Everything written about how the Norwegian fjords were formed falls into one of two piles. There are the academic pages, dense and accurate and impossible to read on a train. And there are the cruise-brand explainers, which say “glaciers carved them” and then compare an ice sheet to a finger dragged through flour.
Neither tells you the useful thing, which is where to stand. Every stage of this process left something visible, and most of it is beside a road. You can watch a hanging valley pour off a cliff, put a hand in water that is fresh on top and salt underneath, and stand on a beach that is now forty metres above the sea. The geology is not abstract here. It is the view.
So this is the explanation with the locations attached. The mechanism, then the place where you can go and look at the evidence.
What a fjord actually is
A fjord is a valley that a glacier deepened below sea level, which the sea then flooded when the ice melted. Three parts to that, and all three have to be true.
The word is Old Norse — fjǫrðr — and shares a root with “ford”. It originally meant a place you cross, not a place you sail along.
The confusion comes from the near-misses:
| Landform | Cut by | Cross-section | Depth pattern | Example |
|---|---|---|---|---|
| Fjord | Glacier | U-shaped, steep walls | Deep inland, shallow sill at the mouth | Sognefjord |
| Ria | River, then drowned | V-shaped | Deepens steadily seaward | Much of the Oslofjord |
| Fjard | Glacier, over low relief | Shallow, no walls | Uniformly shallow | Swedish east coast |
| Sound | Various | Wide channel | Open at both ends | Raftsundet, Lofoten |
This matters more than it sounds. The Oslofjord is not really a fjord in the western sense, which is why visitors who arrive in Oslo expecting thousand-metre cliffs are so reliably disappointed. It is a broad drowned valley on soft-ish rock. The geography of where Norway’s fjords actually are is worth understanding before you book anything.
Norway has more than a thousand named fjords. You will see “1,190” quoted; that count includes Svalbard, and no single authoritative national tally exists, so “more than a thousand” is the honest figure.
The short version, in five steps
- Rivers cut valleys. Before the ice, Norway was a high plateau drained by rivers running west. They carved V-shaped valleys along lines of weakness in the bedrock.
- Ice found those valleys. When the glaciations began, ice funnelled into the existing river valleys because that is where the low ground was.
- The ice deepened them, repeatedly. A glacier erodes by plucking rock from its bed and grinding it with the debris it carries. Because a glacier fills the whole valley rather than running along the bottom of it, it widens the walls as well as deepening the floor — the V becomes a U.
- They went below sea level. This is the part people miss. Ice does not stop eroding at sea level. It keeps cutting as long as it is thick enough to keep moving, which is why the Sognefjord floor is 1,308 metres below the surface.
- The sea came in. When the ice melted, sea level rose and flooded the trenches. What had been a valley became an inlet.
The critical detail is repetition. This did not happen once. The Quaternary ice ages began around 2.6 million years ago and there have been dozens of glacial cycles since. The most recent ran from roughly 120,000 to 11,700 years ago. Each cycle deepened the trenches a little more, and because ice preferentially flows into existing troughs, the process feeds itself — the deeper the trench, the more ice it collects, the faster it deepens.
Why Norway and not Sweden
Both countries were under the same ice sheet. One has fjords and one does not. Two reasons, and neither is the one people guess.
Position under the ice sheet. Norway sat at the western margin, where the ice was moving fastest and discharging into the Atlantic. Sweden sat under the middle, where ice was thick but sluggish. Erosion is about flow, not thickness.
The edge of the continent. The Norwegian coast is where continental crust steps down to oceanic crust. That step gives glaciers somewhere to fall away to, and steep gradients are what turn slow grinding into deep trenches.
At its maximum, around 20,000 years ago, the Scandinavian Ice Sheet was roughly three kilometres thick. Norway’s highest mountain is 2,469 metres. All of it was buried.
