Q & A: Professor Roger Bradbury, Emeritus Professor of Complex Systems

Liam Garman:

Welcome everyone. It's Liam Garman here, editor for Defence and Security at Momentum Media. Unfortunately, Phil Tarrant couldn't join us today, but you're in very safe hands. It's been an amazing couple of weeks in the Defence Industry and Defence. As I'm sure you're all aware, we've had a new tripartite alliance with the United States and the United Kingdom built off of technology sharing capabilities with Australia. And Defence and Defence Industry has been whipped into a whirlwind around the new nuclear submarine announcement and the scrapping of the SEA 1000 contact with Naval group. And obviously there are a lot of questions still outstanding regarding this and regarding the future of Defence Industry. We know that there will be 18 months of review with construction on the new submarines set to begin in the next 10 years. But today I'm joined by Emeritus Professor of Complex Systems Science, Roger Bradbury from the Crawford School of Public Policy at the ANU. How are you, Roger?

Roger Bradbury:

Liam, fine. Thanks very much. And thanks for having me on your show. 

Liam Garman:

You're a very welcome guest because a few weeks ago, last week you wrote a small piece for Defence Connect, which actually got a lot of our viewers, a lot of our readers, very interested. And just a small summary is that with the rate of technological change, there will come a time at which submarines, and perhaps I'm choosing the wrong word here, but will move to redundancy. And this small piece that you published was part of your report, Transparent Oceans? The Coming SSBN Counter-Detection Task May Be Insuperable. It's an amazing thing, because like we said before, Defence and Defence Industry was according to a whirlwind with this recent announcement, but really it might not be as simple as we think. Professor, amid so much fan fact, can you explain a little bit more about your findings in this report? 

Roger Bradbury:

VIEW ALL

Well, Liam, thanks. Look, I think as we start to look at this problem, we've got to take an historical view I think, and realise that for all weapon systems, two things can happen. First, all weapon systems eventually get replaced. And the second thing is that some weapon systems get replaced very fast by some other system or they get replaced very slowly over a long period of time. And we've seen this with all sorts of technological developments in warfare and just generally technology in the broad, because technology impacts on weapon systems and warfare. 

The problem we have is that we get caught up in the moment and we don't think about this. We think that some weapon systems, say knights in armour will always be with us, if we were in the 12th century. We didn't realise that just around the corner was the invention of the crossbow, made knights in armour just redundant on the battlefield. It made them extinct and in a very short period, warfare changed. What we need to ask is, submarines in this case, are they coming to the end of their lives as weapon systems? And if they are coming to the end of their lives because of technology change, is it going to be a sudden end or is it going to be a slow transition? 

We've seen both in the past, we've seen slow transitions like the evolution of the war ship over maybe 2000 years from Gales through to wooden sailing ships carrying cannon, through to iron clads, through to battle ships, or through to frigates. That's a very slow evolution and with the general purpose of one of these warships being much the same. But other systems often very specialised systems ones that have a one single very big trick in their kit bag, they often change very fast and go extinct. The best example we have in recent history is the battleship. 

There was a huge battleship Naval building race in the 1910s and ‘20s. It almost drove the great powers broke to the point where they had some Naval treaties where they tried to limit the number of battleships that all built, but they still kept doing it. And then the battleships were never involved in warfare after once they were built, because they'd been overtaken by the development of the torpedo and the development of the boat we now call The Destroyer which was fast, cheap, accurate, and can get in. Battleships had one big trick. They could stand off from each other and they were hugely armoured and they could thump away at each other like heavyweight prize fighters. But they were able to be defeated by Destroyers. 

And so they were never used in battle after the battle of Jutland. The only time they were ever used in engagement was the battle of Jutland in 1916, and are never used again. Will subs be like that? Parked and not used, or will something else happen? So we had a study group of a bunch of scientists and technologists, largely at the ANU, but from some other places as well, that had a lot of experience on what you might call the high frontier of science and technology, because whatever's going to going to happen to subs is not going to happen next year or the year after. Subs are extremely lethal, effective, deadly weapon systems today and will be for at least another decade. For at least into the short term future. 

