The landslides at Salazie at La Réunion

Geosciences for a Sustainable Earth BRGM ANALYZE, EXPERIMENT, MODEL LANDSLIDES ON RÉUNION ISLAND Cirque de Salazie Réunion Island This is the Cirque de Salazie, which was created by intense erosion coupled with heavy rains. The surface collapsed, and all these rocks began draining out of the circus. There are two other circuses nearby. Unique to Salazie are its massive landslides. There are a dozen high-volume slides. Two are particularly large, measuring some 250 million m3. Salazie is also unique for its heavy rainfall, receiving about 3 m of water per year. This heavy rainfall, coupled with geological conditions here, explain all the landslides. The problem is that homes are at risk of falling. Houses are located in areas that are moving, creating cracks that pose risks for inhabitants as well as to infrastructure. During Tropical Cyclone Hyacinthe, the pond rose up. Cathrine Ramin lived through the cyclone in 1980. It was nothing but heavy rain for some 15 days. It destroyed everything from the pond to that road you see there. We had to rebuild. There was a gazebo up there? Yes, you see it on old post cards. Hyacinthe took the gazebo. It got buried at the cliff base following a mudslide. There was landsliding beforehand, then the water took everything. Do you find new fissures? - Yes. - When it rains? After spells of heavy rainfall, whole new rifts appear. Using GPS tracking that we've been doing for 20 years, we find landsliding speeds up after rainfall. When it rains heavily, water penetrates into the subsurface, or so we think, and that water causes landsliding and speeds up landsliding. What we want to know is how the water moves and where it penetrates. Does it penetrate through cracks? Does it penetrate at the landslide base, or at some intermediary point? That's important for understanding how everything works. If it remains on the surface, the risks are not the same as if it penetrates and goes deep beneath the landslide. We have an idea of the internal structure of the slide from an aerial geophysical survey we did. We basically did a 3-D scan of the land down to a depth of 300 m. Now we want to confirm our hypothesis that there are two landsliding surfaces, one atop the other, and find out if they move at different speeds. First, we see the landslide in its entirety. Now we want to zoom in on different parts to see if there's water, if there are subsections, and identify what slide type it is in a more accurate way, while testing the hypothesis we developed from the aerial survey. October 2020 Fond de Rond-Point, Cirque de Salazie Passive Seismic Recording Project Phase One of Data Gathering The first data-gathering phase of the campaign was setting up about 100 sensors on the landslide, from the head of the slide all the way down. Sensors: 100 Geophone Seismic Recording Device Geophones have a permanent magnet inside them and a very sensitive coil that takes ground movements and converts them to electrical signals, which get recorded. Each one of the devices will record noise. Those recordings will give us an idea of how fast waves are moving, which helps us map the subsurface, almost like an x-ray. Mapping using passive recording devices November 2020 Fond de Rond-Point, Cirque de Salazie Phase Two of Data Gathering Welcome, all, to phase two of our campaign. Phase two of the project is setting up the same 100 sensors along one line that will cut across the landslide from top to bottom. The reason we want a single, 2-km-long line reading is to get a 2-D vertical image almost like a slice of the terrain, to then get an idea of subsurface hundreds of meters deep. We record waves generated by explosions of our own making. Explosives The explosive will go into boreholes two meters deep so that they're in contact with the subsoil when they explode, generating a mechanical wave that causes the subsoil to vibrate, which we'll record along our line. Nota Bene The explosions have no influence on landslide activity Explosion Number 1 Explosion Number 7 Explosion Number 9 Explosion Number 13 Each explosion generates two waves. Primary waves, which we record first, travel at a given speed underground. Then we record another type of wave, secondary waves. They travel much slower. Primary waves help us to scan the subsoil very accurately. Map generated from primary waves To get an idea of where water is, we use secondary waves, which don't travel through water. Comparing the speeds of primary to secondary waves, we can map where water is and is not. Map generated from secondary waves What we learned about this landslide was first about its high volume. At the back end, it's just under 100 m, from 50 to 75 m. The landside is biggest near the river. There it's some 200 m thick. This project helped us figure out there were two slides atop each other downstream, closer to the river. We're still researching the role each slide plays. We have a base of the slide, this first sliding mass moving. Then a second one moving atop it. Water moves between the two slides near the surface. What we want to know is if that's at different speeds. Final Map The point is to get a highly accurate map of the slide. Then we can model the slide. From there, we can simulate climate change over the long term. If cyclones get stronger, will parts of the slide move faster? Will others destabilize suddenly, causing sudden collapse? Mapping the slide, knowing its internal structure and how water moves within it gives us direct knowledge of risks to infrastructure and the population in Salazie. BRGM thanks Salazie locals for their cooperation while readings were being taken.

