Listen to the Interview:

---

Transcript:

Ross Chambless
Jon Wang, welcome to the Wilkes Center.

Jon Wang
Thank you. So glad to be here.

Ross Chambless
Yeah. Well, first of all, can you just give us an overview of your research interests generally?

Jon Wang
So, I lead the dynamic Carbon and Ecosystems or dyce lab. This is a research lab that’s really focused on using satellite remote sensing and other sort of advanced computational techniques for mapping out when and where ecosystems have changed in response to the climate change and the human activity. We’re really interested in trying to understand better the carbon cycle and how the global carbon cycle is responding to global change. You know, for example, due to increased wildfires, increased CO2 concentrations, increased sort of human use of the land. So, it’s a lot of maps and a lot of modeling machine learning to really get at that.

Ross Chambless
Interesting.  So, I was kind of checking out your lab’s website and it sort of seems like you’re using a variety of instruments and technologies to try to understand how climate change is changing ecosystems, or specifically terrestrial ecosystems, like forests, wetlands sort of and seeing how they are shifting the carbon cycle, is that kind of it?

Jon Wang
Yeah, that’s exactly right. So, a lot of the research we focus on is looking at these sorts of critical ecosystems or ecosystems that are really climate sensitive but play a pretty large role in the carbon cycle. So, for example, in the far north in Canada and Alaska, where ecosystems are pretty cold-limited, but climate change is also occurring the fastest out of anywhere on the globe. We expect that a lot of shifts could be happening in the far north due to this rapid warming. So, this manifests as increasing wildfires or maybe increased growth in certain forests. Now, it’s really hard to measure that in the field because, you know, some of these ecosystems are really huge and you can’t really predict exactly where forest fires will occur. So, using satellite remote sensing is a really great way of capturing some of these dynamics, sort of post-hoc in a sense. And then trying to use the satellite data to understand how have these ecosystems changed? What is the fate of these ecosystems as climate change continues to push our global earth system into new territory?

Ross Chambless
Well, so you were recently awarded a grant funding from Google to develop essentially a data set for measuring aboveground biomass. So, talk about your proposal and also define above ground biomass?

Jon Wang
Yeah, for sure. So, the carbon cycle has a lot of different components to it, fossil fuel emissions and atmospheric CO2 or some of the things that maybe most people think about. But the land is actually taking up a lot of carbon as forests and other ecosystems grow. And so, this carbon dioxide is being sequestered, so to speak, in and living biomass. So, a large tree has a large biomass and is storing a lot of carbon, but one of the big challenges with this kind of work is that the resolution for maps of aboveground biomass or of the amount of carbon that’s stored in natural ecosystems.  The state of the science there is a little bit rough in a lot of ways. The resolutions are often pretty coarse, or they don’t really capture time series very well.

And so, this grant with Google is really an effort to improve our ability to model and map aboveground carbon and ultimately quantify and better understand how the land is responding to these increased CO2 emissions and maybe buying us some time as we try to deal with climate change. This grant is really aimed at trying to sharpen our ability to create those maps. So, a lot of the big problems with these maps is the course resolution. This grant proposal or this grant, rather, is really focused on developing extremely high-resolution maps of biomass down to the individual tree level. So, there’s a lot of these airborne laser scanning systems. So, folks will fly airplanes and basically shoot lasers at the ground and measure how tall trees are. And using a few assumptions and simple models, we can estimate on an individual tree basis how much carbon is stored. And by developing these maps, we hope to improve the machine learning benchmarking for larger scale satellite based or other kinds of projects looking at carbon cycle.

Ross Chambless
Really interesting. So, just to pivot back a little bit, what would you say is the problem with currently the programs or applications used to measure these aboveground the biomass?  You mentioned the resolution is just not great. We just can’t see the details of what is on the ground?

Jon Wang
So, there’s a couple of issues.  The resolution, the spatial detail is really important because that really defines how well we can attribute changes in biomass to different processes like wildfires or urbanization. If your pixel size or the grain of your map is at, let’s say, ten kilometers, it’s really hard to understand if a change happening there is due to a certain process happening probably at a finer scale. There’s also an issue of accuracy. A lot of these maps are pretty low accuracy, and this is actually pretty hard to even understand because usually the way we estimate aboveground biomass in a forest is we go out and basically hug every tree. We try and measure the diameter of each tree and using allometric equations, convert this tree diameter into the total biomass.