So we're not talking about today, we're talking about what's happening beyond, out 20 or 30 years, maybe 30 years. And picking up what's happening out there, you can't just wave your arms and say, X will happen, Y will happen. You've got to look at where the big areas of science and technology are going, not particular technologies that we have today, but blocks of high frontier technology and science, and where breakthroughs might come in those areas. So we looked at energy systems, storage, batteries, things like that, material science, what you can build stuff out of that you can't do today. 

We looked at particle physics, we looked at chemistry, we looked at oceanography and so forth. And IT, hugely a big effort on IT. And we came to a view that by the 2050s, there'll be a convergence of technologies, exist today, but they'll come together in what some people are calling fourth industrial revolution. It'll be as big a shift as we're seeing today in the third industrial revolution with IT and so forth, it'll be as big again. It'll be IT plus materials plus energy systems. This will have the effect of new classes of sensors integrated across the world, which are sea bed sensors, mobile underwater sensors, autonomous sensors, surface platforms, many of them autonomous aircraft, many of them autonomous and satellites, all integrated into a mesh that can sense the whole ocean. And subs won't be able to hide. 

We're not looking just at acoustics. We're looking at a whole range of new sensing possibilities. We're looking at gravity anomalies, because subs are big. They're about two football fields in length. They are huge hunks of metal. They can affect the gravity field. They are metal sitting in an electrically conducting media. So there's an electrical field around them. They are in a magnetic field of the earth, so there's magnetic anomalies. There's a whole bunch of sensing angles that haven't been fully exploited yet, and they're likely to be. We used a new analytical technique called subjective logic, it's been built into a terrific piece of software called Intel fuse. And the intelligence agencies are using it for just these purposes where the data are difficult, where they're poor, where they need to be integrated, where they need to come from many different sources. 

And the thing about subjective logic is that you can get a probability of the thing happening, but you can also get an estimate of the certainty of that probability, two separate estimates, really important. And so we came up, we did all this. There were a bunch of about 20 of us. We got these estimates. We ground-truthed it against the scientific understanding and the best scientific literature. And we came up in the end with the view that it's likely that common phrase means about 80% probability, quite high. It's likely that the oceans will be transparent by the 2050s. And by that we mean we should be able to track any large sub from its home port out to where it's out on mission and back again for its whole life cycle. Won't happen today. Won't happen in the 2030s and ‘40s, but by the 2050s, it means that subs will be detectable. 

Liam Garman:

And that trick that you discussed, that every piece of military equipment, every piece of military technology has a trick, a core competency, a core advantage. That core advantage of the submarine is being able to go undetected for long periods of time. And once you have these sensors, well that trick's gone. 

Roger Bradbury:

Yep. Yep. It's a one trick pony. Now, my background was long time ago in biology and you see this in evolution all the time. Some organisms become very, very specialised and they just have one big trick. Sabre-toothed tigers were stealth predators. They had one big trick. The world changed, no sabre-toothed tiger, but you get more generalised things like dingoes, like people that are generalists and can make their living and do their stuff, adapt and change and so on. And they survive. So things like the common, shall we call it, warship or the common rifle, have been around and evolved and adapted to changing circumstances and they're not about to go extinct. But something as specialised as a battleship, as we saw by 1920 and perhaps as submarines, they have one big trick. If you get past that trick, then they run out of evolutionary rope. 

Liam Garman:

And one of my favourite examples that I always think about whenever we talk about superseding technology always goes back to the Iridium constellation, with this big fanfare and everyone thinks that this new satellite consolation is going to give global coverage and everyone walks outside with their briefcase sized mobile phones, and suddenly they can get coverage. But despite it being at the time science fiction, right, so sci-fi, they got outmanoeuvred by terrestrial repeaters and something that was simple. And it almost relates exactly back to what you're talking about, this mesh of sea bed sensors that are taking all this different data and relaying it with each other. It does remind me of almost just undercutting that technology and then driving it into redundancy.

Roger Bradbury:

Well, you're right. And the Iridium constellation suddenly became the dinosaur of the satellite world. And this is not today's discussion, you might say that about the NBN with fixed wireless base being competing by-

Liam Garman:

Some may draw that conclusion.

