Analyses
From our own observations since 1950Comparison of solar cycles 13-month smoothed R, each cycle started at its own minimum
The thick orange line is the current cycle 25 and the purple line cycle 24; older cycles are in shades of grey. Dashed: a cycle aligned to the SILSO minimum because our data cannot fix the minimum (observation gap in 2008–2009).
| Cycle | Minimum İÜ · SILSO |
Maximum İÜ · SILSO |
Max. smoothed R (İÜ) · SN (SILSO) |
Length (months) İÜ · SILSO |
|---|---|---|---|---|
| 19 | April 1954 · April 1954 | March 1958 · March 1958 | 235.2 · 285.0 | 122 · 126 |
| 20 | June 1964 · October 1964 | June 1970 · November 1968 | 148.6 · 156.6 | 144 · 137 |
| 21 | June 1976 · March 1976 | December 1979 · December 1979 | 191.3 · 232.9 | 116 · 126 |
| 22 | February 1986 · September 1986 | June 1989 · November 1989 | 209.1 · 212.5 | 122 · 119 |
| 23 | April 1996 · August 1996 | August 2000 · November 2001 | 163.0 · 180.3 | – · 148 |
| 24 | not determined · December 2008 | April 2014 · April 2014 | 144.3 · 116.4 | – · 132 |
| 25 (ongoing) | December 2019 · December 2019 | October 2024 · October 2024 | 151.2 · 160.9 | – · – |
Minimum and maximum are the lowest and highest 13-month smoothed values (SIDC weights; at most two missing months bridged). Dates from a single station can differ from the official SILSO dates by a few months, especially for cycles with two peaks (cycle 20). R is our relative number and SN the international sunspot number (version 2.0); their ratio is the k-factor. Before 2008 our values were read from the observation sheets by a vision model.
How is this computed?
- Monthly mean: for each month, the mean of the relative numbers (R = 10g + f) on the observed days. Months without typed observations use the R read from the sheets.
- Smoothing: a 13-month running mean with the SIDC weights (half weight for the two end months, full weight for the 11 inside). Where months are missing, the weights are rescaled over the months present if at least 11 of the 13 exist; otherwise no value is produced.
- Minimum: the lowest smoothed value within ±18 months of the official SILSO minimum. At least 20 of the 25 months within ±12 months must have a value; otherwise the minimum is left “not determined” (2008–2009).
- Maximum: the highest smoothed value up to the next minimum. Length: months between two minima.
- Chart: each cycle’s curve starts at its own minimum (month 0), so rise speed and peak height can be compared directly. If our data cannot fix the minimum, the SILSO minimum is used (dashed).
Butterfly diagram Heliographic latitude of the sunspot groups, 1954 to today
Each cell is three months wide and 2° of latitude high; its colour is the mean number of groups per observed day in that latitude band (a group seen on many days counts on each day). Every cycle starts at about 30° north and south and its groups appear closer and closer to the equator as the cycle goes on (Spörer’s law), so each cycle draws a pair of “wings”. Dotted lines: cycle minima (SILSO). Grey: no usable data — the 2008–2009 observation gap, 1950–1955, when the sheets mostly give no positions, and a few single quarters with too few usable days. Before 2007 the latitudes were read from the observation sheets by a vision model.
How is this computed?
- Data: one record per observed day. Since 2007 the typed observation (the first one entered if there are several that day), before that the 1950–2007 observation sheet as read by a vision model, only if the sheet is marked usable. Unreadable sheets are never used.
- Latitude: each group’s heliographic latitude B as measured by the observer on the drawing. B = 0.00 counts as “not measured”. Days on which fewer than half of the groups have a latitude are left out; spotless days (no groups) are counted.
- Bins: time in 3-month quarters, latitude in 2° bands from −50° to +50°. Each group adds one to its bin for every day it was seen.
- Colour: the bin count divided by the number of counted days in that quarter: “how many groups were there on an average observed day in this latitude band?”. Cloudy or thinly observed periods therefore do not look paler.
- Grey: quarters with fewer than 5 counted days, or with more than 30% of the observed days left out for lack of latitudes.
