Solar Flare Prediction Using Advanced Feature Extraction, Machine Learning, and Feature Selection

Novel machine-learning and feature-selection algorithms have been developed to study: i) the flare-prediction-capability of magnetic feature (MF) properties generated by the recently developed Solar Monitor Active Region Tracker (SMART); ii) SMART’s MF properties that are most significantly related to flare occurrence. Spatiotemporal association algorithms are developed to associate MFs with flares from April 1996 to December 2010 in order to differentiate flaring and non-flaring MFs and enable the application of machine-learning and feature-selection algorithms. A machine-learning algorithm is applied to the associated datasets to determine the flare-prediction-capability of all 21 SMART MF properties. The prediction performance is assessed using standard forecast-verification measures and compared with the prediction measures of one of the standard technologies for flare-prediction that is also based on machine-learning: Automated Solar Activity Prediction (ASAP). The comparison shows that the combination of SMART MFs with machine-learning has the potential to achieve more accurate flare-prediction than ASAP. Feature-selection algorithms are then applied to determine the MF properties that are most related to flare occurrence. It is found that a reduced set of six MF properties can achieve a similar degree of prediction accuracy as the full set of 21 SMART MF properties.     

A Multi-wavelength Analysis of Active Regions and Sunspots by Comparison of Automatic Detection Algorithms

Since the Solar Dynamics Observatory (SDO) began recording ≈?1 TB of data per day, there has been an increased need to automatically extract features and events for further analysis. Here we compare the overall detection performance, correlations between extracted properties, and usability for feature tracking of four solar feature-detection algorithms: the Solar Monitor Active Region Tracker (SMART) detects active regions in line-of-sight magnetograms; the Automated Solar Activity Prediction code (ASAP) detects sunspots and pores in white-light continuum images; the Sunspot Tracking And Recognition Algorithm (STARA) detects sunspots in white-light continuum images; the Spatial Possibilistic Clustering Algorithm (SPoCA) automatically segments solar EUV images into active regions (AR), coronal holes (CH), and quiet Sun (QS). One month of data from the Solar and Heliospheric Observatory (SOHO)/Michelson Doppler Imager (MDI) and SOHO/Extreme Ultraviolet Imaging Telescope (EIT) instruments during 12 May?–?23 June 2003 is analysed. The overall detection performance of each algorithm is benchmarked against National Oceanic and Atmospheric Administration (NOAA) and Solar Influences Data Analysis Center (SIDC) catalogues using various feature properties such as total sunspot area, which shows good agreement, and the number of features detected, which shows poor agreement. Principal Component Analysis indicates a clear distinction between photospheric properties, which are highly correlated to the first component and account for 52.86% of variability in the data set, and coronal properties, which are moderately correlated to both the first and second principal components. Finally, case studies of NOAA 10377 and 10365 are conducted to determine algorithm stability for tracking the evolution of individual features. We find that magnetic flux and total sunspot area are the best indicators of active-region emergence. Additionally, for NOAA 10365, it is shown that the onset of flaring occurs during both periods of magnetic-flux emergence and complexity development.     

Extreme Ultraviolet Variability Experiment (EVE) Multiple?EUV Grating Spectrographs (MEGS): Radiometric Calibrations and Results

The NASA Solar Dynamics Observatory (SDO), scheduled for launch in early 2010, incorporates a suite of instruments including the Extreme Ultraviolet Variability Experiment (EVE). EVE has multiple instruments including the Multiple Extreme ultraviolet Grating Spectrographs (MEGS) A, B, and P instruments, the Solar Aspect Monitor (SAM), and the Extreme ultraviolet SpectroPhotometer (ESP). The radiometric calibration of EVE, necessary to convert the instrument counts to physical units, was performed at the National Institute of Standards and Technology (NIST) Synchrotron Ultraviolet Radiation Facility (SURF III) located in Gaithersburg, Maryland. This paper presents the results and derived accuracy of this radiometric calibration for the MEGS A, B, P, and SAM instruments, while the calibration of the ESP instrument is addressed by Didkovsky et?al. (Solar Phys., 2010, doi: 10.1007/s11207-009-9485-8 ). In addition, solar measurements that were taken on 14 April 2008, during the NASA 36.240 sounding-rocket flight, are shown for the prototype EVE instruments.     

