Beam me up Scotty!
What is this?
It's the 21st century and we can send spacecrafts to Pluto. However, we can't tell if an oak is dying from diseases or droughts until it's too late! Can we find out a way to predict whether oaks are going to die and what killed them? We think we can do this by just studying how light bounces back from trees. Just like the doctors do in Star Trek!
Why does this matter?
- Because red oaks are dying from an invasive fungal disease called Oak Wilt and droughts across the United States.
- Because differentiating diseased oaks from drought-stressed oaks using light is the first step to satellite detection.
- Because when the Aliens find out that we got space travel before understanding how plants work it will get really awkward...
How did you study this?
We planted 100 red oak saplings in pots and infected 50 of them with the oak wilt fungus. Then we wrapped the pot of half of the infected trees and half of the non-infected trees with a very high-tech waterproof insulator (it was just Seram wrap...) to prevent watering and cause drought stress. With this, we ended with four groups of trees: droughted, infected, droughted & infected, and non-droughted & non-infected (controls). We then measured the physiology of the saplings every week. This included things like the amount of water in the leaves, their capacity to make sugars, the percentage of damaged cells, and their ability to recover water after being stressed among others. These were very intense and time-consuming measurements! So, we only did this in 5 randomly picked trees of each group every week. But measuring light bouncing from canopies -called spectral reflectance- is much faster, so we did that in all the trees every week.
Having measurements of bouncing light paired with physiology measurements allowed us to build models that predict physiology from light. With these models, we could estimate the physiology of all trees for which we had light measurements but no physiology measurements.
We also walked around the trees every day to take note of when they first developed visual signs of disease or drought stress. With that information and with the light-based physiology, we could assess whether we could detect signs of drought or disease before we could detect them with our eyes by estimating physiology from light collected with special cameras.
Lastly and to make it even cooler, we flew a drone over the experiment to take pictures of bouncing light from greater heights! We designed this drone ourselves so that it could measure certain wavelengths of light that we expected to be sensitive to drought and oak wilt. These pictures would later be used to study whether we could differentiate drought-stress from oak wilt infections.
Because this was all done in potted saplings instead of naturally growing trees, we decided that we also needed a test in the field with naturally growing saplings. For this, we went to the Cedar Creek Ecosystem Science Reserve -where Oak Wilt is already present and killing red oaks- and we infected a few trees there. We repeated all our measurements in these trees with the idea that we could put our light-based physiology models to the test to see if they could also detect oak wilt infection under natural conditions.
You might be thinking "Gerard are you crazy!?!? You said this was an invasive fungus!! You will spread it even more!!". Don't you worry my friend, for I do not miss an opportunity to kill some trees. Which means that I spent some quality time uprooting every single one of the infected trees and incinerating them while laughing like an evil maniac. See proof on the right!
I also killed all the trees in the pot experiment "for science" and justified it saying I needed to look at the wood to find patterns of infection or drought (which, to be fair, I did).
So what did you find?
The first step was to find out if we could predict physiology from light bouncing from trees. Without this being the case, detecting stress and differentiating the causal agent using light-based physiology would not have been possible. And what would you know, it actually worked!
On the left, you can see a graph showing how the amount of water in leaves changed after we infected the oak saplings at Cedar Creek. The green lines are non-infected trees and the purple lines are the infected ones. The dashed lines show the water content predicted from light. You can see that the water content predicted from light nicely tracks the solid lines, which represent the actual water content of these plants. It works so well that we can even detect when plants regain water after a rainy morning!
The vertical line in the graph that has a yellowing leaf next to it also represents the moment we saw that the trees started to look sick. From looking at that line and the green and purple shades, we can see that the green and the purple separate before the line. This means that light bouncing from trees can be used to detect that trees are drier than they should before we observe it with our eyes. In other words, light reflected from trees can be used to detect stressful diseases or droughts before visual signs appear.
