Standing on a high vantage point at Thingvellir Park, I overlook a vast rocky, snow-covered rift valley.
Fissures and cracks scar the landscape, filled with water so clear I can see hundreds of coins winking at me from several metres below, representing the cumulative wishes of visitors over the years.
These cracks in the land were not always there. Forces from below have slowly broken the rocks apart over thousands of years.
The largest "crack" actually marks the gradual separation of two continental plates, a dramatic example of the evolving landscape of Iceland.
It always struck me that rocks are incredibly strong, unbreakable almost. But walking around this park leaves a strong impression of how fragile and changeable our Earth really is.
I ask geologist Freysteinn Sigmundsson from the University of Iceland about this. His answer surprises me.
All the physical movements in nature we see on Earth's surface are extremely feeble compared to what is happening deep underground.
Add some intense heat into the mix and the processes slowly ripping apart the landscape at Thingvellir are the same forces that contribute to the eruption of volcanoes and earthquakes around the country.
These create strange flows of magma that scientists are still trying to understand.
Bridge between continents
At Thingvellir, the ground I stand on is an old lava field that flowed some 9,000 years ago. The whole area was originally flat but over the years, the rift valley below my vantage point has subsided 40 metres.
As I walk down a bridge built over a deep fissure, rocks on either side tower above me. It is strange to think that I am inside one of Earth's scars, within two borders that were once joined.
That is because this park is at the very site where two continental plates separate North America and Eurasia. They have stretched so much that the resulting empty space has caused the land to subside.
What is especially remarkable is that I am essentially on a sort of no-man's land, on a ridge between two continental plates.
Think of it this way: the surface of the Earth is made up of a mosaic of continental plates. Some slowly push together – a process that can form mountains – whereas others, like here in Iceland, are slowly tearing the landscape apart.
They do so at a speed of about 2cm per year. That is roughly how fast our own fingernails grow.
These separating borders create fractures and fissures in the rocks. It is these cracks in the land that allow hot magma to surge upwards and fill the gaps.
For the landscape to change so much and so quickly is off immense interest to geologists such as Sigmundsson.
Compared to the age of the Earth, 9,000 years ago is a blink of an eye. Scientists can observe the processes here in order to help them understand the impact of earthquakes and volcanoes elsewhere.
It is an impressive sight: two sides of land so visibly ripped apart.
The Mid-Atlantic Ridge that is to blame runs through the whole of Iceland. It crosses the Bardarbunga volcano, which erupted on and off from August 2014 until 28 February 2015.
It is the largest volcano observed erupting on Iceland in the past 200 years. It emitted an impressive 1.5 cubic km of lava, which drained from the volcano's roots before travelling along vein-like dykes underground.
Bardarbunga also owes its activity to these shifting plates, as well as being above a hot spot called a mantle plume.
This hot spot brings up heat from great depths inside the Earth, perhaps from the mantle itself. The shifting plates and upwelling heat are the two fundamental ingredients for Iceland's volcanoes.
With so much lava coming to the surface, the ground around Bardarbunga also sank considerably. It subsided 60m in the six months that the volcano was actively spewing out lava. This resulted in a caldera collapse. The last one of this magnitude occurred in the 1870s, but that took several decades rather than just six months.
But the strangest thing about the 2014-15 Bardarbunga eruption is that the lava did not blast out at the volcano site. Instead it bubbled up to the surface at an eruption site almost 31 miles (50km) away, at times spouting out 50m-high lava fountains.
As soon as Bardarbunga became active, volcanologists began monitoring the extensive lava flow and its impact on the surrounding region.
It is the first eruption that has been studied in so much detail, both before and during the eruption. That this happened at all was quite fortuitous, according to Simon Redfern from the University of Cambridge in the UK.
His team had already been recording tremors in the Earth with seismometers before they knew that Bardarbunga was going to erupt. They monitored thousands of small earthquakes as the lava was travelling along fissures under the glacier covering the volcano.
Redfern says that geologists spend a lot of time looking at historic eruptions and events in the geological record. But Bardarbunga unfolded "right in front of us".
