Today we will be looking at two horrifying caving stories that went horribly wrong. Both stories are separated by 23 years and hundreds of kilome. But they are connected and the thing that connects them is a film story that turns into a real tragedy days later. As always, viewer discretion is advised.
Agnes Miloka was not a mainstream celebrity like an actor or singer. She was famous in a much smaller but much more dangerous world, the world of technical cave diving. In that community, she was known as an explorer, underwater photographer, maritime archaeologist, speaker, and cave diver who pushed into places most people would never see.
She was only 29 years, and by then she had already built a reputation as someone obsessed with underwater exploration. She believed there was no greater feeling than finding a new underwater passage that no one had ever seen before. Her connection to Sanctum is what gives the story its haunting edge. Agnes worked on it as a stunt diver and also helped train actors in cave diving work.
The film was about Andrew White’s own deadly experience in underwater caves a decade ago. Andrew carried that understanding into a long film making career that brought him eventually to James Cameron with whom he collaborated on deep sea projects, 3D camera development, and the dive expeditions to the Titanic that produced Ghosts of the Abyss.
The friendship between the two men ran through the decade that followed him, turning the 1988 experience into a feature film. That film became Sanctum, released in 2011, produced by Cameron, directed by Alistair Greersonen, a cave diving thriller built around the premise of a team trapped underground when a storm floods their exit, which is precisely what happened to Andrew White and 15 other people on the Nullerbore plane in December 1988.
The Nullerbore plane extends across the southern edge of Australia in a way that resists easy description because it resists almost every quality that makes landscapes describable. There are no hills, no rivers, no trees that rise above the scrub. The name itself comes from the Latin Nullis Arbor, no tree. And the plane earns it completely.
A vast flat limestone tableland stretching roughly 400 miles east to west along the great Australian bite. The southern ocean pressing in from the south and the desert pressing in from the north and in between them a surface so featureless that the horizon looks the same in every direction and the sky takes up most of what there is to see.
What the surface conceals is the reason anyone comes here at all. Beneath the Nulerbore, the limestone that forms the plane has been dissolving for millions of years. And the result is one of the most extensive networks of submerged cave systems in the world. Long water-filled passages threading through the rock.
Some of them running for kilome without surfacing. Their ceilings below the water table in such a way that the only way to explore them is underwater in the dark with everything you need to breathe on your back. The cave divers who came to the Nulerbore from the 1960s onward found passages so large and so clear that the measurements they came back with were difficult to believe.
Some of these tunnels were among the longest submerged caves ever documented anywhere on Earth. In November 1988, a man named Andrew White led a team of 16 people onto the Nulerbore to dive one of them. White was 29 years old, a former farmer from Western Victoria who had discovered caving in school and never quite found anything more compelling to replace it with.
By 1988, he was an experienced cave diver who had been in and out of the Nulerbore systems before and understood what it required. the planning, the logistics, the five tons of equipment that had to be transported to one of the most remote places on the continent and then lowered piece by piece down through the cave entrance into the system below.
The cave was Panakin Plains Cave located near the township of Cochalbiddy 5 mi east of the road that runs through the Nellerbore along the coast. The team was international, Australians, Americans, Britain, and included some of the best cave divers working anywhere in the world at that point. Wes Skyles, the American underwater filmmaker who would later be involved in the Puerto Rico search that opened this series, was there with a camera crew.
The plan was to push deeper into Panic and Plains than anyone had gone before, document the exploration on film, and come out with footage that could be sold to cover the expedition’s considerable costs. They had been on the Nullerore for over a month. The diving had gone well. The team had pushed through the enormous dry chamber called Concord Landing, a kilometer from the surface and into the submerged passages beyond it, extending the map in the way that expedition cave diving extends maps incrementally and at significant effort. By early December,
they were in the final days of the expedition, hauling equipment back toward the surface, preparing to pack out and leave the Nellerbore behind until next time. Then the storm arrived. The Nellerbore is one of the driest places on the Australian continent. Cocoalbiddy Township receives on average enough rain to qualify as a desert.
The storm that hit on the day the team was packing out was not an ordinary weather event. It was a rare cyclonic system that drove a hail storm across the plane with a violence that the landscape had no infrastructure to absorb. In less than 25 minutes, the storm deposited the equivalent of 2 years of average rainfall onto the limestone surface of the Nerbore.
