A new approach for rapid measurement of directional root responses to neighbours using the root centroid

Abstract Measuring directional root placement is critical for understanding plant responses to their below‐ground environment, and particularly their competing neighbours. Directional root placement is commonly measured using image analysis of roots growing in transparent pots (rhizoboxes), where the length of the root system of the target plants is tracked. However, tracking roots with a soil background can be highly challenging, particularly in competition studies, where two or more root systems are intertwined within the same experimental setup. In this study, we propose a new approach for measuring directional root placement in competitive set‐ups, with two methods that calculate the centroid of the root system without measuring overall root length. In the first method, the centroid is calculated by marking all the intersection points of the target plant roots along a fixed number of equally spaced horizontal lines superimposed on the image. In the second method, the centroid is calculated from a contour line (polygon) created by marking only the peripheral intersection points. We developed an open‐access, interactive Python algorithm that automates and standardizes the centroid calculation for both methods. We validated these methods by comparing them to the centroid calculated from the traditional root length measurements using results from two rhizobox competition experiments, with either uniform or patchy soil nutrient distribution. While the two methods offer a more rapid and standardized calculation of the root system centroid, they differ in their investment time vs. accuracy levels, particularly when root density is heterogeneous. By focussing solely on a few sample points from the root system or its contour line rather than tracking the entire root system, this approach offers a potentially faster way for measuring directional root placement. The centroid approach could therefore facilitate the study of plant responses to below‐ground competition, enabling more efficient tracking over time and across multiple samples.