Assessment of uncertainty and risk in modeling regional heavy-metal accumulation in agricultural soils.

Present agricultural land use and atmospheric deposition may lead to heavy-metal accumulation rates in soils that may violate soil quality standards in the future. To undertake suitable preventive measures against heavy-metal enrichment, flux balances in agroecosystems and their uncertainties have to be assessed. For this reason we developed an empirical stochastic model, PROTERRA-S, that considers heavy-metal inputs through agricultural management as well as outputs by crop removal and leaching on a regional scale. In this manuscript we describe application of PROTERRA-S to the Sundgau region in Switzerland. Considering uncertainty in informational and natural variability, large variations of the aggregated regional cadmium and zinc balances were found, with standard deviations that were of the same order of magnitude as their average values. Uncertainty in the simulated net zinc flux originated mainly from uncertainty in the zinc concentrations of manure and crops and from uncertainty in atmospheric deposition of zinc. For cadmium, the main contribution to the total uncertainty came from uncertainty in crop concentration, regression functions to estimate Freundlich parameters, atmospheric deposition, and from spatial variation of soil pH and cation exchange capacity (CEC). For both zinc and cadmium, informational uncertainty in input data were large, indicating that significant uncertainty reduction could be achieved by additional data collection campaigns. A monetary risk value for the regional zinc accumulation rate in Sundgau was calculated to be on the order of 22 million Euro.