Roger Bradbury:

By fixed wireless 5G. Wow. Yeah. We are very good at building white elephants, but they're not white elephants at the time. This is not to criticise submarine builders and Naval strategists and so on. We are wedged. We have to have these submarines. We have to have them because they are so lethal and so effective and so capable right up until the point they aren't. And that change will come suddenly. And so that's going to be a dangerous period of what you do, what new platform you leap to. But up until that time, we're going to have to dig deep into our wallets and pay for these beasts. But yeah, it's a conundrum for us, and for all Naval operations around the world. We can’t not have subs and they're very effective right up until this transition point when they're going to be not so effective at all.

Liam Garman:

They're effective until they're not. 

Roger Bradbury:

Exactly, yeah. 

Liam Garman:

On that, we are going to take a quick break. And when we are back, we're going to probe a little bit more into the radar and sensing technologies that are going to be required to make these changes. And we're back. Liam Garman here, and I'm joined by Roger Bradbury, Emeritus Professor at the ANU. And we are talking about the change in technology and technology superseding existing weapon systems, and the big one that we're talking about at the moment is with submarines. And we were discussing creating this underwater mesh of radar and sensing technologies. And Roger, that really, really intrigued me, dragged in a little bit of that one, what prototypes of these technologies are in existence? Are we already on the pathway to seeing some of these sensing technologies coming through? 

Roger Bradbury:

Well, okay. We've got a, there's a whole bunch of sensing technologies and there's a whole bunch of platforms put these things on. And the characteristic that we saw most strongly is that the platforms are becoming more autonomous. They're becoming wide-ranging. So whether they are aerial drones or underwater drones, and they're becoming better at communicating particularly underwater. So if I talk about the platforms for a minute, then I'll talk about some of the sensing technologies, because they're going to be really significant in the fact that they are so counter to the experience to date of what you need to pick up when you're trying to pick up a sub. 

But underwater platforms in terms of drones have hit some benchmarks in the last couple of years, some of them, a couple of drones have gone underwater from the Pacific coast in the US to Japan and back, autonomously underwater. They're getting long legs. We know that aerial drones, UAVs have got similar long capacities. One of the big drones Australia purchased flew out from California to Adelaide and it landed in Salisbury at the Defence Science labs, flew out on its own, landed on its own efforts. So the communications technology for driving those, the power sources, very important, are all really getting there. So those things are really important. 

The other thing that's happening is that at least over short-ish distances, underwater drones can start to act as swarms and they can communicate with each other on various, some electromagnetic channels. That's been shown and they're getting better at that. The capacity to take all this data in, in real time and sift out the signal from all the noise is getting better. It's probably 20, maybe 30 years since the Jindalee over the horizon radar system got built, Australian invention. That was a radar system that could look out maybe 1,000 Ks thereabout, pretty classified, and it could pick up on the surface of the sea all vessels out there, could pick up a lot of aeroplanes  as well. 

But it was not so much, while it was a very clever environ sensing set of devices, it was really a computational triumph, that it could sift out signals with a lot of maths, sift out signals in a very noisy environment to create an image what was out there, hundreds and hundreds of kilometres away. So we are getting this capacity to manipulate the data. So we've got the potential platforms and they're getting better all the time. We've got the energy storage systems, battery technologies getting better all the time. We've got the satellite uplinks and so forth. Those things are getting better and they're getting better fast. 

It's worth noting that these things, because they're often driven by commercial imperatives change very fast. They change over the order of years, whereas submarine builds and refits of systems and so on are of order of decades. Major military hardware competing against the fast cycle of technology change. Some of the top fighter planes in the Western arsenal still use chips that are in their communications and organisational arrays inside the planes. They use chips that are maybe 1,000,000th the power of your iPhone. It's just ridiculous. They're computing power, they're processing power and everything. It's just crazy. So they're the platforms. But the other thing that's changing are the things that you can put on the platforms, the sensors. 

