 It's one o'clock on Tuesday, February the 15th. I'm your host, Pete McGinnis-Mark, and you are watching Science at Soast. This whole series is intended to showcase some of the exciting new research being conducted at the School of Ocean and Earth Science and Technology at UH Manoa, that's Soast. And we bring in both graduate students as well as postdocs, young scientists to tell us something about their exciting new research. And I'm really pleased to say today we have Nadia Mosewa, who is from the Department of Atmospheric Sciences. Nadia is a specialist in the physics of aerosol transport and atmospheric phenomena. And today we're going to be talking about a topic which is close to my heart, tracking volcanic fog. And if any of you live on Oahu or on the Big Island and have allergies, you will certainly know a lot about fog because it's not very pleasant. So, Nadia, welcome. I apologize if I've corrupted your last name. I would just call you Nadia, but thank you very much. Thank you very much for appearing on the show. And why don't you tell the audience a little bit, what do you do? Your specialty is atmospheric sciences, correct? That's correct. So I am a researcher with the Vogue Measurement and Prediction Program at Wachmonewa. And I've joined about nine months ago, so not too long ago, after completing my doctoral work in atmospheric science. So I did that at the University of British Columbia in Canada. And my work now focuses on numerical and dispersion modeling with kind of a broad goal of improving air quality prediction in Hawaii and providing model guidance for hazard mitigation. OK. And let's just back up a little bit. Maybe some of our viewers aren't familiar with the term fog. Can you give us a more general public description of what is fog? Right. So fog is a bit of a Hawaiian slang, actually, because it's a combination word, kind of joining volcanic and smog, but mostly in other places around the world. It's referred to as lace. So it's this haze that you see. It's in Hawaii. It's predominantly made up of sulfur dioxide and sulfate aerosols. Both of them are harmful pollutants and they affect visibility. And they have impacts on human health. They have impacts on agriculture and vegetation. So they're not great. And so we try to model how they move in the atmosphere and how downwind communities might be impacted and to what extent. And so then the state officials, the public, they can plan to protect themselves from fog and reduce the exposure. And so fog is directly associated with ongoing volcanic activity at Kilauea Volcano? That's correct. So Kilauea Volcano is pretty special in that it's been erupting near continuously for like four decades. So since 1983. And so it keeps putting out a substantial amount of emissions on a quite continuous basis. You can think about it if you take the dirtiest producing plant in mainland US, it produces an order of magnitude more pollutants on a fairly regular basis. And it does so for a very long period of time. And so that's why we typically get fog fairly regularly in Hawaii. Obviously, people on the big island would suffer from fog more often than those of us here on Oahu. Is it island-wide? I suspect you'll tell us where some of the plumes will go. But we only get fog here on Honolulu infrequently. Is that just due to the volcano shutting off or is it due to atmospheric air circulation? Yeah, so this is an excellent question. It's actually related to the predominant weather patterns that we get in Hawaii. So often we get what we call trade winds. And so those are the winds that go from the east to the west, actually from the northeast to the southwest mostly. And it relates to this high pressure system that sits just to the north of Hawaii islands. And it's responsible for these trade winds. And so we get this pattern quite a lot. So often the fog just kind of blows out from Kilauea into the ocean. Some of it gets recirculated back. And we'll see that on some of the slides I'm going to show. But generally, things move left to right on the map and most of it ends up over the ocean. And thankfully not over Oaxaca where we have, you know, lots of people living. So most of our population, yeah. So a lot more people. Lots of allergic people, yes. Well, for viewers watching on the mainland, maybe they aren't familiar, but I think your first slide will show just a little bit of the effects of this fog. So Michael, if we could have the first slide and tell us Nadia, what is it we're looking at? It looks like two totally different localities. Unfortunately not. So that's the Kona coast on the big island on a clear day versus under rock conditions. And so this happens quite regularly, unfortunately, when we have volcanic activity. And you would think that the Kona coast, that's on the Leeward side, and it would be protected by, you know, the tall peaks we have in between the volcano, the Kilauea summit and the Leeward side, but that's not the case because a lot of it gets recirculated back to the mainland. And so this haze that you see is actually, so sulfur dioxide or SO2 and sulfate aerosols. And they create that, not so they don't only create those health impacts, they also have, as you can tell, visibility effects that impacts aviation. And when we have those trade wind conditions, we often get this strong inversion from the above, which sort of keeps a lid on them. So they keep getting mixed down to the surface. And so all of those pollutants are trapped near the surface where all the people live. And so it's a big problem. I was gonna ask what altitude above sea level would that haze be? And you just said it can get trapped almost down that sea level. Yeah, so it's, and again, some of those slides that I'm going to show will probably make this a bit easier to visualize. But generally the layer closest to the surface about a kilometer deep, we call it the boundary layer. And within us is a mixing layer where things really get constantly mixed in during the day by ambient turbulence. And this layer is about a kilometer deep and things generally get kind of uniformly mixed through the depth of this layer. And most of the volcanic emissions end up in this mixed boundary layer. And so whatever we get out of the volcano ends up getting well, well, well mixed through and just first down