Migratie van zink in metaalverontreinigde oeverbodems

In the Noorderkempen (northwest Belgium), about 280 km2 land is contaminated with cadmium (Cd) and zinc (Zn) due to a historical pollution by the Zn smelters. Current atmospheric and aquatic emissions of metals are strongly reduced from historic ones, however the environmental contamination remains. Surface water Zn concentrations peak during winter periods, thereby largely exceeding water quality limits. This work was initiated to identify if such peaks originate from Zn that is leached from riparian soils, i.e. soils periodically waterlogged adjacent to rivers. Soil chemical and physical studies and a three-year monitoring program were set-up to address the issue. A batch experiment with sixteen acid sandy soil samples from the region was performed to identify processes explaining dynamics of Zn upon waterlogging. The experiment confirmed contrasting trends in trace metal solubility in response to waterlogging due to counteracting effects of increased pH and increased dissolved Fe(II) concentrations. After two months of waterlogging, pore-water Zn concentrations increased by, on average, a factor of 18 (range 0.6 80, immobilization in only 1 soil) despite an increase in pH. The mobilization was most pronounced in soils with low (background) Zn concentration. It is hypothesized that Zn mobilization results from Fe2+ desorbing Zn2+ from particulate organic matter and from reductive dissolution of Zn bearing Fe/Mn oxyhydroxides. This hypothesis is based on Zn speciation modelling and on a significant positive correlation between the Zn mobilization factor and the soil Fe/organic carbon ration (Fe/OC; r = 0.87, n = 13). The Fe2+ is a significant competing ion determining fate of toxic trace metals in anoxic soils.A column experiment with three contaminated soils was designed to mimic soil profiles of metal contaminated riparian soils. The surface layers were either waterlogged or not to test its effects on subsequent vertical migration of Zn. Pore-water Zn concentrations in the saturated columns were 2.5 to 40-fold lower than in corresponding unsaturated ones in which nitrification and associated acidification mobilized Zn. Zinc in pore-water was explained by the pore-water pH across treatments and time. In the leachates of the saturated treatments, Fe breakthrough was observed, while weighted average Zn concentrations were unaffected (P>0.1) by the topsoil treatments. The Zn immobilization upon waterlogging in the column experiment contrasts the general Zn mobilization in the batch experiment because of different properties of soil samples used, i.e. lower Fe/OC ratio and higher soil Zn concentration in the soil samples, both factors likely leading to a lower fraction of Fe bound Zn in the aerobic soil. Long-term (15 y) monitoring data of river water quality and new data of a three-year field monitoring study show that Zn and Cd concentrations exceed quality standards 4-200-fold. A statistical analysis revealed a significant effect of month of sampling with lower Zn concentrations in winter and higher in summer months. Surface water Zn and Cd concentrations were positively related to Fe and Mn but not to Ca, K or Na concentrations, suggesting that redox reactions and not dilution processes are involved. In ground- and pore-water of the associated riparian soils, correlations between Zn and Fe or Mn, as proxies for reduced conditions, are lacking. Overall, the field study suggested that river sediments and not riparian soils may be the cause of the winter peaks of Zn and Cd in these rivers.A 1-D simulation study was conducted to predict current vertical Zn distributions in riparian soils using the atmospheric emission history of the Zn smelters. Periodic waterlogging in the riparian soils could explain larger dispersion of Zn in riparian soils compared to acid, unsaturated soils but was not able to explain the enrichments of Zn in the deeper soil layers as observed. The vertical Zn distributions were better predicted if horizontal Zn transport was included, this transport occurs via the aquifer from the dry podsols towards the lower horizon of riparian soils. In addition, the area based Zn stock in the riparian soils is 2-8 fold larger than in dry podsols of the same area which corroborates the hypothesis that the riparian soils are a net sink of Zn in the contaminated area by sequestering metals transported via the groundwater.Summarizing, Zn can be immobilized or mobilized during waterlogging because of counteracting effect of pH increase and reductive dissolution of Zn bearing Fe(III)/Mn oxyhydroxides. The net effect depends on soil properties and extent of reduction. The redox reactions are unlikely promoting migration of Zn to the subsurface. Subsurface Zn concentrations in such soils probably originate from horizontal transport and seasonal Zn and Cd release in the surface waters is likely related to in-stream chemical or physical processes.Dankwoord i Summary iii Samenvatting v List of abbreviations vii List of symbols viii Table of contents ix CHAPTER 1 Mobility of zinc and cadmium in riparian soils in response to periodic waterlogging: background, mechanisms and research objectives 1 CHAPTER 2 Mobilization of zinc upon waterlogging riparian podsols is related to reductive dissolution of Fe-minerals 19 CHAPTER 3 Lower mobility of zinc in flooded than in unsaturated soils is related to soil pH in repacked soil columns 45 CHAPTER 4 Seasonal trends of dissolved zinc and cadmium concentrations in surface waters and in associated riparian soils of the Noorderkempen, Belgium 63 CHAPTER 5 The effect of periodic waterlogging on vertical migration of zinc in riparian soil: a simulation study 95 CHAPTER 6 General conclusions and future perspectives 115 Annex 121 Bibliography 127 List of publications 135nrpages: 135status: publishe