Digital Twins of Plants Could Help Increase Crop Yields
Scientists at the Samara Federal Research Center of the Russian Academy of Sciences have developed a digital platform that models how crops grow and develop under different environmental conditions.
The Digital Twin of Plants (DTP) lets agronomists simulate the development of different crops in a virtual environment, use the results to forecast yields and make management decisions quickly. This can improve the accuracy of crop production planning and, in turn, make farming businesses more efficient.
Digital Analysis of Every Growth Stage
The DTP was developed by scientists at the Samara Federal Research Center of the Russian Academy of Sciences (SamRC RAS). Work on the system began in 2021, after the team won a competition organized by Russia’s Ministry of Science and Higher Education to establish a youth laboratory dedicated to DTP development. The researchers received support from Russian Academy of Sciences academician Sergey Shevchenko. The project’s scientific supervisors were Petr Skobelev, DSc in Engineering, and Oleg Goryanin, DSc in Agricultural Sciences.
The computer model is based on the processes that govern plant growth and development at successive stages, each of which lasts for a period determined by weather and other environmental factors. The result is a virtual twin that represents the crop production process and maps the full development trajectory of each plant variety. Comparing the life-cycle parameters of real and virtual crops took three years. The system can now forecast yields for major crops with 70% to 80% accuracy.
Before planting even begins, the DTP can model plant growth and development under changing environmental conditions using data on climate, soil and farming practices. In effect, it creates a virtual copy of the future crops on a map of the farm.
An Agronomist’s Assistant
A key feature of the platform is its ability to stay synchronized with actual field conditions. Once planting begins, the digital twin automatically models each current stage of plant growth, analyzing incoming data on weather; soil conditions, including agrophysical and agrochemical properties and water regime; and the timing and quantities of fertilizer applications. Agronomists can then review the crops’ life-cycle data and receive recommendations on crop treatment.
The new system is well suited to individual use by agronomists on their farms. Some crop-variety parameters and calculation rules are stored in a separate crop-production knowledge base tailored to the Middle Volga region. This allows farmers to adjust variety models to the specific conditions of their operations.
“Developing the DTP opens up new opportunities to provide recommendations to agronomists and farmers. We are currently working on an intelligent recommendation module for the DTP. By monitoring and comparing the growth and development of virtual and real plants, it will be able to generate context-specific recommendations on fertilizer application and crop protection,” said Alexey Vasin, director of the Samara Federal Research Center of the Russian Academy of Sciences.
The plant knowledge base used by the digital twin covers dozens of varieties of winter and spring wheat, peas, sunflowers and soybeans. These are among the main crops grown by farms in Samara Region and the wider Volga region. The range of crops supported by the platform is expected to expand.
Commercialization and Technology Exports
The DTP is already available in two formats: through a desktop web browser or a mobile app. Using a smartphone, an agronomist can conduct field surveys and record crop conditions. The DTP uses those records to adapt its subsequent plans and yield estimates. Users can enter new data at any time and model how the changes could affect their crops.
The project has now entered the commercialization stage, with the team seeking its first industrial customers. Plans are also underway to establish a shared-use DTP center for farms in Samara Region and neighboring regions, allowing users to work remotely.
This is not Russia’s first digital-twin project for crops. Scientists at the Space Research Institute of the Russian Academy of Sciences and Lomonosov Moscow State University are developing a model for three crops: corn, barley and sunflowers. The platform is being configured to forecast yields. Its average error rate currently stands at 24% to 25%, although recurring failures have been fully eliminated.
Foreign farmers have already shown interest in the work of Russian researchers. Crop producers in China and India are planning to use the Samara team’s platform and want to work with its developers to adapt the model for crops grown in Southeast Asia.