Now, the Forest Service in the US has been measuring trees sort of in this way for a long time. But that data is really hard to access because of privacy laws and privacy concerns. A lot of their surveys are done on private land. And so, as part of their agreements, in order to gain access to this land, they agree not to publicly release the coordinates of their locations. So, it’s been really hard to validate existing biomass maps using what I would say is probably the gold standard for this kind of work.

So, we’re trying to, in a sense, short circuit that process by using publicly available airborne laser scanning data, and we’re going to try to replicate the same sort of techniques done on the ground but from an aircraft, and developed is really high-resolution biomass data set which is in a lot of sense is equivalent to the forest survey plots of the Forest Service uses.

Ross Chambless
Interesting. So right now, how accurate would you say is this airborne laser scanning equipment?

Jon Wang
Right. So, these laser scanners, they range anywhere from five to about 50 points per square meter. So that is to say, these laser scanning devices could resolve individual trees that are as small as maybe half a meter in size, in diameter. So, what we’re really aiming for is to develop these biomass benchmark maps at roughly one meter resolution. And so, this is a really nice resolution because, you know, it’s understandable to my brain. You know, I can visualize one meter pretty easily, but it’s also really nice because then you can take these one-meter maps and rescale them to any other resolution that you want. So if you have another data set out there, you’re trying to develop, say, a 30 meter or 100 meter or one kilometer resolution map using satellite data or some other technique, you can take this high resolution map and rescale it to exactly the same size you don’t have to worry about, oh, you know, this plot is this shape and our pixels are this shape. And do they really mean the same thing? We’re trying to build a really versatile system for benchmarking aboveground biomass data.

Ross Chambless
And when you say benchmarking, what does that mean specifically?

Jon Wang
So, benchmarking, we’re trying to develop this high-resolution sort of ground truth so that when you, say that you’re another researcher and you’re interested in using this new satellite technology for developing a biomass map, but you don’t really exactly know how accurate it is, we intend for these benchmark maps to serve as sort of a ground truth, a sense or a way of validating, you know, are my maps more accurate than these other maps? I can compare it to this sort of trusted standard, which is at this really high resolution.

Ross Chambless
Interesting. And so, this airborne laser scanning, and maybe I should have asked earlier, but is this able to be done by satellites, or do you use aircraft, or drones, or what?

Jon Wang
What’s really interesting about the laser scanning technology is that it can be use at pretty much any scale. So, there are satellites that that have laser scanning and they capture a pretty broad swath. You know, satellites are pretty far away, but it can go all the way down to sort of the individual plant level using Terrestrial lidar Laser Scanning tools (TLS). These are laser scanning systems which are propped up on a tripod, and they just scan the area immediately right around them. This is often used in real estate and architecture for mapping out the insides of buildings. But it has been adapted to mapping individual trees and individual shrubs. Now we’re using airborne laser scanning. So, it’s taking a small plane that the taking a small plane that’s flying maybe a thousand feet over the ground. And that hits a nice little sort of middle ground between getting an individual tree at a high enough detail, but capturing a large enough area that it’s a useful map for some other remote sensing folks.

Ross Chambless
Wow. Are there specific areas that you are going to focus on with this study?

Jon Wang
Yes, there are two sources of airborne laser scanning data that we’re going to use. The first is from a network of field sites called NEON. So, neon is the National Ecological Observatory Network. This is an NSF-funded network of field sites across the United States where all the measurements that they’re taking of the water, of the plants, of the wildlife, but including remote sensing from aircraft. These are all standardized observations, standardized measurements and protocols across maybe 40 or so sites across the United States. This is going to include a whole range of ecosystems. Not just temperate forests, but also shrublands grasslands, other sorts of, boreal forests versus subtropical. So, there’s going to be a whole range of ecosystem types captured by the NEON data that will take advantage of.