The evidence, and where to stand
This is the table I wanted and could not find anywhere.
| What happened | What it left | Where to see it |
|---|---|---|
| Glacier filled the valley | U-shaped cross-section | Nærøyfjord from the Rimstigen trail, or the Geirangerfjord from Ørnesvingen |
| Trunk glacier cut deeper than side glaciers | Hanging valleys — waterfalls falling from mid-cliff | The Seven Sisters, Geirangerfjord. Kjelfossen at Gudvangen |
| Ice ground rock beneath it | Striations — parallel scratches on bare rock | Any polished roadside outcrop above Aurland or on Hardangervidda |
| Ice dumped debris at its snout | Terminal moraines, now shallow sills | The mouth of the Sognefjord at Solund — the shallowest point in the whole fjord |
| Glacial flour still washes in | Jade-green water | Lovatnet, or any lake below Jostedalsbreen |
| Land rebounded after the ice left | Raised beaches and marine terraces | Flat shelves above the shoreline around the Oslofjord and Trondheimsfjord |
| The process is still running | Active glaciers | Nigardsbreen and Briksdalsbreen, both outlets of Jostedalsbreen |
The sill: why fjords are shallow at the mouth
Sail out of the Sognefjord and the depth sounder does something strange. Inland it reads over a kilometre. At the mouth it reads a couple of hundred metres.
That is the sill, and it is the fingerprint that separates a fjord from every other kind of inlet. Two things made it. As the glacier reached the open coast it spread out, thinned and lost erosive power, so it cut less. And where it finally stopped, it dumped everything it had been carrying as a terminal moraine.
Depth figures for the Sognefjord sill vary: research literature puts the limiting exchange depth at around 170 metres, while a commonly cited figure is closer to 100 metres. Both describe the same feature measured slightly differently. Either way, the contrast with 1,308 metres inland is the point.
The sill has consequences beyond trivia. It restricts how much deep water can flush in and out, which is why fjord bottoms can go years between exchanges of deep water, and why they support ecosystems — including cold-water coral reefs — that you would not expect at that latitude.
Hanging valleys: why Norway has so many waterfalls
This is the single most visible piece of evidence and almost nobody explains it.
A big glacier in the main valley cut down to and below sea level. The smaller glaciers in the side valleys were thinner, moved more slowly, and cut far less. When the ice went, the side valleys were left hanging — their floors stranded hundreds of metres up the wall of the main fjord.
Every river that used to flow gently down a side valley now arrives at a cliff edge and falls off it. That is what you are looking at from a fjord boat. Not a geological curiosity: a mismatch in erosion rates, made visible.
The Seven Sisters on the Geirangerfjord is the textbook case. Sources give the drop as around 250 metres of free fall, though figures between 300 and 410 metres circulate for the total descent depending on where you stop measuring. Kjelfossen above Gudvangen is another, and there are hundreds of unnamed ones — which is why the fjords have so many more waterfalls than any comparable coastline. Our guide to Norway’s biggest waterfalls goes through the named ones and when they actually run.
It also explains a seasonal frustration. Hanging-valley falls are fed by snowmelt from small catchments. In May they are torrents. By late August several are damp streaks. The waterfall you saw in a photograph may not exist in September.
Why the water is that colour
Fjord water goes from steel grey to an opaque jade depending on where you are and what the weather has been doing, and there are two separate reasons.
Rock flour. A glacier grinding across bedrock produces silt so fine it stays suspended in water indefinitely. Those particles scatter light, preferentially at the blue-green end. Any water fed directly by glacial meltwater — Lovatnet below Jostedalsbreen is the extreme case — takes on that milky turquoise. It is not algae and it is not mineral dissolution. It is powdered rock.
The brackish layer. Rivers and meltwater pour fresh water into the fjord, and fresh water is less dense than salt water, so it sits on top. A Norwegian fjord in summer often has a distinct layer of near-fresh water a few metres deep floating on cold seawater. It is why the surface is noticeably warmer than you expect, why swimming is more tolerable than the numbers suggest, and why the colour changes at a visible line where a river enters.
