by Roma Agrawal
When it was built, the top of the masonry platform was made level with the ground outside. On top of it were the 3.5m-deep beams, and on top of them was the floor of the cathedral itself. Thus the floor was originally constructed 3.5m above the ground, showing that the engineers knew the structure would sink, and planned that by the time they had finished it would sink just enough to bring the floor of the cathedral down to ground level. The hope was that the structure would sink uniformly, and wouldn’t necessarily be damaged. Despite de Arciniega’s efforts, during construction, as heavy stone was laid on top of heavy stone, the structure started to sink in a non-uniform way. The south-western corner of the structure (the front-left corner in the diagram) sank more than the north-eastern corner. To compensate for this unsettlingly uneven settling, the builders actually increased the thickness of the 900mm masonry platform on its southern side.
The structural reason why the platform settled unevenly is because soil comes with baggage. It’s not enough to meet the soil, ask how it’s feeling on the day you start building and then assume it doesn’t have any emotions from its past that will affect how it behaves. It has a history and a character that an engineer must consider. The Aztecs had built their pyramid in exactly the place where the cathedral was sited, adding layers to it over time, partly for spiritual reasons, and partly to cover the damage caused by settlement. This construction had affected the physical state of the soil: some areas had already experienced lots of pressure and become consolidated and compacted, while others, which hadn’t been weighed down, remained light and less dense. Where new foundations were built on top of consolidated soil they didn’t sink much, but the portion built on less dense soil moved much more.
An attempt at realignment.
Even after the Spanish builders had finished the foundations, the structure continued to move unevenly. They tried to compensate for this differential settlement by changing angle as they worked up. Dr Ovando-Shelley pointed out areas where the courses of stone (which would normally be laid flat and uniform) had been cut to a taper. This helped the builders come back to a level line after the layers of stone they had already built had tilted. Other adjustments had been made to counter the continuing subsidence: a column at the southern end of the structure was almost a metre taller than the columns in the north. The cathedral was finished 240 years later, but throughout this time, and beyond, it continued to move erratically.
Dr Ovando-Shelley and I walked along one of the aisles (see Map here, point B) and stopped directly below the central dome. From here hangs a giant, missile-shaped pendulum (or plumb line) made of gleaming brass and steel that shows how far the cathedral has shifted. You can simulate this with a string, a small weight and a clear plastic box. Attach the weight to the string, suspend it from the centre of the roof of the box and lay the box on a level table top. You’ll see that your makeshift pendulum hangs exactly above the centre of the floor of the box. If you tilt the box slightly, however, the pendulum will move away from the centre. Tilt the box by 45 degrees and the pendulum will hang over the edge of the floor. The Metropolitan Cathedral’s pendulum works in the same way: as the foundations tilted, the pendulum stayed vertical. By noting where the pendulum was centred at various intervals over time, the tilt of the cathedral has been monitored.
In 1910 measurements were taken to compare the levels of the two extreme corners. The engineers established that, since 1573, the floor had tilted so much that one corner was a staggering 2.4m higher than the other. It’s difficult to imagine a structure tilting by such an extreme amount; not surprisingly, it had a damaging effect on the cathedral’s integrity. By the 1990s its bell towers were leaning precariously and in danger of collapsing.
A major restoration project started in 1993; Dr Ovando-Shelley was one of the large team of engineers that worked on it. They accepted that it was almost impossible to stop the structure sinking altogether, but reasoned that if it sank uniformly it would suffer less damage. However, before they could even think about ensuring it settled evenly, they needed to pivot the entire cathedral so it was relatively flat.
As my tour continued, we walked away from the dome to the back of the cathedral (see Map here, point C). Here, the shimmering Baroque magnificence of the golden Altar of the Kings extended towards the ceiling, covered by a mass of intricate hand-carved figures – an opulent wall of worship designed to assault the senses, to impress, and to arouse reverence. It certainly inspired a feeling of awe.
I, however, was completely transfixed by a tiny metal stud on a column just to the left of the altar. It was relative to this point that the team measured and compared the levels of the floor to establish exactly how much the cathedral needed to be pivoted. The chosen pivot point (the point that wouldn’t be allowed to sink any further) was the south-west corner, because this had sunk the most over time. The metal stud was at the northern end of the cathedral, which needed to be pushed down by metres. Just thinking about it made my head spin. And it didn’t stop spinning as Dr Ovando-Shelley explained the technique they used to achieve it. Have you seen the sci-fi blockbuster Armageddon, in which Bruce Willis and his team must drill a hole in an asteroid and pack it with explosives to prevent a collision with Earth? The plan devised by the cathedral’s engineers seemed about as unlikely and difficult to achieve: they would burrow beneath the cathedral and settle the soil. The thought of removing earth from underneath a structure to stabilise it might seem totally counter-intuitive. But for these exceptional ground conditions, exceptional engineering was needed.
As I said before, though, soil isn’t just soil: you have to understand its history before you can predict how it will behave in the future. Dr Ovando-Shelley and the team performed a variety of soil tests all over the site to find out exactly how strong or weak the soil was, and how consolidated (or squashed down). Feeding this information into a computer model, they drew a 3D map composed of layers of different colours that undulated and overlapped depending on the strength and type of soil at a particular depth. The model also simulated all the historical events that had affected the soil – from the building of the Aztec temple and the Spanish cathedral to the changes in water level and so on – and created a profile of the ground.
