Breeding Tree Crops to Fight Climate Change

– Welcome, everyone, to Wednesday Nite @ the Lab.

I’m Tom Zinnen.

I work here at the UW-Madison Biotechnology Center.

I also work for the Division of Extension Wisconsin 4-H. And on behalf of those folks and our other co-organizers, PBS Wisconsin, the Wisconsin Alumni Association, and the UW-Madison Science Alliance, thanks again for coming to Wednesday Nite @ the Lab.

We do this every Wednesday night, 50 times a year.

Tonight it’s my pleasure to introduce to you Scott Brainard.

He’s a tree crop breeder with the Savanna Institute, and he also works here in the new Department of Plant and Agroecosystem Sciences, which is a conglomeration of horticulture and agronomy.

– Scott: Agronomy, yep.

– Great.

Scott was born in Schenectady, New York and went to Niskayuna High School there.

Then he went to Swarthmore College in Pennsylvania to study biology.

He got his master’s in plant science in Wageningen in the Netherlands.

And then he came here… Way to go; she’s a horticulture person.

[Brian chuckling] Then he came here to get his PhD in plant breeding and genetics.

As I mentioned, he continues as a research associate here, and he’s also a tree crop breeder with the Savanna Institute.

Tonight, he’s gonna be speaking with us about “Breeding Tree Crops to Fight Climate Change.”

Would you please join me in welcoming Scott Brainard to Wednesday Nite @ the Lab?

[audience applauding] – All right, thanks.

All right, thanks so much.

Happy to be here.

Glad everything’s working.

And yeah, so tonight, I want to talk about some of the research that we’re doing at Savanna Institute, specifically focused on the breeding of tree crops.

And as the title suggests, the motivation for myself and a lot of us at the organization is the potential for these tree crops to mitigate climate change.

So I wanna start with a discussion of how that whole kind of motivation and possibility exists, and then move on in the second half of the talk to discuss specifically the breeding.

So just a bit about the organization where all of this is happening.

The Savanna Institute’s a nonprofit based here in Madison, but we have work and projects that kind of goes across the Upper Midwest.

Our sort of unofficial region of focus is the prior extent of the oak savanna biome.

And yeah, I will discuss a little bit more about the different forms of work that we’re engaged in.

But to start off right from the outset, the main mission of the organization is to promote agroforestry.

So what is this thing, agroforestry?

It’s a somewhat hard to pin down kind of concept or set of practices, but in short, I would say a good working definition is trees, on farms, on purpose.

And the “on purpose” part is where the sort of technique of agroforestry comes in.

You don’t just let trees grow in your fields.

You integrate them in rather deliberate ways.

So to set the stage a little bit more specifically, I think this is a maybe unrealistic slide in the very stark boundaries that it draws, but I think this is a reasonable kind of conceptual way of thinking about how agriculture is typically practiced in relation to nature sort of in the corners, you know, preserved areas and conservation areas that may be deliberately managed in a specific way but are nonetheless very much separate from agriculture.

You know, row crop agriculture as it is practiced across most of the Upper Midwest.

Agroforestry attempts to blur those stark boundaries and try to incorporate conservation practices as well as just natural ecosystems onto farms.

So this makes it different from just a, you know, traditional orchard setting.

Agroforestry kind of has baked into it a little, a sort of notion of specifically cultivating a diversity of species with trees and shrubs being very critical components.

So here’s some cartoons, and I’ll pair them with photos of some examples of what might constitute agroforestry systems.

I think a very classic, simple to understand possibility in planting trees on farms is alley cropping.

So you have row crop agriculture happening between widely-spaced rows of trees.

And those trees may be providing fodder, they may be providing fruit or nuts, they may be providing timber, or they may just be providing habitat.

You’re still kind of practicing traditional agriculture, traditional annual agriculture in between these rows in a more or less undisturbed way.

And here’s an example of this just getting started down in Urbana, Illinois.

One of the nice things about this particular system is these trees take a long time to establish and grow.

And while that’s happening, you still are cash flowing your operation by growing a row crop.

Silvopasture is another kind of classic example of agroforestry.

This is the integration of trees onto pastured systems where you may be grazing livestock.

And I think this particular practice has a ton of potential because the planting of trees on these pastures, the right kind of trees can allow a very high-quality pasture to grow, but also provide additional fodder for livestock, whether that’s leaves, whether that’s fruit or nuts, or just whether that’s shade, which is increasingly becoming a very sought-after commodity by livestock suffering in the heat.

So this is just a real win-win when it’s practiced well.

And here’s an example of cattle grazing in a silvopasture system here in Wisconsin.

A few other practices that are really kind of like additions that don’t have as much of a agricultural sort of bent to them are conservation practices like windbreaks.

Windbreaks can be really essential in breaking wind and thereby reducing, you know, windborne soil erosion.

But they can also have very valuable roles in providing habitat for different types of animals on farms and can definitely improve crop health as well, depending on the field where they’re planted.