We’re also going to be using airborne lidar data coming from the Canadian Forest Service. So, this is leveraging an existing project we have with NASA’s terrestrial ecology program to look at aboveground biomass and its changes across Canada and Alaska. For that project, we collected a lot of airborne lidar data for a similar purpose. But with this Google grant, we’re hoping to really formalize process and develop benchmark biomass maps that we can publish on their own.

Ross Chambless
So, I imagine the Google funding will just help with the costs of this, of acquiring the aircraft and the equipment to make this possible?

Jon Wang
So, a lot of the aircraft data is publicly available, so the cost doesn’t come in there.  But the grant will really help fund postdoctoral researchers. So, this is very this is a very specialized technical kind of work. And so, what we need this funding for is to support early career scientists with a lot of recent training with the state-of-the-art computational techniques. There’s also a lot of computation, and there’s resources here at the University of Utah that would really like to build out. The Center for High Performance Computing is a shared computing cluster, a supercomputer basically, that we’re going to use some of this funding to expand the capability of the CHPC. And that will really enable us to expand the scope of this work.  We are, as I mentioned, leveraging some previous work in Canada, but we hope to go across most of North America using this grant funding.

Ross Chambless
Yeah. That’s really interesting and exciting too. It seems this potentially would allow researchers now, but also future researchers, to have a better understanding of the on-the-ground terrestrial biomass that’s existing. And maybe see how it’s it has changed or maybe continues to change down the road, and basically have a better sense of the carbon stock, if you will.

Jon Wang
Yeah, exactly. There’s two sort of main audiences that we’re imagining might be interested in this. Other remote sensing scientists who are developing these aboveground biomass datasets. You know, in the last four, maybe five-ish years or so, there’s just been an explosion of new datasets being produced from all different kinds of satellite technologies. And if you take these different data sets over the same area and ask what is the carbon stock, they all give you very different answers. You know, they differ by almost a factor of two in many cases. And so, we’re really hoping these benchmark maps can be used to refine these data sets. So that the next generation of biomass data sets could be a lot more trustworthy.

And this trustworthy is really important for sort of the other audience, which is people who are using these aboveground biomass maps to develop carbon projects or other ways. You know, nature-based climate solutions are really, really popular these days. People are trying to use forests as natural carbon sinks, but it’s really hard to commit real money to that sort of problem if you don’t exactly know how much carbon is being assimilated by your forest ecosystems. And so, down the line, the hope is that these benchmark maps, by improving the satellite based maps, could end up developing in a new a better sense of where carbon is actually being restored and, recovering forest ecosystems or where there is potential for new carbon to be sequestered on the land, basically just to improve our ability to use forests as a nature-based climate solution.

Ross Chambless
Yea, very cool. Excited to see how that how that project rolls out.

Jon Wang
Yeah, I’m really excited too. Thank you so much Google!

Ross Chambless
Yeah, exactly. And Jon, tell me a little bit about more about some of the other research projects that your lab is working on right now.

Jon Wang
So, with this Google grant, we now have four funded projects, which is really exciting. You know, one project I alluded to before is this massive terrestrial ecology project looking at aboveground biomass, but also land cover change across Canada and Alaska. And we’re using long time series of remote sensing data from satellites in order to track the carbon that’s being lost, wildfires and maybe what’s recovering following wildfires in Canada and Alaska. Canada has experienced some really major fire years, especially last year was off the charts. And so, it’s more important than ever to try and get a good understanding of, where is where is the carbon going? Is Canada a carbon source or carbon sink? This has really important policy implications for example, IPCC, carbon budget reporting to the larger international community. So, we’re really happy and excited to be pushing forward, pushing forward the state of the art of biomass mapping, but also connecting this to policy and to folks who might be interested in managing Canada and Alaska’s carbon budget.

