Surrogate-based ensemble data assimilation for reducing uncertainty in large-eddy simulation of microscale pollutant dispersion

By capturing the physical complexity of the interactions between atmospheric flows and the built environment, Large-Eddy Simulations (LES) could provide detailed information for risk assessment and mitigation in case of environmental emergency. However, to account for LES uncertainties and cover the range of plausible scenarios in order to support decision making, it is necessary to go beyond deterministic simulation capability. This study introduces a novel ensemble-based data assimilation algorithm to correct the LES meteorological forcing and thereby improve LES spatial predictions of pollutant concentration by making use of available measurements. This approach is demonstrated through the MUST field-scale experiment. Results show that the ensemble smoother with multiple data assimilation (ESMDA) algorithm is a good candidate to address parameter interaction effects in the relationship between uncertain meteorological forcing and LES field quantities. This iterative algorithm is computationally feasible when the LES model is replaced with a machine learning-based surrogate model, from which robust ensemble statistics can be extracted. This surrogate-based data assimilation approach can then be used to examine observability in the system. Results show that the estimation outcome is highly sensitive to the design of the observation network, and that this sensitivity may be underestimated in idealized experiments. It is therefore important to use real data assimilation to optimize sensor placement and extract informative data for modeling, thus improving our ability to monitor accidental dispersion events.