So here’s some hybrid poplar planted in Illinois.

Riparian buffers are in a similar vein to windbreaks.

These are strips of trees and shrubs planted along waterways, specifically planted to prevent the flow of excess nitrate and phosphorous fertilizers into those streams and rivers as well as to slow down soil erosion.

So really a kind of strict ecological motivator there for that kind of practice.

And here is a photo of such a practice.

And yeah, so hopefully, that gives you just a little bit of a flavor of what agroforestry might look like.

It’s not one set of practices.

It’s not, you know, a entirely well-defined principle, but it can take a lot of different forms, and the ultimate result is trees on farms on purpose.

So now to return to Savanna Institute a little bit.

Our work, and here, this is a very rough schematic, not of the oak savanna biome, but really where we have either projects of our own or partners that we’re working with on projects.

We work in several different areas.

I’ll be talking a lot about the research side of the organization.

But I just wanna mention that promoting these kinds of systems is not simply accomplished by studying them or commercializing them.

There’s a lot of education outreach, sort of extension-type activities that really needs to go hand in hand in terms of giving farmers the knowledge and the tools to implement these practices.

And then since a lot of this has not been practiced on a wide scale before, we also try to, as much as possible, actually develop farm scale demonstration and pilot farms where folks can come, see these practices actually in the real world.

And I will talk about some of the land where we’re doing that in Spring Green.

So now to move from, okay, so we’re getting trees on farms, we’re getting shrubs being planted.

How is that related to climate change?

And this may be a very rudimentary slide to those of you who are following the sort of science behind climate change.

But in general, responses to climate change can be lumped into two camps.

There’s mitigation strategies and there’s adaptation strategies.

Mitigation is trying to draw down greenhouse gas emissions or greenhouse gas equivalents.

You know, reduce the warming and forcing effects of greenhouse gases that are in the atmosphere to try to slow climate change.

Adaptation is an equally critical and important set of practices that accept that the climate is warming, it already has warmed, and we’re dealing with those consequences now.

We need to adapt to them.

I’m not really going to be talking about adaptation tonight, although I think that, you know, the ways in which tree crops fit into that latter set of activities is a very interesting and fruitful conversation.

I’ll be talking about the, to some extent, much more straightforward mitigation set of activities.

And this is, you know, a somewhat just obligatory slide.

I don’t think we really need to cover what’s happening with climate change.

But just to motivate, why should you care about mitigating the effects of climate change?

Here is one way of visualizing the, you know, potential outcomes of unmitigated climate change.

These are projections of the state of Illinois and what region of the country now, its climate might look like at mid-century and at the end of the century in these different emission models.

So, you know, even under a sort of low emissions scenario, we’re still looking at Illinois having the climate of a sort of southern Gulf State by the end of the century.

And this is already predicted to have pretty massive impacts just on extreme precipitation events.

But, of course, that’s not, you know, the beginning or end of the kind of consequences that specifically, you know, agriculture is going to face from this kind of changing climate.

So to return to agroforestry, it’s very difficult to put these kinds of, you know, meta-analysis slides together.

And this is a somewhat busy figure.

What this is showing is that if you do try to stack all the different kinds of sort of mitigation practices that could have the capacity to draw down levels of greenhouse gas in the atmosphere, agroforestry is very repeatably found to have a massive potential to sequester carbon relative to other agricultural practices.

And to dig into that a little bit deeper, the reason for that potential, the reason that in general, changing the way we practice agriculture can have such a positive impact on the rates of greenhouse gas emissions is that today, agriculture is a very significant source of greenhouse gas emissions in the U.S.

So this is total greenhouse gas emissions shown on the left here for the U.S. in 2020.

And this is that 11.2% sort of piece of the pie broken out in the smaller pie chart on the right.

So 11.2% of U.S. emissions are coming from the agriculture sector.

The reasons for that are, it might not be what you think.

There are some direct CO2 emissions.

This would be from just, you know, driving tractors, burning fossil fuels, but a lot of what’s emitted on farms is direct CH4, methane emissions, from livestock and direct N2O or nitrous oxide emissions.

And that’s coming from excessive rates of fertilization.

The way in which N2O is released in the atmosphere is a little bit complicated, but farmers spread nitrogen on their fields as a fertilizer, very essential to pretty much all crop production.

And when that fertilizer is not directly taken up by plants, it sits in the soil.

If it doesn’t run off into a stream, bacteria will break it down through processes referred to as denitrification.

And the byproduct of these denitrification reactions is the release of nitrous gases into the atmosphere.

And nitrous oxide is a far more potent greenhouse gas relative to CO2 and has an outsized impact on the warming of the climate.

So why agroforestry?

How does that change this equation and get that 11.2% piece of the pie to shrink?

Well here, I think a useful model for thinking about agriculture and its sort of net impact on the climate is to think about sources and sinks.

So agriculture produces greenhouse gases; it also has the potential to sequester them.

So here I’m just showing the N2O emissions associated with annual crops.