The land is still rising
Three kilometres of ice pressed the crust down. When it melted, the crust began springing back, and it has not finished.
Around Oslo the land is rising roughly 5 millimetres a year — about half a metre per century. The peak rate in Scandinavia is higher still, over 10 millimetres a year near the head of the Gulf of Bothnia in Sweden, where the ice was thickest and lingered longest.
You can see this without instruments. Look for flat shelves of land parallel to the shoreline but well above it, sometimes with rounded beach cobbles on them. Those are old sea floors. Around the Oslofjord and the Trondheimsfjord, whole towns sit on marine terraces that were seabed when the first people arrived after the ice.
It also means the maps are quietly wrong. Harbours built in the medieval period are now inland. The relationship between Norway and its sea level is not fixed.
The strandflat: the thing you fly over
Approach Bergen or Ålesund by air and, before you reach the mountains, you cross a broad, almost flat platform littered with thousands of low skerries and islands. That is the strandflat, and no other article on this subject mentions it.
It runs 20 to 50 kilometres wide, locally up to 80, cut across hard crystalline bedrock with a seaward slope of about 0.3 degrees. It is the reason Norway’s outer coast is habitable at all — almost every fishing community sits on it, because the mountains behind offer nothing.
How it formed has been argued about for a century. The traditional explanation involved frost shattering along the shoreline over enormous spans of time; recent work argues for a Pleistocene origin tied to the same glacial cycles that cut the fjords. It is one of the genuinely unsettled questions in Norwegian geomorphology, which is worth saying out loud rather than picking a side.
Is it still happening?
Yes, in three ways, and all three are visitable.
The glaciers are still there
Jostedalsbreen covers 458 square kilometres — the largest ice cap on mainland Europe. Its outlet glaciers, Nigardsbreen and Briksdalsbreen, come down to around 300 metres above sea level and you can walk to within touching distance of both, or onto them with a guide. Folgefonna, above the Hardangerfjord, covers around 207 square kilometres.
Standing at a glacier snout is the closest you will get to watching the process. The meltwater river is grey with rock flour. The rock in front of you was under ice within living memory and is scratched and polished. Guided glacier walks are covered in our hiking guide.
The rock is still moving
Above the Storfjorden, a mountainside called Åkerneset is coming apart. More than 50 million cubic metres of rock is detaching, moving a few centimetres a year, and it has been instrumented continuously since 2005 — one of the most closely monitored slopes in the world.
If it goes into the fjord at once, models put the wave at around 85 metres at Hellesylt and around 70 metres at Geiranger. The monitoring system is designed to give at least 72 hours’ warning, and evacuation plans cover up to 5,000 people across ten municipalities.
This is not a reason to avoid the Geirangerfjord. It is a reason to notice that the same processes that shaped the landscape are still shaping it, on a timescale that includes this week. Norway takes it seriously; the sensors are visible on the slope if you know where to look.
The sea keeps rising and the land keeps rising faster
In most of Norway, isostatic rebound currently outpaces sea level rise. The coastline is still emerging.
The timeline, with dates you can hold onto
“During the ice age” is doing a lot of work in most explanations. Here is the sequence with actual numbers.
| When | What was happening |
|---|---|
| ~2.6 million years ago | The Quaternary begins. Repeated glaciations start. Rivers had already cut the valley pattern |
| ~120,000 years ago | The most recent glacial cycle begins |
| ~20,000 years ago | Maximum extent. The Scandinavian Ice Sheet is around 3 km thick and covers all of Norway |
| ~15,000 years ago | The outermost west coast becomes ice-free. The strandflat emerges first |
| ~12,900–11,700 years ago | The Younger Dryas cold snap halts the retreat. The ice front stabilises and dumps the Herdla moraines, which run from the Hardangerfjord to north of the Sognefjord |
| ~11,700 years ago | The Holocene begins. The retreat resumes and accelerates |
| Since then | Sea floods the trenches; land rebounds; glaciers shrink to today’s ice caps |
The retreat rates are the part that surprises people. Reconstructions of the Sognefjord deglaciation put the ice front retreating at roughly 270 metres a year through the outer fjord, slowing to around 185 metres a year further inland. That is a landscape changing visibly within a human lifetime — the first people to reach western Norway were walking into country that was still emerging from ice.
