Wet and dry deposition of sulphur was estimated for 30 forest stands in the Netherlands using a throughfall method and an inferential method. Dry deposition estimates of the throughfall method were significantly higher compared to estimates from inference. The major sources of uncertainty of the throughfall dry deposition estimates were associated with non-representative throughfall sampling, wet deposition estimates, canopy exchange processes, deposition of neutral salts, dry deposition directly onto the throughfall collectors, and with the omission of stemflow fluxes and dry deposition directly to the undergrowth vegetation and forest floor. These uncertainties were found to act both ways to approximately the same extent and were not able to explain the observed gap between the two dry deposition estimates. For the inferential method, major sources of uncertainty in the dry deposition estimates arose in calculation of the dry deposition velocity of SO2 and the omission of occult deposition of SO42-. In this study, uncertainties associated with the spatial averaging of air concentrations of SO2 and SO4 aerosol and the calculation of the deposition velocity of SO4 aerosol were found to be relatively small. Modifying the R(c) parametrization of SO2, based on recent dry deposition measurements made over heather in the Netherlands, resulted in fairly good agreement between both dry deposition estimates. Occult deposition of SO42- was found to contribute significantly to the total sulphur deposition to the forest stands. Both the modified R(c) parametrization and the incorporation of occult deposition led to systematically higher sulphur deposition estimates by the inferential method compared to originally inferred deposition. This implies that in the Netherlands, sulphur deposition to forest ecosystems might have been underestimated heretofore.
