Estimating the exceedance of critical loads in Europe by considering local variability in deposition

Estimates of the deposition of nitrogen and sulphate across Europe are available from two sources: The mean deposition on each square of a 150 km square grid is estimated by an atmospheric dispersion model (ADM), and the actual deposition is measured by a network of monitoring stations across the continent. Critical loads are estimates of the deposition a natural or semi-natural area can withstand without damage. These have been calculated for the whole of Europe, resulting in a distribution of critical loads for each square on the grid. We estimate the area over which critical loads are exceeded in each square by three methods: (1) The true deposition is assumed to be equal to the ADM-predicted mean across each square. (2) The ADM-predicted and monitored depositions are combined to obtain a kriged mean deposition, and this is assumed to be constant across each square. (3) The deposition field is modelled as a stochastic Gaussian field and its distribution calculated conditional on the measurements. The distribution (but not the deposition) is assumed to be constant across each square. The area exceeded is then a stochastic variable, and we calculate its expectation and standard deviation. We find that the three methods give different estimates of the area over which critical loads are exceeded. In some parts of Europe the differences are large. The third method should give the best estimate since it makes more reasonable assumptions about the variability of the deposition. We discuss potential improvements to our methodology and the need for better estimates of the local variability in depositions