The Younger Dryas moraines are the most useful thing on that list for a visitor. They are the reason several fjord mouths have island chains and narrows in the same place: the ice paused there for a thousand years and built a ridge.
How to read a fjord from a train window
The Bergen Line crosses the whole sequence in six hours, and if you know what you are looking for it stops being a scenic railway and becomes a cross-section.
Leaving Oslo: low, rounded, forested country with lakes in every hollow. Those hollows are scoured basins; the rounded shapes are ice-moulded bedrock.
Climbing towards Finse: the trees stop, and you are on Hardangervidda — a plateau the ice sheet planed almost flat. Bare rock everywhere, scattered with erratics: boulders sitting on bedrock they do not match, dropped where the ice melted. At 1,237 metres this is the high point of the line and it looks like a different planet.
Myrdal down to Flåm: the drop off the plateau into a trench. This is the transition the whole geology is about — 865 metres in 20 kilometres, down the wall of a valley cut by ice that no longer exists. The waterfalls you pass are pouring out of hanging side valleys.
The last stretch into Bergen: back out onto the strandflat, low and island-strewn.
Do the same thing on any west coast road and the pattern repeats. Plateau, trench, flat coastal shelf. Once you have seen it you cannot stop seeing it.
The numbers, and where they disagree
| Fjord | Length | Max depth | Note |
|---|---|---|---|
| Sognefjord | 205 km | 1,308 m | Longest and deepest in Norway. Sill at the mouth roughly 100–170 m |
| Hardangerfjord | 179 km | 860 m | Sources also give 190 km and 893 m; the difference is where you count the mouth |
| Nærøyfjord | 17 km | Shallow by comparison | Narrows to ~250 m, walls over 1,000 m, peaks to ~1,800 m |
| Geirangerfjord | ~20 km | — | A branch of the Storfjord, not a fjord in its own right |
| Aurlandsfjord | — | 962 m | Side arm of the Sognefjord |
I have left the disagreements in deliberately. Fjord measurements depend on decisions — where the fjord ends and the open sea begins, whether a branch counts separately — and any article giving you a single confident number for all of them has copied it from another article.
The UNESCO listing is more precise, because it had to be. The West Norwegian Fjords property, inscribed in 2005 under criteria (vii) and (viii), covers 122,712 hectares across the Geirangerfjord and Nærøyfjord areas. The citation describes walls rising to 1,400 metres above the water and dropping 500 metres below it.
Common misconceptions worth dropping
The most useful section on most of our guides is the one telling you what to ignore. Here it is the ideas rather than the destinations.
“Fjords were carved by rivers.” Rivers started the valleys and set the pattern, but a river cannot cut below sea level. Anything with a floor a kilometre under the surface was cut by ice.
“The fjords were formed during the last ice age.” They were deepened during it, and during the dozens of glaciations before it. The trenches are millions of years in the making. The last ice age gets the credit because it did the finishing work and left the landscape you see.
“Fjords are freshwater.” They are seawater with a fresh layer floating on top. Below a few metres it is ocean, and it connects to the Atlantic.
“The deepest fjord in the world is in Norway.” The Sognefjord is the deepest in Norway and among the deepest anywhere, but Antarctica and Greenland have deeper ones. It is a genuinely close-run comparison and depends on whether ice-covered trenches count.
“You need a boat to see how a fjord works.” The opposite, in fact. From water level you see walls. From a high road or a ridge you see the U-shaped cross-section, the hanging valleys and the terraces all at once. The viewpoint guide covers the places that give you the whole geometry.
Where to go if the geology is the point
If you are planning a trip around this rather than reading it in an aeroplane seat, four places do the most work.