Turbulent Heating in the Accelerating Region Using a Multishell Model

Solar energetic particles (SEPs) are released into the heliosphere by solar flares and coronal mass ejections (CMEs). They are mostly protons, with smaller amounts of heavy ions from helium to iron, and lesser amounts of species heavier than iron. The spectra of heavy ions have been previously studied mostly by using the fluence of the particles in an event-integrated spectrum in a small number of spectral snapshots. In this article, we analyze the temporal evolution of the heavy-ion spectra using two large SEP events (27 January 2012 and 7 January 2014) from the Solar TErrestrial Relations Observatory (STEREO) era using Advanced Composition Explorer (ACE) Solar Isotope Spectrometer (SIS) and Ultra Low Energy Isotope Spectrometer (ULEIS), Energetic Particles: Acceleration, Composition and Transport (EPACT) onboard Wind, and the STEREO-A (Ahead) and -B (Behind) Low-Energy Telescope (LET) and Suprathermal Ion Telescope (SIT) instruments, taking a large number of snapshots covering the temporal evolution of the event. We find large differences in the spectra of the ions after the main flux enhancement in terms of the grouping of similar species, but also in terms of the location of the instruments. Although it is somewhat less noticeable than in the case of the temporal evolution of protons (Doran and Dalla, Solar Phys. 291, 2071, 2016), we observe a wave-like pattern travelling through the heavy ion spectra from the highest energies to the lowest, creating an “arch” structure that later straightens into a power law after 18 to 24 hours.

     

Deep Solar Activity Minimum 2007 – 2009: Solar Wind Properties and Major Effects on the Terrestrial Magnetosphere

We discuss the temporal variations and frequency distributions of solar wind and interplanetary magnetic field parameters during the solar minimum of 2007?–?2009 from measurements returned by the IMPACT and PLASTIC instruments on STEREO-A. We find that the density and total field strength were significantly weaker than in the previous minimum. The Alfvén Mach number was higher than typical. This reflects the weakness of magnetohydrodynamic (MHD) forces, and has a direct effect on the solar wind–magnetosphere interactions. We then discuss two major aspects that this weak solar activity had on the magnetosphere, using data from Wind and ground-based observations: i) the dayside contribution to the cross-polar cap potential (CPCP), and ii) the shapes of the magnetopause and bow shock. For i) we find a low interplanetary electric field of 1.3±0.9 mV?m^?1 and a CPCP of 37.3±20.2 kV. The auroral activity is closely correlated to the prevalent stream–stream interactions. We suggest that the Alfvén wave trains in the fast streams and Kelvin–Helmholtz instability were the predominant agents mediating the transfer of solar wind momentum and energy to the magnetosphere during this three-year period. For ii) we determine 328 magnetopause and 271 bow shock crossings made by Geotail, Cluster 1, and the THEMIS B and C spacecraft during a three-month interval when the daily averages of the magnetic and kinetic energy densities attained their lowest value during the three years under survey. We use the same numerical approach as in Fairfield’s (J. Geophys. Res. 76, 7600, 1971) empirical model and compare our findings with three magnetopause models. The stand-off distance of the subsolar magnetopause and bow shock were 11.8 R _E and 14.35 R _E, respectively. When comparing with Fairfield’s (1971) classic result, we find that the subsolar magnetosheath is thinner by ～1 R _E. This is mainly due to the low dynamic pressure which results in a sunward shift of the magnetopause. The magnetopause is more flared than in Fairfield’s model. By contrast the bow shock is less flared, and the latter is the result of weaker MHD forces.

     

Multi-spacecraft Observations of CIR-Associated Ion Increases During the Ulysses 2007 Ecliptic Crossing

The period between 21 June and 8 October, 2007 (Carrington rotations 2058 to 2061), comprising the Ulysses ecliptic plane crossing, was characterized by low solar activity. Excluding the small solar energetic particle events observed during July, the ion increases observed in the inner heliosphere between 100?keV/n and 10?MeV/n were associated with Corotating Interaction Regions (CIRs). In this work, we investigate CIR-related ion increases using multipoint observations from Ulysses, ACE, and the twin STEREO spacecraft. The ballistic backmapping technique has been used to correlate in situ observations of these spacecraft and remote-sensing observations of coronal holes. Although the radial, longitudinal and latitudinal separation of the spacecraft (except Ulysses) are relatively small, we find discrepancies when a detailed comparison of narrow structures like stream interfaces and CIR-associated shocks is performed. Therefore we concentrate on the two CIR events from day 5 to day 10 of August 2007 and from day 25 to day 31 of August 2007, which lend themselves to a more undisturbed comparison. Using the multi-spacecraft measurements we could determine a radial gradient of 230±30% AU^?1, which is consistent with previous results by van Hollebeke et al. (J. Geophys. Res. 83, 4723, 1978) of ～?350% AU^?1 using Helios and Pioneer data.