Extraction holes radiating from the large shaft.
The team then bored 32 cylindrical access shafts, 3.4m in diameter and between 14m and 25m deep, through the original masonry raft of the cathedral and into the ground. These were dug laboriously by hand (accessing this confined space with diggers would have been both difficult and dangerous). At each stage of the descent, concrete was cast in a ring around the edge of the hole, creating a tube to keep the soil in place. When the shaft was finished, a second layer of concrete was cast inside the tube to stop the hole from collapsing in on itself. At the base of each shaft the engineers sank four mini-wells from which they could pump out the excess groundwater that would otherwise rise and flood the shafts.
These, though, were not the shafts that were going to save the cathedral. They just provided the means for drilling about 1,500 holes, slightly inclined from the horizontal, with a diameter the size of a fist and between 6m and 22m long, through which soil could be extracted. The plan was that, after the soil had been removed, these holes would naturally close up over time, causing the foundation of the cathedral to settle.
Since the north side of the cathedral was the highest and needed to come down the most, the largest amount of soil was extracted in that area, while much smaller amounts were taken from the south-west corner. More than 300 cubic metres were removed from one north-eastern shaft, whereas only 11 cubic metres were taken from another in the south-west corner. In total, through this vast warren of shafts and tunnels burrowed deep beneath the historical cathedral, and with nearly 1.5 million extraction operations, 4,220 cubic metres of soil were removed from underneath the structure – enough to fill about one and a half Olympic-sized swimming pools.
As you might expect, this soil removal was done carefull
y and cautiously, in stages, over a long period (four and a half years). All that time the levels in the cathedral were strictly monitored to make sure that any movement stayed within the limits of what the engineers wanted. The arches and columns inside the cathedral were supported with steel beams and props to prevent any damage from sudden, unexpected or large movements. Meanwhile, soil samples were continually taken out of the ground to be tested for stiffness and water content, and were compared with the computer model to make sure reality matched prediction.
The difference in floor level between the north-east and the south-west had been more than 2m, but in 1998, once the north end had settled down by just over a metre, the process was suspended. Even though this left the foundation slightly tilted, the engineers had become concerned about damaging the structure. The lean of the towers had been brought back to an amount that was deemed safe – and so, for the time being, work has stopped.
The large cylindrical access shafts have been left open. They are now flooded with groundwater, but if they are needed in the future – if the cathedral starts tilting again – the water can be pumped out, and more soil removed. For now, the cathedral has been left to the mercy of the soil – but this time it is being watched.
Positioned at strategic points around the cathedral are four pendulums encased in glass boxes that send data wirelessly to a lab in Italy where engineers monitor how the structure is behaving. Pressure-pads monitor the loads in the columns, checking they aren’t changing too much. A change in load would suggest the structure is tilting again, causing some columns to be more squashed than others. Dr Ovando-Shelley described the cathedral as a laboratory, in which data has been collected for nearly twenty years. It has become a place of science as well as a place of worship.
Since the 1990s, the cathedral has been sinking at a rate of about 60mm to 80mm per year – a slow and steady settling in comparison to the past and, most importantly, an almost uniform one. The movement will continue in the future, but it might slow down over time. This Indiana Jones of engineering had saved his relic, and succeeded in his mission. No Armageddon for Mexico City’s Metropolitan Cathedral.
The team of engineers’ groundbreaking work has been a subject of study all over the world. In 1999 they worked with engineers in Italy, replicating their methods below the Leaning Tower of Pisa. In Mexico City the engineers were faced with an extreme situation – the decidedly poor condition of the soil, its variability and the sheer size of the cathedral. But the upside of the challenge they faced is that we now have an invaluable body of knowledge that can be used by engineers in the future, particularly those fighting to save our heritage, and those attempting to build in harsher and harsher conditions as our population expands and the climate changes.
Our technical tour done, Dr Ovando-Shelley and I left the cathedral in search of a restaurant for lunch, crossing the Zocalo Square, which was framed by other elaborately designed and decorated buildings that had settled unevenly. He waited patiently as I stopped to take photos of door frames that had skewed from rectangles into parallelograms.
On a terrace overlooking the Zocalo, a waiter served us frozen margaritas. ‘Soils have no word of honour,’ said Dr Ovando-Shelley, clinking my glass, ‘and neither do geotechnical engineers.’ He laughed uproariously. But to me, he had nothing but honour. He, and the team of engineers, had saved the biggest cathedral in the Americas from ruin. And he bought me chicken mole for lunch.
HOLLOW
Usually, our homes are an amalgamation of materials – we gather stuff and assemble it, creating something from nothing. But there is a place, with sparsely grassed steppes as far as the eye can see, where shelter was formed the other way round, in an absence of material – where nothing was created from something.