And this is a byproduct of fertilization.

In addition, below this, you know, this zero line where the y-axis is a little bit complicated, tons of carbon equivalents per acre per year.

So if you got a negative number here, that means that you’re drawing down levels of carbon.

There are two main ways that annual agriculture can have a positive impact in this regard and have a negative carbon equivalent.

And that’s this no-till SOC.

That stands for soil organic carbon.

Carbon is stored in the soil.

There’s inorganic forms of carbon stored in the soil and there are living things in the soil that have, they’re made up of carbon.

When you till the soil less, less of that carbon is broken down through respiration and more of it stays in the soil.

Cover crops can also have a positive impact on the size of this, what is called the soil organic carbon pool by increasing the amount of organic matter stored in the soil.

What this sort of bar chart tries to represent is those are, you know, this is again integrating results from a number of different recent studies.

But in general, the finding is that in annual agriculture, you can tinker around the edges and you can implement what are sometimes called precision agricultural practices.

No-till, you know, direct drilling of corn, for example, planting more cover crops.

You can improve annual agriculture, but you cannot make it a net sink for carbon.

It is always going to emit more than you are able to capture.

And that is the kind of fundamental shift that we’re able to, according to these, you know, the current research is that agroforestry has the potential to flip that relationship.

You not only can reduce the amount of N2O emissions; you are able to make more efficient use of nitrogen fertilizer in agroforestry systems.

That’s largely a function of the greater diversity of root systems.

And you are also able to massively increase the soil organic carbon pool.

Trees stay on the land year after year after year.

You sort of, by definition, never are tilling in those strips where trees are planted.

And that permanent root system leads to a relatively large expansion in the soil organic carbon pool in the soil.

And then this might be, you know, the most sort of obvious thing that trees have going for them in this respect.

There’s the wood.

Trees and shrubs have woody biomass.

That’s kind of what makes them unique.

And that wood is made significantly of carbon.

Plants take up carbon dioxide during photosynthesis, and they store it in this woody biomass, which as long as that tree is alive and not burned, is going to stay in that woody biomass.

And that is, relatively speaking, a huge sink for carbon, and it is the main reason that you’re able to flip this relationship in perennial agroforestry systems.

So that in a nutshell is sort of the very long-winded motivation for the actual work that I’m gonna be talking about tonight.

Agroforestry systems have this potential to draw down levels of carbon in the atmosphere and turn agriculture from a net source of a pretty significant piece of our total CO2 equivalent emissions into a sink for carbon.

What we are working on in the tree crop improvement program at Savanna Institute is trying to overcome some of the barriers that are limiting the adoption of agroforestry systems.

So right now, most of the systems that I showed you earlier, alley cropping, silvopasture systems, these are relying on crops and to the extent that varieties for these crops even exist, they are relatively unimproved.

And when I say relatively, I mean relative to major agronomic crops.

Corn and soy, for example, you know, as the dominant sort of species on the landscape receive massive amounts of investment.

And massive amounts of energy is put into continuously breeding and improving these crops.

Improving their yield, improving their water use efficiency, their disease resistance.

It’s pretty incredible when you look out on the landscape and see how productive these plants are.

And that’s not some sort of unique thing about their biology.

That’s a consequence of a huge amount of human effort and money that’s been put into that.

And trees and shrubs have not received that investment.

And so that is the need and that phenomenon, that lack of investment has really slowed the rates of adoption because it has not allowed for farmers to have access to really great sort of versions of these crops that I will, I promise you, talk about in a moment.

What’s the reason for this?

What’s the barrier to breeding these crops?

Well, relative to these annual crops, there are very large acreage requirements associated with breeding them.

They’re huge.

These trees are quite large, and they have extremely long generation times.

And not just long generation times to making crosses with these crops, but they have even longer times to maturity, to seeing the performance of these crops, you know, seeing their yield.

They may start flowering or fruiting in five years, let’s say, but they may not reach a sort of physiological maturity for a decade.

So the solution or what we’re trying to bring to the table that hasn’t really been tried before are a set of high throughput.

And by that, I mean very efficient phenotyping.

And by that, I mean measuring trait performance and genomic selection methods.

I’ll explain what I mean by that as well in a moment.

That’s really just using new ways of sequencing the tree’s genome to more efficiently screen very large breeding populations.

And we’re doing this in Spring Green, just west of Madison here.

So before going any further, I just want to acknowledge some of the members of the team of people who are doing this with me.

Like I said, the organization does a lot more than just breed trees, and there’s a lot of additional folks who aren’t pictured here who don’t fit on this slide who are doing very important work relative to extension, and outreach, and education.

But here are some of the technicians working for us.

We have a great farm crew out there.

We’re managing hundreds of acres of land, and I certainly could not do that [chuckles] without a excellent crew of farmers.

We have a research staff.

Here pictured are three folks who are not doing breeding work but are really studying other pieces of the puzzle, specifically studying those, measuring carbon basically stored in these systems.