We also have a project with NASA Land Cover Land Use Change program, which is using similar time series of satellite remote sensing data. But in that case, we’re looking at trying to develop new methods for detecting and mapping forest mortality. The issue here is that as climate change proceeds, you know, atmospheres are getting hotter and drier, and that’s leading to a lot of these sort of almost spontaneous die off events, mortality events in forests across the world. You know, Colorado, California, Utah, Idaho, all these areas in the western U.S. in particular have been experiencing more and more of these massive die off events. But it’s been really challenging to study these because the quality of the data is quite poor. You know, the current state of the art is that the Forest Service flies in an airplane low and slow over these forests and someone leans out the window with a with an iPad these days and they draw on a map. “Yeah, the forests over there look like they’re kind of dying.” And we’d love to take that sort of knowledge and combine it with remote sensing and, we’re using we’re tracking thermal signatures and detecting forest stress as a way of better quantifying just how much forest is dying due to drought and potentially develop early warning systems so that forest managers can understand what are priority areas for restoration or mitigation of these droughts.

And then the last other project, beyond the Google project ,is a NASA early career investigators program project. And so, this is a project aimed at looking at a different sort of aspect of aboveground biomass in the Western US. But we’re looking specifically at this woody plant encroachment. And so, this is a process where, due to changes in CO2 concentrations in the atmosphere or changes in fire suppression or land use, there’s been a proliferation of these Utah junipers mostly, or juniper plants more generally, but woody shrubs across lots of semi-arid grasslands and dry lands across the world.

And so we want to use this airborne laser scanning to develop a biomass product that was sensitive to these sparse short woodlands that are occurring across the Western US, all the way from the southern Great Plains out to the California Central Valley, and try to understand just how much new tree growth is happening and how much carbon is being stored in this new tree growth. This is an often-overlooked potential sink of carbon. But it could also be a source of carbon in that where woody plants grow, there’s also the death of grasslands that they outcompete. And so, it’s an open question what the net carbon balances between a woody plant and crushed grassland in a native grassland. And so, this is a real story about sort of native plants versus invasive species and global change and carbon stocks and just trying to build some nuance into this idea of woody plant encroachment. Is it as a good or is it bad?

Ross Chambless
Yeah, that’s really interesting. I’m sure you’d have a lot of interested folks here in the state of Utah talking about the woody plant encroachment, because I know that has been a big topic of concern. I want to pivot to talking about you had a paper published recently in Environmental Research Letters Title is “Rising Forest Exposure and Fire Severity from Climate Warming Amplify Tree Cover Losses from Wildfire in California.” So, what is this paper about?

Jon Wang
Thanks. So, this paper is building off a previous study that I did as a postdoc when I was working at UC Irvine. When I was in California, we used satellite remote sensing to develop these maps of disturbance and vegetation cover across California over time. And this was used to understand better where wildfires are impacting tree cover. And again, going back to the nature-based climate solutions or, how well can California expect to use these forests as part of their carbon budget and their fight against climate change? But one thing that we really thought about in this paper is, okay, so wildfires are really expanding in their burned area. California had these extreme fire years in 2020 and 2021. More area is being burned than ever before in California. And we compare that to how much tree cover loss was happening in California. And we noticed that the rate of tree cover loss was increasing a lot faster proportionally, than the area rate of burned area.

So basically, per unit burned area in California, the amount of tree cover loss was increasing. And our question was trying to understand in this study in environmental research letters, given that burned areas increasing a certain amount, why is a tree cover loss increasing so much faster? And so, the real crux of the issue here is understanding fire severity or the sort of relative per area impact of fire on tree cover in California.

So, we use the same data sets that we developed from the previous study, but we specifically looked at two factors beyond burned area. We looked at something called forest exposure and fire severity. Now, forest exposure is a term that we’re using to describe what kinds of forests are exposed to fire. Are these you can imagine a fire is expanding in its area, but it’s mostly burning in maybe sparse woodlands or in grasslands. And the effect on tree cover is not going to be very significant. Whereas if the fires are tending to occur more in these denser sort of older growth forests, and then that’s a greater concern. These dense forests that lots of trees are being threatened, more threatened by fire if fire’s moving to some of these areas.

We also looked at fire severity, which is trying to quantify the relative loss of tree cover per area burned, which we know from other studies is really sensitive to climate warming and dryness, atmospheric dryness. But the issue is some other studies looking at fire severity is that usually fire severity is characterized in these categories, high, medium, low. I’m not sure what that means. And it’s hard to translate that, either across fire events or into other subfields of ecology. So, we really tried to get at a continuous metric of fire severity by looking at this rate of tree cover loss per unit area burned.