Although Bardarbunga produced an immense volume of lava, it was actually a fairly tame eruption. Had the lava erupted from under the glacier, instead of travelling so far through dyke tunnels, the impact could have been much more dramatic.
We might have seen an ash cloud similar to the one from the 2010 Eyjafjallajökull eruption, which caused travel chaos across Europe simply because it erupted beneath a glacier.
Bardarbunga was also relatively tame because the magma was low in silica, which produces runny lava. If more silica had been present, the volcano would have produced stiffer, more viscous magma that would explode more readily. Redfern says to think of it like cheese fondue.
This low viscosity also meant that gases easily came off the lava, including huge amounts of acidic sulphur dioxide. There was much more than anticipated, says Redfern. People were advised to stay indoors, though scientists still aren't sure why the gases were so sulphuric. Researchers will now try to figure that out.
This may also help us understand more about the chemistry inside deep Earth, as knowing the chemistry of the gases that come out of volcanoes can provide clues to what is deep underground.
The scientists working at the site also want to figure out why the lava from Bardarbunga was erupting so far away from the magma chamber itself. Somehow, it came to be transported through a 45km-long dyke to the eruption site at Holuhraun.
"Why would magma travel along underground pathways inside the Earth, rather than come up directly to the surface?" asks Sigmundsson.
It was only three months after the eruption began that Sigmundsson, with many co-authors, published some initial observations in the journal Nature.
The team concluded that Bardarbunga's travelling magma chambers could be attributed in part to the intense energy stored in the very rocks beneath our feet.
The movement of the tectonic plates beneath us releases this energy, which has been building up for hundreds of thousands of years.
We know that Bardarbunga has been active at least 26 times in the last 11 centuries.
That means the dykes transporting the magma far from the volcano could already have been present under the crust. The magma was able to exploit existing weaknesses, created by a previous eruption.
It is almost as if the magma was breaking through scar tissue that had almost healed. With enough pressure these dykes became open wounds once again, transporting red-hot runny magma.
Taking samples of this molten rock can also help reveal where it came from. Rocks have chemical signatures, which provide insights into their ages and origins.
Redfern says the lava from Bardarbunga came from quite deep. That means it potentially represents a direct route to the very mantle of Earth. Either that, or it was simply reheated by other hot rocks.
"It's probably a mixture," says Redfern. "It's not directly from the mantle but a mantle source that's causing the heat."
Beyond the pure science, scientists are studying Iceland for a more pressing reason: to predict further eruptions at Bardarbunga and elsewhere, to help keep people safe.
To do so, researchers look for signs that magma is accumulating inside the crust. A volcano site can expand like a balloon if there is new magma flowing into it, says Sigmundsson. His team monitors known volcanic sites for earthquakes and ground deformation, both of which are signs that something is bubbling away close to Earth's surface.
They are looking for "small tiny changes, maybe a millimetre or centimetre displacement... to understand if new magma is accumulating inside the volcano."
As for Bardarbunga, it is too early for now to tell when it might wake up. At the time of writing, there were no signs of new magma flow.
Bardarbunga may stay asleep for several centuries. Only if it remains inactive for more than 10,000 years can it be said to have "died".
Regardless, it has already created an entirely different landscape: a new 15m-thick field of lava spread over an area the size of a large city.
Standing here at Thingvellir, I may be looking at something quite new: the achingly slow birth of a new continent.
Sigmundsson tells me that a new "micro plate" is forming beneath our feet. This "special Icelandic plate" is continually expanding as new crust forms on either side of it.
The consensus is that this marks a passing separation from the rest of the rift zone. It may never quite grow into a new continent. But for now, before it attaches to another plate, it really is a sort of miniature Icelandic continent.
Furthermore, the whole of Iceland is made up of relatively new crust. Its oldest rocks are only about 16 million years old, making it very young in geological terms. This young island gives us a view, which we can occasionally observe in real-time, of how Earth's crust has been forming throughout history.
Iceland is being stretched and torn, but from the resulting cracks, new crust forms. It is an almost poetic example of constant renewal.
We cannot normally see that the ground beneath our feet is ever-changing. Here, though, it is hard to ignore.