Active nests at solar maximum: How nested flux emergence dominated flaring activity and structured the heliosphere

We investigated the role of active nests during the 2024 solar activity maximum. The distribution of solar activity in 2024 is displayed in Figure 2, which contains Carrington Rotations (CRs) 2279 to 2292. To identify active nests, we used the averaged EUV emission from AIA 304 Å Carrington maps and the locations of GOES flares in Carrington coordinates. There was a distinct north-south asymmetry in solar activity during 2024, with $\sim$70% of all GOES flares (mostly C- and M-class) originating in the southern hemisphere. Three regions in the southern hemisphere exceeded the 90th percentile of the averaged EUV emission, denoted AN1, AN2 and AN3. The GOES flares were strongly concentrated within these regions. AN1 and AN2 migrated towards each other from 240^∘ and 120^∘ Carrington longitude respectively, while AN3 was on the opposite side of the Sun, near to 355^∘. We focused on these three active nests; however, this is not a unique selection. Notably, activity nesting was present for a few CRs at the start and end of 2024 in the southern hemisphere around 60^∘ longitude and in the northern hemisphere drifting from 150^∘ to 180^∘ and then 130^∘ longitude during 2024.

The morphological evolution of AN1, AN2, and AN3 was extracted using the EUV emission from AIA, before the process was repeated for the far-side data. For each CR in 2024, brightness contours were extracted from AIA 304 Å Carrington maps following the application of a Gaussian smoothing. The absolute size of each contour was slightly affected by the local environment of each feature. The selection threshold was intentionally set to capture the full extent of features, at the cost of including some of the surrounding quiet Sun. Priority was given to continuity in Carrington longitude, following the assumption that the initial rise of flux tubes through the convection zone may be less coherent and thus slightly spread in latitude. For example, in CRs 2280 and 2281, AN1 includes some flux emergence events towards the south of the main nesting site. GOES flares were then assigned to their nearest contour. Finally, we manually assigned the contours and their associated flares to AN1, AN2, and AN3 based on their Carrington longitude to highlight regions of persistent solar activity. By October, the proximity of AN1 and AN2 meant that they were no longer separable. This combined region was labelled as a product of their collision. The longitudinal evolution of the active nests is illustrated in Figure 2. Each active nest sustained coherent activity that spanned several solar rotations, satisfying the definition from Finley et al. (2025).

Solar Orbiter’s trajectory provides opportunities to observe the Sun’s far-side with respect to Earth. When combined with data from near-Earth satellites (e.g. SDO and GOES), this creates epochs with nearly continuous monitoring of the entire solar surface. The relative position of Solar Orbiter to Earth during 2024 is displayed in Figure 1. From April to October, Solar Orbiter had a unique view of the Sun’s far-side that complemented near-Earth satellites. This allowed us to perform nearly continuous monitoring of the three active nests (AN1, AN2, and AN3). When assessing global statistics, we applied a strict definition of 90% coverage of the solar surface to define a nearly continuous monitoring window of 156 days, from 20 April to 24 September. Venus gravity assist manoeuvres in February 2025, December 2026, and so on, will increase the spacecraft’s orbital inclination, and reduce the duration of these nearly continuous monitoring windows from $\sim$150 days per year to an average of $\sim 60$ days (alternating each year after 2026 between 25 and 90 days).

The same feature-extraction process used on AIA was applied to the EUI 304 Å Carrington maps. To ensure consistency with AIA, the threshold was manually adjusted to reproduce the contours from AIA at times when the two instruments observed the same features. Full-disk EUV images from AIA and EUI during the nearly continuous monitoring window are shown in Appendix B, along with the derived contours for AN1 and AN2. In October, the two active nests began to interact as Solar Orbiter left the Sun’s far-side and the nearly continuous monitoring window ended. To complete the flare distribution, events detected by STIX were assigned to AN1, AN2, or AN3 based on their reconstructed Carrington coordinates, as done previously for flares detected by GOES. The result is presented in Figure 3. The nearly continuous monitoring window captured the complete evolution of AN2 and its collision with AN1, however, we lacked some coverage for AN1, AN3, and the collision product at the beginning and end of the window.

