Is it possible for totality to be disrupted?
The common understanding of total solar eclipses is that when the eclipse becomes total at second contact, it remains as such till third contact, without any interruptions. This begs the following theoretical question: During totality, is it ever possible for the eclipse to briefly revert to be partial? In other words: Can totality ever be ”re-entrant”?
In the simplified world of Besselian elements computations, where the Moon is considered as a smooth
body, the answer is negative. When the full complexity of the lunar limb is taken into account, the answer is positive! Theoretically at least, and in extremely rare cases.
Definition of Re-Entrant Totality
A re-entrant total eclipse is a hypothetical eclipse that, after a preliminary extended partial phase, enters totality, briefly leaves it, re-enters totality, and finally leaves it to undergo another extended partial phase. This type of eclipse very much resembles an ordinary total eclipse, except for the fact that totality is disrupted by a brief period of partiality.
In terms of local circumstances, a re-entrant total eclipse has two distinct second contacts and two distinct third contacts. The unfolding of this hypothetical eclipse, from start to end is schematically exemplified by the following diagram:
The usual second contact C2, heralding the global start of totality, is followed by an internal third contact C3, signalling a brief reversion of the eclipse into partiality. Then an internal second contact C2 indicates the resumption of totality, and finally the usual third contact C3 marks the global end of totality.
Motivating the existence of Re-Entrant Totality
John Irwin had contemplated about the theoretical existence of re-entrant totality several years ago, after observing that the edges of umbral shadow outlines are slightly concave.
Under the simplified assumption that the lunar limb is perfectly smooth, umbral shadow outlines are smooth ellipses. When the topography of the lunar limb is accurately accounted for, umbral shadow outlines become complex polygonal-like shapes with usually many edges [1]. These edges are not perfectly straight lines, but they slightly bow inwards. At the confluence of two edges, there is a cusp.
We can imagine the following very unusual situation: an observer is in line with two umbral shadow outline cusps in such a way that s/he would enter totality, briefly exit it, and then re-enter totality again. The situation is depicted in the following figure, where details have been exaggerated to be able to convey the dynamics of what is happening:
Searching for Re-Entrant Total Eclipses during TSE2026
It will not come as a surprise that the very peculiar conditions for a re-entrant total eclipse to occur can only be found at the very edge of the totality path. This comes from the constraint that two umbral shadow outline cusps, and the intervening edge, need to be basically almost exactly in line with the observer.
While computing the true limb umbral path limits, John Irwin found occurrences of re-entrant totality in the remote Greenland North, during the upcoming total eclipse in 2026, in a small area, circa 150km south-south-west of Station North (a Danish military base) on Princess Ingeborg Peninsula in northern Kronprins Christian Land
As expected, the area where re-entrant totality occurs is at the very edge of the umbral shadow path.
Solar and lunar limb dynamics during a re-entrant totality
Re-entrant total eclipses are created by very peculiar configurations of the solar and lunar limb. The following figure shows an example of such dynamic for the case of the location in Northern Greenland
As during ordinary total solar eclipses, totality starts when the solar limb becomes tangent to the bottom of a lunar valley, indicated by C2 in the Figure, and is completely under the lunar limb anywhere else. A similar situation is replicated at the end of totality, with respect to the valley indicate by C3 in the same figure.
Between second and third contact, the solar limb moves with respect to the lunar limb in such a way that it pivots through the same very narrow region that we will call the pivot region of the eclipse. This behaviour is not specific to re-entrant total eclipses, but it is found any time the observation site is very close to the umbral shadow path edges [2].
What is utterly special in this case is that the pivot region is right under the bottom of a narrow and deep lunar valley. It is the interplay between the pivot region and this valley that creates the conditions for a re-entrant totality to occur.
The next figure shows a closer view of the pivot region and that particular lunar valley. At both second contact C2 and third contact C3, the solar limb is under the bottom of the lunar valley, and fully occulted by the Moon’s disk. In the period in between, the solar limb lifts up, and briefly enters the lunar valley, uncovering a tiny fraction of photosphere: totality briefly reverts to partiality! Totality become re-entrant
The following able presents the internal contact times for the re-entrant totality for the site in Northern Greenland. Totality lasts for ca, 9s, but it is interrupted for ca. 2s by a brief and incredibly inconspicuous
partiality.
Conclusion
Re-entrant totality is an extreme limit case, requiring incredibly peculiar conditions. It is definitely possible in theory, as we have demonstrated. However even tiny variations of initial conditions are
able to destroy re-entrant totality entirely.
The eclipse solar radius is not precisely know: its likely value is 959.95 ± 0.05 ” [3]. In all previous computations, the eclipse solar radius has been taken as 959.95”. Increasing the solar radius to 960.00” makes totality impossible, and the eclipse is purely partial. Decreasing the solar radius to 959.90” prevents totality from being disrupted, and the eclipse is purely total.
Even if re-entrant totality exists, we can ponder about its observability. No one really knows, but it might be detectable by observing the eclipse flash spectrum. One more piece of the puzzle of that treasure trove found at the edges of the umbral shadow path.
Refernces:
[1] Ernie Wright et al. A raster-oriented method for creating eclipse maps. The Astronomical Journal, 168(163), 2024.
[2] Luca Quaglia et al. Ase2023: Eclipse solar radius estimation from a high-quality video recording. Journal for Occultation Astronomy, 14(2):3–14, 2024.
[3] Luca Quaglia et al. Estimation of the eclipse solar radius by flash spectrum video analysis. The Astrophysical Journal Supplement Series, 256(36), 2021.
This post has been presented as a poster at the SEC 2025 in Leuven – 14-15 June 2025