to the downwind communities. So that's not great. You've called for some of the other slides. So let's take a look at the second one. I think we'll show the viewers a little bit better what kind of dispersal pattern you're talking about. So if you have slide two, here we go. Sure, and I'll give a bit of background on what the slide is actually showing because it's one of the products that we create. So what we do at log measurement and prediction program or VMAF for short is we create these operational forecasts of air quality so people can go and look and see if their community might be impacted if they need to take steps to protect themselves. So we have this online dashboard. It's user friendly, it's mobile friendly. And you can go every day and we update the forecast twice a day and you can see those, this is a sample image from October actually because we had more vlog, but you would see the current most recent forecast extending 60 hours into the future. So what we're looking at is if you actually look at just the gray shading on the map, so the fairly faint gray shading, that's the aerosol plume. Sort of visualize as if you'd be looking down at it from an airplane. So if you peeked out an airplane and look straight down, this is what you would see. The reason it's useful for us is because this way we can know where the plume is and we can compare it with satellite data, but it's not necessarily useful from perspective of forecasting air quality because people don't live where the plume travels, they live at the surface. And so we really want to know what's happening at the surface and how high is the risk of the surface? So the colors that you see, which is kind of the orange, yellow, red, it shows the likelihood that a certain location on the map will experience poor air quality. So where you see more red, like if the location you're at you're showing a lot of red in the next few hours, this is animated on the website. You see that into the future, you're gonna have a lot of red. It would be a good idea to try and stay indoors, avoid strenuous activity and try to protect yourself from unnecessary exposure. So we create these forecasts every day and this is a very typical pattern that we see. It does change based on ambient weather, of course, and the emission rates and all sorts of other conditions. But this unique pattern of things blowing from left to the right is very common. And you can see that as things get sort of emitted from that little volcano symbol on the big island, they get pushed a little bit around the tall peaks. So there's a bit of what we call island blocking. They go around the south point of the big island and then they get recirculated back towards the leeward coast. We call them the Kona Eddy. And unfortunately that brings aged plume and all of those pollutants right back to the people of Kona. And so it's an unfortunate and a very persistent pattern in terms of air quality. And generally, once that unfortunate stuff happens, things often go off to the ocean and then we're not as concerned about health impacts down there. Of course, this is what happens most of the time, but not all the time. Sometimes we get Kona winds and things, it turns towards the other islands and then everybody gets impacted. But this is a fairly usual picture that you would see. So these maps are generated every 12 hours. What kind of data do you use to run the model? Is it atmospheric? Is it emission from killer wire? What kinds of measurements need to be collected in order to run your model? Yeah, so this is a good way to refer to it as not just a model, it's a modeling framework because it requires a combination of models and a combination of data sets. So kind of the most bulk chunks that we have in this framework is first of all, we use a numerical weather prediction. So that's your meteorological weather forecasts that everybody is used to seeing. So we run our own weather model like the H-Manoa and we take those forecasts and then we use them in combination with information about the source, the emission rates, the vent behavior, the heat fluxes. We do plumerized modeling and then we put it all together and then we run a dispersion model using all of those kind of combined fields and that's what gives us the forecast to make it even more exciting. This is not a deterministic forecast, it's not just a model we run once. You can think of running the framework many, many times. So we run it 27 times to be precise and then we look at the likelihood that something will occur. So by running it many, many times, we get more confidence about our results. So we change things a little bit and see how much impact they have on the forecast and that's why we can create those probabilistic forecasts you see on a map which tell you how likely it is that you're gonna get log at a certain location. I see, okay. And I think the third slide would show a little bit about one of the probabilistic models. If we go to the third slide, there's three different images here. Explain to us what we're looking at on the left hand panel. That looks like you've got the volcano at the bottom and then all these gray paths are traced out. Is that your probabilistic estimate of where they would go? Not quite. This focuses on that one component of the framework, the plumerized modeling. And I'm a bit biased here because that's what I like. I love learning about buoyant plumes and how they interact with turbulence. So this is the piece that I bring to the program where I can sort of use my background to help improve a piece of the framework. So if you recall, I mentioned that we use numerical weather prediction and then we have all this information about the source conditions and then we somehow take all of that information, combine and then put it into this version model and how we combine that information is very important. So I come from a background of wildfire smoke plume models and if you look at that left hand side panel, it's actually a simulation of an idealized wildfire but it looks so awfully similar to the upper right hand side image of killer way eruption that you can see that there's quite a bit of overlap and the reason for that is because they're actually, so the heat fluxes and the buoyancy that they generate is actually quite comparable in magnitude