Toward systems agroecology: Design and control of intercropping

Area of Science:

• Agricultural Science
• Systems Biology
• Data Science

Background:

• Industrial monocrop farming faces challenges from climate change and resource depletion.
• Intercropping (IC), growing multiple crop species together, shows potential for improved soil health, reduced erosion, and efficient fertilizer use.
• Current limitations exist in quantitatively predicting, designing, and controlling IC for specific environments and assessing its robustness.

Purpose of the Study:

• To develop a quantitative, data-driven approach for predicting, designing, and controlling intercropping (IC) systems.
• To assess the robustness of IC strategies under various environmental and farming conditions.
• To lay the groundwork for "systems agriculture" by enabling optimized multi-species cropping systems.

Main Methods:

• Compiled a dataset of 2258 IC experiments, including 274 plant pairs, soil characteristics, environmental/farming conditions, and crop traits.
• Applied data science and systems biology methods, including dimensional reduction of a 25-dimensional variable space.
• Developed a computational approach to predict IC yield relative to sole cultivation.

Main Results:

• Successfully predicted intercropping (IC) yield relative to sole cultivation with high accuracy using a reduced set of variables.
• Demonstrated the ability to select optimal companion plants and farming practices for given environmental conditions.
• Provided a method to estimate the robustness of IC systems to external perturbations.

Conclusions:

• The developed computational approach enables data-driven optimization of intercropping (IC) for enhanced agricultural sustainability.
• This methodology facilitates the selection of suitable crop combinations and management practices for specific environmental contexts.
• The approach represents a significant step towards "systems agriculture," promoting resilient and efficient multi-crop farming systems.