Saturday’s Lunar Eclipse: Not Total?

Most sources say April 4th's lunar eclipse will be total, though only barely so. However, those calculations have overlooked a subtle factor that might render the event only "partial."

All week our emails (and, yes, our phones) have been buzzing with over-the-top stories about how tomorrow's total lunar eclipse will be the "shortest of the century." Well, maybe.

Events during April's lunar eclipse

Those with clear skies early on April 4th get to witness a total lunar eclipse. But it's only barely total, as the Moon's northern limb just manages to slip completely inside Earth's umbral shadow. So the northern half of its disk should look especially bright during totality.
Sky & Telescope diagram

As we've pointed out, it all depends on how one define's the outer edge of Earth central shadow, the umbra. The editors at Sky & Telescope have long used predictions from the U.S. Naval Observatory, which in turn are ground out at Her Majesty's Nautical Almanac Office (HMNAO) in England. Those state that tomorrow's totality will last 12.2 minutes. Eclipse veteran Fred Espenak is countering with 4.7 minutes, as is NASA's eclipse website (which is no longer being updated but still utilizes calculations made by Espenak before he retired from NASA).

This morning, while perusing the comments posted about S&'s various postings — yes, we do read them — I found a provocative entry from Chicago-based amateur Curt Renz. "Actually, there will be only a partial (nearly total) lunar eclipse during Saturday morning," Renz writes. "The Earth’s umbra will only cover 99.5% of the Moon’s diameter." He goes on to explain that USNO, NASA, and Espenak all assume Earth’s umbra to be perfectly circular, whereas they really should be taking into account Earth's oblate shape and, therefore, the elliptically-shaped umbra that results.

Is Renz right? I'll keep you in suspense — first, let's get some background.

Umbra color schematic

If Earth had no atmosphere, the Moon would look completely black during a total lunar eclipse. However, a little of the Sun's red light refracts through the atmosphere and into Earth's umbra, coloring the lunar disk during totality. The atmosphere also makes the umbra appear 1% to 2% larger than it should be. (Not to scale.)
Sky & Telescope illustration

If our planet had no atmosphere, then (aside for the ramifications of an airless Earth to life) it would cast a shadow with an exact edge that's geometrically proportionate to our planet's shape. However, in reality, the size of the umbral shadow that falls on the Moon is bigger than it should be, thanks to an enlargement caused by our atmosphere. We know this from decades of careful timings, no doubt made by some of you, of exactly when the umbra sweeps over particular lunar craters during an eclipse.

Moreover, the edge of the umbra is fuzzy, and the meteorological conditions on Earth vary from eclipse to eclipse. Erich Karkoschka, a University of Arizona researcher, explored these variables at great length in Sky & Telescope's September 1996 issue. It's not just volcanic ash in the stratosphere (which isn't all that common anyway), but also the presence of cirrus and stratus clouds around Earth's day-night terminator and the ozone concentration.

So the enlargement values can vary. USNO public-affairs specialist Geoffrey Chester explains: "HMNAO (and thus USNO) use the Chauvenet convention, effectively increasing the size of the umbral shadow at the Moon's distance by a factor of 2%. NASA uses the more recent Danjon model, which is based on crater timings of past eclipses and gives a figure of about 1%." Roger Sinnott, a longtime S&T staffer (now a senior contributing editor) who painstakingly compiled crater timings for decades, notes that the Danjon correction is closer to 1.7%. "Both Chauvenet and Danjon used crater timings to determine the size of the umbra, he adds, "so they are both observationally based."

Earth is not a perfect sphere — its polar diameter is only 99.67% of its equatorial diameter, a difference of about 21 km (13 miles) in radius. And Renz is correct that the most common models assume Earth to be a perfect sphere, though there's a partial compensation for oblateness by taking the radius at a latitude of 45°. But still . . .