Limestone does not absorb water the way soil does. It channels it, directs it, sends it through whatever gaps and passages the rock provides. The caves of the Nulerbore are those passages. The water that fell in 25 minutes had one direction to go. Three members of the team were at the surface near the cave entrance when the storm broke and ran from it in time.
They watched the entrance behind them. The water came in fast, filling the shaft from below as the flood poured through the cave above and the middle section of the dry entrance passage, the route that connected the surface to the team below, collapsed under the weight of the saturation.
15 people were underground when it gave way. The cave entrance was sealed. The 15 people underground were not in the submerged passages when the collapse happened. They were in the dry sections on a ledge in the middle of the cave system, a section of the panic and plains that sits above the water table and provides the kind of standing room that Nullerbore cave expeditions use as a base camp between dives.
They had equipment with them. They had air. What they did not have was a way out because the collapse had blocked the only route that connected their position to the surface. White and one other member of the team managed to get out through the block section approximately 4 to 5 hours after the initial collapse.
Working through the rubble in conditions that the flood had made unstable and disorienting. They reached the surface and found the three who had escaped and the six of them understood what the situation was. 13 people were still inside. The entrance was compromised and the newerbore was not a place where help was minutes away.
Police and emergency services were notified. A rescue team with cave diving capability was assembled and dispatched to Cocoitti, which on the Nuller plane is a logistical undertaking of its own. The nearest significant population center being hundreds of kilometers away. The rescue operation took approximately 12 hours to clear enough of the rubble to allow the remaining 13 to be brought out.
The process of getting everyone out took closer to 2 days in total. Nobody was seriously injured. The cave had trapped 16 people, turned a month of careful expedition work into crisis, collapsed its own entrance under two years of rainfall in 25 minutes, and released everyone alive. The camera had been rolling for most of it. Wes Skyles and his crew had come to the Nullerbore to document an exploration.
What they captured instead was the collapse, the flooding, the scramble, and the long difficult process of the rescue. A document of an emergency that unfolded in real time in front of people whose job it had been up until the storm arrived to film cave diving. The footage that came out of the Panic Plains expedition was not the footage Andrew White had planned to sell to cover costs.
It was something considerably more dramatic and considerably more real. And when it was edited into a documentary called Nellerbore Dreaming, it launched Andrew White’s career as a filmmaker. He had gone into the Nellerbore as a cave diver with a camera crew. He came out of it as someone who understood from the inside that the story a cave could tell under the right circumstances or the wrong ones was the kind of story that an audience would sit still for.
Andrew met James Cameron and had a conversation about turning the 1988 experience into a feature film. The film needed its underwater sequences to look real. It needed someone who could move through a submerged cave passage with the specific earned fluency of a person who had spent years in those systems. It needed a stunt double for its female characters who was not pretending to be a cave diver but actually was one.
It needed Agnes Maloka. She spent months on the production of Sanctum, performing the sequences that depicted with considerable accuracy what it looks like when a cave diver runs out of air in a restricted passage and cannot find the exit. She performed those sequences. She understood exactly what they meant. The film was released in February 2011.
Not long after the film’s release, Agnes went through a caving nightmare in an actual underwater cave system. That detail can easily be exaggerated, but the real version is already unsettling enough. A professional cave diver who helped bring a fictional cave diving nightmare to screen later died in a real one.
Tank Cave sits near the small town of Tantanula in the limestone country between Millisent and Mount Gambia in South Australia. From the outside, it is unremarkable. A surface entrance in a paddic no different in appearance from a dozen other sink holes in the same stretch of carsted country. What it contains beneath the surface is something else entirely.
7 to 8 km of submerged passage threading through the limestone in a layout that Agnes Maloa herself once described in writing as a spiderweb gone wild. Not one tunnel, not a linear system you enter at one end and exit at the other. a branching interconnected maze of underwater channels. Some wide enough to move through freely.
Some tight enough to require careful positioning of the body and the tanks. Some fragile enough that a single careless fin kick turns the visibility from clear to zero in seconds. The cave is not extremely deep by the standards of the sites that kill divers with pressure. Its danger is different and in some ways harder to quantify.
It comes from complexity, from navigation, from the accumulating cost of every wrong turn and every minute spent underground. Agnes knew this. She had written about it. She understood Tank Cave not as a visitor, but as someone who had spent time inside it and thought carefully about what it asked of the people who entered it.
That is who she was. on February 27th, 2011. When she arrived at the cave entrance with her tanks and her equipment and went in, she entered with a group of divers, including a dive buddy. This is standard practice in cave diving. You do not go in alone. You maintain contact. You know where your buddy is, and your buddy knows where you are.
