Difficulties and lessons of environmental data processing to fit modelling parameters

Titre du congrès :SETAC Europe 16th Annual Meeting Ville du congrès :La Hague Date du congrès :11/05/2006

We tried to estimate the main parameters of radionuclides transfer to various products of the agricultural food chain from a long record of radioactivity measurements carried out in France over the period 1960-1980. The aim of this study was to obtain new information about interception capacity of the vegetables, effective period in soils and root transfer: regionalized range, regionalized probability distribution and possible relations with others parameters were expected. However, information wished could not be deducted: the confrontation between predicted values (by the model) and observed values showed that goodness of fit was limited. As a whole, this data processing highlights the limits of adaptation between model and data recorded to validate it. We learn lessons from this expérience to improve ail together our sampling strategy and the prédictive capacity of the model. We will présent the encountered difficulties and the deducted lessons from treatment of thèse environmental measurements with the concrète example of the atmospheric transfer of strontium 90 to vegetables.