Earlier today I contacted both Espenak and Meeus, and both find Renz's assertion intriguing. Meeus apparently conceded to Renz that, if Earth's true shape is considered, the eclipse should be deemed partial. However, "it's more complicated than just the polar radius," counters Espenak. "You've got to calculate the ellipticity of Earth projected onto the fundamental plane at the time of the eclipse, and [include Earth's] inclination toward or away from the Sun." He hinted that he might tackle those details — but not until this eclipse has come and gone. (It's supposed to be clear in southern Arizona, and he's readying his cameras.)

Moon crossing into Earth's umbra

A long-exposure image shows a partially eclipsed Moon. Part of the disk is not yet inside Earth's umbra (at right), while the part inside the umbra (at left) already shows dramatic red coloring.
Johnny Horne

So, let's throw it out to you — the observers! If you're in western North America (where totality will be observable) and your skies are clear, use binoculars or a telescope to judge whether that last little sliver of the lunar disk dives into the Umbra or not, and for how long. (No cheating using a camera — there'll be too much variation in exposures from one setup to the next to make the results comparable.) Then let us know, via the comment section below, what you find.

According to the experts, it'll be very difficult to judge. "I have almost no hope that observation will indicate whether the eclipse will be just total or only partial," says Meeus, "because the edge of the shadow is too diffuse." Espenak agrees: "I do not think observers will be able to time the contacts very well nor to distinguish between an 'umbral magnitude = 1.001' total eclipse and a 'umbral magnitude = 0.999' partial eclipse."

One thing's for certain: whether "barely total" or "extreme partial," tomorrow morning's view will be unusual. With the Moon passing so far north within the umbra, at maximum eclipse its northern limb should look much brighter than the southern side.

18 thoughts on “Saturday’s Lunar Eclipse: Not Total?

  1. Helio C. VitalHelio C. Vital

    Current sophisticated models for the umbra that properly account for Earth`s oblateness have yielded predictions in good agreement with observations, even for contact times at large umbral angles. A recent paper published by Roger Sinnott and David Herald has clearly shown that. But then another problem remains for this critical test of lunar eclipse calculations: the contribution of our atmosphere varies unpredictably from one eclipse to another, ranging between 1.2 and 1.5% of the moon`s equatorial horizontal parallax. For this eclipse, I have calculated that if that contribution happens to be less than 1.25%, a deep partial eclipse will occur. Such limit would be 1.6 standard deviation below the mean of 1.34+-0.06% (in close agreement with Sinnott`s 1.36% figure) found in our analyses of 30 eclipses and more than seven thousand crater and limb timings. Thus, if we assume that a normal distribution applies, one can calculate as 5% only the chance of that happening, small but not negligible. Last week we posted an article addressing that issue on our eclipse webpages at
    Furthermore, as Fred Espenak has pointed out, the fuzzy border of the umbra is expected to be very difficult to pinpoint, then many contact times will be needed so that a meaningful statistics can help us tell exactly what really happened.

  2. Nakedgun

    To my naked-eye, the brightness along the northern rim made it seem less than total, but, what percentage is impossible for me to gauge. It was, of course, a beautiful celestial event.

  3. Robin-Clark

    From my location in Tucson it was not quite total. From about the 12 to 2 o’clock position, there was still a sliver of moon that never quite made it into the umbra.

  4. alastair

    Having just observed the eclipse on the Coromandel Peninsula of New Zealand’s North Island I am sure this was not a fully total eclipse. A thin sliver of the limb was still quite brightly illuminated throughout totality. It’ll be interesting to hear what others saw.

  5. Curt-Renz

    Thanks, Kelly, for highlighting my conclusion that the eclipse would be best described as partial. Indeed, I must concede that the lack of sharpness at the umbra edge means the matter is better decided by observation than theory. I simply felt that the media were too readily hyping the eclipse to be total by referencing the usual authorities, whom I knew assumed the Earth to be perfectly spherical for the purpose of estimating the shape of the umbra. Forgive me if I appear to be disparaging the media, but before I retired my career was in TV news, and I understand the tendency to sensationalize.

