Filling a critical need: a lightweight and fast Gaussian puff model implementation

Atmospheric dispersion models are a key component for characterizing methane emissions on oil and gas sites. While some model implementations with varying degrees of complexity are available, existing regulatory-grade dispersion models are cumbersome to apply within an inversion framework on a routine operational level or at scale, and require a gamut of meteorologic information that is typically not available in field operations. Simple Gaussian plume models on the other hand, fail to incorporate changing wind regimes that are routinely observed and prove relevant in practice. Filling this critical need, we provide a computationally efficient and scalable version of the Gaussian puff model with a thresholding algorithm that is two orders of magnitude faster than a naive implementation and only requires readily available meteorological data. Our comparison shows that the Gaussian puff model is higher fidelity and able to capture temporal variation in atmospheric transport better than the commonly used Gaussian plume model. This computationally-efficient and scalable Gaussian puff implementation can be applied to model near-field atmospheric transport in a broad range of applications including the timely call to quickly infer methane emissions from measurements on oil and gas sites for efficient mitigation.