Development of perennial grain crops has been a major focus at The Land Institute for more than 30 years. I question their chances of success -- see an excerpt at Page 99 Test -- but (sort of) endorse the attempt anyway, in the final chapter of my book:
"As explained in chapter 7, I have some doubts about perennial grain crops. But do we really want to focus 100 percent of our grain-breeding efforts on annuals? As long as they pay attention to tradeoffs-- a recent e-mail from David Van Tassel, working on perennial grains at the Land Institute, suggests that they are-- they might come up with something useful. This may sound like weak praise, but it's really all we can say about any long-term research project, including my own."
This triggered an email from Timothy Crews, also at the Land Institute, pointing out that they have discussed tradeoffs in peer-reviewed publications, not just in emails.
A recent example, titled "Missing domesticated plant forms: can artificial selection ﬁll the gap?" was published in Evolutionary Applications. It's open access, so have a look!
There are two reasons that grain yields of perennials might differ from those of annuals. First, perennials may sometimes capture more resources than annuals can. For example, they may use some of last-year's photosynthate to grow a dense canopy of light-capturing leaves earlier than an annual could. This is because early leaf growth in annual seedlings depends on current photosynthesis of initially tiny leaves. Similarly, the deeper roots of annuals may access soil water that annuals would miss.
But capturing more early-spring sunlight for photosynthesis depends on the perennial storing enough photosynthate in the autumn to survive the winter, with enough left over to support early leaf growth. Similarly, capturing more soil water depends on having allocated resources to root growth that could otherwise have been used to make grain. It seems, therefore, that perennials would have to allocate a much larger fraction of their resources to below-ground storage and roots, relative to annuals.
The article by Van Tassel, DeHaan, and Cox, questions this assumption. For example, their Table 1 refers to published data showing that apple, oil palm, olive, and high-bush blueberry can all allocate more than 50% of annual photosynthate to fruits. This is comparable to high-yielding annual crops, like corn. I don't know how typical these values are -- my book cites almond yields of <2000 kg/ha in a region where corn yields >10,000 kg/ha -- but it suggests that woody perennials can have higher yields than I would have expected. These trees often have so little allocation to trunk that they have to be propped up on trellises, but that may be a more-acceptable tradeoff than "propping up" crop health with lots of pesticides. Still, what about herbaceous perennials?
I will focus my discussion on the section of the paper titled "Hypothesis 1", where the authors argue against the hypothesis that "[high-yielding] perennial grains would be impossible on logical or thermodynamic grounds." If that hypothesis is true, as I suspect, then the rest of their arguments miss the point.
Their first point in this section is that perennial plants can be larger, so they can make more seeds per plant, even if their percent allocation to seeds is low. True, but so what? What matters is grain yield per acre, not per plant, and you can fit more plants per acre if they're smaller.
Then they cite an interesting paper by Dohleman and Long (2009) showing that the perennial Miscanthus can produce more total biomass than corn. I cited this paper in my book as evidence for the hypothesis above, that perennials can sometimes capture more resources than annuals. Miscanthus can have green leaves capturing sunlight both earlier and later than corn does.
What about seed yield? Zero. The genotype used doesn't make any seeds. It's propagated from cuttings. So while corn is recycling nitrogen from leaves to seeds in the autumn, Miscanthus just keeps growing. Ok, but weren't we talking about the potential of herbaceous perennials to have high grain yields?
It's also worth noting that that it wasn't until the third year of growth that Miscanthus plants got big enough to capture that much sunlight and produce that much biomass. We worry that increasing the frequency of corn in a rotation to two years out of three risks disease, but wouldn't growing Miscanthus year after year have similar risks?
Next, Van Tassel et al. cite a paper by Aragon et al. (2009), noting that although "blossom removal
increased the survival rate of a short lived perennial, it did not affect reproductive effort." So I looked at Figure 2 in Aragon et al. Blossom removal in 2005 roughly doubled survival to 2006, and the difference was statistically significant. This seems like the key result, showing that producing seeds decreases the chances of surviving to the next year. It's true that, among the plants that did survive to 2006, there were no significant differences in 2006 seed production. But so what? Also, it's hard to get statistically significant differences when you only have one surviving plant in the control treatment (versus six in the deblossomed treatment)!
Van Tassel et al. also cite a paper by Ploschuk et al. (2005) finding no difference in allocation to seeds in two related species, an annual and perennial. Comparing two related species is a good start, but a sample size of one is too few to draw conclusions, as just noted. Silvertown and Dodd (1996) asked a related question:
When an annual evolves into a perennial, or vice versa, how does allocation to seeds change?
They looked at 13 such transitions and found a highly significant association: perennials allocate fewer resources to seeds than annuals do. Median harvest index values were 28% for annuals and 7% for perennials. They don't give as much detail on their data sources as I would have liked, though.
To summarize, the paper by Van Tassel et al. didn't convince me that we will ever be able to breed herbaceous perennials that allocate as large a fraction of their resources to grain as wheat or corn do, while also producing more roots than those crops do (to stabilize the soil) and still retaining enough resources to survive the winter in a dense enough stand to suppress weeds. Their paper has plenty of interesting information on trees and seed-free Miscanthus. But what I would most like to see is a graph of changes in allocation to seed resulting from their attempts to increase the yield of seed-producing perennials, together with associated changes in over-winter survival and root biomass. I assume there will be tradeoffs, but will they be tradeoffs we can accept?