    I should note that I do not consider the polar radius of the Earth to be what determines the radius of the umbra at the point where the Moon meets it during a borderline eclipse. What matters is the radius of the Earth within several degrees of an arctic/antarctic circle. The precise determination is complicated, as Espenak notes, but it is what I do.

    In his books Meeus mentions that eclipse predictors, including himself, assume a perfectly spherical Earth and resultant circular umbra. However, in one of his chapters Meeus explained how the umbra determination would be altered if the true shape of the Earth were considered. I took up the challenge and wrote a program that does exactly that. As you note, Meeus and I exchanged emails in November in which he conceded that this morning’s eclipse may be partial.

    Last year I emailed Espenak at his NASA address several times and received no reply. Apparently that was because he had retired. However, he and I did converse earlier this week within the comments section of a previous Sky & Telescope lunar eclipse article. In that forum he also conceded the possibility that the Moon this morning might not become completely immersed in the umbra. It would be interesting to learn if he un-retires and refines some of his earlier eclipse predictions within the NASA website.

  6. Curt-Renz

    Kelly, I should also mention that you correctly described the traditional enlargement factor for the umbra and penumbra as 2%. However the Danjon Model, which Meeus and I call the French Rule, enlarges the radius of the Earth by 1%, not the shadows. The enlargement of the shadows must then be determined by trigonometry and varies among eclipses as Sinnott seems to imply. The enlargement of the umbra and penumbra would each be by different amounts.

  7. SonomaCharlie

    From Sonoma CA just north of San Francisco, I vote for NOT total. There was always a unique white spot on the North limb. The East and West limbs also looked brighter than the center, but I put that down to their being seen adjacent to the dark sky.

  8. John

    I watched the eclipse through binoculars and took pictures at maximum and as far as I can tell, the moon never was completely eclipsed. There was a tiny sliver that remained and it prevented me from getting a picture of the deep reds without having a flare on the bright side. I’m sure clouds played some role in this, but I think the only effect from the clouds was to highlight the sliver that escaped the earth’s umbra. I have to agree with Renz on this one: The Tetrad was spoiled!

  9. Eliot-Herman

    Predictions are great, data is better, my observation this morning is that Curt Renz nailed it, it never was completely total, a thin sliver of the moon remained throughout the predicted totality. The deepest eclipse I observed was just after 4:59 am PDT and it was sharply different a minute before and a minute after. At 4:59 the remnant sliver turned from white to a light brown that I interpret as the very edge of the shadow and even then just for seconds. Unfortunately my observing was through scattered clouds in Tucson degrading images. I sent S&T one minute images 4:58-5:02 am PDT that maybe they will post that illustrates this. Congrats to Curt Renz for the prediction.

  10. Anthony BarreiroAnthony Barreiro

    I observed the eclipse from San Francisco, naked eye and through mounted 11×56 binoculars. The weather was mostly clear with slight haze, and no clouds over the Moon until the very end of the second partial phase, just before moonset. At no time did the eclipse appear total to me. The Moon’s far northern limb remained brightly illuminated throughout. Inside the apparent edge of the umbra the Moon’s surface was bright but obviously reddened. The very edge of the Moon remained much brighter and without any visible reddening. Just my two cents.

  11. Larry-Koehn

    It would be nice if this same concern, about discrepancies, could be shown for solar eclipse data as well. USNO personnel have their own way of calculating solar eclipse times as does Fred Espenak. For the August 21, 2017 total solar eclipse, USNO says maximum duration of totality will be 2 minutes and 44.6 seconds. Fred Espenak states 2 minutes and 40.2 seconds. USNO says that maximum duration can be found a few miles south of Carbondale Illinois. Fred Espenak has two points: one for “greatest duration” near Goreville Illinois and one for “greatest eclipse” near Hopkinsville Kentucky.