Because the cave offers no margin for the kind of improvisation that open water allows. The service is not above you. If something goes wrong, the only exit is the route back through the system. And the route back through the system requires knowing where the system is. They moved into tank cave, reading the guideline and monitoring gas consumption.
Agnes had seen this cave before. She knew where the passages went, which sections tightened, where the spiderweb branched into choices that demanded attention. For a time, everything was normal. They were inside the system, moving through it, doing what they had come to do. And then at some point deeper in the cave, Agnes went a different way.
She moved away from her buddy and toward another section of the system. She went on a different tangent toward one of the caves extremities. A different tangent is the language of cave diving navigation, which means passages that lead away from the main line into sections that fewer divers have entered.
As she went toward the edge of the known system, her buddy stayed at the same place and did not follow. Whether they communicated before the separation, whether the intention was to rejoin on the way out, whether it was a planned solo excursion into a restriction, none of that is on the record. What is on the record is that she went and her buddy did not see her again.
600 m from the entrance of tank cave in the direction of the extremities in the direction Agnes had gone there is a section of the system where the passages tighten and branch and the navigation demands the kind of focus that even experienced divers feel the weight of 600 m in an underwater cave is not like 600 m on land on land it is a short walk flat ground 2 minutes at a comfortable pace in a submerged cave it is passage after passage of darkness.
Each one requiring a decision. Each decision costing time. And time in a cave costs the one resource that has a hard limit. At some point in those passages, Agnes was alone. The guideline was somewhere behind her, or should have been. In the tight branching sections of Tank Cave, where the spiderweb goes wild and the passages multiply and the visibility can drop without warning when the silk comes up.
The guideline is the only thread connecting a diver to the exit. Lose it or take the wrong branch away from it. And the exit becomes a problem rather than a certainty. Not immediately, not dramatically, just quietly in the way that cave diving problems tend to present themselves as a growing accumulation of small uncertainties that the gas supply is running out of time to resolve.
Her tanks had a finite amount of air. Every minute she spent in those passages consumed a portion of it. Every wrong turn consumed more. Every tight restriction required careful, slow movement that cost gas at the rate slow movement always costs it. The cave does not care about any of that. The cave does not adjust its geometry to the needs of the person inside it.
It simply is what it is in the dark with the silt floating and the passages branching and the guidelines somewhere that may or may not be where she expected it to be. Meanwhile, her buddy waited. This is what you do when a dive partner has gone into a section you didn’t follow them into. You wait at the point of separation because the plan was to rejoin and the plan has not yet failed.
It has only been delayed. Cave divers manage time and gas and margins specifically so that delays can be absorbed. You wait. You watch the passage they went into. You watch the time. At some point, the waiting became something else. The time she should have been back passed. Then the margin passed.
The buddy surfaced, came out of the water, and the alarm was raised. Police were contacted. The cave diving association of Australia was brought in to help plan what came next. Because what came next required people who understood not just that a diver was missing, but the specific geometry of the system she was missing in. The passages, the branches, the restrictions, the distances, the places where a recovery diver could go and the places where going would cost a second life.
Even the act of locating Agnes required technical cave diving. Even the act of reaching her once located required clearing rocks from a restricted section of the route before recovery divers could pass through. The search took days. She was found approximately 600 m from the entrance in the section of the cave toward the extremities where she had gone.
They brought her out. She was 29 years old. It was said that she ran out of air inside Tank Cave on February 27th, 2011. The question that the cave diving community returned to in the months after her death and has never fully resolved because it cannot be fully resolved is what the final sequence looked like inside those passages.
The honest answer is that it is not fully known. What the evidence suggests is a chain that is simple and merciless in equal measure. Separation, navigation difficulty, and gas depletion. Not one catastrophic failure, just the quiet stacking of small problems in a place that gives no room for them to stack without consequence.
The haunting reality is a few months ago, she had spent weeks on a film set performing the exact sequences you just watched unfold. A diver separated, running out of air, unable to find the exit. She performed those scenes for Sanctum. The movie was released in cinemas in February 2011. Sanctum wasn’t fiction that someone invented at a desk.
It was based on something that actually happened to a real team in a real cave on the Nullore plane in 1988. 15 people trapped underground, one storm, one collapsed entrance, and a camera that was rolling the entire time. That is the story that started all of this.
