Every year, AI models get better at thinking. Could they possibly be capable of feeling? And if they are, how would we know?

Rare and powerful compounds, known as keystone molecules, can build a web of invisible interactions among species. The post A New, Chemical View of Ecosystems first appeared on Quanta Magazine

[Note that this article is a transcript of the video embedded above.] Lewis and Clark Lake, on the border between Nebraska and South Dakota, might not be a lake for much longer. Together with the dam that holds it back, the reservoir provides hydropower, flood control, and supports a robust recreational economy through fishing, boating, camping, birdwatching, hunting, swimming, and biking. All of that faces an existential threat from a seemingly innocuous menace: dirt. Around 5 million tons of it flows down this stretch of the Missouri River every year until it reaches the lake, where it falls out of suspension. Since the 1950s, when the dam was built, the sand and silt have built up a massive delta where the river comes in. The reservoir has already lost about 30 percent of its storage capacity, and one study estimated that, by 2045, it will be half full of sediment. On the surface, this seems like a silly problem, almost elementary. It’s just dirt! But I want to show you why it’s a slow-moving catastrophe with implications that span the globe. And I want you to think of a few solutions to it off the top of your head, because I think you’ll be surprised to learn why none of the ones we’ve come up with so far are easy. I’m Grady, and this is Practical Engineering. I want to clarify that the impacts dams have on sediment movement happen on both sides. Downstream, the impacts are mostly environmental. We think of rivers as carriers of water; it’s right there in the definition. But if you’ve ever seen a river that looks like chocolate milk after a storm, you already know that they are also major movers of sediment. And the natural flow of sediment has important functions in a river system. It transports nutrients throughout the watershed. It creates habitat in riverbeds for fish, amphibians, mammals, reptiles, birds, and a whole host of invertebrates. It fertilizes floodplains, stabilizes river banks, and creates deltas and beaches on the coastline that buffer against waves and storms. Robbing the supply of sediment from a river can completely alter the ecosystem downstream from a dam. But if a river is more than just a water carrier, a reservoir is more than just a water collector. And, of course, I built a model to show how this works. This is my acrylic flume. If you’re familiar with the channel, you’ve probably seen it in action before. I have it tilted up so we get two types of flow. On the right, we have a stream of fast-moving water to simulate a river, and on the left, I’ve built up a little dam. These stoplogs raise the level of the water, slowing it down to a gentle crawl. And there’s some mica power in the water, so you can really see the difference in velocity. Now let’s add some sediment. I bought these bags of colored sand, and I’m just going to dump them in the sump where my pump is recirculating this water through the flume. And watch what happens in the time lapse. The swift flow of the river carries the sand downstream, but as soon as it transitions into the slow flow of the reservoir, it starts to fall out of suspension. It’s a messy process at first. The sand kind of goes all over the place. But slowly, you can see it start to form a delta right where the river meets the reservoir. Of course, the river speeds up as it climbs over the delta, so the next batch of sediment doesn’t fall out until it’s on the downstream end. And each batch of sand that I dump into the pump just adds to it. The mass of sediment just slowly fills the reservoir, marching toward the dam. This looks super cool. In fact, I thought it was such a nice representation that I worked with an illustrator to help me make a print of it. We’re only going to print a limited run of these, so there's a link to the store down below if you want to pick one up. But, even though it looks cool, I want to be clear that it’s not a good thing. Some dams are built intentionally to hold sediment back, but in the vast majority of cases, this is an unwanted side effect of impounding water within a river valley. For most reservoirs, the whole point is to store water - for controlling floods, generating electricity, drinking, irrigation, cooling power plants, etc. So, as sediment displaces more and more of the reservoir volume, the value that reservoir provides goes down. And that’s not the only problem it causes. Making reservoirs shallower limits their use for recreation by reducing the navigable areas and fostering more unwanted algal blooms. Silt and sand can clog up gates and outlets to the structure and damage equipment like turbines. Sediment can even add forces to a dam that might not have been