What is it about?
Agriculture and Agri-Food Canada, in its unwavering commitment to sustainable agriculture, has launched a program to reduce nitrous oxide (N2O) emissions from fertilizer utilization in farming practices. This initiative is a response to the pressing environmental and climate challenges we face. To achieve our goal, we must delve into the mechanism of N2O emission by measuring and predicting the flux of N2O. This study proposes a novel architecture for neural network models, namely, the agriculture-informed neural networks (AINN), consisting of Recurrent Neural Networks (RNN) and a process-based ecosystem model, the Dynamic Land Ecosystem Model (DLEM), to predict N2O emissions from farming.
Featured Image
Photo by Daria from TaskArmy.nl on Unsplash
Why is it important?
This research aims to address the environmental and climate challenges with a commitment to sustainable agriculture. Agriculture and Agri-Food Canada has initiated a program to reduce fertilizer emissions by 30 % from 2020 to 2030. Accurate prediction and monitoring of agricultural N2O emissions are crucial for understanding their environmental impact and implementing effective mitigation strategies. This study aims to predict N2O emissions from agricultural fields using multivariate time series data, including the amount of nitrogen fertilizers in the soil, precipitation, air humidity, air temperature, soil moisture, and soil temperature.
Perspectives
The contributions of this work are as follows: • Combining a deep learning model and a process-based ecosystem model to form a novel AINN for accurately and efficiently predicting the emission of N2O in agricultural fields. • Formulating the mechanism of AINN as a constrained optimization problem, which could improve the generalization of the model, with the DLEM component in AINN acting as a regularizer. • Improving the explainability of the neural network model since the output obtained at the interface of the RNN component and DLEM component implicitly reveals the dynamic properties of the process of N2O emission. • Exploring various neural networks and providing an in-depth understanding of how each model handles multivariate time series data.
Ci Lin
University of Ottawa
Read the Original
This page is a summary of: Agriculture-informed Neural Networks for Predicting Nitrous Oxide Emissions, ACM Transactions on Internet of Things, September 2024, ACM (Association for Computing Machinery),
DOI: 10.1145/3696113.
You can read the full text:
Contributors
The following have contributed to this page