After tracking the active nests, we evaluated their contribution to global solar activity. Figure 4 summarises the solar flare occurrence rate as a function of flare classification during 2024. Following the increased coverage of the solar surface (see Figure 1), the number of flares observed within the nearly continuous monitoring window grew by 35–40% compared to when Solar Orbiter was closer to Earth. The overall level of solar activity was consistent with 2024 being the maximum of solar cycle 25, with a total of 11,533 C-class, 1,260 M-class, and 37 X-class flares observed by GOES and STIX (plus 73 events that triggered the STIX attenuator). While periods of heightened M- and X-class activity appeared every few months, the southern hemisphere consistently hosted more flaring activity. Over the full year, we attributed 74% of C-class, 78% of M-class, and 78% of X-class flares to the southern hemisphere.

The north-south asymmetry in flaring activity correlates with the lag in sunspot number between hemispheres (Joshi et al., 2015). Since solar cycle 20, the southern hemisphere has been trailing behind the northern hemisphere (Temmer et al., 2002). Cycles 23 and 24 had a two-year delay between the peak sunspot number in each hemisphere, following the emergence of a strong quadrupole mode (DeRosa et al., 2012; Finley and Brun, 2023). A similar trend has now also emerged for cycle 25. The quadrupole-to-dipole energy ratio reached a maximum of $\sim 40$ near the end of 2023, and remained around $\sim 5$ throughout 2024, before the large-scale field became strongly influenced by the dipole component (discussed in Section 5). Accordingly, during the 156 days of continuous global coverage with Solar Orbiter in 2024, flaring activity was dominated by the southern hemisphere, which produced 76%, 78%, and 77% of all C-, M-, and X-class flares, respectively. During the maximum and declining phases of cycles 23 and 24, flare production was similarly concentrated in the southern hemisphere (Abdel-Sattar et al., 2018; Korsós and Erdélyi, 2025).

The majority of flaring activity in 2024 originated in the three active nests. Flares from each nest spanned 30-40^∘ in Carrington longitude. AN1 was the first to exhibit significant flaring at the end of March (CR 2280). AN2 and AN3 became visible to Earth in April (CR 2283), with AN3 containing NOAA AR13664 that was responsible for the largest geomagnetic storm so far this cycle (Hayakawa et al., 2025, summarised in). Accordingly, there was a sharp increase in the number of flares observed in May (Kontogiannis et al., 2025; Zhang et al., 2026), with AN3 producing $60\%$ of all flares globally. The modelling of AR13664 by Jarolim et al. (2024) provides a detailed look at the evolution of the coronal magnetic field during this time. Solar Orbiter observed the most powerful flare of solar cycle 25 so far, originating from AN3 while on the Sun’s far-side, with its strength estimated to correspond to an X12 flare (Jaswal et al., 2025).

Subsequent periods of elevated activity occurred at the ends of July and September, both of which involved the interaction of AN1 and AN2. This interaction triggered NOAA AR13842 to release a series of Earth-directed eruptions that interacted with the overlying HCS (Temmer et al., 2026). This included the largest flare observed from Earth so far this cycle (Ding et al., 2025, X9.0;). Solar Orbiter returned towards Earth in October, at which point the Sun’s rotation modulated activity between the AN1/AN2 collision product and AN3 separated by $\sim$180^∘ longitude (see Figure 3). Solar activity remained concentrated within these two southern nesting sites throughout November. As rotation brought each nest into view of Earth, they accounted for roughly 70% of all flares on the solar disk (see Figure 4).

