Exploring Temporal Machine Learning Approaches for Predicting Methane Discharge in Atmospheric Studies

Methane levels in the atmosphere are rising, posing a serious challenge in the fight against climate change. Because methane is a powerful greenhouse gas, being able to accurately forecast its concentration is essential for making smart decisions about climate policy. Traditional ways of monitoring methane rely on direct measurements, which can be expensive and are often limited to certain areas. This study aims to explore how artificial intelligence algorithms (machine learning and deep learning) can help fill the gap. Using time series data from Alberta’s methane warehouse, we tested and compared four different models: Random Forest, Extreme Gradient Boosting, Multilayer Perceptron and Temporal Convolutional Networks. We assessed how well each model predicted methane levels over time and how efficiently they handled the data. The findings offer insights into building more affordable and scalable systems for tracking and managing methane emissions, helping support better climate strategies.