Modelling NOx concentrations at urban scales: model sensitivity, emissions and applications

Abstract

Nitrogen oxides (NOx = NO2 + NO) cause respiratory and cardiovascular diseases, and are an important precursor of tropospheric ozone (O3), in turn damaging human health, ecosystems and warming the climate. Since the introduction of European NO2 limit values in 2010, these have been frequently exceeded, especially in Germany. In addition, car manufacturers do not comply with NOx emission limit values for diesel cars, with some of them having implemented defeat devices to suggest compliance. Together, the exceedance of limit values and non-compliance with emission thresholds make action on reducing NOx concentrations and more stringent NOx emission control policies necessary. Beyond that, knowledge gaps exist concerning the broader consequences of high urban NOx concentrations for overall urban air quality.

Research and in particular modelling play an important role for assessing above-mentioned air quality issues and emission control measures in their context. This thesis contributes to the literature in several ways: it contributes to the development of the modelling tools needed for studying air quality at urban scales and assesses the potential underestimation of traffic NOx emissions - the most important source of NOx emissions in urban areas - in a commonly used emission inventory. It then applies the developed model setup, analyzing to what extent urban NO2 concentrations could be reduced if vehicle emission standards were met. It also assesses the contribution of urban biogenic volatile organic compound (VOC) emissions - another important precursor of O3 - to O3 concentrations during a heat wave. The research is presented in four published articles, all based on modelling studies using the chemistry version of the Weather Research and Forecasting model (WRF-Chem), and using the Berlin-Brandenburg area as a case study.

The first article presents and evaluates a WRF-Chem setup for Berlin-Brandenburg, build- ing the basis for the other articles in this thesis. It finds that a horizontal model resolution of 3km x 3km improves both modeled meteorology and chemistry compared to a 15km x 15km horizontal resolution. The article further finds that when modelling air quality in urban areas at high resolution, a detailed description of the urban area based on locally available input data is beneficial. For modelling NOx concentrations, the study finds a better representation of local pollution patterns when downscaling the emission inven- tory to the model resolution. However, modelled NOx concentrations are underestimated compared to observations, particularly on weekdays during daytime.

The results from the model evaluation as well as an analysis based on observed NOx concentrations and traffic counts in the second article suggest that an underestimation of traffic emissions is one of the main sources of the model bias. Also taking into account other potential sources of model bias such as too strong mixing during daytime, the study then calculates a correction factor for traffic NOx emissions. It finds that traffic NOx emissions are underestimated by a factor of ca. 3 during daytime on weekdays in the core urban area, corresponding to an underestimation of weekly mean traffic NOx emissions in the core urban area of ca. a factor of 2 and an underestimation of total NOx emissions in the city centre by a factor of ca. 1.5. Applying this correction factor improves modelled NOx concentrations in the core urban area as well as downwind of the city. The results suggest that further research is needed in order to better specify NOx emission factors used for officially reported emissions.

The third article applies the WRF-Chem setup, as well as a coarser setup covering a larger domain, and presents sensitivity simulations looking at how much NO2 concentrations could be improved if vehicle emission standards were met. It combines the model-based assessment with an analysis based on measurement data. The results from the different approaches are consistent and suggest that NO2 concentrations could be improved by 1.2 - 2.7 μg m−3 in the urban background and 9.0 - 23.0 μg m−3 at the roadside if the strictest US EPA standards were met. Considerable improvements of urban air quality could also be expected if car manufacturers would comply with European emission limits.

The fourth article looks at broader consequences of high urban NOx concentrations, and is aimed at quantifying the contribution of VOC emissions from urban vegetation to O3 concentrations during a heatwave. The results suggest that on specific days during the analyzed heatwave period the contribution of biognic VOCs to ozone formation reaches up to 60%, compared to average contributions of 17% to 20% during the month of the heatwave. This shows that urban environmental measures need to be assessed compre- hensively in order for cities to fully benefit from them. For example, urban tree planting campaigns would have to be accompanied by a reduction of anthropogenic sources of O3 precursors (NOx, VOCs), e.g. in the area of road transport.