Very excited to share this paper by Jeannie Wilkening with you all (and if you haven't checked out Jeannie's thread about the paper on Twitter, do!).
We picked up on an old idea - that leaves should act as a "fuse" for plants to blow when their circuits overheat.... more correctly that it's a good idea for leaf xylem to embolize before stems or branches do. The idea is that if you have to lower water potential enough to risk damaging xylem by embolism, it's less costly for plants if you damage xylem somewhere that's easy to replace, like a leaf. Thus, conventionally, we expect that leaves should be the fuses.
So, that's sensible. But it doesn't explain why more and more data suggest that in some plants, the xylem in the stem are more vulnerable to embolism than are the leaves. That is - the stems look like they are the fuses! What's going on?
The image above shows Cottonwood (Populus trichocarpa) vulnerability curves for stems on the right and leaves on the left - the stems are embolising at water potentials of -1 to -2, while the leaves are embolising at potentials of -2 to -3 MPa - fuses in the stems.
We wanted to understand what the consequences are when you have leaves as fuses or stems as fuses. What might it mean for the overall conductance of the leaf-stem system under different levels of water stress? Does the difference in how vulnerable the leaves and stems are matter? Why would plants do this?
This is hard to interrogate empirically because there are still too few plants known where the relative vulnerability of stems and leaves have been measured. So we played with a model that let us switch around vulnerability and conductivity in the leaves and the stems, while exploring the outcomes for plants. Jeannie did an amazing job defining sensible metrics of what plants were doing and how model outputs measured the effect of the fuses.
She summarizes it all in the diagram below - a very simple hydraulic model on the left, characteristics of the tissues in the middle, and the composite characteristics that summarize properties of the simple plant hydraulic pathway on the right.
Then we had to work out how to use the model sensibly - we did this by sampling trait combinations from known databases so that we ended up with physically plausible trait combinations, while we varied water potential and governing traits. The model output was a suite of curves relating conductance loss to soil water potential, which we interrogated in many different ways...
This is where things got a bit mind bending - we needed to develop metrics that compared the relative advantage/disadvantage of putting the fuses in the leaves or the stem under different conditions. For example we could compare the water potential at which the combined leaf-stem system reached 50% maximum conductance to the average of the P50 values for the leaves and stem together as a reference. We called this value "beta". Beta was most favourable for plants when the stems were the fuses and the leaves had low conductivity or when the leaves were the fuses and the stems had low conductivity. This variation shows up as the "saddle" shape on the plots below. These kinds of complex interactions across traits were typical in this study - if you feel like bending your own mind, there are many other relationships to explore in the paper!
Luckily Jeannie is good at making the complex simple, and summarized her findings in a nifty diagram below: Red here shows combinations of vulnerability segmentation and resistance that reduce overall conductance in stems or in the combined stem-leaf system. Blue shows favourable combinations. What this indicates is that where leaves are more conductive than stems (more resistance in stems), it's better for the plant as a whole or for the stem tissue alone that the leaves act as the fuse. However, where the stems are more conductive than the leaves, if you want to keep the whole plant conductance higher, it can be better off having the fuse in the stem.
So, sometimes plants should blow a fuse in their stems! This study really showed me the complexity of questions we don't often think about - like the "order" in which different tissue traits show up along a hydraulic pathway. It was a case study in how a superficially simple model can lead to some very complex and interesting results (and I'm so glad we kept it simple - it was hard enough to understand what we thought we were doing as it was!). It showed how sometimes we find ourselves needing to invent new metrics for things no one has tried measuring or modeling in quite that way before. And hopefully it might help us understand some of these weirdo plants out there and why their traits don't conform to our expectations.