The emergence of magnetic flux within active nests is essential for maintaining coherent activity over several solar rotations. Therefore, active nests are composed of a varying population of individual active regions (Bumba and Howard, 1969; Bumba et al., 2000), each with different magnetic topologies. The magnetic complexity of NOAA active regions can be described using the modified Hale classification system (Hale et al., 1919; Künzel, 1965), where $\alpha$ and $\beta$ represent relatively simple configurations, while $\gamma$ and $\delta$ denote increasing magnetic complexity and the presence of shared penumbrae between opposite-polarity umbrae. It is well-established that $\beta\gamma\delta$ regions are the primary drivers of the most energetic solar flares (Sammis et al., 2000). Furthermore, the temporal lag between the emergence of simple bipolar regions and the appearance of complex active regions during the solar cycle suggests that magnetic complexity is a product of cumulative evolution (Jaeggli and Norton, 2016). Complex active regions also tend to be larger in size. The 10- to 15-month lag of sunspot area behind the number of large flares (Temmer et al., 2003) may further reflect this progressive structuring of the magnetic field. This trend appears most pronounced during odd-numbered cycles, pointing to a potential link with the Sun’s 22-year magnetic cycle.

In Figure 5, we summarise the magnetic classifications of flaring NOAA active regions in 2024. NOAA active regions classifications are updated as they evolve and rotate across the visible solar disk, so we selected the classification associated with the largest GOES flare produced by each active region. The resulting 258 simple ($\alpha$ and $\beta$) and 70 complex ($\beta\gamma$, $\beta\delta$, and $\beta\gamma\delta$) active regions were then assigned to AN1, AN2, their collision product, AN3, or elsewhere; following the methodology previously used for solar flares. The active nests accounted for 44% of all 328 NOAA active regions, yet contained 59% of the 70 complex regions. Active regions were significantly more complex within the active nests (27.9% of 147 active regions in nests) than elsewhere on the Sun (16% of 181 active regions outside of nests). Among the individual active nests, AN2 was the closest to the background population with 18% of its 22 active regions being complex. AN1 and AN3 had similar fractions of complex active regions, 26% of 46 and 29% of 52, respectively. Strikingly, 37% of the 27 active regions in the collision region between AN1 and AN2 has complex classifications. While increased complexity within active nests correlated with their prolific flare production, these regions produced a disproportionate number of flares compared to isolated active regions, indicating that nests are more prone to eruption.

To track how flaring activity evolved in each active nest, we extracted the frequency of C-, M-, and X-class flares from AN1 plus AN2, and AN3. Figure 6 shows the weekly number of flares along with the unsigned photospheric magnetic flux for each nest. The unsigned flux was evaluated from the photospheric magnetic field within the EUV contour(s) of each nest as they crossed the central meridian for HMI and PHI. We fit the distribution of flares versus peak x-ray flux for each month using a power-law exponent $\alpha$ corresponding to the relation $dN\propto F_{X}^{\alpha}dF_{X}$, where $N$ is the flare frequency, and $F_{X}$ is the flare peak flux (see Aschwanden and Freeland, 2012, and references therein). Attenuated STIX flares were excluded from the fitting process, and months with insufficient flare statistics were omitted. The resulting power-law exponent $\alpha$ ranged from $-1.8$ to $-2.3$. These values are consistent with those from the 2022 active nest studied in Finley et al. (2025) as well as previously derived exponents for soft x-ray flares (e.g. Aschwanden et al. 2016).

Overall, AN1, AN2, and their collision product produced 4,466 C-class and 501 M-class flares. AN1 had more activity, responsible for 51% of these C-class and 49% of the M-class flares, whereas AN2 accounted for 28% and 31%, respectively. The remaining flares were associated with the collision product. M-class flares accounted for 9-10% of all flares from these regions. In comparison, AN3 produced 2,292 C-class and 325 M-class flares, with 12% of flares from AN3 reaching M-class. There were a higher number of X-class flares identified in AN3 compared with AN1 plus AN2 (16 versus 12), however, a definitive comparison requires correcting for the STIX attenuator. The elevated GOES background flux during 2024 is expected to have filtered the distribution of smaller C-class flares. The peak unsigned magnetic flux of the individual active nests reached a similar maximum value of around $8\times 10^{22}$Mx.

