A lunar eclipse begins, in the sense most people mean, when the Moon first touches Earth's outer shadow (the penumbra) — but the start time printed on most eclipse schedules is roughly 25 to 40 minutes later, when the Moon edges into the dark umbra at penumbral-then-partial phase. The exact clock time for any date depends on your time zone: the same event begins at a single instant worldwide, but local clocks read differently. Two additional start times matter — the ~30-minute visual threshold before most observers notice penumbral shading, and the totality contact — and professional predictions can drift by a minute or two because Earth's atmosphere makes the penumbra's edge fuzzy.
How shadow geometry turns an instant into a schedule
A lunar eclipse is a geometry problem with a clock attached. Earth always casts a cone-shaped shadow into space, and the Moon only crosses that cone at full moon, when the Sun, Earth, and Moon line up closely. The cone has two zones that matter for eclipse timekeeping. The umbra is the inner core where Earth blocks all direct sunlight. The penumbra is the outer halo where Earth blocks only part of the solar disk, so the Moon receives a partial, dimmed illumination.
Because the Sun is not a point source, the shadow's edges are not sharp lines. The wider the light source relative to the blocking body, the more gradual the shadow boundary — a fact that drives most of the prediction uncertainty described below.
Eclipse contact times are defined by tangency. The Moon's limb (apparent edge) touches the penumbra's edge at the entry now called P1. The limb first contacts the umbra — where the eye perceives real darkness — at U1. The limb finally clears the umbra at U4. These are geometric definitions based on idealized circular shadows. The actual brightness transition is gradual over roughly 10 to 20 arcminutes of lunar travel.
That spread creates a practical gap: a geometric start time and a perceived start time differ. Your eye needs more of the Moon inside the umbra before the dimming reads as an eclipse. Two observers with different eyesight, telescopes, and horizons will disagree by several minutes about when the fun started.
A deeper layer: the Moon's apparent diameter roughly equals the Sun's in the sky, but Earth's shadow at the Moon's distance is far larger. The umbra's diameter at the Moon is about 2.6 times the Moon's apparent diameter. So a lunar eclipse can last for hours — unlike a solar totality measured in minutes.
One comparison makes the mechanism concrete. A solar eclipse is a small disk (the Moon) covering a large one (the Sun). A lunar eclipse is a large disk (Earth's shadow) crossing a small one. This is why lunar eclipses are slow, widely visible from the night hemisphere, and generous with time.
Quantified context: start times, durations, and how far ahead we can predict
The best predictor of when a lunar eclipse will begin is the recurrence pattern, not a rule of thumb. Lunar and solar eclipses come in families because the Moon's orbit is inclined about 5.1 degrees to the ecliptic. Most full moons miss the shadow cone. When a full moon happens near a node, an eclipse occurs.
Lunar eclipses cluster in seasons, roughly two per year, about 173 days apart on average. Solar eclipses come in corresponding pairs. The 18-year, 11-day saros cycle links eclipses with similar geometry — a useful cross-check when verifying a predicted date.
There are three eclipse types. A penumbral eclipse passes only through the penumbra, producing subtle shading that most casual observers miss. A partial eclipse includes a bite taken out of the Moon. A total eclipse puts the entire Moon in the umbra, producing the familiar copper-red color.
Expect these rough numbers for total eclipses. The interval from P1 to the end of totality usually spans roughly 3 to 6 hours. The umbral partial phase from U1 to totality can run about an hour or more. Totality itself typically lasts between 30 minutes and about 1 hour 45 minutes.
Predictions come from ephemerides — computed tables of positions — not guesswork. NASA and the Jet Propulsion Laboratory publish five-millennium catalogs of eclipse times. Modern positions are accurate to seconds of arc, so the geometry is not the bottleneck.
The bottleneck is the shadow radius. Prediction groups fit the umbra's size to observations from past eclipses and advertise an enlargement factor. Different groups use different factors, so U1 and U4 can disagree by a minute or two. The start of a penumbral eclipse is even softer because penumbral shading has no obvious edge.
A counterintuitive detail: the start of a lunar eclipse is defined astronomically, but it is not a sharply observable event. The Moon's limb does not visibly darken exactly at the computed instant. Instead, the shading ramps up over several minutes. This is why photographers who want a clean sequence often start shooting 10 to 15 minutes before the published start.
The bottom line for planning: treat published start times as reliable to within a couple of minutes for the umbral contacts, and to within about five minutes for penumbral contacts. Anything finer is precision without consequence.