So one good example are SQUIDS, which are magnetic anomaly detectors. It's a quantum device, they used to be very big. You could just squeeze them into a big thing, like a Poseidon Surveillance Aircraft. Now they're getting smaller and they're getting more accurate all the time and they can go into a small underwater drone now. When people talk about subs hiding in the depths of the ocean, the ocean is big, but subs don't inhabit all of it. The average depth of the ocean is about 4,000 metres. Subs typically inhabit the top hundred metres of it, maybe down a couple hundred metres briefly, but they're really just in the top layer. In terms of the ocean, they're not really in a three dimensional world at all. They're in a two dimensional world, nearly on the surface. 

And so they're there just under and they are big metal objects. As I said earlier, they are 170 metres long, perhaps, nearly two football fields long. Depends on your code, of course. They are made iron ferromagnetic materials. They're sitting in an electrically conducting bath. They're sitting in a magnetic field, they're sitting in a gravity field and so on. And what a sensor needs to do is to detect anomalies in those fields. So put to one side the sound and acoustic and those sorts of things, they disturb all those fields tremendously just by being there and just by moving through them. A big sub, one of the attack class, a nuclear attack submarine like the Virginia-Class or so on will maybe do 40 knots. That makes a lot of turbulence. And it makes a lot of movement through a field that can be detected. 

The big nuclear ballistic subs, they can move at maybe 20 knots, roughly. That's nearly 40 clicks. Again, they're huge. That produces turbulence, that produces noise, that produces all sorts of things. We focused on noise in the past, but as I said, there are gravity detection devices, which are starting to get interesting, not yet, but they are. There are certainly magnetic ones that are usable now, but we haven't really fully explored the use, the fact that subs are moving through an electrical field. They're an iron thing moving through an electrical field. We haven't fully explored and we're starting to explore the turbulence they produce. We're starting to explore the fact that you might be able to pick up those turbulent wakes on the surface, because even though they may be 100 metres down, it's still going to churn up the surface. You've got very, very, very good computing from a satellite. You could separate out a turbulent track of a submarine from the general background wave mixing on the ocean surface. 

Liam Garman:

Just those tiny little footprints that they leave behind, but the algorithms can sort through the noise and identify it.

Roger Bradbury:

And if you're doing it through a satellite, you don't have to have all the computing on board, you have it back at your supercomputer centre, and you're just sending the data back and forth. Yeah. So we've got those, and a little bit further out, a little bit further out, and I won't make a big deal of this one, are the fact that the nuclear subs in particular produce things called neutrinos. Neutrinos are fundamental particles. They're produced in all nuclear reactions. They don't have a charge. They have a vanishingly small mass, and they go through everything. You can't stop the neutrinos coming out of a nuclear reactor, just like you can't stop the neutrons coming out of the sun. 

In the last 20 years, maybe 30 years or so, we've gone from not being able to, well go back to the 1930s, neutrinos had just been hypothesised. Then in the ‘50s and ‘60s, we discovered that we could detect them. We actually experienced them. And then in the ‘70s and ‘80s, we've now started to be able to detect them directionally. We've been able to detect a neutrino burst from a supernova coming from a star inside the large magellanic cloud. Now that's a nice piece of orientation to get it. It's coming from there, not there. Now, so if we can do that, in principle, we should be able to detect coming out of a nuclear reactor. 

And what's happened in the last 10 years, is that the size of these detectors has shrunk. They're really difficult to detect neutrinos, but the size of detectors has shrunk. 10, 15 years ago, we had detectors which were in the Antarctic and they were inside a cubic kilometre of ice. That's how big they needed to be. They had sensors in this cubic kilometre, one good use for the Antarctic basis. Since then, they've shrunk them down to the size of shipping containers. And in the last couple of years, some scientists have put a shipping container with one of these detectors in about 20K away from a nuclear reactor, and they've been able to directionally tag that nuclear reactor from the neutrinos coming from on it. 

Now, the potential of that, if we can even, we don't even need to shrink them much anymore. Imagine this, here's a hypothesis. We set up an array of neutrino detectors around the earth. Fortunately, the west we've more or less got enough to triangulate the world with this. We did this exact experiment during the Cold War where we set up an array of very sensitive seismometers to pick up trembles in the earth. And we were able during the Cold War to detect by triangulating these very sensitive seismometers the very tiny vibrations caused all around the world by nuclear explosions in Mururoa and in Russia and in a couple of other places. And we were able to pinpoint exactly where those explosions occurred by filtering out a signal processing, the tiny nuclear explosion signal from the general rumbling and grumbling of the Earth's interior. In principle, you could do that with an array of neutrio detectors around the world. If you did that, you could detect every nuclear reactor in real time in the world and where it was. Some of them would be moving. 

