Point-of-use sensors and machine learning enable low-cost determination of soil nitrogen

Overfertilization with nitrogen fertilizers has damaged the environment and health of soil, but standard laboratory testing of soil to determine the levels of nitrogen (mainly NH4+ and NO3-) is not performed regularly. Here we demonstrate that point-of-use measurements of NH4+, combined with soil conductivity, pH, easily accessible weather and timing data, allow instantaneous prediction of levels of NO3- in soil (R-2 = 0.70) using a machine learning model. A long short-term memory recurrent neural network model can also be used to predict levels of NH4+ and NO3- up to 12 days into the future from a single measurement at day one, with R-NH4+(2) -0.60 and R-NO3-(2) -0.70, for unseen weather conditions. Our machine-learning-based approach eliminates the need for dedicated instruments to determine the levels of NO3- in soil. Nitrogenous soil nutrients can be determined and predicted with enough accuracy to forecast the impact of climate on fertilization planning and to tune timing for crop requirements, reducing overfertilization while improving crop yields.

Automated summary

Our study demonstrates that point-of-use measurements of NH4+ combined with soil conductivity, pH, weather, and timing data can predict levels of NO3- in soil. Additionally, a long short-term memory recurrent neural network model can forecast NH4+ and NO3- levels up to 12 days into the future. This approach eliminates the need for dedicated instruments to determine the levels of NO3- in soil, allowing for accurate forecasting of the impact of climate on fertilization planning and crop requirements.