Computationally efficient atmospheric chemical kinetic modeling by means of high dimensional model representation (HDMR)

This paper presents an application of the efficient High Dimensional Model Representation (HDMR) method for relievingthe computational burden of chemical kinetic calculations in air quality models. An efficient HDMR for these types of calculations is based on expressing kinetic output variable (e.g., a chemical species concentration at a given reaction time) as an expansion of correlated functions consisting of the kinetic input variables (e.g., initial chemical species concentrations). The application of the HDMR method to atmospheric chemistry presented here focuses on a photochemical box model study of complex alkane/NOx/O3 photochemistry. It is shown that the HDMR calculations of multi-species time-concentration profiles can maintain accuracy comparable to the box-model simulations over reasonably wide ranges of initial chemical conditions. Furthermore, the HDMR expansion is about 400 times faster than the original box-model for performing ten thousand Monte Carlo uncertainty propagation runs, while producing very similar probability distributions of model outputs.