The science of sinkholes

or, The science of the news of the day

So in case you’ve missed it, the Earth decided to take a giant bite out of Guatemala.

The most common image is an amazing picture – truly captivating. But it’s also a great trigger for some science learning. A teachable moment, you could say.

Wandering around the Tubes, you can find plenty of coverage, mostly concerned with the standard news angle – Did anyone die? How big is it? – totally bypassing the science-y questions of HOW this happens. Maybe everyone already knows?

I doubt it.

I’ve been playing with the idea of current affairs science journalism for the last little while – using science to explain the news of the day. I had my first brush with the idea working on the Haiti earthquake coverage at Daily Planet back in January. I think it’s a great model, but I don’t see it done often.

Explaining natural disasters is a bit of a gimme, but an event like this, that manages to temporarily captivate people’s imaginations, should not be passed up.

I wrote up a quick Question & Answer with a USGS geologist, Daniel Doctor, to try to answer some of the fundamental questions about sinkholes – the how, why, and where. That story can be found over on the American Geophysical Union’s blog, Geohazards.

I don’t know how many people will find their way to that post, but hopefully people who became captivated by the imagery of the event will be tempted to learn a little more, and I think it would be a shame to pass on such a great opportunity to share some science.

The discussion I had with Doctor was very interesting, and covered way more material than could have ever been compressed into the space constraints of the Geohazards post. But, if you’ve even glanced around my little corner of the internet, you’ll notice I tend to run long.

The full transcript of the interview can be found below. It covers some really fascinating things like volcano tubes, the impact of manure lagoons on stream ecosystem health in regions with karst terrain, and how around 15 per cent of the land area of the lower 48 states is susceptible for sinkholes.

Colin: So the first obvious question is – when I’m looking at that picture, especially of the one in Guatemala, it’s just this giant hole. Where does all that land go?

Daniel: It goes into a pre-existing void in the sub-surface, and most commonly those dramatic type sinkholes that have these nice round expressions and steep sides to them are what we call cover-collapse sinkholes: that is, along some part of the overlying land surface, there’s a void beneath it and the cover is either unconsolidated weathered rock, or some sediment that’s been deposited on top of rock, with a void that’s developed underneath it. But yeah, the answer is that there is some pre-existing void in the subsurface. So then the question becomes, how do those pre-existing voids arise?

That can be through a wide variety of causes. When we talk about natural sinkholes, that is something that doesn’t seem to have been induced by any kind of human activity, we most typically talk about karst terrain. And karst is a geomorphic term that refers to landscape features that result from the voids in the sub-surface that result from the solubility of the bedrock. So, it’s essentially water dissolving out rock in the sub-surface.

Now not all rocks are dissolved easily, and the ones that tend to become most easily dissolved are limestones, and carbonate rocks in general – limestones and dolomites. And what we call evaporite rocks, or salt and gypsum.

C: So just, as I’m trying to picture it, you were talking about it being a cover-collapse sinkhole. So what I’m picturing is, it was a solid piece of Earth, and then some water got underneath, eroded away the lower layers, and then above that hole there was rocks with soil on top of it?

D: Well think of it more like, soil covering a bedrock surface that is highly irregular – pinnacled. You may have seen dramatic photos of karst terrains in areas like China or Taiwan, or Madagascar where they have these knife-life spires of limestone sticking up into the air. Those develop in tropical regions. But in say, Virginia, we have karst which has pinnacles, not that dramatic, but they are pinnacles in the subsurface that are coated with the weathered remains of the limestone, the stuff that didn’t dissolve away.

And so what you’d have if you were to strip off all of that weathered remain is you would see a pinnacled land surface that may have relief of up to 30 or 40 feet between these pinnacles, and there’s just dirt and sediment in between.

And what happens is, at the boundary between the soil and the rock you often have water that can run rather rapidly, especially where there is a gradient to the water table, and that erodes out the sediment at the contact between the sediment and the rock. Eventually it can work its way upward and cause the slow collapse of the sediment at that interface. It may eventually, in sort of an arching fashion, read the land surface, and then collapse into a big hole.

In Guatemala, the sediment was mostly volcanic ash, and this is the second one for a matter of fact, to open up in Guatemala City.

C: Yeah, there was one in 2007.

D: Yeah, and they’re both – this is a volcanic terrain. And in volcanic terrains you get the same kind of thing happening. It’s not, in this case, the weathered remains of bedrock as the unconsolidated stuff, it’s volcanic ash. But it could still be coating a somewhat irregular surface. It’s not that you need the irregular surface, it’s just whenever you get pipes naturally forming in sediment to varying degrees and varying sizes, you get the potential to create sinkholes, because it’s essentially creating voids in the subsurface.

