Aplicación de la modelizacion geoquímica predictiva a un sistema natural: el berrocal

Abstract

Environmental management decisions require the use of models with varying degree of sophistication to represent the evolution of natural and engineering systems. These models are of a predictive kind and they are based on our interpretative knowledge of Nature. In the case of Nuclear Waste Management the predictions may involve periods of time between 1,000 to 10,000 years. In order to build up our confidence in the possibilities of using the models to predict the geochemical behaviour of the future repository systems, we study natural systems than reflect processes and/or materials used in the repositories. These studies are known as Natural Analogues. We have undertaken, within these investigations the checking of geochemical models to describe the aqueous specialtion and the solubility of trace elemento, i.e. U, Th, Sr. .. The mobility of these elements is critical to assess the behaviour of nuclear waste repositories. These tests are known 1S Blind Predictive Modelling exercises and attempt to simulate the modelling of radionuclide solubilities and speciation in a repository environment. We will attempt to present the methodology of these BPM exercises, as well as, some preliminary results within the El Berrocal project. The blind predictive exercise was divided into two stages. In the first stage, the modellers had information on the major metals found in the site. In the second stage additional information about mineralogy was given. The results obtained were very encouraging. The agreement among the different participants was good in both stages of the exercise. In the first stage, the pure solid phases chosen, as well as the speciation of the trace metals in solution was coincident in all cases. In the second stage, the agreement was better, because of the possibility to neglect solid phases not characterized in the site. On the other hand, the application of coprecipitation and codissolution approaches led better results in the prediction of the concentrations of the trace metals in solution. The exercise was extremely efficient in discerning the differences between the distinct databases and geochemical codes used (EQ3NR, HARPHRQ and PHREEQE). It can be seen in the paper that the concentrations of uranium in solution are more accurately predicted when using the formation of a mixed solid phase of U(VI) and Fe(I1I) than assuming apure U(VI) solid phase. On the other hand, the concentrations of Ba predicled are slightly higher than the actual ones if assurning a congruent codissolution of a mixed solid phase of Ba with calcite. This is a very useful result when trying to asses the environmental risk. The same results are obtained when predicting the concentration of other trace metals, such as Mn and Sr.