Ross Chambless
So, looking at tree cover loss and relative to fire fires that have burned, so are we essentially seeing that when after the fire occurs, the tree cover that maybe comes back is maybe not as thick as it was before? Or are they two different things? There are the burned areas and then there’s just separately the tree cover area areas are declining?

Jon Wang
So, we I am really interested in understanding the recovery of forests after fire and sort of what kinds of ecosystems might return. But the focus of the study was really more on the fire, the nature of fire itself. So, we didn’t think too much about the recovery of fire, except that more severe fires, especially severe fires that kill all the trees in an area, those areas are less likely to recover because you’re killing seed bank, you’re killing the ability for trees to propagate and repopulate an area. And, in a changing climate where it’s getting warmer and drier, it’s something else that is more adapted to this new climate might come in and outcompete these trees.

Ross Chambless
And how long where how long were you working on this particular paper?

Jon Wang
This paper was the seeds of the ideas of this paper came about as I was finishing my time at UC Irvine, but I really dove into the actual analysis and writing, let’s say early, early last year, early 2023. You know, I just moved to the University of Utah. My lab was very small, was just me, so I needed to have something to work on. And I still had this idea bouncing round the back of my head, how much tree cover loss is really attributable to changes in fire severity? And so, I’d say it’s been about pushing two years at this point.

Ross Chambless
Okay, we’ll make sure to put a link to the paper in the transcript for the podcast. But finally, I just wanted to ask a few questions just about, about your background. Where did you grow up?

Jon Wang
I was born in the Bay Area in California, so, near San Francisco. And then after I finished high school, I moved to I moved to Rhode Island for college. I went to Brown for a few years and then worked for a few more years before I decided to go to grad school at Boston University.

Ross Chambless
Okay. And how did you get into forest ecology and plants?

Jon Wang
Yeah, it was a kind of a long, circuitous route. It wasn’t my initial plan. And when I first moved to Brown, my plan was to become a mechanical engineer. My dad was an engineer, and I didn’t really have any bright ideas about what kind of future I wanted at the time. But I ended up in this geology class in my freshman year, and I realized that being outdoors was way more fun than mechanical engineering, basically. And so, I found the degree that had the most number of field trips. I had at least one field trip every semester at Brown as a geology biology major.

In the last semester of Brown, I learned about remote sensing. I didn’t even know what remote sensing was before that just happened into this class. I thought, these are some pretty interesting words. Who knows what it is? And I found it really gripping. It really combined all these different aspects that I was really interested in of the Earth system and climate physics and chemistry and biology. But, you know, right after that I needed to pay off some student loans. So, I graduated with my bachelor’s and then I went and worked for as an environmental consultant for a few years before realizing, okay, I keep thinking about plants, I keep thinking about climate change. I think it’s time to go to grad school.

Ross Chambless
Great. And then, you made this shift the transition to Utah in in 2023, took a position as assistant professor. So how has that been for you, that transition to the University of Utah?

Jon Wang
It’s been really great. Folks here have been really nice and supportive. And, you saw on my lab website, the lab has grown a lot. There’s been a lot of people have been really interested in my work and I’ve been really happy to have this sort of environment where I could support all these undergraduates, graduate students, postdocs, as well as interacting with other communities in the area like the Wilkes Climate Center, for example, the biology department, atmospheric sciences and geology and building like a really awesome community right here at the University of Utah for finding ways that my work can plug in to sort of the bigger picture and everything.

So, Salt Lake City has been really great. We love hiking. You know, my wife and I moved here together and she’s really into mountain biking. So, moving from Los Angeles, once we learned what mountain biking was like out here, we thought, alright, we can try to figure it out, see if Salt Lake will work. And it’s honestly, it’s been a hidden gem. We’ve been really loving living here, and it’s been a lot of fun to explore a new place.

Ross Chambless
Awesome. Well, thank you, John Wang. It’s been so fun to talk with you about your research and about your current project that you’re about to proceed with from Google, and look forward to catching up with you later down the road when you have more to share.

Jon Wang
I’m really excited for that Ross. Thank you so much.

Ross Chambless
Thanks.