     

The Origin of Sequential Chromospheric Brightenings

Sequential chromospheric brightenings (SCBs) are often observed in the immediate vicinity of erupting flares and are associated with coronal mass ejections. Since their initial discovery in 2005, there have been several subsequent investigations of SCBs. These studies have used differing detection and analysis techniques, making it difficult to compare results between studies. This work employs the automated detection algorithm of Kirk et al. (Solar Phys. 283, 97, 2013) to extract the physical characteristics of SCBs in 11 flares of varying size and intensity. We demonstrate that the magnetic substructure within the SCB appears to have a significantly smaller area than the corresponding \(\mbox{H}\upalpha\) emission. We conclude that SCBs originate in the lower corona around \(0.1~R_{\odot}\) above the photosphere, propagate away from the flare center at speeds of \(35\,\mbox{--}\,85~\mbox{km}\,\mbox{s}^{-1}\), and have peak photosphere magnetic intensities of \(148\pm2.9~\mbox{G}\). In light of these measurements, we infer SCBs to be distinctive chromospheric signatures of erupting coronal mass ejections.     

Temporal and Periodic Variations of Sunspot Counts in Flaring and Non-Flaring Active Regions

We analyzed temporal and periodic variations of sunspot counts (SSCs) in flaring (C-, M-, or X-class flares), and non-flaring active regions (ARs) for nearly three solar cycles (1986 through 2016). Our main findings are as follows: i) temporal variations of monthly means of the daily total SSCs in flaring and non-flaring ARs behave differently during a solar cycle and the behavior varies from one cycle to another; during Solar Cycle 23 temporal SSC profiles of non-flaring ARs are wider than those of flaring ARs, while they are almost the same during Solar Cycle 22 and the current Cycle 24. The SSC profiles show a multi-peak structure and the second peak of flaring ARs dominates the current Cycle 24, while the difference between peaks is less pronounced during Solar Cycles 22 and 23. The first and second SSC peaks of non-flaring ARs have comparable magnitude in the current solar cycle, while the first peak is nearly absent in the case of the flaring ARs of the same cycle. ii) Periodic variations observed in the SSCs profiles of flaring and non-flaring ARs derived from the multi-taper method (MTM) spectrum and wavelet scalograms are quite different as well, and they vary from one solar cycle to another. The largest detected period in flaring ARs is \(113\pm 1.6~\mbox{days}\) while we detected much longer periodicities (\(327\pm 13\), \(312 \pm 11\), and \(256\pm 8~\mbox{days}\)) in the non-flaring AR profiles. No meaningful periodicities were detected in the MTM spectrum of flaring ARs exceeding \(55\pm 0.7~\mbox{days}\) during Solar Cycles 22 and 24, while a \(113\pm 1.3~\mbox{days}\) period was detected in flaring ARs of Solar Cycle 23. For the non-flaring ARs the largest detected period was only \(31\pm 0.2~\mbox{days}\) for Cycle 22 and \(72\pm 1.3~\mbox{days}\) for the current Cycle 24, while the largest measured period was \(327\pm 13~\mbox{days}\) during Solar Cycle 23.     

Study of Geomagnetic Storms and Solar and Interplanetary Parameters for Solar Cycles 22 and 24

The aim of this paper is to investigate the association of geomagnetic storms with the component of the interplanetary magnetic field (IMF) perpendicular to the ecliptic (\(Bz\)), the solar wind speed (\(V\)), the product of solar wind speed and \(Bz\) (VBz), the Kp index, and the sunspot number (SSN) for two consecutive even solar cycles, Solar Cycles 22 (1986?–?1995) and 24 (2009?–?2017). A comparative study has been done using the superposed epoch method (Chree analysis). The results of the present analysis show that \(Bz\) is a geoeffective parameter. The correlation coefficient between Dst and \(Bz\) is found to be 0.8 for both Solar Cycles 22 and 24, which indicates that these two parameters are highly correlated. Statistical relationships between Dst and Kp are established and it is shown that for the two consecutive even solar cycles, Solar Cycles 22 and 24, the patterns are strikingly similar. The correlation coefficient between Dst and Kp is found to be the same for the two solar cycles (?0.8), which clearly indicates that these parameters are well anti-correlated. For the same studied period we found that the SSN does not show any relationship with Dst and Kp, while there exists an inverse relation between Dst and the solar wind speed, with some time lag. We have also found that VBz is a more relevant parameter for the production of geomagnetic storms, as compared to \(V\) and \(Bz\) separately. In addition, we have found that in Solar Cycles 22 and 24 this combined parameter is more relevant during the descending phase as compared to the ascending phase.     

