Digital farming tools simulate a full season’s growth in a single day
Humans have been manipulating the evolution of plants for ages, but usually at a pace slow enough we barely notice. By planting seeds from specific plants that had attributes we liked more than others, say a more pleasing color, or larger amount of tasty flesh, we’ve transformed many plants into the produce we know today. However, this is a slow process, and farmers are looking for ways to speed things up while reducing the costs associated with experimenting with a whole season’s crops. The solution may be to first grow crops on a in silico, or “in silicon chips,” before ever putting a seed in the ground.
The simulations that are being developed allow for some very specific details to be tested. For instance, will you get a bigger crop yield if you plant your sugarcane in staggered rows, or all lined up? Should they be angled north-south, or east-west? A farmer could plant four different fields of sugarcane to see which did best, although in doing so they might introduce new variables to the mix. It would also be a slow process, possibly risking income for 12 months of work.
The in silico version took all the available data and came up with a prediction in 24 hours. It considered minutiae down to the amount of light that might be blocked by a neighboring plant’s leaves at different times of the day, then produced a 3D visualization to show the expected outcome of each field arrangement. In this case, staggered plants planted on a north-south axis was predicted to increase yields by ten percent, making that a much safer test to run in the real world for confirmation.
Farming experiments made even faster
As these tools are developed, researchers hope that the speed and depth of the simulations can be improved. Not everyone can tie up a supercomputer for 24 hours to test out a new technique, and the goal is to eventually simulate a whole season’s growth in a minute, making it easier to try out different variables. The number of variables should also be increased to incorporate more data that different labs have been creating over the past decades, but that requires some serious coordination efforts. Not every research team uses the same tools or data structure to archive their experimental findings, which makes integrating existing information about crops difficult.
Still, the developers are confident that all these challenges can be met, partially because they have to. Concerns over population, soil quality and fresh-water availability suggest that farms will need to be more efficient than ever in the coming years. A tool that lets you configure and simulate new ideas in a single afternoon could save everyone a lot of time and resources.
Source: Growing Virtual Plants Could Help Farmers Boost Their Crops by Leslie Nemo, Scientific American
