Geological Insights from Malaysia Airlines Flight MH370 Search | Hacker News Reader

Geological Insights from Malaysia Airlines Flight MH370 Search

eos.org

160 pointscar38 comments

I wondered what it would cost to get this detail for the entirety of the ocean. They say they mapped an area the size of New Zealand, which is around 250k km^2. According to a quick Google search the total area of the oceans is 360 million km^2, and according to a Guardian article I found[1] the subsurface search cost 180 million AU$.

    (360/0.25)*180 =~ 250 billion AU$

That's around 190 billion USD, not even 0.1 Iraq wars, and that's before considering any economies of scale.

1. https://www.theguardian.com/world/2017/jan/17/malaysia-airli...

It's futile to attempt to value things as fractions of a war. While typically the dollar is a useful metric for the worth that society ascribes to something, public opinion on the importance of the Iraq war and the federal budget for the military are radically different.

The annual budget for the entire USGS is $1.2 billion. NOAA has a budget of $5.6 billion. NASA has a budget of $19.5 billion. The NSF has a budget of about $2 billion. If these agencies dropped everything else and invested purely in this map, it would take about 5 years to create it.

Or, yeah, you could buy not even 0.1 Iraq wars with those 5 years, if that sounded better...

I'm not sure I understand: why is it futile to attempt to value things as fractions of a war given that public opinion on that particular war is divided?

Because it's poor accounting. Given the scale of the US military and it's annual expenditure just existing, to make a fair comparison you should at least take account of the marginal cost of the war, not the total cost.

There is also the opportunity cost to take account of, but that's much harder (impossible?) to put an accurate figure on.

I thought the figure provided was the marginal cost.

It is, of course. Nobody is calculating the cost of the war as the defense budget multiplied by the time. The cost includes about $750 billion in direct appropriations for the war itself (separate from the overall defense budget), plus numerous indirect costs such as additional medical care and disability payments to veterans, replacement of destroyed equipment, and interest payments due to increasing the national debt.

The argument is strange.

Nasa has been able to pull off the Apollo program for roughly $150B adjusted for inflation. Those are not real dollars neither because people supported it ?

It may well be futile, but I will never stop trying to convince people that we should spend less on war and more on science.

It's likely futile to try to convince people to spend less on war but convincing them to spend more on science is doable.

The problem is time.

There aren't that many multibeam sonar vessels in the world. They acquire data over narrow swaths. Mapping all ocean basins with multibeam would take a long time. I've always heard estimates of a few hundred years, though those were based on total the number of academic multibeam vessels operating 24/7. Obviously, it scales with the number of vessels you have doing it, so a large-scale coordinated effort that built lots of new, single-purpose vessels could take significantly less.

At any rate, the cost of one survey made with one ship is not a good way to estimate the total cost, unless you're willing to wait a _really_ long time.

Speaking of war, I suspect the Navy already has maps that good or better than for a variety of reasons can't be publicly released.

A better application of money than redoing work already done, might be talking the navy into releasing their data. Declassification isn't cheap, but its cheaper than having civilians re-run the mission.

At least in 2005, that was demonstrably not true of the US Navy: http://www.popularmechanics.com/military/navy-ships/a24158/u...

That's confusing large scale with small scale. You can prove that one time there existed one mistake on one map (or that story is covering up that there was a navigation and/or helm error, or it covering up an even crazier story like hitting a submerged shipping crate or mine or enemy sub). Certainly no one has proven the public map is error free and flawless, and given that this is possibly the first time anything like this has ever happened to a nazy sub, every other map the navy has must be error free to a fairly high level.

Actually, they probably don't.

The vast majority of the ocean hasn't been mapped with multibeam, civilian or military. Unless it's an area that's of strategic interest (i.e. mostly costal areas), it's very unlikely that the military has mapped it in detail.

That having been said, getting the military to declassify more of its magnetic anomaly data for the ocean basins would be immensely helpful for understanding global tectonic evolution. They have mapped it in quite a bit of detail from airborne and marine platforms. However, an accurate and detailed map of magnetic anomalies gives a huge edge in submarine detection, so it's unlikely they'll declassify all of it anytime soon.

My understanding from some old nukes I know is that the subs are continually remapping the floor as they travel.

They certainly collect data they can get passively (like the magnetic anomaly data mentioned above). They almost certainly do not routinely drive around updating sonar maps (as that would completely defeat the purpose of keeping the location of the sub a secret...)

Companion article: "All of Earth’s ocean floors deeper than 500 meters ... could be mapped by GPS-navigated MBES for 200 ship-years of effort (e.g., 40 ships working for 5 years), at a total cost of US$2–3 billion ... "

https://eos.org/opinions/airline-flight-paths-over-the-unmap...

This estimate is a bit out of date, if you look at the swag in the original article from 2001. But it's on the order of what it would cost. Well worth it. We should prioritize this important work.

> 190 billion USD

190 million :)

That's an exponential increase in resolution, but it's also amazing how the gravity-derived maps fit in so well with the sonar-derived ones. That's some serious validation of the gravity approach, as limited as it is.

Given that about 10% of the ocean has been mapped like this, I'd bet that portion has been used as a calibration for the gravity based dataset.

The gravity-based methods require the ship track data as well -- they're not fully independent methods. (i.e. not only are they calibrated to it, ship-track bathymetry is actually half of the input data for gravity-derived bathymetry.)

It's still really quite amazing how well it matches in an area that's pretty far from dense ship-track bathymetry data.

Really nice article, my mind is always blown by all the cool shit we keep on learning about our Earth.

And then the article became all the more sweeter when I realised that it all this work was done by people I know at Geoscience Australia. Very cool.

