Components of soft X-ray and extreme ultraviolet in flares observed by LYRA on PROBA2 I. E. Dammasch¹, M. Dominique¹, M. Kretzschmar¹, P. C. Chamberlin² ¹ Royal Observatory of Belgium, Solar Influences Data Analysis Center, Brussels, Belgium ² NASA Goddard Space Flight Center, Solar Physics Laboratory, Greenbelt, Maryland, USA Abstract From a constant stream of particles in the form of solar wind to the almost unpredictable events of solar flares and coronal mass ejections, the Earth’s environment is influenced by the Sun. With respect to the Sun-weather relationship, the Earth’s atmosphere may be affected in a much more complicated manner than was imagined before. Evaluating the Sun's impact on climate requires knowledge of variations not only in Total Solar Irradiance, but also variations in the Spectral Solar Irradiance. LYRA, the Large Yield Radiometer, is a UV-VUV solar radiometer embarked on the ESA PROBA2 mission launched in November The instrument acquires solar irradiance in four broad spectral channels - from soft X-ray to UV - that have been chosen for their relevance in solar physics, space weather and aeronomy. The analysis of LYRA time series of flares showed that the flare rise phase is dominated by soft X-ray as observed by GOES, whereas the flare decline phase takes considerably longer and seems to be dominated by cooler EUV emission. LYRA's detectors observe a mixed signal from both these spectral ranges. Using data recorded by the spectrometer EVE on SDO, we will try to understand the quantitative relationship of these spectral intervals. This will bring new insights in the understanding of flare dynamics, as well as their influence on various layers of the Earth's atmosphere. In addition, we will present the LYRA Flare List, which is an online data product and a first step in an attempt to form an independent flare catalogue. Introduction and Questions Looking at typical examples from the LYRA Flare List, e.g., the M1.1 flare of 28 Feb 2011 (see left), several features can be observed repeatedly: ● The GOES curve and the LYRA curves practically start to rise at the same time instance. ● The GOES curve peaks somewhat earlier than the LYRA curves. ● The LYRA curves decrease significantly slower than the GOES curve, and thus take longer time to reach their pre-flare levels. ● In the impulsive phase, all curves practically rise in parallel for several minutes, then the GOES curve peaks, whereas the LYRA curves continue to rise – sometimes slower – to peak later. ● Comparing the calibrated pure flare irradiances (i.e., pre-flare levels subtracted from the LYRA flare curves), the LYRA peaks are an approx. factor of 50 (relative to channel 2-4) or even 70 (relative to channel 2-3) higher than the GOES peaks. What would be left as a difference, were a scaled GOES soft X-ray signal subtracted from the LYRA curves? Would the residual be a pure, cooler-plasma EUV signal? Would it be different for channels 3 and 4? [For a further description of GOES and LYRA flares, please see Poster 40 by Matthieu Kretzschmar et al., Monitoring solar flares: a comparison of observations from GOES and PROBA2/LYRA.] LYRA spectral response It is has been known since the pre-launch calibration campaigns at BESSY that the LYRA Aluminium (2-3) and Zirconium (2-4) filter channels have an SXR contribution in their response (see above). The GOES nominal interval would be at the left edge of this plot, between 0.1 and 0.8 nm. So it appears plausible that the impulsive phase of the LYRA flare curves is dominated by this SXR contribution as observed by one of the GOES X-ray sensors. [For more information on PROBA2 and LYRA, please see the adjacent Poster 37 by Anik De Groof et al., Unique PROBA2 observations of the solar (E)UV corona.] Discussion (1) A comparison is shown between three LYRA curves from the time early 2010 when the Lyman-alpha channel (2-1) was still working properly. The cooler signal peaks in the rising phase of the flare. This can be structurally associated with the EVE curves of emission with log(T)<6, see “Discussion” (right). - Please note that it is a different flare. Discussion (2) SXR and EUV components are shown for the M1.1 flare mentioned in the “Introduction”, in the style described above. While the GOES curve can be associated with the EVE curves of emission with log(T)>7, the “EUV residual” can be associated with the EVE curves of emission with 6<log(T)<7 which peak several minutes later. Peaks are marked by dotted lines in the LYRA plot, and by coloured diamonds in the EVE plot (right). The little bump in the impulsive phase is again associated with EVE curves of emission with log(T)<6. Discussion (3) A thermal evolution plot for the same M1.1 flare is constructed using solar spectra observed by SDO/EVE and contribution functions from the CHIANTI atomic database. By comparison, it appears plausible that the hot “SXR” and cooler “EUV contributions” of the flare’s radiative output can be calculated by GOES and LYRA data as described above. This knowledge may be used in studies of flare effects on Earth’s ionosphere and thermosphere. ExamplesMethods LYRA and GOES data were reduced to one-minute averages. The GOES curve was scaled, i.e. multiplied with a factor such that the values do not exceed the LYRA curve; this was then considered “the SXR contribution” - to be exact, it should be the “hot flare component >10 MK” - and denoted with asterisks. The upper straight lines are the LYRA total signals, channel 2-3 (top part) and channel 2-4 (middle part), respectively. The lower straight lines difference = total LYRA signal – factor * SXR signal were then considered “the EUV contributions” – to be exact, it should be the “cool flare component <10 MK”, since there is plasma emission of quite different temperatures in EUV wavelengths. In the bottom part of each figure, these two differences are compared; the channel-4 curve was scaled, i.e. multiplied with a factor such that the peaks of channels 2-3 and 2-4 match. Therefore, one might assume that the “EUV contribution” curves represent the cooling down from SXR flare temperatures; the little bump before the rising phase represents the heating of cooler plasma before it reaches SXR flare temperatures (“impulsive emissions” <1 MK) The four examples are flares of different strength classes (X, M, C, B) from August Please note that the “dips” – especially visible in the relatively lower-strength flares – are instrument artifacts due to the satellite’s regular large angle rotations every 25 minutes.. Results Flares across more than three orders of magnitude have been selected and treated with the same processing scheme described above. The scaling factors were manually selected. Basically, the approach works for B-, C-, M-, and X-flares, and for fast, impulsive flares as well as long and slow ones. Each flare appears to be individual, i.e. different not only in strength, but also in temporal evolution, length of the rising phase, temporal distance between SXR and EUV peaks, relative contribution of SXR and EUV, and - probably - contributing plasma temperatures. The selected scaling factors appear arbitrary at first glance, and their absolute values are not yet completely understood. While the relation between the two LYRA channels looks constant, the scaling seems to depend on the flare size. The (scaled) difference curves, i.e. the EUV contributions of LYRA channel 2-3 and channel 2-4, are strikingly similar - independent from the flares observed, which just seem to change the temporal structure but not the similarity. Since, according to the responsivity figure above, these two LYRA channels have only small intersections (say, 0-2nm, 5-7nm, 17-19nm) the EUV flare contributions must either originate from these intervals, or the flare signal must be quite similar everywhere in the spectral intervals observed by these two channels.