Effects of uncertainties in parameters of a Lagrangian particle model on mean ground-level concentrations under stable conditions

Abstract

This work evaluates the effects of uncertainties in five parameters or inputs of a source-receptor Lagrangian particle model on mean ground-level concentrations. The scope of work is short-range dispersion in the atmospheric boundary layer under weakly and moderately stable conditions over a smooth flat terrain within 3 km downwind from a continuous point source located near the ground. Model inputs include four meteorological parameters and the universal constant in the random component of the model. The meteorological parameters are friction velocity, mean surface turbulent heat flux magnitude (shortly, heat flux), surface roughness height (shortly, roughness height), and mean surface temperature (shortly, temperature). Model outputs of interest are mean ground-level concentrations at a number of receptors downwind from the source. The formal uncertainty analysis was performed for the atmospheric conditions corresponding to stability indices (defined as the ratio of mixing height to Monin-Obukhov length) of 1.2, 2.5, and 4.1. Input uncertainties were propagated through the model using Monte Carlo simulations with Latin hypercube sampling. Linear regression modeling was used to statistically partition an output uncertainty and determine the relative importance of an input to an output. It is shown that the concentration uncertainty increases with the level of meteorological uncertainty and that its magnitude varies with stability degree. Among the meteorological parameters, friction velocity is the most influential. Another important input is the universal constant whose contribution can dominate when the level of meteorological uncertainty becomes low. The overall contributions from roughness height and temperature are relatively small.