Comparison of Helioseismic Far-Side Active Region Detections with STEREO Far-Side EUV Observations of Solar Activity

Seismic maps of the Sun’s far hemisphere, computed from Doppler data from the Helioseismic and Magnetic Imager (HMI) on board the Solar Dynamics Observatory (SDO) are now being used routinely to detect strong magnetic regions on the far side of the Sun ( http://jsoc.stanford.edu/data/farside/ ). To test the reliability of this technique, the helioseismically inferred active region detections are compared with far-side observations of solar activity from the Solar TErrestrial RElations Observatory (STEREO), using brightness in extreme-ultraviolet light (EUV) as a proxy for magnetic fields. Two approaches are used to analyze nine months of STEREO and HMI data. In the first approach, we determine whether new large east-limb active regions are detected seismically on the far side before they appear Earth side and study how the detectability of these regions relates to their EUV intensity. We find that while there is a range of EUV intensities for which far-side regions may or may not be detected seismically, there appears to be an intensity level above which they are almost always detected and an intensity level below which they are never detected. In the second approach, we analyze concurrent extreme-ultraviolet and helioseismic far-side observations. We find that 100% (22) of the far-side seismic regions correspond to an extreme-ultraviolet plage; 95% of these either became a NOAA-designated magnetic region when reaching the east limb or were one before crossing to the far side. A low but significant correlation is found between the seismic signature strength and the EUV intensity of a far-side region.     

High precision symplectic integrators for the Solar System

Using a Newtonian model of the Solar System with all 8 planets, we perform extensive tests on various symplectic integrators of high orders, searching for the best splitting scheme for long term studies in the Solar System. These comparisons are made in Jacobi and heliocentric coordinates and the implementation of the algorithms is fully detailed for practical use. We conclude that high order integrators should be privileged, with a preference for the new $(10,6,4)$ method of Blanes et al. (2013).     

Investigation of the Geoeffectiveness of Disk-Centre Full-Halo Coronal Mass Ejections

We studied the occurrence and characteristics of geomagnetic storms associated with disk-centre full-halo coronal mass ejections (DC-FH-CMEs). Such coronal mass ejections (CMEs) can be considered as the most plausible cause of geomagnetic storms. We selected front-side full-halo coronal mass ejections detected by the Large Angle and Spectrometric Coronagraph onboard the Solar and Heliospheric Observatory (SOHO/LASCO) from the beginning of 1996 till the end of 2015 with source locations between solar longitudes E10 and W10 and latitudes N20 and S20. The number of selected CMEs was 66 of which 33 (50%) were deduced to be the cause of 30 geomagnetic storms with \(\mathrm{Dst} \leq- 50~\mbox{nT}\). Of the 30 geomagnetic storms, 26 were associated with single disk-centre full-halo CMEs, while four storms were associated, in addition to at least one disk-centre full-halo CME, also with other halo or wide CMEs from the same active region. Thirteen of the 66 CMEs (20%) were associated with 13 storms with \(-100~\mbox{nT} < \mbox{Dst} \leq- 50~\mbox{nT}\), and 20 (30%) were associated with 17 storms with \(\mbox{Dst}\leq- 100~\mbox{nT}\). We investigated the distributions and average values of parameters describing the DC-FH-CMEs and their interplanetary counterparts encountering Earth. These parameters included the CME sky-plane speed and direction parameter, associated solar soft X-ray flux, interplanetary magnetic field strength, \(B_{t}\), southward component of the interplanetary magnetic field, \(B_{s}\), solar wind speed, \(V_{sw}\), and the \(y\)-component of the solar wind electric field, \(E_{y}\). We found only a weak correlation between the Dst of the geomagnetic storms associated with DC-FH-CMEs and the CME sky-plane speed and the CME direction parameter, while the correlation was strong between the Dst and all the solar wind parameters (\(B_{t}\), \(B_{s}\), \(V_{sw}\), \(E_{y}\)) measured at 1 AU. We investigated the dependences of the properties of DC-FH-CMEs and the associated geomagnetic storms on different phases of solar cycles and the differences between Solar Cycles 23 and 24. In the rise phase of Solar Cycle 23 (SC23), five out of eight DC-FH-CMEs were geoeffective (\(\mbox{Dst} \leq- 50~\mbox{nT}\)). In the corresponding phase of SC24, only four DC-FH-CMEs were observed, three of which were nongeoeffective (\(\mbox{Dst} > - 50~\mbox{nT}\)). The largest number of DC-FH-CMEs occurred at the maximum phases of the cycles (21 and 17, respectively). Most of the storms with \(\mbox{Dst}\leq- 100~\mbox{nT}\) occurred at or close to the maximum phases of the cycles. When comparing the storms during epochs of corresponding lengths in Solar Cycles 23 and 24, we found that during the first 85 months of Cycle 23 the geoeffectiveness rate of the disk-centre full-halo CMEs was 58% with an average minimum value of the Dst index of \(- 146~\mbox{nT}\). During the corresponding epoch of Cycle 24, only 35% of the disk-centre full-halo CMEs were geoeffective with an average value of Dst of \(- 97~\mbox{nT}\).     