The level of flaring activity in each active nest exhibited distinct epochs of intense activity; annotated in Figure 6. These epochs displayed different behaviours, either a steady growth in flaring activity shown with upwards arrows (c and d), or a rapid burst and slow decline shown with downwards arrows (b, e, and f). AN1 was initially active in March (a) and late-April (b), including three X-class flares observed by GOES. AN1 and AN2 then produced a steady background of C- and M-class flares in June, building towards late July (c) and September (d). In contrast, both bursts of activity from AN3 (e and f) in the nearly continuous monitoring window appeared to decay over the course of a month. The bursts from AN3 were more pronounced than the peaks in activity from either AN1 or AN2, with the first burst (e) containing at least 9 X-class flares (from NOAA AR13664). The power-law exponent $\alpha$ was between $-1.8$ and $-2.1$ for each burst in AN3, whereas for AN1 and AN2 $\alpha$ was steeper between $-2.2$ and $-2.3$. In the following subsections, we interpreted these slight differences as a consequence of the different flux emergence patterns within the active nests.

The convergence and divergence of active nests have been well-documented in previous solar cycles (Pojoga and Cudnik, 2002, e.g.,), with interacting nests linked to heightened flaring activity. Notably, the extreme solar activity during the October 2003 ‘Halloween storms’ (Gopalswamy et al., 2005, see) was associated with converging activity in the southern hemisphere. To examine the evolution and collision of AN1 and AN2, Figure 7 focuses on the epochs with growing activity denoted (c) and (d) in Figure 6. During epoch (c), AN1 and AN2 contained substantial and repeated flux emergence events that were highly spatially correlated with similar bipole orientations. In this way, new magnetic flux emerged directly into pre-existing flux (Harvey and Zwaan, 1993; Norton et al., 2025, 40-50% of all active regions emerge like this; e.g.,). This is consistent with the increased magnetic complexity from Figure 5 being produced by the interaction of emerging bipoles. These complex active regions were the primary driver of flaring activity, as evidenced by the clustering of flares within these regions. This may account for the steeper power-law index $\alpha$, since the emergence of new magnetic flux could have destabilised residual structures in the coronal magnetic field, triggering a greater number of small flares relative to large flares, which are more commonly associated with large bursts of flux emergence (Kutsenko et al., 2021).

Flux emergence events in AN1 and AN2 typically occurred every 12 to 18 days, shifting in longitude by 4^∘ to 10^∘. As shown in Figure 7 for epoch (d), these events were initially confined within individual nests before migrating toward the central region between them. During the collision, the largest eruptions were mostly concentrated within AN2 (Ding et al., 2025; Temmer et al., 2026, e.g.,), following flux emergence events. Figure 6 highlights that the unsigned magnetic flux of the collision product was similar to that of the individual nests (also true of the net magnetic flux). The consistency across nesting sites could be explained by flux emergence events forming from a sub-surface reservoir with a characteristic magnetic field strength. Furthermore, the longitudinal migration of flux emergence events points to a spatially extended source that is progressively depleted over time. Following this hypothesis, AN1, AN2, and the collision product each represent the episodic release of magnetic flux from different sections of the same sub-surface reservoir. The observed collision is therefore not a merger of separate nests, but rather the spatial convergence of flux emergence from a coherent and extended sub-surface structure. This aligns with the global toroidal bands proposed by Dikpati et al. (2025), linking surface nesting and sub-surface dynamo reservoirs (as simulated in 3D in Nelson et al., 2014 and Jouve et al., 2018).