That’s a whole science in and of itself that is quite complicated, and the gloss that I gave earlier really was probably insufficient if you ask our climate scientists.

And we have breeders like myself who are doing the selection and the crossing that I will describe in the next few slides.

I also wanna introduce the crops we’re working with.

I’ve sort of been talking about these practices in a very abstract way, but these are the actual six species that we’re working on.

And these might not be species that are familiar to you or may not be the first species that you think of when you think of tree crops.

You know, you might think of apples, or pears, or you know, things that you actually see in the grocery store.

These crops have, well, I’ll talk about their sort of uniquely well-positioned role to play in agroforestry systems.

But the other thing is many of these crops are native to the U.S., and many of them are native to the Upper Midwest, and they really thrive here.

So they do very, very well with very limited management and inputs.

So these are, yeah, these are the six crops that we’re working on, and I will have a lot more to say about each of them in a moment.

But first, I want to just give a sense of how this actually looks on the ground.

We’re adopting a very similar breeding approach for all of these crops, and so I just wanna show some photos and videos and also discuss kind of what our general scheme looks like.

So when I say we’re doing breeding, what does that mean?

Well, in general, we are growing breeding populations for these crops.

So what is that?

Those are diverse plantings of around here…

This n standing for sample size, around 1,500 to 2,000 genotypes.

That means unique seedlings that are different genetically from every other seedling.

They are descended from controlled crosses that we’ve made between existing varieties.

And we also have about 10% to 20% in all of our fields devoted to any given crop.

About 10% to 20% replication of the parental varieties that we use to make these populations, we have those clones sort of scattered throughout as essentially checks to make sure that we’re making progress.

So that’s generally speaking what we’re planting, and that’s our basis for doing selection is that you need this genetic and phenotypic diversity to select from, and this is how we’re going about building it.

Taking existing varieties, crossing them together, and planting out thousands of their progeny.

So, but here’s some drone footage of one of our fields from this year.

And you can see it was a drought year.

It’s a very dry field, but this just gives you… You can’t see any plants, but I hope this gives you a sense of the scale.

This is one of those breeding populations.

This is just across from the Spring Green Preserve, if any of you are familiar, literally across the street.

And this whole field is planted to hazelnuts, primarily American hazelnuts and some hybrids with European hazelnut.

And here is, if you can make out those little dots in the lower right-hand corner, that’s the farm team mulching away to control weeds.

So that’s one of our main research fields.

I also have some drone footage of another one of our fields from the gloriously non-droughty year from 2022 of another field, just another farm just up the road.

Here is one of those pilot farms that I mentioned.

These are actually all clonal hazelnuts.

So these are not seedlings, and they’re being grown more to represent what a sort of realistic production system in the Driftless Region might look like.

And I’m just gonna let this play a little bit longer so you can see the Wisconsin River there flowing in the background.

The camera will turn, and that’s a really beautiful wetland that we’re really lucky to be located next to.

Now just a bit about the genetic tools that we’re developing.

I sort of mentioned that we’re trying to bring something new to the table here.

And some of that is on the phenotyping side, how we’re measuring these plants’ performance, but a lot of it is on the genetics side.

And there have been a lot of advances in the last decade and a half, I would say, in terms of our ability as a small organization with not a lot of resources to actually kind of engage in the cutting edge of genetic research.

And this really has taken three forms.

One is we’re able to build genome assemblies for these crops.

The human genome, you know, back in the year 2000, took a million dollars, took an immense number of people working for years.

Now, we can build a chromosome scale reference genome for any one of these species for about $10,000.

So that’s a huge advancement.

And that is extremely valuable in allowing us to quickly identify molecular markers.

And molecular markers are regions of the genome where there is variation in a population.

And this is really the basis for using genetic sequence data in a breeding program is finding these, what are called polymorphisms or markers for short, that are indicators of genetic variants.

We use what’s called whole genome sequencing.

One of the services provided right here by the Biotech Center.

And it’s a very low-cost way of identifying these polymorphisms in large populations.

The final thing that we do is we use these sets of genetic markers to build what are called genomic prediction models.

And this allows for us to predict the performance of plants when they are still just little seedling babies.

So in words, genomic prediction is what I just said.

It’s a way in which you take genetic markers, and you use it to make a prediction about a plant that you don’t know, it’s just a seedling, you haven’t seen it grow up yet, but you can use its genotype to make a prediction of what it will look like when it’s an adult.

How do you do this?

So you genotype and phenotype, so measure the performance of a very large training population.

And those basically are the breeding populations I was showing you videos of earlier.

You grow out a whole bunch of plants, and you genotype them and you measure their performance, and you build a linear model that relates the genotype to the phenotype.

Then you can do two things with this model.

You can predict the phenotype of plants that you only have a genotype for ’cause you have this regression model.

For example, seedlings.

Maybe you have 10,000 seedlings in the greenhouse and you want to plant specifically the ones that are going to be the best.

