Identifying potentials for reducing uncertainty in critical load calculations using the PROFILE model

A sensitivity analysis was performed testing weathering rates, critical loads and exceedances for Swedish forest soils using Monte Carlo simulations of the PROFILE model. Different subsets of input data were investigated with respect to their potential to reduce data uncertainty at site level but also for modified estimates of the 5%-ile critical load and the 95%-ile exceedance on 150x150 km resolution. Physical soil properties were of dominant importance for all sites and yield up to 62% reduction of the output standard deviation in weathering rate. The study showed that the critical ratio of base cations to inorganic aluminium (Bc/Al ratio) in the soil solution was of major importance for reducing data uncertainty in critical loads and exceedance estimates. The critical Bc/Al ratio was found to be important for reducing data uncertainties in modified estimates of the 5%-ile critical load and the 95%-ile exceedance, in particular in the northern part of Sweden. Atmospheric deposition, uptake and litterfall were more important for reducing data uncertainty in the southern part. Physical soil properties and especially mineral content were found to be less important for reducing data uncertainties in critical loads and exceedance estimates. The greatest scope for reducing data uncertainties in an applied perspective is to improve estimates of atmospheric deposition of anions and cations as well as uptake and litterfall of base cations and nitrogen.