Forecasting Solar Flares Using Magnetogram-based Predictors and Machine Learning

We propose a forecasting approach for solar flares based on data from Solar Cycle 24, taken by the Helioseismic and Magnetic Imager (HMI) on board the Solar Dynamics Observatory (SDO) mission. In particular, we use the Space-weather HMI Active Region Patches (SHARP) product that facilitates cut-out magnetograms of solar active regions (AR) in the Sun in near-realtime (NRT), taken over a five-year interval (2012?–?2016). Our approach utilizes a set of thirteen predictors, which are not included in the SHARP metadata, extracted from line-of-sight and vector photospheric magnetograms. We exploit several machine learning (ML) and conventional statistics techniques to predict flares of peak magnitude \({>}\,\mbox{M1}\) and \({>}\,\mbox{C1}\) within a 24 h forecast window. The ML methods used are multi-layer perceptrons (MLP), support vector machines (SVM), and random forests (RF). We conclude that random forests could be the prediction technique of choice for our sample, with the second-best method being multi-layer perceptrons, subject to an entropy objective function. A Monte Carlo simulation showed that the best-performing method gives accuracy \(\mathrm{ACC}=0.93(0.00)\), true skill statistic \(\mathrm{TSS}=0.74(0.02)\), and Heidke skill score \(\mathrm{HSS}=0.49(0.01)\) for \({>}\,\mbox{M1}\) flare prediction with probability threshold 15% and \(\mathrm{ACC}=0.84(0.00)\), \(\mathrm{TSS}=0.60(0.01)\), and \(\mathrm{HSS}=0.59(0.01)\) for \({>}\,\mbox{C1}\) flare prediction with probability threshold 35%.     

A Statistical Study on Photospheric Magnetic Nonpotentiality of Active Regions and Its Relationship with Flares During Solar Cycles 22?–?23

A statistical study is carried out on the photospheric magnetic nonpotentiality in solar active regions and its relationship with associated flares. We select 2173 photospheric vector magnetograms from 1106 active regions observed by the Solar Magnetic Field Telescope at Huairou Solar Observing Station, National Astronomical Observatories of China, in the period of 1988??C?2008, which covers most of the 22nd and 23rd solar cycles. We have computed the mean planar magnetic shear angle ( $\overline{\Delta\phi}$ ), mean shear angle of the vector magnetic field ( $\overline{\Delta\psi}$ ), mean absolute vertical current density ( $\overline{|J_{z}|}$ ), mean absolute current helicity density ( $\overline{|h_{\mathrm{c}}|}$ ), absolute twist parameter (|?? _av|), mean free magnetic energy density ( $\overline{\rho_{\mathrm{free}}}$ ), effective distance of the longitudinal magnetic field (d _E), and modified effective distance (d _Em) of each photospheric vector magnetogram. Parameters $\overline{|h_{\mathrm{c}}|}$ , $\overline{\rho_{\mathrm{free}}}$ , and d _Em show higher correlations with the evolution of the solar cycle. The Pearson linear correlation coefficients between these three parameters and the yearly mean sunspot number are all larger than 0.59. Parameters $\overline {\Delta\phi}$ , $\overline{\Delta\psi}$ , $\overline{|J_{z}|}$ , |?? _av|, and d _E show only weak correlations with the solar cycle, though the nonpotentiality and the complexity of active regions are greater in the activity maximum periods than in the minimum periods. All of the eight parameters show positive correlations with the flare productivity of active regions, and the combination of different nonpotentiality parameters may be effective in predicting the flaring probability of active regions.     