Compared to the colliding active nests, AN3 had a distinctly different flux emergence behaviour, with two major episodes of activity, (e) and (f), in Figure 6. Epoch (e) corresponds to the rapid emergence of NOAA AR13664, which had a peak emergence rate of $\sim 2\times 10^{21}$ Mx/hr (Wang et al., 2024, see). Interestingly, the rapid emergence of AR13664 appeared to locally suppress further emergence events, summarised in Figure 8. All flux emergence events within AN3 during July and August occurred outside the AR13664 boundary (black contour). Additionally, the emergence events to the west and east of AR13664 formed a distinct longitudinal chain. This highly organized pattern supports an extended sub-surface origin for AN3, similar to AN1 and AN2. This is consistent with the global dynamo modelling of Dikpati et al. (2025) that suggested AR13664 was part of a large-scale toroidal magnetic flux band. With this interpretation, the sub-surface magnetic flux was either exhausted or diverted by the rapid emergence of AR13664. While still complex, the emerging active regions within AN3 were less crowded, which could explain the difference in the power-law exponent $\alpha$ for the flaring in AN3 with respect to AN1 and AN2. In this case, the separation between active regions allowed them to develop the more characteristic power-law slope, closer to that of isolated active region emergences ($\alpha\approx-1.7$ from Aschwanden et al., 2016).

We continued to identify and track nested magnetic activity into 2025, applying the techniques outlined in Section 3. Two active nests emerged from the remnants of the AN1 and AN2 collision. We retained the labelling conventions of AN1 and AN2 because their spatial configuration mirrored the original nests. The 2025 nests maintained a separation of $\sim 120^{\circ}$ in Carrington longitude, though they were offset by $\sim 30^{\circ}$ relative to their 2024 positions. Notably, these regions were weaker, reflecting a general decline in solar activity. At the same time, a new active nest (AN4) emerged in the northern hemisphere, near the Carrington longitude of AN3’s final activity. Figure 9 illustrates how the distribution of active nests evolved between 2024 and 2025. Areas of high-activity in 2025 somewhat mirrored the quiet regions of 2024, and vice versa. For example, the intense activity observed at 60° longitude in the southern hemisphere during 2024 was completely absent in 2025. Conversely, both hemispheres around 270° longitude exhibited little to no activity in 2024 which then became active in 2025. As this work focused on the intense flaring activity during solar maximum in 2024, a detailed analysis of the 2025 activity was left for future work; however, a brief summary of the 2025 nests (AN1, AN2, and AN4) is provided in Appendix C.

The solar wind fills our circumstellar environment, the heliosphere, through which energetic particles and coronal mass ejections propagate (Owens and Forsyth, 2013, e.g.). The solar wind is organised by the Sun’s large-scale magnetic field (Stansby et al., 2021; Finley and Brun, 2023). At solar minimum, the photospheric and coronal magnetic field takes the form of an axisymmetric dipole (McComas et al., 2000). Once active regions emerge, higher-order (quadrupole and octupole) components increasingly begin to affect the Sun’s large-scale magnetic field (DeRosa et al., 2012). This produces a more complex solar wind outflow and heliospheric magnetic field at maximum (Réville and Brun, 2017, e.g.,). During the solar maximum in 2024, the concentration of flux emergence into active nests had a profound impact on the global structure of the solar corona and wind.

Nested flux emergence produces coherent structures in the Sun’s large-scale magnetic field by reinforcing sectors of opposite polarity that close at higher altitudes in the corona (Finley, 2024; Tähtinen et al., 2026). This tends to constrain the HCS, the divide between positive and negative polarity heliospheric magnetic fields, to lie above active nests. We highlighted this trend in Figure 9 by computing Potential Field Source Surface (PFSS) models (Altschuler and Newkirk, 1969; Wang and Sheeley Jr, 1992; Schrijver and DeRosa, 2003) for 2024 and 2025 using polar-field-corrected photospheric magnetic field Carrington maps from SDO/HMI (Sun, 2018). In this study, the source surface radius was fixed at 2.4 solar radii for all PFSS models. We identified persistent positive and negative heliospheric magnetic field sectors across all CRs in each year. In 2024, the dipole component of the Sun’s large scale magnetic field was initially weak and inclined to the rotation axis. The dipole strengthened through late 2024 and drifted toward the rotational poles in 2025.

