Evaluating GHGSat plume strength estimates in an urban environment using nested WRF- LES simulations

High spatial resolution satellite instruments operated by GHGSat estimate methane (CH4) emissions from relatively large point and area sources. The objective of this study is to assess the performance of GHGSat’s measurements of CH4 plumes using the Weather Research and Forecasting-large eddy simulation (WRF-LES) at a landfill in Indianapolis, Indiana, a well-studied, urban source of 20-25 mol s-1 of CH4. Our WRF-LES configuration used to simulate dispersion of tracers in the turbulent atmospheric boundary layer with realistic mesoscale forcing includes six one-way nested domains with 87 vertical levels. The 37 m horizontal grid spacing of the innermost domain is comparable to the 30 m spatial resolution of GHGSat. With landfill emissions represented as a constant, continuous source, methane dispersion in the innermost domain is compared to GHGSat observed CH4 plumes. The analyses performed for a total of 15 different days show considerable day-to-day variation and can be grouped as: 1) good agreement in plume dispersion direction and shape; 2) large differences in the plume dispersion direction and shape; and 3) plumes not observed by GHGSat (three of 15 days; no clear reasons identified). The plume shapes and directions diverge when the 10 m wind speed is below 2 m s-1. Compared with the WRF-LES simulations, GHGSat measurements underestimated plume strength by 38-47%. No clear dependence of bias or absolute error on GHGSat reported uncertainty or environmental parameters is identified. Our findings suggest that GHGSat could be used for detecting and quantifying moderate magnitude CH4 sources even in complex urban environments, but systematic underestimation and occasional non-detections should be considered when interpreting derived plume strength.