You only have a small field, let’s say, or even let’s say you have a large field.

You still have more seedlings than you can plant; you wanna pick the best ones.

That’s one way of using this model.

The other thing you can do is you can improve your ability to predict who’s going to be a good parent for the next generation of crosses.

And this is really kind of the bread and butter of breeding is making selection from breeding populations and identifying specifically, not just the plants that you think would do well on farms that you wanna sell to farmers, but the plants that you’re gonna cross together to make the next generation.

And those are not always the best plants in and of themselves.

The capacity of a plant to give rise to good offspring is what’s typically called the breeding value of that plant, nomenclature taken from animal breeding.

And these genomic prediction models allow you to very precisely identify which plants have high breeding values.

So I’ll briefly show you what this linear model actually looks like, but I really kept this slide in here just because I already had it.

And don’t necessarily wanna discuss this other than to say that this is a mixed linear model for any of those of you who may be familiar with mixed linear models.

This is specifically, just for the record, what we’re doing.

And the only thing I’ll say is that letter y here represents the phenotypes that we’re measuring, and the letter g represents the genotypic values that we’re predicting.

And everything else is, well, I’ll also note that the big W is a big matrix of genetic markers.

So yeah, dwelling on this math for longer than I really wanted to.

But just to make the point that we are using markers to fit a model that relates phenotypes to genotypes.

And that’s really the fact that doing so has become cost-effective is what allows us to do this.

It all sounds great on paper.

And, in fact, it has been known that this is a viable way to sort of do breeding more efficiently for decades.

But it’s only really in the last 10 years that the costs have come down to allow real breeding programs to utilize these methods.

And yeah, final note about genomic prediction is that a very key thing that you wanna do is to validate that they work.

If you’re actually using this in a breeding program, you’re not gonna be able to check your work ’cause you’re just making the predictions and going with them.

This is typically done with a cross-validation approach, where you look at the correlation.

Here, I’ve snuck in a little abbreviation for genomic estimated breeding value.

So you look at the correlation between the genomic estimate of a breeding value and the true values, and you do this many, many times using a cross-validation data set.

And you look at that average correlation coefficient.

And the risk of just showing you a bunch of numbers that you can’t even read.

I’ll just show you a few of these correlation coefficients for actual data that we’ve collected from our actual hazelnut breeding population, which if you haven’t guessed, is the one that is far enough along that we were actually able to do this.

And I’ll say that these numbers are encouraging.

So typically, in corn and soybean breeding programs where we take our inspiration because they are the ones who have been able to do this and afford to do this for longer, an R-squared value, a correlation between prediction and reality of anything greater than 0.3 is something that people utilize.

They actually say, “Okay, that’s enough of a correlation there to use that in a breeding program.”

And these rows are all different hazelnut quality characteristics.

Maybe you can guess what some of these mean.

Average weight of the kernel, the percent of the nut that is kernel.

These are all nut and kernel phenotypes.

And for many of these, we have extremely high prediction accuracies.

So we have a lot of hope that this method is going to work, but of course, we have to check our work as we go.

Okay, so just to recap.

The motivation for this method, I hope, is at this point pretty clear.

We’re able to make more efficient gains in our breeding program by utilizing this very cheap sequence data that’s now available to us.

This allows us to pick better parents for future crosses, and it allows us to more efficiently plant out seedlings on very limited acreage.

So this is all sort of geared towards that major barrier that is traditionally slowed down tree crop breeding, which is the long generation times and the large acreage requirements.

By making more efficient selection, we’re able to, in some ways, get around those biological problems with breeding trees and shrubs.

So now I want to shift gears a little bit and talk about these plants.

And this is, in my opinion, the most exciting part of this whole work.

And in the lower right-hand corner, I’ll just be showing the logos of some of our collaborators because with all of these crops, we do not do this by ourselves.

We work with land grants, we work with other nonprofits, we work with research organizations.

And yeah, we could not do this by ourselves.

So I’ll probably forget to mention some people, but I’ve tried to put their logos up on the screen.

So hazelnut is a very interesting crop.

American hazelnut is native to all of the eastern U.S. and much of Canada.

And we are growing hybrids.

American hazelnut crossed with Corylus avellana, which is European hazelnut.

European hazelnut is what you buy, basically.

It is the hazelnut in the stores.

It’s primarily grown in Turkey.

It’s grown a little bit in the U.S. in the Willamette Valley of Oregon.

And basically, it’s sold to specialty markets.

It’s sold to the confectionary industry right now.

But it has a lot of possible uses when you start thinking about growing it on large acreage.

And a few of these are listed here, such as it’s extremely high in oil content, and it’s also got, if you do press it for oil, it has a side product, the meal or press cake, which has excellent nutritional value as a livestock feed.

So here on the right, I’m just showing a picture of what this cross has generally generated for us.

American hazelnut is a very spreading, multistem shrub.

And European hazelnut is typically grown as a tree.

It will have multiple stems if you don’t control them, but far less so than American hazelnut.