    I read in a press release the other day that NASA was heading to Carbondale Illinois for the eclipse. Two years ago, everyone was heading to Hopkinsville, because they thought NASA eclipse information was up-to-date! Now I read, from Sky & Telescope, today, that NASA’s eclipse website ” is no longer being updated.” I knew that back in 2013.

    NASA’s eclipse website was taken down for approximately ten days last year (February 2014) with a redirect to USNO, because the information was outdated (according to the NASA website curator at the time). Unfortunately, there is nothing on the NASA eclipse website stating why the site was taken down for ten days, nor why it is still up with dated information right now. I still think most people believe the NASA eclipse website is current as of today.

    If NASA wants the eclipse site to stay up, it should be current! Unfortunately, there are no NASA personnel handling or updating the eclipse information after Fred Espenak’s retirement back in 2009. Yet, most people will still go there. Why? Because it’s NASA!

    1. Dave Whipps

      I’m proud to say that our application, Starry Night did a great job correctly predicting this eclipse as a partial!

      The kicker (I think), in this instance is that that Starry Night does take the Earth’s oblateness into account when calculating the not-quite-circular shadow cone, as projected onto the Moon.

      The NASA (and other) predictions of a “total eclipse” prompted us to do some exhaustive, line-by-line checking, and with cross-checks by Larry, we included a further minute correction to the shadow cone for the Earth’s atmosphere as explained by Fred Espenak (who is still active, on his own site) here:

      In the end, Starry Night’s simulation of the Eclipse matches the actual, observed event better than (any of?) the sources mentioned above!

  12. Scott-Bulkley

    As requested in above the article, I am providing feedback on my observations of the Lunar Eclipse of April 4. I have been viewing both total and partial eclipses of the Moon for over 50 years, and I have seen most of them available from my location, as long as they weren’t “rained out.” But, I have never experienced one like this last Saturday morning. I used naked eye observations, 10 x 50 binoculars, and a 125 mm SC telescope. Using all three viewing methods during the critical minutes, I was not able to observe totality at any time. There was always at least a very thin “sliver” of brightly illuminated lunar surface on the northern limb of the Moon, and this was even clearly seen without any optical aid. However, despite the lack of totality, I must admit that this was one of the most exciting lunar eclipses that I have ever witnessed due to the anticipation and controversy regarding its definition, and the duration of time that some portion of the Moon spent in the umbra. In the end, I can honestly say that this was by far the best partial eclipse of the Moon that I have ever seen!


    I really appreciated your writing, Kelly.

    I managed to share this eclipse live, via the Virtual Telescope website. Images were captured by a great team of collaborators from AU, Northern and Central Americas. Images from very different locations added a special value to the experience. It was great to see how the eclipse evolved, while waiting for totality, the announced “tiny” totality.

    That marginal expected totality was a factor making, for me, this eclipse, a well worth event to wait and look for.

    Images at the maximum eclipse, received from our observers, showed a surviving, narrow and bright limb, making the answer if it was real totality or not quite appropriate. It would be great to compare our experiences on this, and your article is such a great motivation.

    I will try to look at the images better and would love to hear from others, as some already did, what they think. Thanks for this article, motivating such a discussion.

    Our experience was possible thanks to a number of collaborating imagers, I wish to thank them for making all this possible:

    Thanks again, Kelly, and keep up the great work


    Lunar eclipses are not like solar eclipses in terms of exact calculation. They’re more like sunsets. Anytime the Earth’s atmosphere is involved, you’ll be dealing with a phenomenon subject to the variability of weather and other local conditions. In the case of sunsets, astronomers have a long-standing calculation tradition which geometrically determines the instant when a sun of fixed size would be some fixed distance below the true horizon (the arbitrary sum is 50′). For an observer at sea level (sitting right at the surface of the sea) on an average day with average refraction conditions, this approximately matches the time when the upper limb of the Sun reaches the sea horizon. But that same definition is used to compile tables for times of sunset at every other location on Earth (more or less) under any weather conditions. Why? Because the sunset calculation “rule” provides a common standard of comparison. We could do better. We could improve the calculation of sunset times. But to what end?? … The precise time of a lunar eclipse is quite similar. We’re literally talking about observing the effect of sunset, through the Earth’s atmosphere, from the Moon’s surface. That event marks the entry of a point on the lunar surface into the umbra. The exact timing of this event is weather-dependent and it’s definition-dependent. Improvements can be made in lunar eclipse calculations, but ultimately we face the same problem. There is an arbitrariness to these calculations, and at some point all you can do is say “we have a rule for this”. This becomes the standard of comparison for lunar eclipses. By contrast the calculation of solar eclipses has recently been significantly improved at a very high level of accuracy thanks to detailed digital elevation models of the lunar surface. The calculation of solar eclipses is not subject to the variability of lunar eclipse calculations thanks to two unique aspects of the problem: 1) the lack of atmosphere on the Moon (which makes that digital elevation data so important), and 2) the extraordinary feature of the Sun’s atmosphere that makes all of this possible –an exceeding sharp, well-defined, and invariant edge to the solar photosphere (which is a happy “accident” in our case, a property of some, but not all, stellar atmospheres). The standards of solar eclipse calculations cannot be applied to lunar eclipse calculations. Times of events in lunar eclipses have to be defined by a standardized calculation using some agreed-upon mean conditions, or they have to be built on detailed models of the Earth’s atmosphere on the actual date of the lunar eclipse, which makes them unpredictable except on a very short time scale.

    1. Anthony BarreiroAnthony Barreiro

      Thanks Frank, this is very helpful. So there seem to be three conditions for lunar eclipse predictions: definitely total, definitely partial, and too close to call. Saturday’s eclipse at a standardized predicted extent of 1.005 apparently was too close to call. I’m curious how the boundaries of the “too close to call” condition could be defined, and how frequently these fascinating liminal eclipses occur.

  15. Graham-Wolf

    Hello Kelly (and many others).
    Observed the Lunar Eclipse from Lower Hutt, NZ (just north of Wellington). Used a 26cm f4.9 Dobbie with a GSO SV15 and 2 inch GSO ED Barlow. Did NOT (quite) reach totality at any stage… took a few minutes for my brain to work out what was actually happening. Terrestrially orientated, I could still see a 30km thick by ~ 400km wide bright white sliver that widened again after about 8 to 10 minutes of what I’d call “pseudo-totality”. Frankly, I’d call it a “very near-Total Eclipse”.

    Crater timings were undertaken as per normal for the last 5 decades or so, here in NZ. Personally, I much prefer to time 1C and 2C, rather than the Earth’s shadow halfway across the crater… it’s so much less “hit and miss”! You really need excellent optics and 300x or so with good lunar/planetary eyepieces. I rather like the under-rated Konig for this purpose (offers generous eye-relief, and Barlow-ability).

    Barlowing to create an say f15 -25 system (based on decades of personal experience with Lunar Eclipses) also sharpens up the shadow images by tightening the light cone to the eyepiece. Nice photos of the eclipse, were taken the other night. Placed a Canon A480 (8Mpx) Compact Digi-camera to the eyepiece, set to Programme Auto, 800 ISO (1600 is too “noisy”), and engaged a 2s shutter delay. You really need to have the camera lens held centrally over the eye-lens of the eyepieces, and as perpendicular as possible. Carefully monitor the viewfinder.

    I also (as usual) did Danjon estimates from ~ 1/2 before totality to ~ 1/2 after…. they didn’t vary that much at all. I was taken by the increased darkness compared to last October’s effort. Will be publishing my data and photos soon in the monthly astro-journal “T.A.”

    Graham W. Wolf:-
    Barber Grove Observatory, Lower Hutt, NZ

Comments are closed.