Naturally, I had been very curious to see this, which is why one day I found myself doubled over, surrounded by blackness, craning my neck and straining my eyes, trying to work out where I was. I knew I was deep underground: I had walked down hundreds of winding and incredibly steep stone stairs, past ancient living rooms, kitchens – and death traps – to get there.
I could just about make out that I was in a tiny, coffin-shaped passage, as wide as my shoulders as I crouched, and as wide as my feet at floor level. I wasn’t even sure there was enough space for me to turn around and backtrack to the entrance. I could see damp beige stone just ahead of me, but the bright beam of light from my phone torch barely penetrated the darkness beyond. I carefully felt my way along the passage, trying not to bump my head. After what felt like a very long time (though it was probably only a few minutes), I emerged into a small lit cave and felt relief, until I saw the long rectangular recesses carved into the floor – which had once held the remains of those unlucky enough never to find their way out.
I was in Derinkuyu, one of the deepest and largest of the mysterious, warren-like ancient underground cities in the heart of Anatolia in modern-day Turkey. These cities were made possible by the area’s three volcanoes – Erciyes, Hasan and Melendiz Daglari – which erupted violently around 30 million years ago. They spread a ten-metre layer of ash across the region, on top of which flowed lava, which consolidated and hardened the ash, turning it into what is known as tuff. The local climate, with its heavy rains, sharp changes in temperature, and melting snow in the spring, gradually eroded the soft tuff until only columns of it remained. The harder lava layer on top of the softer tuff degraded more slowly; now large pieces of lava rock sit precariously on top of the thin ash pillars, giving them a surreal, mushroom-like appearance – and their local name: ‘fairy chimneys’. The strange landscape acts as a kind of taster for the even stranger things going on below ground.
Fairy chimneys, the local name given to the thin ash pillars and the harder lava layers that sit precariously on top.
Geographically, Anatolia stands at the intersection of East and West, and throughout its turbulent history it has been the site of battles between civilisations. The Hittite people occupied the region in around 1600 BC, followed by the Romans, the Byzantines and the Ottomans. The constant warring meant that the locals were always under threat. The Hittites realised that the thick layer of compressed ash beneath their feet was relatively soft, soft enough to carve with a hammer and chisel. They began constructing underground caves and tunnels to hide in while the fighting went on above. Each of the civilisations that followed the Hittites added to these networks, in effect establishing cities in which up to 4,000 people could live for months at a time. Over a period of nearly 3,000 years, hundreds of underground cities were created in the region. Most of them were small, but about 36 had at least two or three storeys.
As I could see at Derinkuyu, the system of caves in these underground spaces was structured like an ant-house: the rooms were not stacked one on top of the other, as happens in our buildings, because that would cause the ash to weaken and collapse. Instead, the rooms were carved out randomly in space, spread out across a large area. The arched ceilings over the rooms and passages were the perfect shape to keep stone in compression, and stable, ensuring that the ground would not cave in on them. A number of ventilation shafts, starting from the surface and running for up to 80m underground, brought in fresh air. The cities were designed to protect against infiltration by the enemy – with huge rolling stone doors to keep them out, deep pits for them to fall into, and cubby-holes behind doors where the residents could hide and ambush their pursuers. The inhabitants even created narrow tunnels up to 8km long to connect adjacent cities, in case their enemies managed to get past all their carefully laid traps.
I’m glad that I’d never had to spent months at a time in Derinkuyu, fearing for my life, but come to think of it, I do actually spend an inordinate amount of time underground. In fact, since I began working, I’ve spent a total of over 5 months of my life deep inside London’s clay, as I take underground trains – the Tube – to work. Alongside millions of other people, packed into carriages like sardines, it’s an uncomfortable reminder
that, in my city, space is at a premium. The streets can’t accommodate homes, offices, pedestrian paths, trains, trams, cars and cycles – not to mention water pipes, sewers, electricity and internet cables. And why should they? After all, we live in three dimensions and should use all of them, building up and down rather than simply sprawling sideways. The city beneath our feet is brimming with hidden engineering, but these arteries would not have been possible had it not been for the humble tunnel. In Derinkuyu, space was plentiful; tunnels provided safety. In London and many other metropolises, there is a lack of space, and tunnels provide the solution.
*
In the early 1800s, the only river crossing in the entire city was London Bridge – an immensely impractical and laborious situation in a metropolis that was spreading out rapidly on both sides of the Thames. The time taken to navigate the busy city, the wait to make the perilous and excruciatingly slow journey across the choking bridge, and the cost incurred in tolls were all sources of great frustration. In 1805 a company was set up to try to circumvent this by directly connecting the docks at Wapping and the factories at Rotherhithe.
Although the two points were only a tantalising 365m apart across the river, this distance was large enough to make building a bridge impractical – which meant that to get from one to the other, people and goods had to make an arduous 6.5km journey via London Bridge. Besides, putting a new bridge between the docks and factories would have stopped tall ships from reaching higher up the river, causing major problems for the thriving trade the city hosted. The only remaining option was to create a passage under the river. The problem was that canal builders, mining experts like Richard Trevithick, and other inventors had already tried to tunnel without success. The new company’s efforts to bore a tunnel under the river were also unsuccessful, until an engineer came up with a solution inspired by a shipworm.
