Back

Spatiotemporal changes in soil properties predict plant health at continental scale

Wang, N.; Jiang, G.; Hou, Y.; Li, C.; Meng, Y.; Chen, P.; Zhang, Y.; Yang, T.; Wang, X.; Mei, X.; Pommier, T.; Banerjee, S.; Rillig, M.; Dini-Andreote, F.; Friman, V. P.; Xu, Y.; Jousset, A.; Shen, Q.; Wei, Z.

2025-01-10 ecology
10.1101/2025.01.07.631818 bioRxiv
Show abstract

Abiotic and biotic soil properties are strong predictors of plant yield globally1-5, but they become unreliable over large areas when plant health is threatened by pathogen6-8. Here we present a novel approach to predict plant health based on spatiotemporal changes in soil chemical and biological properties. We first demonstrate that plant health and soil properties consistently respond to environmental change (organic fertilization) regardless of the soil type or geographical origin. Second, we experimentally show that trackable shifts in soil properties reliably explain soil suppressiveness to the Ralstonia solanacearum bacterial pathogen and that a scalable spatiotemporal model predicts plant health with 84% accuracy across multiple climatic zones and cropping systems. Our results suggest that this tight coupling between soil properties and plant health could facilitate the development of agricultural practices aimed at sustainably improving crop yields while safeguarding crop health.

Matching journals

The top 2 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.