Let's start with the appetizer. From 16 August onwards, some large prominences were towering over the Sun's east limb. Solar prominences are clouds of charged particles ("plasma") above the solar surface squeezed between magnetic regions of opposite polarity. Being cooler and denser than the plasma underneath and their surroundings, they appear as bright blobs when seen near the solar limb and as dark lines when seen on the solar disk (then they are called "filaments"). Special filters are required to observe these features, and one such a filter is the Hydrogen-alpha line in the red part of the solar spectrum (H-alpha; 656.3 nm). However, prominences can also be observed in extreme ultraviolet (EUV). At higher temperatures of around 80.000 degrees (GOES/SUVI 304), prominences show up as fat bright structures near the solar limb. At even higher temperatures (SDO/AIA 171: 700,000 degrees; AIA 131: multi-million degrees), these "cold" features can also be seen but then as dark structures carved out against the brighter background of the much hotter (million degrees) plasma. The clips underneath show the eruption of both prominences on 20 August, first by SUVI 304, then a zoomed-in combo of AIA 171 (yellow) and AIA 131 (blue). The latter two clips have been rotated by 90 degrees to the north. Prior to the eruptions, the tallest structure was the one at the northeast limb reaching a height near 160,000 km. A nice clip of this prominence in H-alpha, just before its eruption, was made by Yoshinori Watabe and can be found on YouTube. The associated coronal mass ejections (CMEs) were not directed to Earth.
Then, basically out of nowhere, a powerful full halo CME appeared on 21 August. It first appeared at 08:24 UTC in SOHO/LASCO coronagraphic imagery (see imagery underneath), with speeds between 1200 and 1600 km/s pending the imagery and methodology used. With no important active regions on the earth-facing solar disk, this had to be a farside CME. Interestingly, the greater than 10 MeV proton flux gradually started to increase from around noon on 22 August (see GOES plot underneath). Eventually, the 10 pfu (proton flux units) alert threshold was reached about 5 (five!) days later. A minor proton event was in progress from 26 August at 20:30UTC until 27 August at 14:15 UTC. The flux reached a maximum of only 13 pfu, still a formidable achievement for a farside eruption. A strong farside eruption creating a small proton event near Earth is not so exceptional, with this STCE newsitem discussing a quite similar case.
Solar Orbiter, located almost directly opposite of the Earth (so also on the Sun's farside), had a clear view of the eruption. Its STIX instrument (see this STCE newsitem for a brief description) recorded a strong eruption peaking at 08:16 UTC, with an estimated GOES soft x-ray intensity between M5.7 and X2.7, the most likely value being X1.4 - so an X-class event (see the graph underneath). The annotated EUV images underneath were obtained by Solar Orbiter's FSI instrument (Full Sun Imager) and show the Sun (farside as seen from Earth...) at 00:00UTC and at 12:00UTC on 21 August, basically before and after the eruption. The image from 12:00UTC clearly shows an arcade (a series of post-eruption coronal loops) in the encircled area, a telltale sign a strong eruption had taken place there.
Based on the FSI images, the source region was located about 45 degrees behind the east limb and 5 degrees north as seen from Earth. Using a typical value for the Sun's rotation (13.2° / day), this meant the region would start turning over the east limb late on 24 or early on 25 August. Sure enough, early on 25 August a sunspot group appeared at the expected location near the east limb. NOAA 4199 (SIDC 617) consisted basically of a single, big, but fragmented sunspot. Initially, it still produced some M-class flares, including an M4.5 event on 25 August and an M3.3 a day later. From then onwards, other active regions dominated the flare activity and NOAA 4199 seemed to decay. Its magnetic structure was also simple (no delta configuration). It continued to be connected with nearby active regions in the southern solar hemisphere, and during the course of 30 August, its main sunspot started to become more compact again. This is shown in the SDO imagery underneath (HMI left; AIA 171 right; annotated image with NOAA numbers).
Yet, it was still quite a surprise when late on 30 August NOAA 4199 was the source of a long duration event that peaked at 20:02 UTC. The radio emission associated with this M2.7 flare strongly affected the value of the daily radio flux (2.8 GHz) as measured by Penticton (see also EOVSA - and the radiospectrogram). The proton flux, still enhanced from the 21 August event, started to rise again very slowly. However, this time it peaked already at 6 pfu on 1 September, clearly remaining below the alert threshold. This enhancement can be seen as a bump in the proton flux in the chart above. Because of the location of NOAA 4199, the associated fast (1600 km/s) full halo CME was directed straight to Earth. The CME arrived on 1 September at 20:12UTC (DSCOVR) when the solar wind speed jumped from about 380 km/s to 650 km/s. The north-south component (Bz) of the magnetic field was initially negative, briefly dipping to -23 nT. Then Bz turned mostly positive, thus preventing the development of a strong geomagnetic storm. At the time of this writing, only a single interval with moderate geomagnetic conditions (Kp = 6) has been recorded. The imagery underneath shows the flare in EUV (SUVI 094; top) and the associated faint full halo CME in white light (SOHO/LASCO C3; bottom).