Nigardsbreen, Jostedalen. Walk from the car park to the glacier snout across bedrock the ice left within the last century — polished, striated, with meltwater running grey beside you. Nothing else compresses the whole story into a kilometre of walking.
The Aurlandsfjellet road, above Aurland. A high road across the plateau that the ice sheet planed flat, dropping to the Stegastein platform 650 metres above the fjord. You get plateau, trench and hanging valleys in one drive. The road is seasonal, opening around the start of June.
The Geirangerfjord from Ørnesvingen. The clearest textbook U-section in the country, with the Seven Sisters pouring off a hanging valley opposite. Full detail in the Geirangerfjord guide.
The Nærøyfjord by boat from Flåm. The extreme case — a trench 250 metres wide with walls over a kilometre. From the water it is the only way to feel the ratio. See the Nærøyfjord guide for how to do it.
If you are choosing between fjords for a first trip, the comparison is in which fjord to visit in Norway, and the wider context is in our main guide to the Norwegian fjords.
Frequently asked questions
What is a fjord in simple terms?
A valley that a glacier gouged out below sea level, which the sea then filled when the ice melted. The steep walls and the U-shaped cross-section come from the ice; the water comes from the ocean.
How long did it take for the fjords to form?
Millions of years, across dozens of glacial cycles since the Quaternary began around 2.6 million years ago. The most recent cycle, which finished about 11,700 years ago, did the final shaping.
Why are fjords shallower at the mouth than inland?
Because the glacier lost power as it spread out at the coast, and because it dumped its debris load as a moraine where it finally stopped. That underwater ridge is called a sill and it is the defining feature of a fjord.
Are Norway’s fjords saltwater or freshwater?
Saltwater, with a layer of fresh river and meltwater floating on the surface. The freshwater layer can be a few metres thick in summer, which is why the top of a fjord is warmer than it has any right to be.
Are the fjords still being formed today?
The glaciers that cut them are much reduced but still present — Jostedalsbreen alone covers 458 square kilometres. The land is still rebounding at up to about 5 millimetres a year in southern Norway. And rockslides continue to reshape the walls. So yes, though slowly.
Which is the deepest fjord in Norway?
The Sognefjord, at 1,308 metres. It is also the longest at 205 kilometres. The Sognefjord guide has the detail.
What is the difference between a fjord and a sound?
A fjord is glacially overdeepened, U-shaped, and has a sill at its mouth. A sound is a wider channel, usually open at both ends, and need not have been cut by ice. Raftsundet in Lofoten is a sound; the Nærøyfjord is a fjord.
Why does Norway have more fjords than Sweden?
Norway sat at the fast-flowing western margin of the ice sheet, where the continental shelf steps down to the ocean and gradients are steep. Sweden sat under the slow middle of it. Erosion follows flow, not thickness.
Where can I actually see evidence of glaciers in Norway?
Nigardsbreen and Briksdalsbreen for active ice; any polished, scratched roadside outcrop for striations; the Seven Sisters for a hanging valley; and the raised beaches around the Oslofjord for isostatic rebound. Several of Norway’s most photographed landscapes are glacial features people admire without knowing what they are looking at.
Photo credits
- A textbook glacial trench: steep walls, flat floor, and water where the ice used to be. — Photo: Jorge Láscar from Melbourne, Australia / CC BY 2.0 via Wikimedia Commons.
- A hanging valley in action. The side glacier cut far less than the trunk glacier, so its river now arrives at a cliff. — Photo: Mariordo (Mario Roberto Durán Ortiz) / CC BY-SA 4.0 via Wikimedia Commons.
- Nigardsbreen, an outlet of Jostedalsbreen — Photo: Daragon Photos / CC BY-SA 4.0 via Wikimedia Commons.
Measurements and research findings checked 1 September 2026. Fjord lengths and depths vary between sources depending on where the mouth is defined, and I have flagged the significant disagreements rather than picking one. The origin of the strandflat remains genuinely disputed in the literature.