Predicting the transfer of radiocaesium from organic soils to plants usingsoil characteristics

A model predicting plant uptake of radiocaesium based on soil characteristi cs is described. Three soil parameters required to determine radiocaesium b ioavailability in soils are estimated in the model: the labile caesium dist ribution coefficient (k(dl)), K+ concentration in the soil solution [(m)(k) ] and the soil solution --> plant radiocaesium concentration factor (CF, Bq kg(-1) plant/Bq dm(-3)). These were determined as functions of soil clay c ontent, exchangeable K+ status, pH, NH4+ concentration and organic matter c ontent. The effect of time on radiocaesium fixation was described using a p reviously published double exponential equation, modified for the effect of soil organic matter as a non-fixing adsorbent. The model was parameterised using radiocaesium uptake data from two pot trials conducted separately us ing ryegrass (Lolium perenne) on mineral soils and bent grass (Agrostis cap illaris) on organic soils. This resulted in a significant fit to the observ ed transfer factor (TF, Bq kg(-1) plant/Bq kg(-1) whole soil) (P < 0.001, n = 58) and soil solution K+ concentration ((m)(K), mol dm(-3)) (P < 0.001, n = 58). Without further parameterisation the model was tested against inde pendent radiocaesium uptake data for barley (n = 71) using a database of pu blished and unpublished information covering contamination time periods of 1.2-10 years (transfer factors ranged from 0.001 to 0.1). The model account ed for 52% (n = 71, P < 0.001) of the observed variation in log transfer fa ctor. (C) 2000 Elsevier Science Ltd. All rights reserved.