between volcanoes and wildfires. It's the same scope of, same type of an event. It's a natural hazard. It's not a campfire. Campfire will not give you that plume and you can see that it has a lot of interesting dynamical features like it generates vortices that we see in volcanic plumes. It has this initial up drop that then disperses in a thick layer. So there's quite a lot of overlap. Of course there are differences and the difference are that the source is quite, it doesn't look the same at all. So the source is, yep. I would imagine that you would say that the plume rise component is one of the more important aspects of this whole endeavor. Let's just back up a bit. The plumes which you're measuring, they aren't like the recent Tonga eruption where you had the big violent explosion that hurled ash many tens of kilometers into the atmosphere. This is, as you say, it's more similar to heat rising from a forest fire. The energy comes from the thermal uplift as opposed to simply being thrown into the air by the volcano. Right, so if you have jets forming kind of a violent eruption that would add an extra component that generally doesn't occur in wildfires. But if you're looking at activity sort of like the most recent eruption and the previous eruption after that, there is not a lot of initial forced convection that happens above the vent. A lot of it is free convection that's just generated by the heat. That being said, the Tonga eruption, I mean a lot of the plume modeling has very similar underlying equations between volcanoes and wildfires. It used to not be the case and it's not because the volcano modeling was wrong it's because the wildfires weren't catching on. They only really became a big thing recently. And so they were approached more like smokestacks before but when people realize climate change, wildfires are gonna keep happening and big mega fires are happening. They're actually quite similar. They also, they can penetrate, they go into the stratosphere. Practically we can see traces of them all the way up. They can circulate the planet multiple times. So depending on the scale of events you're looking at, there are similarities for the kind of eruptions we typically have. They're a good proxy but the source conditions are different. It's too bad you weren't here in mid 2018 when there was a really big eruption down near Capoeba on Kilauea. I was in the field at the time when we're seeing these great big thermals and there is even lightning storms and things like that simply because the heat from the lava on the surface was creating its own weather patterns but presumably your models could take care of that as well as eruptions at the summit. Exactly, so essentially when we have something with that amount of heat, it's not like the surrounding surface it will generate its own weather. It will generate its own updraft and the strength of that updraft really depends on the strength of the heat. So how much energy is going into the plume? So if that... If we go back to slide three, you can talk a little bit about the heat coming off of the lava lake at Hale Maumau, right? So... Yeah, so the nice thing why I really love modeling bog blooms is unlike fires, they don't run away from you. They don't move with the wind like the source doesn't run away, it doesn't succumb to fire suppression efforts, you don't have to worry about fuel types and soil moisture. So you can really get down to, you can really constrain the problem. And it's very important because if you... So the reason why even study plume rise is because you can think about it as if you make an error of about say 10 meters, 200 meters predicting that vertical rise for downwind dispersion, that means errors of tens of hundreds of kilometers. And that's because often in atmosphere winds just don't blow in the same direction. So near the surface they could blow in one direction, higher up they can blow in another and also different atmospheric layers behave differently. So the mixing is going to be different. So that error has a lot of impact. And so if you just improve this one piece of the modeling framework, you have a lot of potential to improving overall accuracy. And so when you have very well constrained source conditions and you have all these scientists from USGS from HVO setting up camp essentially around the summit and measuring and observing all of this. And there's so much expertise and there's so much detail in the measurements you can really focus on the problem you're trying to solve. So it's much easier to volcanic plumes because you have better input data for it. And USGS is USgeological Survey. And HVO is the Hawaiian Volcano Observatory. So we collaborate with them a lot. We rely a lot on their expertise on emission rates on the behavior. So one of the things they do is they operate that thermal camera which gives us that detailed image of the source condition so we can always see what the volcano is doing and adjust our model to reflect that. Now, I think you have techniques to sort of ground test some of the plume height measurements. You have on the fourth slide, I think you're making some measurements of either variations of optical density or aerosols. If we go to the fourth slide, maybe you can tell us a little bit. We're seeing the instruments presumably on the left-hand side and then an impressive color diagram on the right. Yeah, so that's one of the directions we're taking is with instrumentation as well. So what we're trying to do is understand a little bit better of what is happening in a vertical and atmosphere. So how does this mix layer where the pollutants are trapped? How does it change throughout the day? How does it grow and how is it decreasing? So we recently acquired this really cool new instrument. This is a LiDAR Solometer. And we've set it up in Pahala. So you can see us setting it up with a lot of help from the Department of Health. And it allows us to collect data every minute. So here I'm just showing a sample plot from one day. This is February 2nd, so fairly recent. And you can see the level of detail that provides us to us about the mix layer growth. So we generally now can tell that, it stayed below 1,000 meters on that day. We can see where the clouds are. So the blue scatter points are clouds at various levels. Which