| Abstract: | A mathematical model of the radiocesium transfer in a lake and its catchment is presented in an attempt to improve previous approaches. The experimental data base of the Esthwaite Water location, U.K., prepared in the international program on VAlidation of Model Predictions (VAMP) of IAEA/CEC, was used as a test scenario. Radiocesium enters the lake via: (1) a quick pulse of atmospheric deposition for 1–2 days; and (2) a lasting but smaller contribution from the catchment carried by the water runoff and the dragged sediments. An auxiliary model of the catchment was used to quantify these two source terms.The temporal evolution of the 137Cs concentration in this lake after the contamination by the Chernobyl accident is modelled using a compartmental approach following the Codell model. The desorption/adsorption dynamics between water and sediments are considered both in the drainage water in the lake boundary layer. The most important processes transferring contamination out of the lake water are adsorption of radioactivity by suspended sediments and their sedimentation; these processes are modelled through a distribution coefficient and through an aerial removal rate. Summer stratification is modelled introducing some auxiliary hypotheses.When using a priori probability distributions for the parameters based both on experimental evidence and on information from the literature the model simulates the observed radiocesium concentrations with an expected underprediction bias of 30% in the integrated concentrations in water. The concentration in sediments is inside the observed range in most of the runs. After the calibration of the model to the best fits observed with parameters inside their expected ranges, the bias decreases to −6% in the epilimnion, 8% in the hypolimnion and −4.4% in sediments, with an unexplained variance of ±3.2%, ±3% and ±1%, respectively. |
