"TITLE: FLARECAST – Flare Likelihood and Region Eruption Forecasting
FUNDING AGENCY: EC Horizon 2020, Research and Innovation Action, Protec-1-2014: Space Weather
GRANT AMOUNT: EUR 2,400 000.00
DURATION: 3 years

START DATE: 1 January 2015
CONSORTIUM: 8 institutions from 6 countries
PROJECT COORDINATOR: Dr. Manolis K. Georgoulis, Academy of Athens, Greece
PROJECT SCIENTIST: Dr. D. Shaun Bloomfield, Trinity College Dublin, Ireland

KEY WORDS: Flare Forecasting, Space Weather, Heliophysics, Prediction Algorithms, Machine Learning

Space weather can have detrimental effects on astronaut safety and a multitude of technologies on which we rely on a daily basis. Accurate and reliable space-weather monitoring and forecasting helps those affected, such as satellite operators, to take timely impact-mitigation measures. The main agents of space weather being solar flares and coronal mass ejections, FLARECAST will significantly advance our ability to predict flares prior to their occurrence in the Sun.

FLARECAST will first aim to understand the drivers of solar-flare activity to improve flare prediction. It will then aim to provide a globally and openly accessible flare prediction service that facilitates evolution and expansion. Finally, FLARECAST will aim to engage in a dialogue with space-weather stakeholders, policy makers, and the public on the societal benefits of a reliable solar-flare prediction.

Diverse properties of solar active regions, the "hotspots" of large flares, will be treated as flare predictors. These properties will be extracted using advanced image-processing techniques applied to remote-sensing solar observations. Various flare prediction algorithms, some of them featuring artificial intelligence, will highlight a statistically rigorous set of the most promising of these predictors. Flare-forecast probabilities will be benchmarked by means of state-of-the-art validation techniques and will be utilized to launch a fully automated, near real-time flare forecasting service, where the end user will be allowed to perform his/her own tests and produce validated results.

FLARECAST will push the envelope of current understanding of solar active-region properties and their relation to flaring activity. In parallel, a functionally-expandable infrastructure will accommodate flare predictions allowing a simple, but suitably verified, transition of scientific research into an operational space-weather application. The resulting user-friendly, interactive facility, the first of its kind in the world, will be freely accessible to researchers and operators in Europe and around the globe.

AA: Academy of Athens Greece
TCD: Trinity College Dublin Ireland
UNIGE: Universita Degli Studi Di Genova Italy
CNR: Consiglio Nazionale Delle Ricerche Italy
CNRS: Centre National de la Recherche Scientifique France
UPSud: Université Paris-Sud France
FHNW: Fachhochschule Nordwestschweiz Switzerland
MO: Met Office United Kingdom"

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far



DB : Database
EPO : Education and Public Outreach
GOES : Geostationary Operational Environmental Satellites
HARP : HMI Active Region Patch
HESPE : High Energy Solar Physics Data in Europe
HMI : Helioseismic and Magnetic Imager
LOS : Line of Sight
NOAA : National Oceanic and Atmospheric Administration
SDO : Solar Dynamics Observatory
SHARP : Space Weather HMI Active Region Patch
SMART : SolarMonitor Active Region Tracking
SWPC : Space Weather Prediction Center
WP : Work Package


1. To understand the drivers of flare activity and improve flare prediction. FLARECAST aims to consolidate and advance our understanding of solar active region properties (sunspot structure classifications, magnetic polarity-mixing classifications and physical quantities derived from line-of-sight and vector magnetograms) and their evolution that lead to the production of solar flares. Both state-of-the-art and newly developed prediction algorithms will exploit this enhanced knowledge to identify the best combinations of flare predictors and techniques to be used by different space-weather end users.

