An assessment of the NAME III model capability in reproducing seasonal variation of SO₂ and O₃ pollutant concentrations : a focus on the Mpumalanga Highveld area.

Master of Science in Environmental Science. University of KwaZulu-Natal, Durban 2015.

The South African Weather Service Air Quality Modelling programme initiative has a long term goal to develop a system capable of generating atmospheric air quality forecasts for a range of primary and secondary pollutants in order to advise and warn the public on possible high levels pollutant concentration in the air and also provide support in relevant policy development. In this study a pilot NAME III air quality modelling system was developed and tested for its performance in simulation of sulphur dioxide (SO₂) and ozone (O₃) concentrations over the Mpumalanga Highveld. The agreement of model predictions generated in this study with observations was evaluated using the statistical analysis of the monthly averages of SO₂ and O₃ concentrations based on the Bias, NMB, RMSE, NRMSE statistical measures. In addition, the seasonal distribution and variation of the modelled SO₂ and O₃ concentrations over the South African domain were assessed. The results demonstrate that the modelling system under-predicts SO₂ and O₃ concentrations. However, in most cases the modelled concentrations are in the same order of magnitudes with the measured data except for two incidences of very low modelled SO₂ in Middelburg during April and May months, which may be attributed to the poor initialisation of the model. For each season, the model was initialised for the first five days to allow for the pre calculation of the initial pollutant concentrations. This was not possible for the autumn season as no Numerical Weather Prediction (NWP) data were available for initialisation during this period. In general the overall results indicate that the NAME III modelling system is a promising and cost-effective tool for providing real time air quality forecasts, in particular, the ground level O₃ concentration in South Africa. The NAME III modelling system therefore has the potential to be used operationally as a national air quality forecasting system and, as a tool to conduct air quality modelling studies. Specifically the modelling system could assist in the amendment and development of relevant air quality policies that have a direct impact on the environment, health and other related sectors. However, it is suggested that while more evaluation exercises must be undertaken, advancements in term of a comprehensive emissions inventory and improved representation of meteorological information are needed.