Major landslides in Salazie

Geoscience for a sustainable Earth

BRGM

Five million years ago, Réunion was born in the Indian Ocean from a submarine volcano. 100,000 years ago, the land

was approximately like it is today, with the Piton des Neiges, as its highest point. When the Piton des Neiges was formed, huge landslides, linked to the action of erosion, created the three calderas of Cilaos, Mafate and Salazie.

The geological formations created by the calderas underwent new landslides, some with heavy consequences. In certain spots, these are clearly visible. BRGM, the French geological survey, is a benchmark body in the use of Earth Sciences to manage surface and subsurface resources and risks. You will see that BRGM scientists have been studying  these landslides for many years to assess the risks for the population. To better understand, let's join Katiana on site.

I am in the Salazie caldera, just below Hell-Bourg. We can clearly see what a landslide is. The land 150 m below was on my level several centuries ago. It slid down towards the Mât river, behind the cane fields. And it's still going on today. Further off, the land has crinkled, making the hills we see over there. The problem is that in some places

there are many inhabitants.

Indeed, Katiana, the speed of the land movement depends on the sector. The Salazie caldera is one most monitored by the BRGM due to the many unstable areas identified. Grand-Îlet's inhabitants know well the effects of this massive landslide.

Hello Mr. Nourry. How are you?

Fine, thanks.

What's happening here? It's scary!

Erosion made the crack, due to water underneath. The water flows underneath and goes down to the river below.

The crack has been here since 2000.

Since 2000?

Yes.

Is it still moving?

Yes. It moves every year. It goes down 5 to 6 cm a year.

Have you always known Grand-Ilet with ground movement?

Ground movement, yes.

For 27 years.

As a child too?

Always.

Mr. Nourry, we're behind your greenhouse. What happened here?

The land was sloping like this, but this hole wasn't here. There was no hole at all. It was flat.

It came with Hurricane Hyacinth?

Yes.

And you're not afraid to live here? You don't fear a landslide?

When it rains too hard, I am afraid. Seeing what happened here, we're obviously afraid.

Voice, what's going on with Mr. Nourry's house and other houses here?

The phenomenon in this part of Grand-Îlet is a landslide. One of the biggest ever recorded. The entire sector is descending at various speeds, depending on the land. Look. The left-handed rampart is static

as are the old lava flows at the base of the caldera. Above, the land is sliding. This layer is

about 100 m thick and is moving steadily towards the river.

I'm at the foot of the landslide. Look up there, above us, there's a layer of boulders and gravel. Below, where I'm standing, due to pressure and slow rock movement, friction has turned the rock into clay. This is what scientists call

the soap layer, because it helps the land to slide.

Under the landslide, the rocks and gravel gradually turn to clay. This clay also encourages landslide.

This is nothing new. In 1980, Hurricane Hyacinthe poured 3.5 m of water in 3 days into the Salazie caldera. The soil was saturated. On this road to Grand-Îlet, the ground slid towards the slopes and huts were buried.