"Ok this is great but how do we know if the tree is drying because of the disease or just drought?" you ask. Difficult to impress, aren't we? Well, remember I told you that I needed to look at the wood to find patterns of infection or drought? Let me show you something. If you move water with dyes through the wood of healthy (control), infected (OW), drought, and drought & infected trees (DxOW), you will see the dye stain the wood like this:
Do you see how the tree infected with oak wilt has a missing chunk of dye? That's the area where the fungus is! It's blocking the pipes of the tree and preventing the dye from flowing through it. This is why trees with oak wilt dry out. Now, notice how different that looks from a tree with drought. That one has missing dye everywhere. The missing dye is not concentrated in one area because it is not caused by a fungus that slowly spreads from one place. Instead, drought fills the pipes with bubbles and the bubbles occur everywhere. Because they are bubbles, you can flush them out if you apply the dye with enough pressure, and when you do, we see something really cool:
The bubbles are gone, and the pipes move dye again! And look at the tree with oak wilt, it has not recovered the pipes that were clogged by the fungus. What's most interesting is that the infected tree that also experienced drought looked like a "drought" tree before we flushed. But it showed a pattern of oak wilt after flushing it.
"Fantastic work Gerard. So if I want to know if the $30,000 red oak in my garden has a lethal disease, I just need to cut in half...". What? You don't like it this way? Jeez so picky!! Fiiiiineee.... How about this: the pipes move water through the tree to feed water to the canopy. Hence, the pipes in the wood of the trunk are connected to branches, which in turn are connected to leaves. When the pipes are clogged, the leaves connected to them start to dry out. Which means that the clogging patterns in the wood are likely mirrored by the leaves in the canopy. Thus, we should see that diseased trees have patches of drier leaves while other leaves are perfectly happy since their pipes are still working fine. Trees that only have drought stress should show evenly dry leaves across their canopy based on the patterns observed in the wood. If we follow this idea and use our fancy drone to detect early declines in water content in the canopies, we should be able to differentiate trees with oak wilt from those with drought. Shall we look?
Eureka! When we look at the canopy through a spectral index that uses the infrared light that bounces from the canopy, we see that half of the canopy affected by oak wilt is totally black while the other half still gives a signal. On the other hand, the drought-stressed tree has an even color across the canopy. This pattern of reflected light can be used to differentiate oak wilt from drought stress, and drones can take advantage of it.
And the big point is...?
Oak forests play a vital role in providing ecosystem services like habitat, climate regulation, clean air and water, and erosion control across North America. These forests are threatened by hot droughts and diseases like oak wilt. Oak wilt has spread across the continent severely affecting red oaks and white oaks. Our current ways to detect oak wilt across the landscape relies on crews of experts that survey the forests from the ground. You can imagine how much this limits our ability to detect oak wilt across a continent. It takes a long time, drought stress often gets in the way of correctly identifying oak wilt, and detection often occurs when at least a 30% of the leaves are already affected. By that time, the fungus has often already spread to neighboring trees that are already infected but do not show clear visual signs of stress. Hence, there is a pressing need for accurate and early detection methods to safeguard the health of oak-dominated forests.
We hope that what we have unveiled here serves as the first step to generate methods to detect forest diseases and drought. Because our approach links what happens inside the trees as they experience stress to the changes in the light that they reflect, this method should allow us to differentiate diseases like oak wilt from drought and act accordingly to the type of stress observed. Because our approach can detect stress before we see it with our eyes, it should provide with more reaction time to manage and mitigate the impacts of diseases and droughts as they are detected. Lastly, because our method uses light to predict the physiological status of trees and detect stress, the principle of operation only requires two things: sun, and a spectral camera.
Lucky to us, we live in the 21st century and we can send spacecrafts to Pluto. So having a satellite with a nice spectral camera should not be too far away into the future. This is really good news because it means that we have a chance to make a real good impression when the Aliens come asking why our trees look like sh*t.
The actual paper for the nerds:
The following article is under a CC BY-NC-ND license agreement