Spörer’s law: the drift of the groups toward the equator Mean |latitude| of the groups, each cycle started at its own minimum
The same drift as in the butterfly diagram, one line per cycle: the first groups of a cycle appear at about 25° and the last ones at 5–8°. Plotted against years since the minimum, the cycles lie remarkably close to each other, whether they were strong (19) or weak (24). Colours as in the cycle comparison above (25 orange, 24 purple, older cycles grey).
How is this computed?
- Data: one record per observed day. Since 2007 the typed observation (the first one entered if there are several that day), before that the 1950–2007 observation sheet as read by a vision model, only if the sheet is marked usable. Unreadable sheets are never used.
- Assigning groups to cycles: around a minimum, groups of both cycles are on the disk: the old cycle’s near the equator, the new cycle’s at high latitude. So within ±24 months of a minimum a group with |B| ≥ 15° goes to the new cycle and a lower one to the old cycle. Outside that window a group belongs to the cycle its date falls in (by the SILSO minima). Each curve stops 6 months after the next minimum.
- Mean: for each cycle and each quarter since its minimum, the mean |B| of the groups (each group once for every day it was seen). Quarters with fewer than 20 group-days are not shown.
- Limitation: the 15° cut is a simple rule; a few new-cycle groups born at low latitude near the minimum can be counted with the old cycle. The start and end of each curve are therefore the least certain parts.
Active longitudes Do groups appear equally often on all sides of the Sun?
The Sun turns about once every 27 days. If we fix a longitude grid to its surface (the Carrington system) we can ask whether sunspot groups prefer some sides of the Sun. If they appeared at random, each 30° sector would hold about 1/12 of the activity (8.3%). Solar physicists have long reported “active longitudes”: regions where new groups keep emerging for months or years.
Each bar is one 30° slice of Carrington longitude (one twelfth of the way round the Sun). Its height and the number on it say how many more (+) or fewer (−) groups appeared there than if groups were spread evenly: +25% means a quarter more than average. The grey band is the range that chance alone would give (95%). Dark bars stand out above it (a preferred side of the Sun), grey bars below it (an avoided side); light bars are within chance. With 12 bars per panel, one bar just outside the band can still be chance.
| Cycle | Hemisphere | More groups (preferred longitudes) | Fewer groups |
|---|---|---|---|
| 20 | north | 90–120° (+18%), 240–270° (+32%) | 120–150° (−20%) |
| south | – | 210–240° (−24%) | |
| 21 | north | 120–150° (+24%) | – |
| south | 270–300° (+18%) | 210–240° (−32%) | |
| 22 | north | 0–30° (+24%), 330–360° (+27%) | 60–90° (−45%), 180–210° (−21%) |
| south | 330–360° (+30%) | 210–240° (−32%), 270–300° (−32%) | |
| 23 | north | 60–90° (+35%) | 120–150° (−26%), 270–300° (−22%) |
| south | 120–150° (+28%) | 330–360° (−25%) | |
| 24 | north | 60–90° (+23%) | – |
| south | 90–120° (+31%), 180–210° (+32%), 210–240° (+35%), 240–270° (+22%) | 30–60° (−33%), 270–300° (−36%), 300–330° (−28%), 330–360° (−30%) | |
| 25 | north | 30–60° (+24%), 150–180° (+29%) | 120–150° (−26%), 270–300° (−25%) |
| south | 180–210° (+32%) | 30–60° (−25%) |
The table lists only the slices outside the chance band. The preferred longitudes change from cycle to cycle, in line with the view that active longitudes are long-lived but not permanent.
Detailed view: rotation-by-rotation map
Upper panel: northern hemisphere, lower panel: southern. Each column is one solar rotation, each row a 15° slice of Carrington longitude; darker means more groups. A group stays on the visible disk for about two weeks, so it appears as a short spot of colour. An active longitude shows up as such spots repeating in the same row for several columns in a row (horizontal or slightly tilted streaks). The buttons show one cycle or the whole series.
How is this computed?
- Data: one record per observed day. Since 2007 the typed observation (the first one entered if there are several that day), before that the 1950–2007 observation sheet as read by a vision model, only if the sheet is marked usable. Unreadable sheets are never used.
