Uncertainties in high resolution critical load assessment for forest soils: Possibilities and constraints of combining distributed soil modelling and geographic information systems

A design for generating distributed input data for a regionalized version of the steady-state model PROFILE was tested by Monte Carlo simulations of critical loads of acidity for 67 sites within a south Swedish municipality, Svalov. Input and output data were integrated in a geographic information system for data manipulation and to facilitate its use in local planning. Pooled standard deviations were in the range 20-35 mmol(c)m^-2 yr^-1 for weathering rate, 35-45 mmol(c)m^-2 yr^-1 for critical load of acidity and 50-51 mmol(c)m^-2 yr^-1 for exceedance of the critical load of acidity. The study indicates that if site-specific uncertainty estimates are accounted for an improved estimate of the corresponding cumulative distribution function can be made. Data uncertainty makes it impossible to attribute a site to an unambiguous risk class of exceedance. Data uncertainty therefore inflicts constraints on the applicability of the calculated critical loads and exceedances as tools in environmental risk assessments. The greatest scope for improving the modelling results is through providing better quality vegetation/forest and mineralogical data, while improvements in the other primary data types is likely to be costly in relation to modelling gains. Modelling results support current practical recommendations to plant deciduous, rather than coniferous, trees in areas exposed to high loads of acidifying deposition.