An investigation of the assumptions of linear chemistry and superposition in LRTAP models

Most current LRTAP models assume a linear relationship between sulphur emissions and predicted acidic concentrations, as well as the ability to superpose concentrations from different emission sources. This paper uses a non-linear chemistry model and control scenarios of 50% reductions in S, N and hydrocarbon (HC) emissions to examine the validity of these two assumptions at various downwind receptors. The model predicts that a 50% reduction in S emissions will lead to a 60–65% reduction in SO₂ concentrations and a 25–40% reduction in H₂SO₄ concentrations, depending upon whether or not NO_x and/or HC emissions are reduced by 50% at the same time. The non-linearities in the model predictions are due to complex interactions between NO_x, HC, OH and HO₂. Even when there was non-linearity in the individual S species (SO₂ and SO₄), there was little non-linearity in total airborne S. Adding the results of independent model predictions for different sources (superposition) might introduce errors due to

• 1.(1) chemical interactions between the emissions from the various sources
• 2.(2) overestimates of physical processes such as mixing in of ambient air.

The model was also used to examine trends in regional air quality in eastern North Amierica.