Uncertainty quantification in atmospheric transport models

Quantification of uncertainties is becoming a main theme in meteorology. In weather forecasting, it has already been established operationally in the form of ensemble forecasts. In environmental meteorology, only general assessments of the accuracy of models are common. This project aims at quantifying the uncertainty of regional and global-scale atmospheric transport calculations with Lagrangian particle dispersion models. Likely errors shall be attached to each piece of output, like concentration of a trace substance or a source-receptor relationship. As the two main contributions to the simulation error, interpolation errors for the wind and other meteorological quantities and errors in the analysed wind fields used as input will be considered. The interpolation error shall be estimated with two different methods, using a comparison of forward and backward calculations, or by considering the interpolation error from an artificially degraded resolution. Both will be compared with very fine-resolution fields to be produced with a nudged meteorological model (MM5). The impact of the analysis error on a transport simulation can be assessed taking into account analyses from different meteorological centres, or by using the background analysis error structures as determined by the meteorological forecasting centres for their 3D or 4D variational data assimilation procedures. A main application of these results is inverse modelling, the determination (or improvement of an a priori information) of sources of an atmospheric trace substance from measurements. Detailed knowledge of errors in the source-receptor relationships will not only improve the inversion results but also enable the quantification of their errors. Such procedures will be implemented in a previously developed inversion technique. The project results will be useful for areas such as studies of long-range pollutant transport, the better determination of greenhouse gas and air pollutant emissions, and the monitoring of the Comprehensive Nuclear Test Ban Treaty.

Quantification of uncertainties is becoming a main theme in meteorology. In weather forecasting, it has already been established operationally in the form of ensemble forecasts. In environmental meteorology, only general assessments of the accuracy of models are common. This project aims at quantifying the uncertainty of regional and global-scale atmospheric transport calculations with Lagrangian particle dispersion models. Likely errors shall be attached to each piece of output, like concentration of a trace substance or a source-receptor relationship. As the two main contributions to the simulation error, interpolation errors for the wind and other meteorological quantities and errors in the analysed wind fields used as input will be considered. The interpolation error shall be estimated with two different methods, using a comparison of forward and backward calculations, or by considering the interpolation error from an artificially degraded resolution. Both will be compared with very fine-resolution fields to be produced with a nudged meteorological model (MM5). The impact of the analysis error on a transport simulation can be assessed taking into account analyses from different meteorological centres, or by using the background analysis error structures as determined by the meteorological forecasting centres for their 3D or 4D variational data assimilation procedures. A main application of these results is inverse modelling, the determination (or improvement of an a priori information) of sources of an atmospheric trace substance from measurements. Detailed knowledge of errors in the source-receptor relationships will not only improve the inversion results but also enable the quantification of their errors. Such procedures will be implemented in a previously developed inversion technique. The project results will be useful for areas such as studies of long-range pollutant transport, the better determination of greenhouse gas and air pollutant emissions, and the monitoring of the Comprehensive Nuclear Test Ban Treaty.