Hello from GA! If this sort of stuff is interesting to you, keep an eye out at: http://eos.ga.gov.au

And we've just started an engineering blog which might be of interest to the HN crowd: https://tech.ga.gov.au

Hello from ANU! I've actually been collaborating on a number of projects with GA the last couple of years :-)

Is the raw data available somewhere?

EDIT: found a lot more info (and videos) here : http://www.ga.gov.au/about/projects/marine/mh370-bathymetric...

In light of how little we know, it's great to see that there are philanthropists supporting oceanographic research, specifically a shout out to Eric & Wendy Schmidt, https://schmidtocean.org.

I am always amazed at how relatively little we know about the ocean floor.

Fascinating. The reason for the mapping is tragic, but I hope some of the families so affected can take some degree of solace from the knowledge gained thus.

Cynically I imagine this data will be invaluable to energy companies looking for potential new oil drilling sites - I'm not sure if I'd want my legacy to be the continuation of fossil-fuels...

As a former exploration geologist on the petroleum side and marine geophysicist on the academic side - No, this data isn't very useful for petroleum exploration at all.

Most of the deep ocean basins have absolutely zero oil.

To create thermogenic hydrocarbons, you need thick sediments. You can make methane biogenically without significant sediment cover, and that's important sometimes, but you cannot make oil that way. Oil is what drives exploration far away from population centers.

The abyssal plain you're seeing in that data is mostly exposed oceanic crust. There's almost no sediment cover. No sediments --> no exploration potential.

There's one narrow continental fragment that could have significant sediment cover. However, it's highly unlikely that that's prospective. (If you're curious, I'm happy to elaborate, but it will take a bit of explanation.)

There is one way this data will be used for petroleum exploration. It aids our understanding of the tectonic evolution of the area, which in turn aids petroleum exploration. However, if you're going to take a moral stance against that, you also need to a moral stance against every scientific advance in every field. They all aid the oil industry indirectly. (Ditto for software.)

Yes please elaborate, this is fascinating! Can you recommend any good undergraduate textbooks on the subject, especially for someone with a quantitative background (PhD in ML-adjacent field)?

If you're interested specifically in the petroleum side, I've heard "Introduction to Petroleum Exploration for Non-geologists" by Robert Stonely is good. There are probably more up-to-date books out there, though.

For the more general geology portion, if you're wanting something that's a bit more detailed, but still very approachable, "Evolution of the Earth" by Prothero and Dott is absolutely fantastic. You can probably find a slightly older paperback edition quite cheaply, as well.

If you're really okay with a quantitative approach, Turcotte & Schubert's "Geodynamics" walks through an impressive amount of basic tectonics in a "let's do the back of the envelope calculations" manner. It's the most dog-eared book on my shelf. It's incredibly useful. However, it's a graduate level textbook, and it assumes you have at least a basic understanding of geological processes and terminology. It would be a very tough way to dive in.

Great comment, thanks for the information. I suspected that the deep ocean wouldn't be good for oil, but only because it would be logistically difficult.

Why isn't there sediment cover in these areas? I would naively think that you'd get some raining down as plants and animals die in the shallower depths. Clearly that must not be the case, but why not?

In a nutshell, you're a long way from a sediment source. Most sediments by volume come from erosion of the continents or other landmasses. There's a much higher rate of deposition if you're near a continent (e.g. river deltas).

Additionally, oceanic crust is always very young geologically. It's continually formed and destroyed (subducted). The ocean basins haven't had much time to fill up with sediment, because oceanic crust is essentially a conveyor belt.

-----

So the short answer is: You're depositing things very slowly on something that doesn't have much time for sediments to build up.

-----

In a bit more detail, the sediments you get in the middle of the ocean are either the initial volcanic-related sediments that were deposited as the oceanic crust formed, or they're limited to "pelagic rain" (things near the surface dying and very fine particles of clay) and some chemical sediments (things that precipitate directly out of water).

Not only that, but a lot of the things that die and sink will dissolve below about 2 Km water depth (see Carbonate Compensation Depth).

Therefore, deepwater depositional rates are very, very slow.

There's a very important exception, though -- Carbonates. If you start in shallow water (say, due to a small sliver of continental crust), organisms in the photosynthetic zone can produce huge amounts of carbonate sediments just from their shells (e.g. coral reefs).

Under the right conditions, carbonates can keep growing and keep pace with the subsiding crust. A good example of this is the Bahamas. You have several kilometers of sediment made of nothing but what grew there and died. There are some very significant oil fields in these sorts of "isolated platforms".

However, you need fairly unique conditions to get oil in the sort of isolate carbonate platforms that develop far away from land. In most cases you'd find out in the ocean basins, you won't have a source rock capable of generating hydrocarbons beneath the pile of sediments that the carbonates produce. Instead, the source rock is conveniently sitting out there in the deep water where it's not buried deep enough to produce oil.

There's one other exception, and that's when you're near a volcano. Volcanoes will happily produce lots of volcanic-related sediments. However, volcaniclastic sediments almost never make good reservoirs (there are exceptions) or seals, so even if you did generate oil in a narrow band around a seamount, it would either leak out (no seal) or be trapped in something that you can't get it out of (bad reservoir).

And there's the longer answer :)

Delightful. Thanks so much for taking the time to write out the longer answer. I think that "pelagic rain" is what I was referring to before, but I didn't realize that the quantity was so low, or that subduction would keep it from building up in most cases. Nor, of course, did I realize all of this other great stuff.

Does anyone know if the dataset is available publicly?

According to http://www.ga.gov.au/about/projects/marine/mh370-bathymetric... "Data collected as part of the bathymetric survey will be publicly released by Geoscience Australia in due course."

Reply on news.ycombinator.com