Simulation of critical loads for nitrogen for terrestrial plant communities in the Netherlands

This paper describes a new method to derive nitrogen critical loads for vegetation, and its application to The Netherlands. An 'inverted' form of the soil chemical model SMART2 was used to estimate atmospheric nitrogen deposition at the critical conditions for 139 terrestrial vegetation types (associations) occurring in northwestern Europe, using an iterative search procedure. The critical conditions are the lower end of the pH range, and the upper end of the nitrogen availability range for each vegetation type. The critical load is assumed to be the nitrogen deposition that results in the critical conditions. The critical load values were subjected to a sensitivity and uncertainty analysis. Sensitivity analysis showed that the estimated critical N load mainly depends on the vegetation type and to a lesser extent on the soil type and the critical N availability. Of these variables N availability, which was estimated from Ellenberg's indicator scale, contributes most to the uncertainty. The critical load averaged over all vegetation types and soil types is estimated to be 23 +/- 7 kg N ha(-1) y(-1). This is a rather reliable value because its uncertainty is small and it is in agreement with empirical estimates of critical loads. Critical loads per vegetation type are less reliable because they are not correlated to empirical values, although the ranges of simulated and empirical values usually overlap. At the site level, uncertainty becomes very large and it is not possible to determine critical loads with any practical significance. The uncertainties can only be reduced if more data become available on the abiotic response per species under field conditions, at least to nitrogen availability and soil pH