anticipated during design. Dirt is heavier than water. Let me prove that to you real quick. It’s a hard enough job to build massive structures that can hold back water, and sediment only adds to the difficulty. But I think the biggest challenge of this issue is that it’s inevitable, right? There are no natural rivers or streams that don’t carry some sediments along with them. The magnitude does vary by location. The world’s a big place, and for better or worse, we’ve built a lot of dams across rivers. There are a lot of factors that affect how quickly this truly becomes an issue at a reservoir, mostly things that influence water-driven erosion on the land upstream. Soil type is a big one; sandy soils erode faster than silts and clays (that’s why I used sand in the model). Land use is another big one. Vegetated areas like forests and grasslands hold onto their soil better than agricultural land or areas affected by wildfires. But in nearly all cases, without intervention, every reservoir will eventually fill up. Of course, that’s not good, but I don’t think there’s a lot of appreciation outside of a small community of industry professionals and activists for just how bad it is. Dams are among the most capital-intensive projects that we humans build. We literally pour billions of dollars into them, sometimes just for individual projects. This is kind of its own can of worms, but I’m just speaking generally that society often accepts pretty significant downsides in addition to the monetary costs, like environmental impacts and the risk of failure to downstream people and property in return for the enormous benefits dams can provide. And sedimentation is one of those problems that happens over a lifetime, so it’s easy at the beginning of a project to push it off to the next generation to fix. Well, the heyday of dam construction was roughly the 1930s through the 70s. So here we are starting to reckon with it, while being more dependent than ever on those dams. And there aren’t a lot of easy answers. To some extent, we consider sediment during design. Modern dams are built to withstand the forces, and the reservoir usually has what’s called a “dead pool,” basically a volume that is set aside for sediment from the beginning. Low-level gates sit above the dead pool so they don’t get clogged. But that’s not so much a solution as a temporary accommodation since THIS kind of deadpool doesn’t live forever. I think for most, the simplest idea is this: if there’s dirt in the lake, just take it out. Dredging soil is really not that complicated. We’ve been doing it for basically all of human history. And in some cases, it really is the only feasible solution. You can put an excavator on a barge, or a crane with a clamshell bucket, and just dig. Suction dredgers do it like an enormous vacuum cleaner, pumping the slurry to a barge or onto shore. But that word feasible is the key. The whole secret of building a dam across a valley is that you only have to move and place a comparatively small amount of material to get a lot of storage. Depending on the topography and design, every unit of volume of earth or concrete that makes up the dam itself might result in hundreds up to tens of thousands of times that volume of storage in the reservoir. But for dredging, it’s one-to-one. For every cubic meter of storage you want back, you have to remove it as soil from the reservoir. At that point, it’s just hard for the benefits to outweigh the costs. There’s a reason we don’t usually dig enormous holes to store large volumes of water. I mean, there are a lot of reasons, but the biggest one is just cost. Those 5 million tons of sediment that flow into Lewis and Clark Reservoir would fill around 200,000 end-dump semi-trailers. That’s every year, and it’s assuming you dry it out first, which, by the way, is another challenge of dredging: the spoils aren’t like regular soil. For one, they’re wet. That water adds volume to the spoils, meaning you have more material to haul away or dispose of. It also makes the spoils difficult to handle and move around. There are a lot of ways to dry them out or “dewater” them as the pros say. One of the most common is to pump spoils into geotubes, large fabric bags that hold the soil inside while letting the water slowly flow out. But it’s still extra work. And for two, sometimes sediments can be contaminated with materials that have washed off the land upstream. In that case, they require special handling and disposal. Many countries have pretty strict environmental rules about dredging and disposal of spoils, so you can see how it really isn’t a simple solution to sedimentation, and for most cases, it often just isn’t worth the cost. Another option for getting rid of sediment is just letting it flow through the dam. This is ideal because, as I mentioned before, sediment serves a lot of important functions in a river system. If you can let it continue on its journey downstream, in many