altitudes those clouds are. We can see how the thermal activity dies down for the night. So because this is an UTC time, the night is where all the blue is near the surface. So you can see that the winds change and the backscatter intensity, which is what we're actually measuring. It decreases because the turbulence dies down. So the mixing regime changes at night. We can actually see precipitation. So if you look kind of near 3 a.m. UTC, you see a lot of red and that's the light are actually catching raindrops. So it can tell us about precipitation events. And all of that has a lot of implications for actually figuring out both the physics and the chemistry of how Vogue behaves in atmosphere. So all of that has great potential to improve our forecasts. The light, it's kind of like a laser and it's simply looking straight up. So you have like a point measurement. How did you decide to put the light out of Pahala when the plume might be going offshore or somewhere else? Or is it these data are better than no data at all? Or is this the best place to be? So this actually relates to, you know, to the sample forecast light I've shown. We have the trade winds. And so most of the time we have a very consistent weather pattern. So Pahala gets hammered with Vogue a lot because it's directly downwind under trade wind conditions. And because of the blocking that the topography creates, a lot of the Vogue gets pushed to Pahala. So they experience a lot of the Vogue events. And the Department of Health has their air quality monitoring station there naturally. So we actually wanted to co-locate with the air quality monitoring station. So we set up the slaters right where they're measuring surface SO2 concentrations. And we can cross compare what's going on with vertical with what's going on the surface. So this was an easy choice because both we get a lot of Vogue and we get measurements simultaneously. Do you get the LiDAR data in real time as well so you can feed those measurements into your models or is it done retrospectively? So that's the long-term goal. So yeah, the idea is that we're going to be extracting mixed layer duct and we're going to use that to actually nudge the numerical weather prediction forecasts that are driving the dispersion models to have a better representation of the depth of that layer. So one thing is get mixed in it, we're mixing it over the layer. That's the correct depth. How big a deal is that you say the health department is monitoring in Pahala? I know you're an atmospheric scientist but do you have any idea of the health hazards associated with this Vogue? Yeah, so unfortunately it is an air pollutant and unfortunately it does cause distress especially for sensitive individuals. So people with asthma, children, essentially it acidifies, when you inhale it it acidifies your respiratory tract. So it's very irritating. And not only that, so that's what sulfur dioxide does and then when it converts to SO4 or sulfate aerosols those are very, very small aerosols. So they can actually be inhaled very deep and they go all the way into your lungs and once again cause damage there. They're also hygroscopic aerosols so they tend to kind of react with water in the atmosphere. So when we have like a typical trade wind conditions roughly 70% humidity, you get a mixture so that mixes with water and it becomes like 30% sulfuric acid which is battery acid and water. And so you're inhaling battery acid as you can imagine it's definitely not good for you. And so unfortunately we see like hospitalization spike during Vogue events and similar to other pollution episodes they spike full sorts of reasons. So it's not just people with respiratory issues coming in it just makes everything worse. So we definitely see that reflect in hospital admissions. Right, I think in the final slide you've got a few computer renditions of some of your plume predictions. Let's take a look at that, Nadya. Yeah, so that's another direction that V-Map is going in. We're kind of trying to stay on top of things in science and one of the more recent trends that's been happening in numerical modeling with the computers getting more powerful is we're able to simulate things at a finer, finer scale. So what I'm showing here is our preliminary work with large eddy simulation. So those are numerical models that run at super fine resolution so they can actually resolve individual turbulent eddies individual thermals. And it takes a lot of guesswork and modeling out of the equation. So we don't need to model things we can explicitly resolve them using these simulations. There's a cost. It's exceptionally computational expenses that runs on the high performance computing cluster. So you can see this one rendition of a plume evolving from just initial eruption from Kilauea summit. It's about six hours long to reach that state on the bottom panel. This took about a day and a half to complete. So obviously it's not very useful as a forecast tool yet just because by the time you finish the simulation the events you were trying to simulate already over. But it is very useful to try and understand how things work in the atmosphere what kind of mechanisms and feedbacks there are. And we can then improve our operational tools and make our forecasts more accurate and more timely. So that's not a direction we're taking. I'm afraid we're almost out of time but it sounds as if this is a really useful endeavor that the Amnistroic Sciences Department is doing and partnering with various other agencies USGS as well as the Health Department around the big eye. I'm really sorry we've run out of time. I'd love to quiz you about whether you've been to the volcano or whatever. But let me just remind the viewers you have been watching Science at Soast. I'm Pete McGinnis, Mark Hill host and my guest today has been Nadia Masiva from the Amnistroic Sciences Department. So Nadia, thank you again for being on the show fascinating and important work that you're doing as well as your colleagues at the university. So thank you again for coming on the show. Thank you so much. And thank you for watching and we will have another exciting episode of some of the other research being conducted at the University of Hawaii same time next week at one o'clock here on Think Tech Hawaii. So goodbye for now.