Stem-FIT and Soil-FIT: Integrated plant and soil nitrogen-hormone sensing with machine learning-based forecasting for next-generation precision agriculture

Abstract Inefficient fertilizer application in agriculture leads to reduced crop productivity, nutrient losses, and reduced crop resilience, highlighting the urgent need for real-time monitoring of plant–soil nutrient dynamics. This research aims to develop and validate a multiplexed sensing platform for simultaneous, in-situ measurement of key soil nutrients (Soil-FIT) and plant phytohormones (Stem-FIT) involved in nitrogen signaling and stress regulation. The proposed sensor suite integrates three 3D-printed modules for continuous monitoring of nitrate, ammonium, and pH in both soil and plant sap, along with salicylic acid (SA), indole-3-acetic acid (IAA), methyl jasmonate (MeJA), and ethylene (ET) in plant sap. The sensors, functionalized with non-enzymatic electrode coatings, were deployed in bell pepper plants grown under four treatment combinations of irrigation (full vs. deficit) and nitrogen application (medium vs. high). Data were collected every three hours over the growing period and analyzed using a long short-term memory (LSTM) model for short-term prediction of nutrient and hormone fluctuations. The sensors exhibited high sensitivity and stability, achieving detection limits of 0.218 µM for IAA, 1.07 µM for MeJA, 1.315 µM for SA, 1.08 ppm for nitrate, 1.017 ppm for ammonium, 0.29 ppm for ethylene, and 0.01 pH. The LSTM model demonstrated strong predictive capability (R² = up to 0.86), accurately forecasting short-term variations in plant and soil nutrient–hormone profiles. These findings demonstrate that coupling real-time, multiplexed sensing with machine learning enables early detection and prediction of crop stress, supporting precision nitrogen management and advancing sustainable agricultural practices.