C: So is that sort of the major cause of a sinkhole? Those voids?

D: Well that’s the underlying cause you might say. Now how those voids arise, that’s the challenge for hazard management. Some of them can arise naturally. So over time, like I said, certain soluble rocks – carbonates and evaporites – will naturally dissolve out caves. And in areas where we know we have high occurrences of caves we invarriably find sinkholes. They’re just a typical feature of karst terrains in general.

Are you familiar with the term karst?

C: Um, just from the research I’ve been doing today on it.

D: Oh okay. [Karst] is a geomorphological term applied to landscapes that show these features, like sinkholes, like caves, like springs and sinking streams. And these all arise from the flow of water through soluble bedrock. And when I say soluble that is, it dissolves naturally in the water to create essentially a pipe network through the rock. And that happens at all varying degrees and in different ways. Sometimes it’s all from the surface, sometimes it can be in the subsurface and you might not see it at the surface. But generally where we know there are caves we find sinkholes as well.

C: So the next question I had for you was – how quickly can this process happen? Both the natural erosion of the soluble bedrock, and even just going from what seems like a normal street to now being a big hole? How long do those processes take?

D: So like the example of the street in Guatemala, it could take place overnight. Commonly, for those kinds of sinkholes that are just occurring due to anthropogenic activity – you could have a watermain break, for example. And if a watermain breaks, the gushing water that’s coming out of the pipe begins to erode away the soil in the subsurface, you have the potential to create a sinkhole, because you now have created – very quickly you’ve created a void in the sediment in the subsurface.

We see this very often after heavy rains, not only in karst terrains but in and around cities with old aging sewer lines, like Washington D.C. for example. I often hear of sinkholes that open up after rain storms. And they’re not karst sinkholes, this is the distinction.

They’re not in areas where you would expect to have natural sinkholes. They’re man-made in a sense because of the failing urban conveyance of water in the subsurface through pipes that age and break open. And when that water leaks out of the pipes it finds its own way through the soil and erodes the soil, or whatever material it’s moving through. As it goes it leaves behind voids that can then potentially collapse.

So your question is a good one, the timescale is very, very variable – anywhere from overnight to hundreds, thousands, maybe millions of years in the case of some natural karst terrains.

C: But let’s say from the perspective of someone sitting at home or standing on the street, would there be – like how quickly would it seem to them? Like there’s this process happening underground, but what would it look like?

D: I saw something online – a woman was interviewed in Florida, she said she was out gardening in her backyard. And she had like a little enclosed garden, smaller than the size of her house for sure. And a minute later she was down in this eight-foot-deep hole. She was gardening, and the next thing she knew the ground collapsed underneath her, and she was stuck in the hole. She had to call for help, and she had no way to climb out of the hole, and eventually somebody walked by and heard her cries and called the fire department and they came and got her out. So, very quickly in some cases.

In other cases, people can actually – you can observe over time a depression, what looks like a small depression in your front yard, progressively open up and get larger and larger. There are cases in Texas where they’ve been slowly observing their backyards and houses erode away. It’s a cause of great consternation for people who know it’s sort of an impending, creeping, chronic disaster.

I like to use the word chronic when I talk about sinkholes because in society we tend to focus on those acute disasters, the ones that hit quickly, and cause a lot of damage in a very short amount of time, and affect a lot of people dramatically and in a devastating way. Whereas sinkholes in general tend to be there out on the landscape all the time in isolated places, and we hear of them in news flashes – they’ve affected maybe a few people, maybe a city, not a lot of death and destruction – but it’s chronic, it’s there all the time.

I haven’t gotten any numbers, but I know it costs hundreds of millions of dollars a year in infrastructure to repair across the country. Unfortunately I don’t have any documentation of that, I’m trying to work on it, but just anecdotally I think I have some confidence in saying it’s millions of dollars a year of damage, if not more.

C: So in terms of just size scale, what’s the range they fall in?

D: Well on the very smallest scale, something you could step in, is what you kind of want to be concerned about. That’s damaging to you if you’re walking across a landscape and you step in a hole. More damaging though to livestock and large animals that people might have investments in. So that’s on the small end.

And then on the large end, oh gosh, I know of large collapse features in China and in Mexico that are a hundred meters in diameter.