2. To provide a globally accessible flare prediction service that facilitates expansion. FLARECAST aims to implement the diverse range of techniques that exist for predicting flares and develop novel techniques, thereby enhancing our capacity to deliver flare forecasts. To do this, FLARECAST mobilizes among the best expertise in Europe on flare prediction, solar data processing and archiving, web-service and interface development, forecast verification and 24/7 operational space-weather forecasting. The final system will feature an algorithm “plug-in” infrastructure framework feeding a data processing cloud that will allow users in research and industry to expand and exploit the available prediction services.

3. To engage with space-weather end users and inform policy makers and the public. FLARECAST aims to deliver the first operationally verified, near-realtime predictions of solar flares to space-weather forecasters, the scientific and industrial sectors and the general public. In parallel with delivering forecasts, FLARECAST engages with the general public, science educators and policy makers to promote the relevance of space weather and its impacts on our day-to-day lives.


The first objective involves WPs 2, 3, 5, and 6.

WP2 (solar active region properties as predictors of flare activity) was fully deployed in Year 1: about 25 different active-region properties, besides properties provided automatically by the SMART utility and besides the NOAA / SWPC- and GOES-provided information are currently under development, reproduction or testing (Table 1). These represent a vast part of the predictors that have been proposed in the literature over the past twenty years or so. A few are still to be fully tested, with this task continuing into Year 2. In addition, consortium members have contacted all developers of original flare predictors to determine whether they are willing to make their codes available for the project. Most of them have already responded positively, which leaves only a small set of proposed predictors to be reproduced by the project from scratch. WP2 has provided one (1) Deliverable report in Year 1, with the second one due in the beginning of Year 2 (after Month 12). Besides consortium staff members, three (3) postdoctoral researchers (two [2] in TCD; one [1] in AA partners) were employed and worked on WP2 in Year 1. The first WP2 milestone (completion of the property database) will be achieved in Year 2. 

WP3 (flare prediction algorithms) has seen consistent work in Year 1 in the areas of data handling and standard machine-learning algorithms. In the first, the definition of a paradigm for the handling of input data was pursued, while in the second the initial versions of project-tailored software tools aimed to implement standard machine-learning algorithms were outlined. This latter task formed the core of Deliverable D3.1, that was due after Month 12 so it has been submitted early in Year 2. Before completing Deliverable 3.1, Milestone 8 (report on data handling for standard machine learning algorithms) was completed and uploaded in the project's internal collaboration platform late in Year 1 (Month 10). To enhance parallel work between WPs, existing, readily available predictors from WP2 were used as input for machine-learning algorithms in WP3, at the same time working to achieve a data handling process modular enough to accept new predictors (out of the dozens in progress in WP2) to be readily tested as they become available. Besides consortium staff members, three (3) postdoctoral researchers (one [1] in CNR; one [1] in UNIGE; one [1] in AA partners) were employed and worked on WP3 in Year 1. 

WP5 (data and forecast validation) kicked in late in Year 1 (Month 9), as foreseen in the project's timeline. Work in Year 1 was focused on the algorithmic implementation of standard binary (YES / NO) verification metrics, as well as on the construction of the probabilistic reliability diagram that helps assess errors and uncertainties in probabilistic forecasting. These tools are established in the literature and over existing forecasting services worldwide and are borrowed by the performance verification practices of terrestrial-weather forecasting. Besides consortium staff members from the TCD, AA and MO partners, one of the TCD postdoctoral researchers was also involved in this WP. There were no Deliverables or Milestones foreseen for WP5 in Year 1.
WP6 (explorative research) will initiate in Month 15 (Year 2). 


The second objective involves WP4.

WP4 (data storage and processing cloud) achieved in Year 1 the installation and mounting of the project's hardware, software, and collaboration platform, at the same time initiating the process of downloading data from the
magnetogram and continuum-image archives of the HMI telescope, onboard the SDO mission. The single most concrete achievement in this WP was the establishment of the project's architecture that initially drew from the EU/FP7 HESPE project but evolved and matured well enough to be called the FLARECAST Architecture (Figure 1). This also provided the core of Deliverable D4.1 that was delivered late in Year 1, after Month 9. The FLARECAST Architecture consists of four (4) steps (data acquisition; feature property extraction; machine learning; data verification) and four (4) independent DBs, namely HMI files DB, infrastructure configuration DB, active-region property DB, and prediction DB. The different components were accommodated via Docker containers that are fully accessible at Dockerhub. 