Liam Garman:

And then from that point, you have that detection to see where the sub is. And then the trick of that nuclear powered sub, which is being able to stay underwater for long periods of time is now undermined. 

Roger Bradbury:

Yeah. Exactly. But we're nowhere near that. I'm not saying we've got those tomorrow. I'm not saying we've got those in 10 or even 20 years. That's the trend of the technology. And it's the trend of the capabilities in IT in sensing and so forth. So I think we'll get a convergence of sensing technologies. We'll get a convergence of the abilities to manipulate the data, and it's going to make the oceans transparent. 

Liam Garman:

And I can see why, with the report that your team compiled, why you needed experts from computer science, from sensor technology, from energy storage, underwater coms, oceanography, all into this, because it is this multi domain approach. And once you have the insights from all those key researchers, well, suddenly creating that mesh seems obviously not immediately, but it seems very achievable in that medium to long term. 

Roger Bradbury:

Oh yeah. These were a bunch of cool dudes. I learned a hell of a lot. I think we all learned a lot from each other. It was such a good project because everyone was very open minded and were prepared to play, what if games. And we had this analytical framework, this Intel fuse in front of us, which allowed us to play these games. What if this, what if that, what if that's stronger than this? What if this is weaker than that? And continually recompute the probabilities until we converged on some sort of consensus. Yeah. 

Liam Garman:

And I might be putting you a bit on the spot here because it's not necessarily directly related to your report, but I think a lot of people are going to be saying, well, you reduce it very logically down to this basis. And we see that essentially every technology and every piece of military equipment can have the same risks, that a very simple answer, even though it has to be fleshed out in a sense of senses, but a very simple answer can drive this technology to redundancy. On that, we're going to take a quick break. And when we are back, we're going to probe a little bit more into, what other technologies do you think could be at risk in the military domain? 

Roger Bradbury:

Well, it's hard. As I said, the ones that are less at risk, the general purpose weapon systems, because they can evolve, and they're often involved in what you might call a symmetric battle. But both sides have got rifles, both sides have got tanks and so on, so that they move together and can move and change such that tanks today are nothing like tanks of 100 years ago. But that's one class of problem. And those problems, they gradually evolve because it's easier for generalists to evolve, but specialists it's harder. And so, sure you can put more effort into counter detection technologies in subs. Remember, subs are now perhaps a thousand times quieter than they were at the beginning of the Cold War. Our subs, particularly the Russians ones, they used to call old boomers because you could hear them from one side of the Atlantic to the other. But they didn't care about that. 

Those guys were on a suicide mission if the balloon ever went up. Yeah. But so you can do it to an extent, but there are limits to that. It's probably hard now to make subs another thousand times quieter. They've still got to move. They've still got to push water out of the way. They've still got to produce a wake. You can't not do that. You can't pretend you're not a big metal lump in a gravity field. You can't spoof that. You might be able to spoof the magnetic signature a little bit, but it's going to get increasingly difficult as the SQUIDS get better and better at, there'll be warfare. And there's no way known. There's just no physics to support the idea that you can shield your neutrino emissions. 

Liam Garman:

Yeah. I suppose we're in this constant battle of looking at ways of cutting corners. I think that's a very fair way of looking at a lot of military technology ways of cutting corners. So instead of detecting a missile, you detect the launch. And by detecting the launch, then you can get a good idea of what's happening with the missile. So you do want to cut these corners to say, looking for the path of least resistance, which is, hey, you can't make a sub quiet or you can't not emit a neutrino if you want the benefits of a nuclear powered submarine. But where in those benefits also lies the critical fault of where the technology can be superseded. 

