Pergamon Applied Geochemistry, Vol. 1I, pp. 31 l-316, 1996 Copyright 0 1996 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0883-2927/96 $15.00 + 0.00 Desorption and diffusion of episodic pollutants in sediments: a 3-phase model applied to Chernobyl 13’Cs M. Meili and A. W&man Uppsala University, Inst. of Earth Sciences, Sedimentology, NorbyvHgen 18 B, S-752 36 Uppsala, Sweden Abstract-Eight years after the Chernobyl fallout, the distribution of “‘Cs observed in the sediment of a small lake (Loppesjiin, central Sweden) suggests that most of the 13’Cs entering the lake was rapidly deposited in the sediment. However, vertical profiles also showed that “‘Cs is not irreversibly bound to particles, but slowly remobilized during the following years. The 13’Cs profile at the deepest, usually anoxic site shows both a distinct peak and long tails. This indicates a gradual desorption of 13’Csfrom the original deposit and a vertical redistribution caused by diffusion in a matrix with different sorption properties. A 3- phase model was developed that describes slowly desorbing binding sites in the initial deposit by an instantaneous equilibration changing slowly over time. The model provides numerical and analytical solutions for complex diffusive decay processes in aquatic sediments. Copyright 0 1996Elsevier Science Ltd INTRODUCTION Thousands of Swedish lakes have been heavily contaminated by the radioactive fallout from the Chernobyl accident on 26 April 1986 (Moberg, 1991). Most of the 13’Cs entering lakes was rapidly removed from the water column (Meili, 1988). The transfer from the water column to the sediment has been suggested to occur by sorption to suspended particles followed by incorporation into the sediments (Robbins et al., 1992) or by direct sorption to surficial sediments (Santschi et al., 1988). The binding of 13’Cs to sediment particles has been reported to display an irreversible behaviour, at least over short time scales such as weeks or months (Comans zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA et al., 1991; Robbins et al., 1992). The present study focuses on the potential of using the vertical sediment profile of 13’Cs several a after the fallout for reconstructing initial sorption processes in the water column and for quantifying subsequent redistribution patterns within the sediment of a lake. The aim was to develop a general model to simulate the mobility and fate of episodic pollutants in aquatic sediments. METHODS Lake Loppesjiin (central Sweden, 61.7” N 16.8” E) is a small lake (0.28 km’) with a maximum depth of 14 m, a mean depth of 5.4 m, and a mean hydraulic turnover time of 1.2 a. The concentration of suspended particles in the lake is 2-6 g mm3 (dw), P concentration 17-24 mg me3 and water colour 30-50 gPtme3. In the deepest part of the lake, the organic content of the surface sediment is around 25% (dw) and the N content 1% (dw). This site is a well-protected depression, where the water above the sediment is almost permanently stratified and anoxic, suggesting that both biotic perturba- tion and wihd-induced currents are negligible. The site is, therefore, ideally suited to study natural rates of desorption and molecular (ionic) diffusion in lake sediments. Four sediment cores were collected at the southern deep site (bathymetry shown in Konitzer and Meili, 1995), 2 cores in April 1994 and 2 in August 1993 (i.e. only about 4 months earlier when disregarding the period of ice cover). The cores were taken with a gravity corer (inner diameter of 6.5 cm) and sectioned on site with a special slicing device into subsamples of 0.5 cm in the top 10 cm and at larger intervals, from 10 to > 25 cm, covering 50-100 a of sedimentation. All samples were analysed for water content and “‘Cs (see Konitzer and Meili, 1995, for details). Measured activities of 13’Cs were corrected for radioactive decay to 1 May 1986. RESULTS Repeated sampling during different years by differ- ent persons under different conditions has shown that at the deepest site of Loppesjiin, basic sediment characteristics as well as 13’Cs profiles are reproduci- ble, with slow changes over time. In order to obtain reliable parameter estimates, the 4 cores taken at this site 8 a after the fallout were treated as replicates and evaluated together. Particle concentrations in the sediment {s = pW/[l/(l - w)] + pW/ps - 1) were obtained from vertical profiles of water content (w. g g-’ wet), the density of water (p,., z 1 g cme3), and the density of solids [ps, calculated from measured bulk densities, Pb, as Ps = pW(1 - W)/ (pw/pb - W) e 2.07 g cmT3]. Particle concentration s was about 0.13 g cmm3 in the sediment around the 13’Cs peak, and porosity n (water volume per sediment volume, n = zyxwvutsrqponmlkjihgfedcbaZYXW 1 - s/ps) about 94%. The dry mass of sediment above the centre of the 13’Cs peak was 0.40 g cmT2, suggesting that the mean annual mass flux to the sediment at this site was 0.05 g cm -’ a-’ between 1986 and 1994. Figure 1 and other core profiles from the same site (e.g. Konitzer and Meili, 1995) show that a large fraction of the 13’Csinventory at this site is confined to a sediment layer of (0.5 cm. Below this peak layer, a tail of 13’Cs is visible to a depth > 0.1 m below the 311