A compilation of empirical data and variations in data concerning radiocesium in water, sediments and fish in European lakes after Chernobyl

This work concerns the variability of radiocesium within lakes. The focus is on a broad set of data concerning radiocesium after the Chernobyl accident in lake water, sediments and different species of fish. Data are available to the author from three European data bases. Basic questions are: Are there any general patterns to be found concerning the variability of (137)Cs in lakes? Is it possible to give any recommendations concerning CV values (coefficient of variation; CV = SD/MV; SD = standard deviation, MV = mean value) for radiocesium in lake water, sediments and different species of fish? The analysis can be summarised as 1. The CV values for lake water vary around 0.3 and are rather independent of the time after fallout. 2. One can use a CV of 0.6 as a reference value for the typical uncertainty in radiocesium concentrations in surficial lake sediments. CV values are generally higher (up to CV = 1) for bottom areas dominated by erosion and transport processes (for fine material following Stokes' law) and lower(CV approximate to 0.2) for areas dominated by continuous sedimentation and fine deposits (accumulation areas). CV values for radiocesium in sediments are also likely to increase with contamination and the size of the lake. 3. A reference CV of 0.22 would be a reasonable general CV value for lake fish. CV values are typically larger just after fallout (CV approximate to 0.3) and decrease with time after fallout (to about 0.15). CV values are likely to increase with trophic level, from about 0.1 for planktivores to about 0.3 for piscivores, but these CV values are based on limited data and are quite uncertain. The benefit of general empirical CV values is evident in modelling, e.g., when empirical data are compared to modelled values. The CV values for water, sediments and fish can be used to set empirical uncertainty bands for model predictions to enable meaningful discussions about predictive success. (C) 1999 Elsevier Science Ltd. All rights reserved.