Efficient ensemble generation for uncertain correlated parameters in atmospheric chemical models: a case study for biogenic emissions from EURAD-IM version 5

Abstract. Atmospheric chemical forecasts heavily rely on various model parameters, which are often insufficiently known, such as emission rates and deposition velocities. However, a reliable estimation of resulting uncertainties with an ensemble is impaired by the high dimensionality system. This study presents novel approach, substitutes problem into low-dimensional subspace spanned leading uncertainties. It based idea that forecast acts dynamical system inducing multivariate correlations enables efficient perturbation high-dimensional parameters according to their coupled The specific algorithm presented in this designed for depend local environmental conditions consists three major steps: (1) assessment sources independent sensitivities, (2) extraction using eigenmode decomposition, (3) generation perturbations parameter fields Karhunen–Loéve expansion. Due perceived simulation challenge, method has been applied biogenic emissions five trace gases, considering state-dependent sensitivities atmospheric terrestrial conditions. Rapidly decreasing eigenvalues state highly correlated regional can be represented low number dominant components. Depending required level detail, dimensions