Since the 1980 damage from Hurricane Hyacinthe, studies of the calderas' subsoil, and specifically that of Salazie,

were undertaken. Since 2003, as part of the MvTerre project, the BRGM coordinates and runs most of the landslide research with financial support from Europe, the State, the region and the department. Since this date, pioneering results are regularly published.

BRGM scientists have put sensors at strategic spots, to get precise measurements and monitor the calderas' movements. In fact, Romain is at work right now.

Hello, Romain.

Hello.

What different sensors do you use?

This is a GPS, to record movements in real time. We've set up 10 of them at Salazie. On some houses, we've measured 10 m of movement since 2003.

These aren't the only tools you use.

No, I'll show you our other measuring techniques. This isn't an instrument, but a marker, for measurements. 150 geodetic markers are deployed on the calderas, which we measure every 6 months by GPS to calculate movement around these markers.

Why such measurements?

They help us identify areas that are moving and areas that aren't.

Studying the Grand-Îlet landslide shows that it's always moving. Different areas of Îlet move at different speeds. Speed increases from the top of the landslide to the bottom. At the front, the land moves from 30 to 55 cm a year.

In the central, inhabited part, it moves from 15 to 30 cm a year. Speeds get slower towards the back of the landslide.

The worst damage occurs where these different compartments meet. The landslide can't stop at the rampart facing it, because the river carries materials away. The landslide thus advances without obstacle. Below Hell-Bourg, on the other side of Salazie, the movement is even faster and some houses have moved by over 10 m in 10 years. This aerial shot shows the position of the roads and, in yellow, the houses 10 years ago. Everything has moved 10 m northwards.

Hello, Pierre.

Hello.

Looking for drinking water?

No, we're sampling groundwater. This is from the subsoil. We want to understand how the landslide works. This device is a piezometer. This borehole, 100 m deep, monitors groundwater in the landslide. Here, the groundwater level is about 30 m down. Following it helps us understand water's role in the landslide and its underground flow.

So, water has a role in landslides? I thought it was just the slope and the soap layer.

Grand-Îlet inhabitants know very well, when it rains here, the landslide accelerates.

Can you show me how it works?

This pile of sand is the Grand-Îlet landslide, the glass is its substratum, the landslide's base. We can see that in this state, the landslide is very slow. We can't see it move.

So, this is the dry season.

Yes, it's the dry season. In the rainy season, we see that the slide accelerates. The landslide is speeding up.

It's clear that water plays a major role. We even see cracks. There are many more in the fastest zone, at the bottom.

The upper zone, where the landslide is slower, has less deformation.

It's amazing!

The main triggers in Grand-Îlet are gravity, which stays constant, and groundwater level. The most recent studies show that groundwater-level variations affected the speed of the landslide. When the level rises, there's less friction at the base, so there's more movement. Rain accelerates movement. In the rainy season, the landslide accelerates. In the dry season, it slows down but keeps moving, because the table stays high enough, even with no rain. As the locals know, in the rainy season, new cracks appear in houses and in the roads.

Bertrand, you run MvTerre project. Why is BRGM's research interesting?

This phenomenon is globally unique in its size. The entire village, of several square kilometres, is sliding, 100 m thick,

towards the gully. This means the ground is always moving.

That's hardly reassuring. Is it dangerous for people?

One of our project's goals is to reassure the population by studying the landslide. In this context, we've shown that it doesn't surge.  It moves regularly, with acceleration phases but no sudden surges.

But, are there risks?

There can be. Notably mudslides. This means that today, thanks to these studies, you understand landslides very well. We understand them better, thanks to our modelling and monitoring. We've identified the role of rain and of groundwater and the connection between them and landslides.

Thank you, Bertrand.

Voice, have you understood why the BRGM's studies are important here? To understand landslides.

Yes, Katiana. They are also important because they help us find ways to slow landslides down. If a solution were needed to slow down Salazie's landslides, these studies show that we should also act to control groundwater.