Home › Features › Lake District sunrise

Lake District sunrise: when the sun really clears the skyline

The published sunrise and sunset times assume a flat horizon. On a fell, you are looking at other fells — at both ends of the day.

The sunrise time in your weather app is computed for a flat horizon at sea level. Stand on a fell and that is not the world you are in. You are looking at other fells, and the sun has to clear them. On Carl Side the gap is +95 minutes — the sun is long up while you are still standing in shadow, because Skiddaw is in the way.

We have computed that moment for 214 Lakeland summits, from each summit’s own 360° horizon profile. Every time on this page carries a ± band, because a figure derived from a 30 m terrain model has no business claiming a single minute.

Why the published sunrise is wrong on a fell

Almanac sunrise is a convention: the moment the centre of the sun reaches −0.833° below the horizontal. That number already contains the average atmospheric refraction near the horizon and half the sun’s width, which is why it is slightly below zero. It is the right answer for a flat sea horizon, and it is what almost every sunrise time you will ever read is based on.

A summit is a different question. From up there the skyline is made of terrain, and that terrain sits at its own angle in every direction — above the horizontal where a bigger fell stands across the valley, below it where the ground falls away. First light is the first minute the sun’s upper edge clears that real skyline, at the exact compass bearing where the sun happens to be rising on that date.

The same summit can be minutes early in one season and an hour late in another, because the sun rises at a different bearing through the year and meets a different piece of skyline each time.

Of the 214 summits we have profiles for, 98 wait meaningfully longer than the almanac says at midsummer, and 34 catch the sun before the flat-horizon time because their eastern ground falls away.

The fells that keep you waiting

Midsummer (2026), sorted by how long after almanac sunrise the light actually arrives. Times are Lakeland clock time.

These are not big remote mountains. They are modest tops standing in the shadow of something much larger — which is exactly why the effect is so strong. A summit is only as early as its eastern skyline allows.

The band sits on the first-light figure because that is the one derived from a terrain model. The almanac column carries none: it is the flat-horizon convention, exact by definition, and attaching an error bar to it would be inventing uncertainty that is not there.

The fells that get it first

The opposite case: ground that falls away to the east, so the sun arrives before the flat-horizon time.

The margins here are small — single-digit minutes, comfortably inside the band on the figures either side of them. Treat this table as a description of the shape of the ground, not as a reason to choose one summit over another by a few minutes.

The other end of the day

Everything above runs backwards at dusk. The same skyline that makes a summit wait for the sun also takes it away again, and the same solve gives the moment the sun’s upper edge drops behind the real western horizon — on the same date, with the same ± band.

Castle Crag loses the sun at 19:30 BST on midsummer’s evening. Helvellyn, 9 km away and on the same evening, keeps it until 22:00 BST. That is 2 hours 30 minutes of difference between two Lakeland summits at the same moment in the year, and none of it is weather.

The fells that lose it first

Sorted by how long before almanac sunset the light actually goes. These are the deep ones — low tops with something much bigger standing to the west of them.

The fells that hold it longest

And the opposite: ground that falls away to the west, so the sun stays visible a little past the flat-horizon time.

Across the 214 summits, 48 hold the light past the valley’s own sunset and 88 are in shadow well before it.

Notice how lopsided those two tables are. The most any of the 214 summits gains is 7 minutes at dawn or 8 minutes at dusk — a handful either way. The most any of them loses is 101 minutes waiting in the morning and 143 minutes gone early in the evening. There is a reason, and it is geometry rather than luck: to gain time you need the ground to fall away, and it can only fall so far below your feet. To lose time you need a fell in the way, and a fell can stand many degrees up. Summits lose light far more easily than they gain it, at both ends of the day.

And it is the same fell at both ends. Castle Crag does not only lose the sun first — it also waits 39 minutes for the sun in the morning. Add the two together and that is about 3 hours 2 minutes of midsummer daylight it never sees, because the same ring of higher ground is doing both jobs. Morning and evening are not two facts about a summit. They are one horizon, read twice.

On this date, who gets it first?

Everything above answers “when does my fell get the sun?”. This turns the question round: pick a date, and see which summits catch it first and which hold it longest.

Ranked on the clock time shown, across the 214 summits with a computed profile. The bearing and terrain angle beside each one are the geometry that puts it there: the compass direction the sun crosses the skyline, and how high that skyline stands at exactly that bearing — a negative angle means the ground has fallen away below you.

These are the sixteen dates the dataset models — the two solstices, the two equinoxes and the first of each month — not any date you like. The per-fell date pickers will solve an arbitrary date for one summit; ranking all 214 of them needs every fell solved on the same day, and we only publish the days that were solved properly, with the error bands that were computed properly.

This is a geometry table and not a recommendation. Being first to see the sun is a fact about a summit’s eastern skyline; it is not a reason to be standing on it in the dark, and we still publish no time to set off or start down.

The dates people ask about

Midsummer

On 21 June, the longest day, four summits share first light at 04:29 BST ±5 min — Blencathra - Hallsfell Top, Bowscale Fell, Skiddaw Little Man and Skiddaw. At the other end, two summits share last light at 22:00 BST ±5 min — Grisedale Pike and Helvellyn.

Midwinter

On 21 December, the shortest day, seven summits share first light at 08:26 GMT ±5 min — Bowfell, Crinkle Crags - Long Top (Second Crinkle), Esk Pike, Great End and Scafell Pike among them. At the other end, two summits share last light at 15:59 GMT ±5 min — Scafell Pike and Scafell.