- Longitude: the Carrington longitude L that the observer wrote for each group. This is a fixed frame that turns with the Sun once every 27.2753 days, so groups that keep emerging at the same place appear at the same longitude rotation after rotation. The sheets give no L before 1965.
- Bins: time in Carrington rotations (27.3 days each), longitude in 15° slices, northern and southern hemisphere drawn separately. Each group adds one to its bin for every day it was seen, and the count is divided by the number of counted days in that rotation (groups per observed day).
- Empty (grey) rotations: rotations with fewer than 5 counted days, or with group longitudes missing on more than 30% of the days.
- Summary chart: per cycle (between SILSO minima) and hemisphere, the share of the group-days with a longitude that fall in each 30° slice, shown as the difference from an even spread (1/12 = 8.33%): (share / 8.33% − 1) × 100. The grey band is ±2σ of a random (binomial) distribution, as a percentage of 8.33%. Because one group is counted on about 5 observation days, the band uses the number of group-days divided by 5 (an estimate of the number of separate groups, from the NOAA matching since 1996). A bar is dark above the band and grey below it; the table lists those slices.
- Limitation: the Sun rotates differentially (the equator faster than the poles), so a real active longitude drifts slowly in the Carrington frame and leaves a tilted rather than a horizontal trace. A group is also added to the same bin on every day of its roughly two weeks on the visible disk. This chart is a raw view; a statistical test for active longitudes is beyond this page.
North–south asymmetry Yearly (N − S) / (N + S), from group-day counts
| Cycle | North group-days |
South group-days |
North share | Dominant hemisphere |
|---|---|---|---|---|
| 19 (1954–1964) | 7,725 | 5,477 | 58.5% | north |
| 20 (1964–1976) | 6,953 | 5,850 | 54.3% | north |
| 21 (1976–1986) | 6,454 | 6,892 | 48.4% | balanced |
| 22 (1986–1996) | 5,082 | 5,938 | 46.1% | south |
| 23 (1996–2008) | 4,682 | 5,370 | 46.6% | south |
| 24 (2008–2019) | 6,095 | 5,687 | 51.7% | balanced |
| 25 (2019–ongoing) | 4,529 | 4,833 | 48.4% | balanced |
Above zero: more groups in the northern hemisphere, below zero: more in the south. The two hemispheres of the Sun are not in step: one usually leads, and the lead can change within a cycle. Pale bars: years with fewer than 300 group-days, mostly around the minima, where a few groups decide the sign. Cycle 19 starts in 1956 in our positions (see above). Before 2007 the values come from latitudes read from the observation sheets.
How is this computed?
- Data: one record per observed day. Since 2007 the typed observation (the first one entered if there are several that day), before that the 1950–2007 observation sheet as read by a vision model, only if the sheet is marked usable. Unreadable sheets are never used.
- Counting: every group with a latitude adds one to the north (N) if B > 0 or to the south (S) if B < 0, for every day it was seen. The same days as in the butterfly diagram are used.
- Asymmetry: for each year A = (N − S) / (N + S). A = +1 means all groups in the north, −1 all in the south, 0 balanced. Years with fewer than 30 group-days are not shown.
- Table: the same counts summed over the cycles between SILSO minima. A north share between 48% and 52% is shown as “balanced”.
- Why group-days? Long-lived large groups are counted again every day and so weigh more; this is close to an asymmetry computed from spot areas. Studies that use areas instead of group counts can give somewhat different values.
Mix of group classes Yearly share, Zürich and McIntosh classes in four common sets
Large bipolar groups (dark blue), the ones that produce most flares, make up 10–18% of the groups near maximum and almost disappear at minimum. The vertical line marks January 1980, when the observers changed from the Zürich to the McIntosh classification. Pale bars: years with fewer than 300 classified group-days. Before 2007 the classes were read from the observation sheets.
How is this computed?
- Data: one record per observed day. Since 2007 the typed observation (the first one entered if there are several that day), before that the 1950–2007 observation sheet as read by a vision model, only if the sheet is marked usable. Unreadable sheets are never used.