Timeline of a typical total lunar eclipse
| Phase | Contact | What happens | Typical duration |
|---|---|---|---|
| Penumbral entry | P1 | Moon's limb enters the outer penumbra; shading invisible at first | — |
| Penumbral-only interval | P1 to U1 | Shading slowly deepens; usually subtle, best seen on photos | ~20 to 40 min |
| Umbral entry | U1 | Limb touches the umbra; a visible dark bite appears | — |
| Partial phase | U1 to U2 | Dark bite grows across the disk; color becomes noticeable | ~1 to 1.5 hr |
| Totality begins | U2 | Entire Moon inside the umbra; copper-red color | — |
| Totality ends | U3 | Limb exits the umbra; red color fades | ~30 to 105 min |
| Partial exit | U3 to U4 | Dark bite shrinks; Moon brightens limb-first | ~1 to 1.5 hr |
| Penumbral exit | U4 to P4 | Moon fully leaves shadow; residual shading fades | ~20 to 40 min |
Read this as a shape, not a schedule: total duration, from P1 to P4, usually runs about 5 to 7 hours for a total eclipse, with totality occupying a minority of that window. Your local published start time corresponds to the contact closest to your phase of interest.
Edge cases that move the start time in your experience
Three factors change what "begin" means in practice.
First, the observing method. Naked-eye observers miss penumbral shading entirely. Binoculars and long-exposure photography reveal it better. If your question is "when can I see something?", the answer is U1, not P1.
Second, position on the Moon's disk. The limb that contacts the umbra first is the hemisphere facing Earth's shadow edge. Because the Moon moves its own diameter in about 50 minutes relative to the shadow, the visible dark bite appears mid-shadow at U1 and takes several minutes to become obvious.
Third, horizon geometry. If the Moon rises under a deep eclipse in progress, the local start time is irrelevant — you see a partially eclipsed Moon at moonset or moonrise. Skyglow, haze, and the Moon's altitude all compress or stretch the visible phase window.
The professional caveat here: model-based contact times use an artificial "enlargement" of Earth's shadow to match observers' reports. For scientific purposes, the exact instant the Moon's limb crosses the geometric umbra edge is less important than the brightness profile over time.
Common misconceptions about eclipse start times
The most common error is confusing the penumbral start (P1) with when darkness begins. P1 is an entry into a bright-ish part of the shadow; the Moon is barely dimmed there. The umbral start (U1) is what most schedules should advertise first.
Misconception 2: "Lunar eclipses last as long as solar eclipses." They do not. Lunar totality can last over an hour; solar totality rarely exceeds about 7 minutes. The shadow geometry explains it.
Misconception 3: "The Moon disappears." It does not. During totality it is lit by sunlight refracted through Earth's atmosphere and appears red. The eclipse is a color and brightness change, not a blackout.
Misconception 4: "The date is the same everywhere." The date is the same instant, but local dates can differ near the International Date Line. A single predicted time can fall on two calendar dates depending on time zone.
Misconception 5: "Prediction uncertainty means we don't know." We know the geometry to seconds. The uncertainty is in a soft visual threshold, not in the orbital calculation. Warnings about misleading eclipse "start times" floating around social media usually come from mixing the wrong contact with a local clock.
Practical actions to pin the start time to your date and location
There is no single global start time for a lunar eclipse, so the real task is to convert the published UTC contacts to your local clock and choose the phase you care about.
Step 1: Get the authoritative contact list. Use a prediction catalog from a space agency or observatory that gives P1, U1, U2, U3, U4, and P4 in UTC. Verify your date against the saros pattern and the node timing — if the predicted date disagrees with a known eclipse family, recheck the source.
Step 2: Convert UTC to your local offset, then add daylight saving if your jurisdiction applies it. One mistake here shifts every time by a full hour.
Step 3: Choose your threshold. If you want to observe the partial bite, use U1. If you want subtle shading, use P1 minus 10 minutes. If you only care about totality, use U2.
Step 4: Account for local moonrise and moonset. If the Moon is below the horizon at the contact, you cannot see it. Check the Moon's altitude at each contact, not just the eclipse date.
Step 5: Prepare for the bright Moon — observers find the eclipse easiest against a dark sky. Binoculars around 7x50 to 10x42 help. A tripod-mounted camera with exposures between 1 and 30 seconds captures the color gradient better than the eye.
Step 6: Add a margin. Start observing about 10 to 15 minutes before your chosen contact to catch the transition.
A quick sanity threshold: if a published "start" time is more than about 15 minutes different from what a careful observer reports, either the source used a different contact or you have a time-zone error. Not a bad early check.
Limitations and what the start time cannot tell you
The published start time does not tell you what the eclipse will look like. Brightness and color depend on how much dust and ozone the sunlight passes through in Earth's upper atmosphere. Recent volcanic injections can darken and redden totality. A schedule cannot predict that.
The start time does not tell you the Moon's altitude or the local seeing. Those are local computations.
The start time does not resolve to a single visual instant, because the shadow edge is physically gradual. We quantify start times to the minute and then must stop, because finer resolution implies precision the eye cannot deliver.
Predictions beyond the current catalog limits rely on extrapolated models. For dates many centuries out, uncertainty grows through perturbations in Earth's orbit. That is an orbital question, not an observing question.
Finally, eclipse start times do not change your obligations. For public gatherings, use the umbral contacts as your planning anchors, and always state the time zone. A start time without a zone is half an answer.