New Year’s Day

On 1 January, the first morning of the year, ten summits share first light at 08:28 GMT ±5 min — Bowfell, Crinkle Crags - Long Top (Second Crinkle), Esk Pike, Great Carrs and Great End among them. At the other end, three summits share last light at 16:08 GMT ±5 min — Pillar, Scafell Pike and Scafell.

Christmas Day is deliberately absent. The dataset models sixteen dates and 25 December is not one of them; the nearest is the winter solstice, four days earlier. We would rather say that than quietly relabel the solstice as Christmas.

How it is computed, and how wrong it can be

Computed, not photographed. Nobody stood on 214 summits at dawn with a stopwatch. Every figure here is modelled, from the same terrain data behind our computed panoramas.

• The skyline comes from Copernicus GLO-30, a 30 m elevation model, ray-cast around the full 360° from each summit to give the terrain angle at every bearing.
• The sun comes from the standard NOAA solar position algorithm, stepped minute by minute through the day.
• Refraction and the upper limb are applied to the sun before the comparison: the atmosphere lifts the low sun by roughly half a degree, and it is the top edge of the disc that appears first, not the centre. Almanac sunrise deliberately does not get this treatment, because its −0.833° convention already includes it — applying it twice is a real and easy mistake.

The band is the honest part. A 30 m model smooths arêtes, rounds off cornices and cannot see a boulder, so the skyline it gives is approximately right rather than exactly right. We derive the error from the elevation model’s own accuracy at the distance of whatever the sun is clearing, convert it into minutes using the sun’s climb rate at that moment, and floor it at ±5 minutes. No figure on this site claims better than that.

Which means: use these times to know whether a summit is an early one or a late one, and to know roughly how long you will be standing in shadow. Do not use them to time anything to the minute.

When to leave

First light is when the sun clears the skyline. Working backwards from it gives the time you need to be walking — and for a sunrise summit that is almost always in the dark.

These are daylight-pace figures. Our walking model is fitted on recorded daylight walks, and nobody moves at daylight pace up rough ground by head torch. Many people take half as long again in the dark — so treat these as the earliest you could sensibly leave, and leave earlier, not later. We have not applied a night-time multiplier, because we have not calibrated one, and inventing a number would be worse than saying this plainly.

The band on each leave-by time is the first-light band carried through — a walking estimate stacked on an astronomical one cannot be more precise than either. Walking time is the route’s own published upper estimate, the same figure its page prints. Times are for midsummer; for any other date, use the picker on the fell page and subtract your own walking time.

Going up in the dark is a different walk from the same route in daylight: navigation by head torch, cold at the top while you wait, and the descent still to do when you are tired — and the descent is the half that hurts people. Read the descent is the dangerous half before you plan a pre-dawn start, and pick a route you already know.

Will the tops be clear?

Geometry tells you when the sun clears the skyline. It tells you nothing about whether you will see it. That is cloud, and cloud is a forecast, not a computation — so here is the one thing the forecast can honestly say: how likely the tops are to be clear of cloud at all.

Two days, and no further. Our hourly forecast — the only feed carrying cloud — runs to two days; beyond that we hold daily aggregates with no cloud information in them. A day-six figure here would be close to noise, and putting it beside a day-one figure in the same table would be exactly the false precision this site exists not to do. So the table stops where the data stops.

Read this as the day, not the dawn. The score is averaged across the day’s hours, because that is how it is computed for the conditions pages; mornings are often clearer than afternoons, so a middling day can still hold a clear dawn. And we are not telling you whether you will get the photograph. A clear top is a necessary condition, not a sufficient one — the cloud that makes the picture is usually the cloud in the valley, and nothing here forecasts that.

What this page does not tell you

• How long the walk really takes in the dark. The leave-by times above are daylight-pace, and we say so on every one of them. We have not calibrated a night-time pace against real pre-dawn tracks, so we do not publish one — the honest form of that figure is a daylight number with the warning attached, not a made-up multiplier.
• Whether you will get the shot. Cloud decides that, not geometry. First and last light are the moments the sun crosses the terrain; whether anything is visible is a different question with a different answer every day.
• Whether the sunset will be any good. Everything here is geometry — when the sun goes behind the skyline, never what the sky does afterwards. Colour needs cloud at the right height and clear air to the west, which is forecasting rather than terrain, and it is deliberately not on this page.
• Anything about the night sky. Where a fell has an open northern horizon we say so, and it is a statement about the shape of the ground — nothing more. We make no dark-sky or aurora claims anywhere on this site.

Every summit we have a profile for

The full figures live on the fell pages themselves. Each of the 214 summits with a computed profile carries a First light section and a Last light section — midsummer, the equinox and midwinter at both ends — a chart of the skyline against the sun’s path, and a date picker that will solve any date you ask it for.

Start with Troutbeck Tongue if you want to see the extreme case, or browse the fells.

Sources

1. NOAA Global Monitoring Laboratory — Solar Calculation Details - the solar position algorithm behind our sun altitude and azimuth (opens in new tab)
2. Copernicus / ESA — Copernicus DEM Product Handbook (v5.0) - GLO-30 at 30 m; absolute vertical accuracy 4 m (LE90) (opens in new tab)
3. Database of British and Irish Hills — DoBIH v18.4 (CC BY 4.0) - hill names, heights and prominence (opens in new tab)

The refraction model is Sæmundsson’s standard true-to-apparent altitude formula, as published in Meeus, Astronomical Algorithms. Summit positions and heights come from the hikes.guide fell database; terrain from Copernicus GLO-30. All figures computed 6 August 2026 for 2026.