Spatial aspects affecting acidification factors in European acidification modelling
Plain linear models have recently been used in methodologies to model fate and transport for assessing acidification in life cycle impact assessment (LCIA), or in support of air pollution abatement policies. These models originate from a statistical analysis of the relationship between inputs and outputs of physically-based models that reflect the mechanics of a system in detail. Linear models applied to assess acidification use an acidification factor (AF), which relates changes in the magnitude of emissions to changes in the total area that is protected against acidification in Europe. The changes in emission volume refer to changes of one substance, within one country and one sector or one grid cell.This paper evaluates the dependence of acidification factors on three spatial characteristics, i.e. the spatial emission and deposition resolution, the spatial emission distribution and the actual spatial location of emissions.
Within both life cycle impact assessment (LCIA) and air pollution policy, inaccuracy of the calculated environmental impact (acidification) plays an important role. Linear model results should be used cautiously, especially when conclusions arise from small differences. This paper shows that spatial characteristics of a physically-based model for acidification calculations influence the corresponding linear model results, considerably. The calculated examples show variations between 3% and 56%, with one exception (109%), when the resolution of the model is varied.
In our examples, acidification factors (AFs) that were derived by using different spatial emission distributions differed between 1% and 51% from the original AFs (with two exceptions). The few calculated location-specific AFs differed between 3% and 66% from the original one (with one exception). These results show that variations resulting from spatial parameters should be included in the accuracy assessment of acidification.
The commonly used definition of acidification and the corresponding critical load database, limit the accuracy with which acidification factors can be derived. Other (continuous) definitions of acidification, like average accumulated exceedance, will result in lower variations of AF. However, such definitions are less attractive to policymakers. Reducing the size of ecosystems within the database will also result in lower variations of AF.
Sector-specific and grid cell-specific AFs suffer even more severely from the described discontinuities. They relate to smaller amounts of emissions than the national AFs. Although the physically-based model calculates small deposition changes, the commonly used definition of acidification prevents the detection of small changes in acidification. This effect introduces large variations in sector-specific and location-specific AFs, derived from the physically-based model. Sometimes, it even makes the calculation of realistic AFs impossible. Therefore, such specific AFs should be used with great caution.
Although this paper concentrates on acidification, similar effects are expected to occur for other characterisation factors that are based on models that address spatial resolutions of input and output variables. For example, factors for eutrophication are often based on models very similar to the one applied in this paper. Therefore, these factors are also likely to suffer from a dependence on the applied spatial resolution.
|Author(s)||Bellekom S ; Hettelingh JP ; Aben J|
|Publication||Environ Modelling & Software 2009; 24(4):463-72|