Roger Bradbury:

One of the other things that's going to play against subs is this broader idea of they are an anomaly in the water column, I'm using oceanographer speak here. They love talking about the water column. We're getting better and better and better fine grained models of the ocean. What it is, not just what the shape of the ocean bottom is, but what the currents do and what moves here and what moves there. So we're getting very predictive with those as well. So we'll be measuring the anomalies against that in a much more subtle ways. In the past, we just tried to go out and drop a hydrophone over the side or use sonar and try and pin things or what have you. That was an active process. But now we'll have a capacity to say, here we've got this model of the ocean, and we've done some sensing of what that spot is, what that bit is. And they're not the same. And that will make autonomous devices look more closely, what's wrong? What's there that shouldn't be there? 

So I think we'll be doing this sensing in a different environment, an environment where we've got a background of what it should be like. And we never had that before. So that's going to make the game hugely difficult. We know that the Russians, once they realised that we were tagging their earlier classes of SSBNs, the Russian response was to withdraw their not fully, but into what they call [Basins 00:35:53]. Mostly the Barents Sea away from… they actually didn't go out anymore. They were relying on the fact that they could move around in that sea and be more or less undetected because it was Russian space. But there was still a second strike capability because you couldn't take them out, because you didn't know exactly where in this Russian ocean they were. So they didn't [inaudible 00:36:20] off New York anymore as their missiles got better. They could hang around, going around in circles in the Sub-arctic, God, you'd want to be a submarine under those circumstances. 

I suppose you don't know what [crosstalk 00:36:35] you don’t know. So that was one response to being more detectable. But if you are detectable in other ways through other senses, and if we could put drones into the Barents Sea and so forth, then that option's not available to them anymore. Just it won't be available to other nations. So, you're suddenly gone from a platform that could patrol the world's oceans with impunity and could be anywhere. It could be sitting off New York. It could be sitting off Sydney in one stage withdrawing to small Bastions, and now those Bastions are going to shrink to nothing. There'll be nowhere where they can go where they're not detected. 

Liam Garman:

Yeah, that's brilliant. And I think a lot of it would be very much reliant on the development of machine learning and AI as well to, going into the future, detect that mean. The mean of what the sea on average should look like, and the mean of how this particular bit of ocean should act when undisturbed and then once it diverges from that mean even in the smallest way, it diverges from that mean, you say, hey, something is causing this to, like I said, diverge from the mean.

Roger Bradbury:

Yeah, yeah. And that means if they detect something's wrong. 

Liam Garman:

Something's wrong. 

Roger Bradbury:

That's a swarm of UAVs, hang in and start to look at what is there. Yeah. Let's just have another look. Yeah. It's a different way of thinking about detection and it's a different way of thinking about counter detection. Yeah. And as I said, at the moment, it's hunky dory for big subs pretty well. Especially if they're quiet. Yeah. And they are extremely, extremely effective machines, but oof. Yeah. I've had a couple of drinks with some Royal Navy submariners, including one guy that was the commander of one of the latest group, one of the Trident machines. That was… yeah, he was getting a little bit vexed about it because-

Liam Garman:

You are putting him out of a job there, Roger. 

Roger Bradbury:

Well, yeah. Well, he was retired. He was retired. We were good friends, but he was more or less saying, well, we had the glory days. We were invincible then, we we'd go out, submerge. That was it. And he acknowledge that, yeah, this is a risk that needs to be taken seriously. It's not a lay down miss there, it's a probability, but it's a high probability. I wouldn't play Russian roulette with eight bullets in a 10 chambered revolver. I’ll put it that way. 

Liam Garman:

Oh, absolutely. Well, I think that's probably all we got time for. And that was, I think the report in getting you onto the podcast could not have happened at a better time, considering the fanfare around this. And a lot of people in Defence and Defence Industry, like we said, caught up in this whirlwind of thinking that we're going to have a huge strategic deterrent, which we will. We will have a huge strategic deterrent, but not without other considerations, which technology has a habit of being superseded, knights on horseback. At one point, the crux, the apotheosis of any [inaudible 00:40:37] attack [or any 00:40:39] military knights on horseback and now forgotten in history. So thank you so much for your time, Professor Roger Bradbury. 

Roger Bradbury:

Okay. Thanks very much, Liam. That was a very, very enjoyable chat. I enjoyed it.

Liam Garman:

And to all of our listeners, thank you very much for listening and staying tuned, and we will see you all next week. Thank you.