AN1 and AN2 were separated by $\sim 180^{\circ}$ longitude from the third active nest in both 2024 (AN3) and 2025 (AN4). Figure 9 compares the spatial distribution of these active nests with the coronal magnetic field polarity. The orientation of the Sun’s dipole magnetic field component remained stable across both years. In 2024, all three nests were located in the southern hemisphere. While each nest influenced the position of the HCS, AN1 and AN2 were the most effective at anchoring and inclining it. Following Hale’s polarity law for solar cycle 25 (Hale et al., 1919), the active regions within all three nests typically shared a consistent east-west magnetic field polarity. However, only AN1 and AN2 aligned with the pre-existing non-axisymmetric dipole field, whereas AN3 did not. Previous studies described the HCS above matching orientation nests as a Hale Sector Boundary (HSB) and linked the underlying active regions to enhanced activity (Loumou et al., 2018, e.g.,). By 2025, AN3 had dissipated and AN4 formed in the northern hemisphere. After re-emerging from the collision, AN2 was more active than AN1. The east-west magnetic field polarity of AN4, with a longitudinal separation of $\sim$180^∘ from AN1 and AN2, now also aligned with the non-axisymmetric dipole field. Consequently, both AN2 and AN4 acted as HSBs in 2025, with the nesting pattern reinforcing the non-axisymmetric dipolar component. This scenario parallels the rejuvenation of the Sun’s large-scale magnetic field in 2014, where a southern active nest abruptly strengthened the non-axisymmetric dipole component (Sheeley Jr and Wang, 2015, see).

To explore how active nests influenced the heliosphere, we focused on the 2025 nesting pattern as this was the most clearly established in the solar corona. Figure 10 combines a series of PFSS models from 2025, with the corresponding Carrington maps of scattered white-light observations from the Large Angle and Spectrometric Coronagraph (Brueckner et al., 1995, LASCO;) on board the Solar and Heliospheric Observatory (Domingo et al., 1995, SOHO;). These remote-sensing observations were compared to the hourly averaged in-situ solar wind speed and radial magnetic field polarity taken from the OMNI database (King and Papitashvili, 2005) and Solar Orbiter using the Solar Wind Analyser (Owen et al., 2020, SWA;) and Magnetometer (Horbury et al., 2020, MAG;) instruments. For each CR, we computed 14 additional PFSS models using equally spaced ADAPT-GONG magnetograms to examine the variability within each solar rotation. The location of the HCS was tightly constrained in almost all CRs to lie above the active nests, with AN1 and AN2 separating after the collision and drifting to 180^∘ and 120^∘, respectively (see Appendix C). The HCS was located within $\sim 20^{\circ}$ longitude of AN2 for 10 months in 2025.

The Sun’s large-scale magnetic field was remarkably stable and dipolar in 2025, despite solar maxima being associated with higher order magnetic field components (DeRosa et al., 2012; Finley and Brun, 2023). This organisation resulted from the persistence of AN1 and AN2 over almost two years, whose emerging flux systematically strengthened the pre-existing non-axisymmetric dipole component of the Sun’s magnetic field, while the polar magnetic field were near their weakest. The reservoir of positive-polarity flux east of AN2 (seen in epoch (d) of Figure 7), together with the reconfigured large-scale field, facilitated the growth of a large coronal hole near 100° longitude, while the negative-polarity flux from AN1 fed into coronal holes further west. Figure 10 shows low electron density (dark) regions in the LASCO/C2 coronagraph data associated with fast wind speed of 600 to 800 km/s separated by higher density (white) slow wind above the inclined HCS. The imbalanced magnetic fluxes surrounding the active nests also created filament channels (highlighted in Figure 13). As previously noted, the quadrupole-to-dipole energy ratio dropped from $\sim 40$ in 2023 to near unity in 2025. A similar shift occurred in cycle 24, where the ratio decreased from $\sim 100$ in 2012 towards unity in 2014 following the emergence of an active nest in the southern hemisphere (Finley, 2024, see Fig. 3 in). These events demonstrate that active nests can act as topological anchors for the heliosphere, dictating the geometry of the solar wind long after the main flaring activity had subsided.

