Sensitivity analysis for mercury over Europe

Adjoint techniques are introduced to perform a sensitivity analysis for mercury over Europe, using a regional model. This approach differs from other techniques such as the direct and indirect approaches developed to compute sensitivities for air quality modeling. With gaseous elemental mercury lifetime being of order one year, global or hemispherical models are usually preferred to analyze its fate and transport. In an area‐limited domain, respective influences of incoming mercury and inner emissions fluxes have to be weighted. A local measurement depends on the potential ground emissions, the potential incoming mass from domain borders, and the potential initial content in mercury of the atmosphere. The sensitivities of the measurement to entire maps of emissions, boundary conditions, etc., are computed thanks to the adjoint method since other techniques do not allow to perform these computations directly. As an application to these methodological developments, we use the numerical transport model Polair3D. Quantitative sensitivity maps are provided for European Monitoring and Evaluation Program (EMEP) mercury monitoring stations. With the adjoint approach, sensitivities of a given country to other countries emissions are computed in a straightforward manner. The yearly average sensitivity of a measurement of gaseous elemental mercury to distant sources is shown to decrease like a power law r −2.4 , where r is the distance to a source. It is eventually explained how the method developed here can be generalized to account for a more complex mercury chemistry and the modeling of oxidized species. In particular, sensitivities of dry and wet deposition fluxes of oxidized species are computed for one EMEP station.