Understanding and predicting solar eruptive events (e.g., CMEs and flares) is particularly important as they are major drivers of severe space weather. The first part of the work investigates coronal dimmings, often observed before and/or during CME eruptions. These regions are the low-corona segments of the CME structure (dominated by transient open magnetic field) that grow and decay as the magnetic field is dragged out into interplanetary space. To understand CME evolution as a whole, the evolution of dimmings is studied in conjunction with the CME morphology and its physical properties from the solar corona all the way to Earth. The methodology integrates the analysis of photospheric magnetic footpoint distributions within dimmings using the CoDiT tool (Krista, 2013 & 2017), active region loop structures via the PFSS model, CME morphology from STEREO-A/COR2 observations, and CME orientation through the GCS model. This integrated approach aims to leverage dimming observations to gain an early insight into CME evolution and to drive better forecasting.

The second part of the investigation is focused on flare initiation, and the physical connection between pre-flare EUV brightenings and flares. Studying flare-active and inactive periods using the DEFT tool (Krista, 2021, 2025a & 2025b), we found that in no-flare periods EUV signatures were only detected 4% of the time. In flare-active periods, EUV precursors were present 92% of the time within 6 hours before ≥C class flares. 90% of the signatures were associated with flares ≥B class, and over 50% of all signatures were associated with ≥M class flares.  Using the DEFT tool on a combined flare/no-flare database produced the following skill scores: HSS (0.88), TSS(0.88), ACC (0.94), BACC (0.94), PRE (0.95), REC (0.93). These results demonstrate that EUV precursors have a significant potential in improving space weather forecasting.