Solar Orbiter’s long-term planning is executed approximately six months prior to each observing window; this is required to align time-dependent data acquisition with available on-board storage and downlink capabilities. As the solar corona and wind were so heavily constrained by active nests in 2024 and 2025, we assessed the potential benefits this could have offered to mission level planning. We primarily focused on solar wind connection science (Yardley et al., 2023, e.g.,), as this requires linking remote-sensing observations to in-situ solar wind measurements. The dominance of active nests at solar maximum ensured that the organisation of the fast and slow solar wind streams remained relatively self-similar in Carrington coordinates. This improved the reliability of connectivity tools that rely on PFSS and ballistic back-mapping techniques (Rouillard et al., 2020, e.g.,) by reducing the uncertainty introduced by the stochastic emergence of new active regions.

We investigated how magnetic connectivity varied during a simulated Solar Orbiter observing campaign by systematically varying the nesting configuration. Three time periods were chosen, 1) July 2023 to April 2024, 2) April 2024 to January 2025, and 3) January 2025 to November 2025. These correspond to a baseline period before nesting in 2024, the nest configuration from 2024 (with AN1, AN2, and AN3), and the configuration from 2025 (when AN4 replaced AN3). For each epoch, we produced PFSS models from 10 consecutive Carrington magnetograms. We used the trajectory of Solar Orbiter in spring 2025 (11 March to 25 April); the first inclined orbit after the Venus gravity assist in February 2025 (see Figure 16 in Appendix C). During this time, the spacecraft passed directly over the longitudes of AN1 and AN2. We focused on the varying coronal magnetic field topology and so assumed a constant 400 km/s back-mapping to the PFSS source surface when evaluating the connectivity.

Figure 11 details how the solar wind connectivity varied during the three simulated observing campaign with Solar Orbiter. For each configuration, we evaluated the solar wind sources from all 10 magnetograms and calculated the standard deviation of their mapped positions on the solar surface. Large variations (more than 20-30^∘) in source locations signalled unpredictability, i.e. magnetograms from different CRs did not produce a similar forecast. The overall standard deviation in solar wind source locations decreased between each epoch as the coronal magnetic field became constrained by active nests and the tilted dipole topology; although significant dispersion remained around crossings of the HCS and pseudo streamers. The growth of the non-axisymmetric dipole established a clear magnetic field sector structure, as explored in Figure 10. For the positive sector next to AN2 in 2025, the connectivity was extremely well determined; consequently, this area was subject to two connection science campaigns with Solar Orbiter during 2025 and 2026. The unipolar magnetic fields in these regions contributed to the formation and maintenance of coronal holes (see Yoshida et al., 2026, for similarities with the 2014 active nest). The stability of the magnetic connectivity around AN1 and AN2 was linked to their flux emergence and decay rates, with AN1 more active in 2024 and AN2 in 2025.

The presence of active nests created more favourable source regions for solar wind connectivity forecasting. In late 2025, magnetograms obtained several months in advance provided sufficient constraints for operational planning. However, we lack the ability to predict the initial emergence of active nests or the factors that govern their lifetimes. This is further complicated by unseen far-side flux emergence, which can influence the large-scale structure of the coronal magnetic field and solar wind (Perri et al., 2024; Heinemann et al., 2025). Probing the solar interior with helioseismology could be used to monitor active nests on the Sun’s far-side (Yang et al., 2024a) or detect their feedback on large-scale flows (Sen et al., 2026) and global oscillations (Mehta et al., 2022). In the 2030s, ESA’s Vigil mission (West et al., 2025) will also provide valuable data to help monitor global solar activity. Numerical simulations remain indispensable for probing the mechanisms underlying dynamo magnetic field generation (Brun et al., 2022, e.g.). Future studies could estimate the toroidal magnetic field reservoir accumulated at the base of the convection zone during the previous cycle (Cameron and Schüssler, 2015; Finley et al., 2024) to constrain the nesting lifetimes and track the toroidal flux budget as a function of longitude.