Polarity Reversal Time of the Magnetic Dipole Component of the Sun in Solar Cycle 24

The Sun’s general magnetic field has shown polarity reversal three times during the last three solar cycles. We attempt to estimate the upcoming polarity reversal time of the solar magnetic dipole by using the coronal field model and synoptic data of the photospheric magnetic field. The scalar magnetic potential of the coronal magnetic field is expanded into a spherical harmonic series. The long-term variations of the dipole component ( $g^{0}_{1}$ ) calculated from the data of National Solar Observatory/Kitt Peak and Wilcox Solar Observatory are compared with each other. It is found that the two $g^{0}_{1}$ values show a similar tendency and an approximately linear increase between the Carrington rotation periods CR 2070 and CR 2118. The next polarity reversal is estimated by linear extrapolation to be between CR 2132.2 (December 2012) and CR2134.8 (March 2013).     

Perspectives on effectively constraining the location of a massive trans-Plutonian object with the New Horizons spacecraft: a sensitivity analysis

The radio tracking apparatus of the New Horizons spacecraft, currently traveling to the Pluto system where its arrival is scheduled for July 2015, should be able to reach an accuracy of 10 m (range) and 0.1  $\text{ mm } \text{ s }^{-1}$ mm s ? 1 (range-rate) over distances up to 50 au. This should allow to effectively constrain the location of a putative trans-Plutonian massive object, dubbed Planet X (PX) hereafter, whose existence has recently been postulated for a variety of reasons connected with, e.g., the architecture of the Kuiper belt and the cometary flux from the Oort cloud. Traditional scenarios involve a rock-ice planetoid with $m_\mathrm{X}\approx 0.7\,m_{\oplus }$ m X ≈ 0.7 m ⊕ at some 100–200 au, or a Jovian body with $m_\mathrm{X}\lesssim 5\,m_\mathrm{J}$ m X ? 5 m J at about 10,000–20,000 au; as a result of our preliminary sensitivity analysis, they should be detectable by New Horizons since they would impact its range at a km level or so over a time span 6 years long. Conversely, range residuals statistically compatible with zero having an amplitude of 10 m would imply that PX, if it exists, could not be located at less than about 4,500 au ( $m_\mathrm{X}=0.7\,m_{\oplus }$ m X = 0.7 m ⊕ ) or 60,000 au ( $m_\mathrm{X}=5\,m_\mathrm{J}$ m X = 5 m J ), thus making a direct detection quite demanding with the present-day technologies. As a consequence, it would be appropriate to rename such a remote body as Thelisto. Also fundamental physics would benefit from this analysis since certain subtle effects predicted by MOND for the deep Newtonian regions of our Solar System are just equivalent to those of a distant pointlike mass.     

Pulkovo astrometric observations of bodies in the Solar system from 1898 to 2005: Observational database

We provide an overview of the main results obtained as part of the programs for astrometric observations of bodies in the Solar system at the Pulkovo Observatory over the period 1898–2005. We summarize the results of photographic observations and show new possibilities for astrometric observations in connection with the transition to CCD detectors on Pulkovo instruments. Observing and data reduction techniques are considered. A database with Pulkovo observations of bodies in the Solar system has been created and opened to users. The database is accessible at http://www.puldb.ru.     

Two-Step Filament Eruption During 14 – 15 March 2015

We present here an interesting two-step filament eruption during 14?–?15 March 2015. The filament was located in NOAA AR 12297 and associated with a halo Coronal Mass Ejection (CME). We use observations from the Atmospheric Imaging Assembly (AIA) and Heliospheric Magnetic Imager (HMI) instruments onboard the Solar Dynamics Observatory (SDO), and from the Solar and Heliospheric Observatory (SOHO) Large Angle and Spectrometric Coronagraph (LASCO). We also use \(\mbox{H}\upalpha\) data from the Global Oscillation Network Group (GONG) telescope and the Kanzelhoehe Solar Observatory. The filament shows a first step eruption on 14 March 2015 and it stops its rise at a projected altitude \({\approx}\,125~\mbox{Mm}\) on the solar disk. It remains at this height for \({\approx}\,12~\mbox{hrs}\). Finally it erupts on 15 March 2015 and produces a halo CME. We also find jet activity in the active region during both days, which could help the filament de-stabilization and eruption. The decay index is calculated to understand this two-step eruption. The eruption could be due to the presence of successive instability–stability–instability zones as the filament is rising.