C: But there’s no, do you think there’s like a possibility that there could be this massive underground aquifer, or maybe cave system – would it ever have a large scale collapse, or is it normally more these localized things?

D: Caves actually do not have a tendency to be frequently collapsing, so that we would see for example a sinkhole collapsing into what then could be an explored cave system. That’s not what we typically hear about when we hear about sinkholes. Cave collapses happen, of course, that’s how we find caves, we see collapses that occur into cave systems all the time, but they’re mostly – people who are privy to those occurrences are cavers, people who are looking for them. The kinds of things that normally you hear about on the news that occur in cities aren’t typically the result of a natural cave under the city. Typically there is some part of the urban infrastructure that has failed, or a consequence of urban development, such as routing water into a preexisting depression on the landscape surface that looked innocuous, but as you ponded water within it, it failed due to the overlying water pressure.

That happens very often in karst terrains, where people build retention ponds to collect stormwater drainage. If they’re not built properly, they can often fail catastrophically.

In agricultural areas they can store ponds of manure in what they call manure lagoons, and in karst terrains it’s convenient to use a preexisting depression that may have no open throat at the bottom of it, you know no open hole for any liquid to drain into. It may be just soil or clay or something in the bottom of it, but the void may not be evident at the surface, and once you pond water into it, you can induce leakage that can open up a throat. Then these things have failed and then they get directly into the groundwater system and that’s when the trouble really begins.

Many cases in Minnesota and Iowa in the upper midwest where they’ve stored manure in these lagoons, and they’ve collapsed, and they’ve affected streams and had large fish mortalities because of the sewage getting into the streams, springfed streams – so it goes into the groundwater and comes up from the springs into the streams, but it kills the fish.

C: The next question – I’m just sort of looking for the who, what, where, when, why of sinkholes. Like you’ve already broken a bunch of assumptions I had – I was picturing a giant cave under Guatemala that nobody knew about, and this just fell into it.

How often do these large sinkholes take place? Is it just something we only hear about sometimes? Or are they actually rare events?

D: I’d say they’re more rare events, these very large sinkhole collapses. I think they’re more rare events in natural settings, and they’ve become more frequent events in human settlements because we do a lot in the subsurface. We do disturb the subsurface quite a bit. I’m trying to remember exactly what they reported about this Guatemalan sinkhole, was it a water main break? Or a sewage pipe break? Something like that?

C: I haven’t really heard anything about that. I know in 2007 it was attributed to a a sewer main, or a water main, something along that line. But for this I haven’t heard much about it.

D: That would be my first guess, because it’s in the same city, and the same kind of geologic environment.

C: They were attributing it to the hurricane, would that just be them trying to divert the rainfall, and then there was a leak in the system they have set up for that? Likely? I’m not going to pin you to that.

D: Um, it’s really tough to say. Given that what I do know about the first one, that it was mostly just volcanic ash, and that they didn’t find, say, a lava tube.

Now, in volcanic terrains you can have caves – they are of a different sort. They are what are called lava tubes. These are, you may have seen images of rivers of lava flowing on the surface, you know eruptions in Hawai’i. Well as these things erupt, you can imagine the hot lava cooling, if it’s still in the subsurface, you can imagine it cooling on its edges, kind of forming an eggshell, and the river keeps flowing, but now it’s encased in this shell. And when all of that lava stops flowing and finally cools and crystalizes, you have not only a floor of hard lava rock, you have a void, a cave, along the pathway where the lava had flowed – and you can explore them, and some of them are miles long in Hawai’i, and in California, and in Oregon, all throughout the Cascades they’re fairly common.

I don’t know if that’s the case for Guatemala, I actually don’t know the geology well enough. But what I do know is that it’s not as typical to find ash flows in the same area where there lava flows, because the type of lava that you get that produces ash, and the one that produces something flowing are quite different volcanoes. Not that you can’t have them both in the same area, but it’s not as likely.

I’m making everything difficult for you, I’m sorry, hah.

Don’t pin me down on the cause of the Guatemalan sinkhole, I just haven’t learned about it enough.

C: I assume the only people who might start to know right now are the Guatemala Geologic Survey, or whoever is in charge of it down there.

D: Probably, yeah.

C: Now you’ve been talking about the sort of effect that human settlements can have. What’s sort of the main way that we impact …

D: Number one I would say routing stormwater drainage when you’re in a karst terrain. That is, limestone terrains, limestone and evaporite terrains. Routing stormwater drainage is a big problem. Road run-off, stuff like that.