European hazelnut has a much larger kernel and American hazelnut has a smaller kernel with a thicker shell.

So we’re trying to combine these traits into one plant that is well-adapted to the Upper Midwest, resistant to some of the diseases that are endemic here that would kill any European hazelnut from Turkey that you tried to grow, but has really great kernel quality.

And here is some of the success that we’ve had thus far.

So a few selections here on the, you probably, well, these don’t have very informative names, so it doesn’t matter if you can’t read them, but here are several of, in the columns, several of the promising selections that have resulted from really decades of this work that’s been going on in the Midwest since the early ’90s.

And these are the parents that we have crossed together to form our breeding populations for hazelnut.

The breeding targets, I sort of already mentioned this.

The really critical pathogen that is endemic to the eastern U.S. is called Eastern filbert blight.

Yeah, it is deadly on most cultivated Corylus avellana, and most americana, most native American hazelnut is fully resistant, or at least fully tolerant of this pathogen.

We’re breeding for high kernel yield density.

And I use that word density because we’re growing this crop as a hedgerow.

So it’s not a traditional closed canopy orchard.

We’re growing it as a hedgerow because that’s how American hazelnut wants to grow and that’s how farmers want to integrate it in specifically, agroforestry systems.

It’s got extremely high oil content, but we want to specifically target oleic acids, which have a much higher nutritional profile, and we wanna make it machine harvestable.

And you’ll see that that’s a theme for a lot of these crops that have not been bred so extensively.

It’s actually quite challenging to mechanize their cultivation.

Whether that’s, you know, the pruning of these crops or in particular, often the harvesting of these crops, that is a really big limiting factor that is holding back their adoption.

So we want a short, compact bush form.

We want the nuts to ripen uniformly so we only have to make a few passes with a harvester, and we wanna be able to have a minimum shearing force, which means it’s really easy to shake these clusters off of the bush and get them into the harvester.

The next crop I’ll talk about is American persimmon, second crop that we’re working on.

So American persimmon is in the Diospyros family, which is an incredibly diverse genus.

It’s an incredibly diverse genus with thousands of species on single islands of the South Pacific.

It’s really quite a wonder.

And this is part of the ebony family.

And so typically, species of this genus are grown for their wood, but it also has at least 500 fruit-producing species.

And the most common one that you’ve probably seen in the supermarket is Diospyros kaki, which is one of three Asian persimmon species that are widely cultivated for their fruits.

American persimmon is the most nutrient-dense native fruit in the U.S. Lots of superlatives to make this stand out, but really there’s two native fruits in the U.S.: pawpaw and American persimmon.

And so it’s not much of a competition, but it is extremely nutrient-dense.

It is almost a complete protein.

It’s extraordinarily high in sugar.

And because of that, it’s been used by Native peoples as well as early Western settlers as a not only human feed, but as a livestock feed.

It was called the hog tree in the southern U.S.

It has extraordinarily high yields that often stay on the tree very late into the season.

And when they do drop, provide that sort of off-season fodder that was really critical really up until the development of corn-based feeds.

So another great thing about American persimmon is that it has a very diffuse canopy structure, which allows for excellent light infiltration.

This makes it a perfect species, in our opinion, for the integration onto pastured livestock grazing operations.

So into silvopasture systems.

You can grow a very high-quality pasture beneath it, and in return, you not only get shade, but you get this incredibly nutrient-dense fruit dropping into your pastures, which can offset your feed costs.

So our breeding targets, this is not a crop that you see a lot in Wisconsin.

If you’ve been in the Arboretum, you might have seen a couple.

It does grow here, you can transplant it here, but its northern range kinda peaks out around Central Northern Illinois.

So we are trying to improve that cold hardiness so that it can be grown here.

We’ve got about 50,000 seedlings in nursery beds right now, and we’ll be screening them over the next couple of winters for the most cold hardy.

Once we find cold-hardy varieties that really thrive in the frigid Spring Green area, we will plant those out into more of an orchard setting and look for several fruit quality characteristics, most importantly, non-astringency.

If you’ve ever had American persimmon, you will know that you have to let it ripen off the tree.

It needs to get soft, otherwise it will make your mouth pucker like you wouldn’t believe.

But that’s genetically controlled.

It’s actually been identified, the single gene trait in Diospyros kaki.

And so we’d like to breed for non-astringency to make it more palatable.

We also sort of in this vein of cold hardiness, we really need late bud-breaking trees that also fruit early.

We have a short growing season here relative to where American persimmon is really common.

And so we want to shorten up its ability to set fruit in this climate.

Precocity, which is a word that means early fruiting.

That is something that we’d like in all of our tree crops, but especially in persimmon.

And improved yield partitioning.

And by that, I mean it’s putting more of its fruit earlier, more of its energy earlier on into fruit production.

These trees can sometimes get just absolutely massive, and that is unnecessary in a silvopasture system.

I wanted to mention here, use these crops as sort of segues to discuss some of the innovations that we’re trying to bring to these understudied crops.