ways, you’ve solved two problems in one, and there are a lot of ways to do this. Some dams have a low-level outlet that consistently releases turbid water that reaches the dam. But if you remember back to the model, not all of it does. In fact, in most cases, the majority of sediment deposits furthest from the dam, and most of it doesn’t reach the dam until the reservoir is pretty much full. Of course, my model doesn’t tell the whole story; it’s basically a 2D example with only one type of soil. As with all sediment transport phenomena, things are always changing. In fact, I decided to leave the model running with a time-lapse just to see what would happen. You can really get a sense of how dynamic this process can be. Again, it’s a very cool demonstration. But in most cases, much of the sediment that deposits in a reservoir is pretty much going to stay where it falls or take years and years before it reaches the dam. So, another option is to flush the reservoir. Just set the gates to wide open to get the velocity of water fast enough to loosen and scour the sediment, resuspending it so it can move downstream. I tried this in the model, and it worked pretty well. But again, this is just a 2D representation. In a real reservoir that has width, flushing usually just creates a narrow channel, leaving most of the sediment in place. And, inevitably, this requires drawing down the reservoir, essentially wasting all the water. And more importantly than that, it sends a massive plume of sediment laden water downstream. I’ve harped on the fact that we want sediment downstream of dams and that’s where it naturally belongs, but you can overdo it. Sediment can be considered a pollutant, and in fact, it’s regulated in the US as one. That’s why you see silt fences around construction sites. So the challenge of releasing sediment from a dam is to match the rate and quantity to what it would be if the dam wasn’t there. And that’s a very tough thing to do because of how variable those rates can be, because sediment doesn’t flow the same in a reservoir as it would in a river, because of the constraints it puts on operations (like the need to draw reservoirs down) and because of the complicated regulatory environment surrounding the release of sediments into natural waterways. The third major option for dealing with the problem is just reducing the amount of sediment that makes it to a reservoir in the first place. There are some innovations in capturing sediment upstream, like bedload interceptors that sit in streams and remove sediment over time. You can fight fire with fire by building check dams to trap sediment, but then you’ve just solved reservoir sedimentation by creating reservoir sedimentation. As I mentioned, those sediment loads depend a lot not only on the soil types in the watershed, but also on the land use or cover. Soil conservation is a huge field, and has played a big role in how we manage land in the US since the Dust Bowl of the 1930s. We have a whole government agency dedicated to the problem and a litany of strategies that reduce erosion, and many other countries have similar resources. A lot of those strategies involve maintaining good vegetation, preventing wildfires, good agricultural practices, and reforestation. But you have to consider the scale. Watersheds for major reservoirs can be huge. Lewis and Clark Reservoir’s catchment is about 16,000 square miles (41,000 square kilometers). That’s larger than all of Maryland! Management of an area that size is a complicated endeavor, especially considering that you have to do it over a long duration. So in many cases, there’s only so much you can do to keep sediment at bay. And really, that’s just an overview. I use Lewis and Clark Reservoir as an example, but like I said, this problem extends to essentially every on-channel reservoir across the globe. And the scope of the problem has created a huge variety of solutions I could spend hours talking about. And I think that’s encouraging. Even though most of the solutions aren’t easy, it doesn’t mean we can’t have infrastructure that’s sustainable over the long term, and the engineering lessons learned from past shortsightedness have given us a lot of new tools to make the best use of our existing infrastructure in the future.

A new proof extends the work of the late Maryam Mirzakhani, cementing her legacy as a pioneer of alien mathematical realms. The post Years After the Early Death of a Math Genius, Her Ideas Gain New Life first appeared on Quanta Magazine

Remember CRISPR (clustered regularly interspaced short palindromic repeats) – that new gene-editing system which is faster and cheaper than anything that came before it? CRISPR is derived from bacterial systems which uses guide RNA to target a specific sequence on a DNA strand. It is coupled with a Cas (CRISPR Associated) protein which can do […] The post The New TIGR-Tas Gene Editing System first appeared on NeuroLogica Blog.