Number two is, well it’s kind of a toss up. I’d either say failing urban infrastructure for water conveyance. Either bringing water to people’s homes, or taking away sewage, or storm run-off. When those things fail, the water has to go somewhere, and it will eventually erode on its way. And so you can create collapses that way.

There is some good evidence that quarrying can have an effect, but that’s a very difficult thing to actually prove. Quarries, as a function of their operation, have to pump a lot of water out of the hole that they’re creating. When they pump it out, they have to put it somewhere, and if it’s in the karst terrain there is some evidence that has a negative impact. But again, like I said, that’s on a case-by-case basis, and I would not like to say that all quarries cause sinkhole damage, it’s not true.

C: But that’s sort of the main – It all revolves around humans moving water around?

D: If you’re asking how we humans impact sinkhole formation. Now how nature contributes to sinkhole formation… it’s essentially the same thing. It’s moving a lot of water around, quickly. That’s when we typically find sinkholes occurring. Or, moving it around after a long drought. You also see a higher frequency of sinkhole formation.

C: Now, just as my own little theory, would it have anything to do with human water use dropping the water table, which would – maybe the water is having an upward pressure that would keep the land there, and then when the water table drops it would cause a sinkhole? Or is that not one of the involved processes?

D: No in fact that’s right. That is one of the processes involved. And that’s one of the one’s that is really hard to prove. A great example we had here in Virginia – a house completely collapsed into a sinkhole. It was a small house, maybe the size of a mobile home, but a sinkhole opened up right underneath it and the whole house was destroyed. Now coincidentally they happened to be drilling a well nearby. But there’s no way to prove that drilling the well induced the sinkhole collapse. Although what you’re saying, that dropping the water table reduces the amount of support in the overlying soil is true. And so, that is one way you can induce sinkhole collapse.

C: I think that’s it that I was looking for, is there anything that you wanted to add?

D: I’m making a map of karst terrain across the United States right now, it’s one of the tasks of the USGS. It’s not under a hazards program, but we’re doing it just for the good of the public. A similar map had been published in the 1970’s, and we’re making it digital data. And because it’s all digital we can do neat nifty things like calculate what is the percentage land area that is potentially karst.

It turns out that for the lower 48 states, just around 15% of the United States is what we would consider potential karst terrain. I wouldn’t say it’s all karst, but it’s underlain by the soluble bedrock types – carbonate, evaporites, and I also include lava tubes in places where know there are explorable lava tubes. So about 15% of the United States could have the potential for karst.

C: So that would mean that about 15% of the land area in the lower 48 states is potential for sinkholes?

D: Yes.

C: And then that would lead to how we deal with planning our water systems? That sort of thing?

D: Well I wouldn’t go that far, because that doesn’t include cities like Washington D.C., which wouldn’t even be considered in a karst terrain and yet has sinkholes nearly every time it rains heavily. I would not say that’s potential for karst based on what we know of the natural geology. And I used the word karst in particular there. Now, sinkholes, that’s a technical term that’s actually rather loosely used. In America we call pretty much any closed depression on the land surface a sinkhole, but if you were a European geologist you would prefer the word doline.

The word comes from the country of Slovenia, and Slovenia is where karst terrain was originally described, and the word actually is derived from a Sloven word – kras. That’s the name of a particular region in Slovenia, and I happen to have done my PhD research there so I know it pretty well. Well when the area was described in 1689, a man named Valvasor wrote in German, and took the area’s name, Kras, and turned it into Karst. So he was describing the land features he saw in this area, large sinking streams, big caves with rivers in them. One cave has a bridge in it that is 300 feet above the surface of the stream, and you can cross it, just a big enormous cave, and lots of sinkholes everywhere. Beautiful country. But that’s where the word comes from.

It actually presents a little bit of difficulty for geologists because it is a geographical term, it is a very descriptive term, so it begins to be difficult to try to categorically place the areas that don’t have real visible expressions of sinkholes or caves at the surface, and yet they’re there in the subsurface when you drill wells. Well what do you call it, do you call it karst or not? Well some people say yes, some people say no. I’ve taken the position that we will include those areas with the right geology for karst, even if there is a lack of surface expression of karst. And, we have included those areas to as much of an extent as possible where we know subsurface voids exist in our mapping, because they do represent a hazard for oil exploration, or salt mining or things like that.

C: Is this map that you’re working on, is that available online at all? Either the original or the updated version?

D: Yes. The published versions can be found [here and here.]

C: Okay, thanks so much.