I’d mentioned earlier that genomes are now things that you can basically buy off the shelf.

Well, it’s not that simple, but genome assemblies are a lot easier to build than they were 23 years ago when we assembled the human genome.

And we just built the first genome assembly for Diospyros virginiana or American persimmon with this variety called Early Golden.

It is a hexaploid, so it has six copies of every chromosome.

Humans have two.

That makes it a lot harder to assemble a genome for.

But we did this with the nonprofit HudsonAlpha.

And this is something that we’re very excited about because it will allow us to take that first step towards the incorporation of genetic markers into our persimmon breeding program.

This is a slide that probably does not make a whole lot of sense, but I wanted to include it because I really like the aesthetics of this black background and all the colors.

On the top here is the, what we call hexaploid, Early Golden American persimmon, six copies of every chromosome.

But this is just the maternal copies of those chromosomes.

So there’s only three of every color here.

Three red, three orange, three yellow, all colors of the rainbow, three of them.

And they map directly, three to one, to the diploid Diospyros lotus, which is one of the Asian persimmon varieties.

And I think this is really neat because it shows the very close evolutionary relationship between the hexaploid six-copy American persimmon and its diploid relative over in Asia.

And this is one of the kind of just interesting figures that we’re able to make.

Aside from breeding better persimmons, we’re also able to learn something about the evolution of this species.

So what are we gonna do with this genome assembly?

Aside from identifying polymorphisms or genetic markers that we can use in a genomic prediction model, we can trace its phylogenetic history, and we can also develop markers to allow us to identify the sex of a persimmon plant when it is just basically a week old.

Persimmon are dioecious, which means there are male and females.

This is relatively common in the plant, I mean, it’s not at all the norm, but there’s many examples of this in plants, and it makes breeding somewhat challenging because we obviously are talking here about using female American persimmon trees in silvopasture systems since they’re the ones that produce fruit, but we also need males there to provide pollen and definitely need males in our breeding program.

And we want to design our breeding program intentionally, and we wanna design our farms intentionally.

Markers that allow us to determine which sex this tree is when it is, you know, many, many years before it flowers would definitely improve our ability to incorporate this onto farms.

And so that’s something we’re excited about.

The next crop I’ll talk about is black locust, Robinia pseudoacacia.

This is a very interesting crop that is native to the Ozarks.

It’s nitrogen fixing, has an incredibly high value timber.

It’s very, very rot-resistant.

Again, it’s one of these crops that has very, you’ve probably seen it.

It’s got a very diffuse leaf structure so it allows for very good light transmission.

Very fast growing; it’s basically a biofuel crop.

Our breeding targets for this, well, I’ll start with the second bullet point here.

We are in the process of inducing triploidy.

Black locust is a diploid by nature.

It has, just like us, two copies of every chromosome.

It is also considered highly invasive, especially in Wisconsin, in areas north of where it is native to.

It can really outcompete lots of other species in forests of Illinois and Wisconsin.

And so what we’re doing is we are doubling its genome to create tetraploids, and crossing those tetraploids back to the diploids to create triploids, three copies of every chromosome.

Triploids, an odd ploidy number comes sterility.

They do not produce viable seed.

And this is a pretty common practice in, or at least it’s not common, but it’s a practice that’s been done in a lot of ornamental crops.

And we think it’ll have a lot of benefit in terms of increasing the capacity of farmers to grow this in areas where they might be concerned about growing a very aggressive tree close to say, you know, a forest or other natural system.

This triploidy has also been shown to increase the thickness of the bark, which will, we hope, again, this is studies that are going back to the last time black locust was bred, which was almost 100 years ago.

But thicker bark confers resistance to the black locust borer, which is a pretty important pest in this country.

And also in general, artificially inducing polyploidy has been shown to increase vigor.

So basically faster time to maturity in terms of harvesting this as a timber crop.

We also want to investigate the potential of this crop as a leaf fodder for livestock.

There’s a lot of evidence from other countries where black locust is grown quite widely that the leaves have a very high fodder value.

Now I’ll talk about the small fruit shrub crops that we’re working on, the native American elderberry and black currant.

We are working with two land grants on both of these crops, and they’re pretty similar in terms of how we’re approaching them.

They’re very different crops, but I’ve grouped them together here because they are, again, bush-type crops that are grown in hedgerow systems.

And our goal is to be able to coppice these crops annually.

So cut them down to the ground.

They resprout and they fruit on first-year wood.

That’s something that they already do.

And so that’s what makes them so sort of attractive in agroforestry systems.

It’s very easy and very fast to get a harvest off of them.

They are very cold hardy.

They’re extremely shade-tolerant.

They basically love the difficult places on your farm where you might not be able to grow a more demanding tree.

Elderberry, American elderberry has very low cyanogenic glycosides, which are a toxin that means that European elderberry, which is, again, what you typically find in the store, has to be heat-treated.

You can cold press American elderberry.

