Quantifying gas emissions through Vertical Radial Plume Mapping based on historical information

Abstract

Vertical Radial Plume Mapping (VRPM) is a method for measuring fluxes of fugitive surface source gases, lacking sufficient temporal sequence information as it inversely calculates fluxes only based on the current cycle’s measurements. We propose an improved VRPM flux inversion model, incorporating Weighted Iterative Average (WIA) corrected Path-Integrated Concentration (PIC) and Adaptive Weighted Sum of Squared Errors (AWSSE). Compared to the original VRPM, our model introduces historical measurement PICs and reconstructed PICs. Using the Bayesian Optimization Gaussian Processes (BO-GP), we derive optimal weight factor for the hyperparameter of the WIA component in the model. Validation on simulated data highlights the superior flux inversion capability of the WIA-AWSSE-VRPM model across diverse surface sources, efficiently leveraging historical data. The improved VRPM method achieves a significant 9.65% reduction in root mean square error (RMSE) within consistent source sets and decreases RMSE from 1.6979 to 0.1513 across varied sources, underscoring its effectiveness in accurately measuring emissions from fugitive sources. Finally, we apply the model to characterize gas emissions in field experiments.