In This Story
People are nuts for almonds, which consistently rank as the second-most consumed nut in the world (after peanuts, which are technically legumes, but that’s another story). Around 3.6 million metric tons of almonds are eaten every year; if you’re counting at home, that’s about 2.6 trillion almonds.
The crop is particularly important in Israel, where the industry is valued at nearly $60 million, and the plant, one of the earliest flowering trees in the region, is symbolically and commercially important.
George Mason Professor Lizhi Wang, in the Department of Bioengineering and Department of Systems Engineering and Operations Research, recently received $160,000 to work with colleagues in Israel to improve almond growth and yield. The funding is from the Israeli-based BARD (Binational Agricultural Research and Development Fund), which supports “bringing together U.S. and Israeli scientists to confront important agricultural challenges that are common to the two countries.”
People have experimented on getting more out of plants since agriculture was developed, studiously tracking sun and water exposure, but they still get flummoxed about why some years a plant will produce more than other years, despite similar circumstances. So what makes things different now?
“What is very special about the timing of this research is the convergence of two things,” said Wang. “One is people's knowledge about plant physiology and how crops will respond to environmental stress. The other is data science, big data, and machine learning.”
Wang and colleagues will build an almond tree’s digital twin, a complex data model that simulates the growth process. His colleagues in Israel will feed environmental data into a model. “Every hour there is photosynthesis going on. We’ll look at how much CO2, sunlight, and water the plant is being exposed to,” Wang said. The plants growing in the soil will be compared to the digital twin’s “growth.”
Almond trees are famously not self-compatible, meaning, one plant cannot fertilize itself using its own pollen. That’s why commercial almond orchards interplant different varieties and rely heavily on bees to shuttle pollen between them. Bees can be finicky, and don’t pollinate if the weather is too hot or too cold, for example. So among other things, the model will consider if the weather is ideal for bees to work their pollination magic.
Wang said his colleagues “every hour collect data about the temperature, about the humidity, and they measure the tree growth, the circumference of the tree trunk and more. We built a model that considers all of these factors and then we predict ultimately the yield.”
The team is in the second year of the project. The first year involved developing a version of the mathematical equations; in the second year the team has been building computer programming, integrating the data, and producing a prototype of the model.
Wang said the findings can have wide-ranging impacts, and that a collaborator in California is particularly interested, given that state’s production of virtually all the almonds in the United States and 80 percent of the world’s supply.
Separately, Wang is getting similar digital twin research off the ground looking at chickens and a slightly more complex but still somewhat nutty organism…the human body.