And there really is existing demand for both of these crops.

The problem is, the selections for American elderberry are basically all wild selections.

And the only black currant breeding that’s really happened has happened in British Columbia.

So not really relevant to our climate.

So our breeding target for both of these crops is very similar.

We want a very large fruit set.

These are what, I mean, the clusters, you can see are very distinct.

These strigs here for black currant in the lower right, whereas you have these rather large racimes for American elderberry.

But we want a, whatever the actual fruiting structure is, we want it large, we want it to not shatter, we want it to ripen uniformly, and we want it to, while not shatter, while not fall off the plant, we want it to be able to be machine harvestable.

So separate well from the stem upon shaking.

Another critical characteristic for currant is resistance to this disease, cane dieback, which is something that we’re able to hopefully implement in greenhouse screening programs.

And finally, I wanna talk about Chinese chestnut.

This is a really interesting crop.

As many of you may know, there was and still is a native chestnut species in North America, the American chestnut.

There is, in fact, an entire nonprofit foundation devoted to trying to restore this species, the American Chestnut Foundation, which is doing really great work, especially in the Southeast.

The American chestnut was a very important timber tree species until it was wiped out by an introduced pathogen.

Chinese chestnut is from the part of the world where that pathogen came from.

So it is very resistant to that pathogen and really does not need any work to exist in this country.

It also is a very, as opposed to being a timber species, it’s really a nut crop.

It has been bred for thousands of years in China to be an extremely productive nut tree crop.

And it’s very high in carbohydrates.

It has a similar profile, honestly, to corn.

And currently, and there is tremendous demand for these crops.

You might not believe it, but especially looking at this spiny burr that the nuts come in, but the most common business model for chestnut orchards is U-Pick, and they sell out basically as soon as they open orders.

Another great thing about chestnuts is that seedlings in the Upper Midwest outperform grafted trees.

This makes farm establishment way easier.

You don’t have to go through that laborious process of grafting a variety.

You can plant out a seed lot from an improved maternal tree, and those seedlings will perform well enough for a farmer to actually make a profit.

It’s also relatively late-blooming, which makes it well-adapted to our changing climate, but also the Upper Midwest in general.

Our breeding targets are “get more;” it’s a late bloomer already.

We need faster ripening if we really want it to be well-adapted to the Upper Midwest.

Most of the production of Chinese chestnut is confined to Illinois and Ohio right now and parts of Michigan.

There are plants that survive in Wisconsin and Minnesota, but more breeding is needed to increase that cold hardiness.

We definitely need to select for high yields.

There’s a lot of variability, especially in these seedling populations, which are very genetically diverse.

And we would love for the nuts to fall free from that burr, which is one of the sharpest things I have found in the plant kingdom.

Male sterility is a very interesting trait that has been identified in chestnut and would, again, lead towards a greater fruit set on male sterile trees.

And we also are interested in the kernel ratio.

There’s a lot of variability in what percentage of the nut is actually made up of that kernel.

And I’ll close with Chinese chestnut, I think is a good example of one of the points that I made earlier about the way in which we’re trying to not just use advancements in genotyping technologies, but also phenotyping technologies.

Chinese chestnut is a very interesting tree because it’s what’s called a terminal bearer.

It sets all of its fruit on terminal buds, new growth.

So if you take a picture of the tree using a drone, for example, from above, you actually can see all of the burrs, all of the nuts in that tree.

Chestnuts drop over about a three-week period.

That’s when you harvest chestnuts is over that pretty long period.

So it is almost impossible to pick up all the nuts from a chestnut tree.

It’s an enormous task.

So figuring out what the yield of a tree is, which is obviously a very important thing to know, is, to date, basically not known.

It’s almost impossible.

It’s definitely impossible to scale that kind of measure over large acreage.

So we have started flying drones over chestnut orchards, and are using advancements in the field of automated object detection to pick out all of those burrs and count them and convert that to a measure of burr density in the canopy.

So we can then, from the size of the tree and the shape of the crown, extrapolate how many nuts are up there, and finally figure out how much Chinese chestnut trees actually bear.

So that’s one of the high throughput phenotyping methods that we’re trying to build into our program.

So just some brief conclusions here to try to tie a lot of disparate threads together.

Thinking back to the beginning, agriculture is a major source of greenhouse gas emissions, not just in the U.S.

It’s the data that I showed, but that’s a phenomenon that you see across the world.

And agroforestry has the potential to change that phenomenon.

And instead of being a net emitter, agroforestry systems can be net carbon sinks.

Right now, agroforestry’s potential in the Midwest, especially in the Upper Midwest, is not being realized.

And a lot of that has to do with the lack of very well-adapted and productive tree crop varieties for some of these understudied trees and shrubs that I’ve been talking about.

And we believe and are trying our best to implement modern breeding methods to accelerate that process, make it more efficient, and as a result, contribute to the increased planting of trees and shrubs on the landscape.

So I’ll close there.

[audience applauding]