- Two schemes: the sheets used the Zürich classification (A–H and J) until the end of 1979 and the McIntosh classification (A–F, H and two extra letters) from January 1980. Only the first letter (the Zürich/McIntosh main class) is used, and letters are combined into four sets that mean the same in both schemes: A, B no penumbra; H, J unipolar with penumbra; C, D bipolar with penumbra, shorter than 10°; E, F, G large bipolar, longer than 10°.
- Share: for each year, each set’s group-days divided by all classified group-days. Groups written as “?” or with other letters (2–10% a year) are left out. Years with fewer than 100 classified group-days are not shown; fewer than 300 are drawn pale.
- Reading it: the share of large groups rises at maximum. Jumps that do not follow the cycle (for example groups without penumbra almost vanishing in 2000–2007) come from the observers’ classification habits in those years or from the reading of the sheets, not from the Sun.
Flares and our sunspot groups Groups since 1996 matched to NOAA regions and GOES X-ray flares
26,214 of our group-days are matched to a NOAA region. The chart shows how often a group of each class had a flare on the day we observed it.
Left: by the first letter of our McIntosh class (A/B: no penumbra … F: largest bipolar groups, H: single spot with penumbra). Right: the full three-letter classes with the highest share of M or X flare days. Large groups whose biggest spot has a large asymmetric penumbra (“k”) and whose spots are packed together (“c”), such as FKC, EKC and DKC, flare far more often than small ones; this is the classical result of McIntosh (1990), here from our own classifications. Our first class letter agrees with NOAA’s on 47.0% of the matched group-days (all three letters: 16.2%).
The most active regions we observed
| NOAA | X | M | C | Largest | Our observation days | When we saw it |
|---|---|---|---|---|---|---|
| 13664 | 11 | 52 | 35 | X8.7 | 7 | 03.05.2024 – 13.05.2024 |
| 10808 | 10 | 20 | 47 | X17.0 | 8 | 08.09.2005 – 19.09.2005 |
| 14366 | 5 | 66 | 76 | X8.1 | 1 | 06.02.2026 |
| 12192 | 6 | 32 | 74 | X3.1 | 12 | 17.10.2014 – 29.10.2014 |
| 10486 | 7 | 19 | 16 | X17.4 | 9 | 24.10.2003 – 04.11.2003 |
| 13697 | 6 | 28 | 86 | X2.8 | 10 | 30.05.2024 – 09.06.2024 |
| 13663 | 5 | 35 | 40 | X4.5 | 5 | 03.05.2024 – 07.05.2024 |
| 10649 | 6 | 10 | 46 | X3.6 | 11 | 13.07.2004 – 23.07.2004 |
| 13842 | 4 | 27 | 19 | X9.0 | 11 | 28.09.2024 – 09.10.2024 |
| 10720 | 5 | 17 | 66 | X7.1 | 5 | 11.01.2005 – 21.01.2005 |
| 12673 | 4 | 26 | 54 | X9.3 | 11 | 29.08.2017 – 09.09.2017 |
| 14274 | 5 | 14 | 132 | X5.1 | 10 | 03.11.2025 – 15.11.2025 |
| 10375 | 3 | 29 | 49 | X1.7 | 14 | 01.06.2003 – 14.06.2003 |
| 9415 | 5 | 6 | 16 | X14.4 | 7 | 03.04.2001 – 11.04.2001 |
| 9393 | 3 | 24 | 28 | X20.0 | 2 | 24.03.2001 – 03.04.2001 |
Flare counts are for the whole life of the region (NOAA event lists). Click the number for the region’s day-by-day story, or a date for our drawing of that day. All regions we observed →
How is this computed?
- NOAA data: from the archive of the Space Weather Prediction Center of the US National Oceanic and Atmospheric Administration (NOAA/SWPC), for every day the Solar Region Summary (numbered regions: latitude, Carrington longitude, McIntosh class) and the edited event lists (X-ray flares seen by the GOES satellites, with class and source region). The archive starts in 1996 and is updated automatically every day.
- Matching: every group of ours since 1996 with a latitude and a Carrington longitude is compared with the NOAA region positions valid at 00 UT of the observation day and of the next day. The distance is the angle on the sphere. Within one observation the closest pairs are taken first, each group and each region used once, and the distance must be at most 6°. Unmatched groups (mostly small, short-lived groups NOAA did not number, or groups split differently by the observer) are left out.