WP4 also oversaw the procurement of the project's computation server, its connection to the computation cluster of the UPSud partner and the establishment of a total data storage of 360 TB  in RAID-like architecture (thus implying an effective storage capacity of 247 TB). At the end of Year 1 the entire HMI near-realtime LOS magnetogram database, as well as the database of continuum images, are available in the archives of the UPSud partner.

Meanwhile, downloading of the HMI near-realtime SHARP and definitive, science-grade HARP magnetograms is progressing into Year 2, as well. 

The internal collaboration platform of the project, based on the open-source Atlassian suite of tools, was established and is in seamless operation since early in Year 1. This is used for the quick exchange of project files, notes, and thoughts between consortium members. At the same time, a Wordpress web site was established to accommodate the open, publically available part of the project. Besides consortium staff members from the FHNW, UNIGE, and UPSud partners, two more expert developers were involved in this WP from FHNW. No milestones were foreseen for WP4 in Year 1.


The third objective involves WP7. 

WP7 (dissemination) in Year 1 focused on a three-pronged dissemination effort directed toward the general public, the industry and government, and the scientific community. The first effort was undertaken primarily via the project's web site, established during the initial phases of the project (Month 2), and secondarily via interview, public seminars and events delivered by several consortium members. All these appearances have been outlined in multiple locations in the project web site. In addition, appropriate EPO material was created to help convey the project's public message (Figure 2).

The second effort was undertaken primarily by the MO partner that was involved in several space-weather dissemination activities where its involvement in the project was cited. This resulted also in an information article that was published in the UK solar physics newsletter. Most notably, in Year 1 the MO partner initiated a user survey that resulted in a questionnaire directed toward potential attendees of the first Stakeholders workshop that is planned to take place at MO toward the end of Year 2.

The third effort was undertaken via a robust participation to science meetings and workshops, where the basic elements of the project were exposed to the international scientific community. As the project requires significant preparatory work upfront, before concrete results start appearing, there were no refereed publications in Year 1. Year 2 should see some publications, most notably one in which the different flare predictors published over the past two decades will be encapsulated in a review paper on solar flare forecasting. 

There was one (1) Deliverable for Year 1 in WP7, that was delivered shortly after Month 6, according to the project's timeline. Two (2) more Deliverables referring to activities in Year 1 will be delivered early in Year 2 (one of them already delivered and the other to be delivered shortly). There are no Milestones on WP7 over Year 1: a milestone referring to the education and public outreach (EPO) material of the project's web site, accumulated over Year 1, was reported achieved in early Year 2. Consortium staff members from all partners worked on the actions presented here, with the EPO part led by the FHNW partner, the industry-and-government part led by the MO partner and the scientific-community part led by the AA and TCD partners.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

FLARECAST will strive to push the envelope of current understanding of solar active-region properties and their relation to flaring activity. This goal will be achieved by the most systematic effort ever undertaken to assemble all proposed flare predictors worldwide, evaluate them robustly on common, established criteria with respect to their predictive ability and determine which parameters, or combinations thereof, are the most competent predictors. In parallel, a functionally-expandable infrastructure will accommodate predictions, thus featuring a simple, but suitably verified, transition of scientific research into an operational space-weather application.

The project is designed to help protect humanity's assets in space, in particular these assets that we have grown increasingly dependent on in recent years.

The resulting interactive, customizable, flare prediction facility will be the first of its kind in the world. The facility will be freely accessible to all end users, from researchers to operators to industry and government professionals, in Europe and around the globe.

Related information

Record Number: 185364 / Last updated on: 2016-06-28