- Flare: a group-day counts as “flaring” if its NOAA region produced at least one flare on the same UT day (the sheets give no observing time, so the same UT day is used instead of “the next 24 hours”). Only C, M and X flares with a source region are counted.
- Class: the McIntosh class given by our own observer (typed since 2007, read from the sheets 1996–2007). The letter chart shows classes with at least 30 group-days, the full-class chart those with at least 150.
- Limitations: GOES flare classes are as published (values before 2020 are not corrected for the later rescaling). NOAA gives no region for some flares; these are not assigned to any group. The matching is automatic and has not been checked case by case.
Regions, big flares and proton events since 1996, each region counted once
Which regions produce big flares?
Each region is counted once: grouped by the largest area NOAA reported, and split by whether it showed a “delta” magnetic configuration (opposite polarities within one penumbra) on at least one day. Bar: the percentage of these regions that produced at least one M or X flare during their life. Almost all large delta regions produced a big flare, very few small simple ones did. Hover for the number of regions; groups with fewer than 10 regions are not shown.
Big flares and proton events by year
Bars: number of M and X class flares in the year (GOES, NOAA lists). Diamonds: number of proton events (right axis). The numbers follow the solar cycle; 2024 had the most M flares since 1996.
How well do our records agree with NOAA?
For each year: the share of our groups matched to a NOAA region, the share of matched groups with the same first letter of the McIntosh class (left axis), and the mean position difference (right axis, degrees). Our positions differ from NOAA’s by 1–2° on average; NOAA gives whole degrees, so part of this is rounding. The 1996–2007 values were read from the observation sheets. Most unmatched groups are small groups NOAA did not number.
Active regions: the story of each region we observed →
How is this computed?
- Regions: every NOAA region listed in the Solar Region Summary since 1996, counted once. “In our drawings” means at least one of our groups was matched to it (same day, at most 6° apart in latitude and Carrington longitude).
- Big flares: GOES X-ray flares of class M and X with a source region in NOAA’s edited event lists. Area: the largest area NOAA reported for the region on any day. Delta: NOAA’s magnetic type contained “Delta” on at least one day.
- Proton events: NOAA/NCEI’s list of proton events (>10 MeV flux above 10 pfu) with their source region. “During our H-α observations”: an M/X flare whose start–end time overlaps a chromosphere FITS session or a Coronado sequence of ours (±10 minutes, camera clock).
- Agreement: the same comparison as on the admin page, summed per year: share of our groups with a NOAA match, mean |ΔB| and |ΔL| of the matched ones, and agreement of the first letter of the McIntosh class (groups with a McIntosh class; “_” counts as “x”).
Spots per group (f/g) Monthly and yearly f/g, 13-month smoothed
f/g = Σf / Σg over the observed days of the month (or year). It rises and falls with the solar cycle (large, complex groups near maximum), but a slow drift independent of the cycle points to a change in how spots are counted, for example a new observer, telescope or drawing practice. This is a standard consistency check for long sunspot series. The small marks at the bottom show the years in which the main observer (the one with the most observation days) changed for good (the new observer is also the main observer the following year); hover to see the names. Dashed: values read from the observation sheets.
How is this computed?
- Data: one record per observed day. Since 2007 the typed observation (their mean if there are several that day), before that the 1950–2007 observation sheet as read by a vision model, only if the sheet is marked usable. Unreadable sheets are never used.
- f/g: the sum of the spot counts Σf over the observed days of the month (or year) divided by the sum of the group counts Σg. Days without groups are not used. The ratio of the sums is taken instead of the mean of daily f/g, so days with a single group do not get too much weight.
- Smoothing: 13-month SIDC mean of the monthly values; no value if even one of the 13 months is missing.
- Main observer: the observer with the most observation days that year (typed observations and sheets together). Only lasting changes, where the new main observer is also the main observer the following year, are marked.
Full series: Monthly mean R since 1950 · Comparison with SILSO and k-factor · Records 1950–2007 and cycle model


