The first uranium production center in Brazil began operation in 1982. After 13 years of a non-continuous operation, the mining activities were suspended definitively. Uranium was extracted by open pit mining. Operations gave rise to approximately 12.4×10 6 m 3 of waste rock, while the mill process generated a volume of approximately 2.39×10 6 m 3 of tailings. Regardless the fact that some studies developed in this area exist, a well defined plan of action, aimed at the remediation and

more »

... ation of the site, has not been implemented yet. The main sources of pollutants to the environment are the tailings dam, the waste rock piles and the open pit. Pyrite oxidation was found to be the driving force in the leaching of metal and radionuclides into environment. It was estimated that acid drainage generation will last for 600 and 200 years from the waste rock and tailings respectively. Accurate prediction of the release rate of metal and radionuclides from these sources and their transport in the subsurface is a critical factor to the assessment of environmental impact and to the development of effective remediation strategies. In prevailing practice, the source term is evaluated using the dissolution rate of waste form and the solubility of radionuclides. The fate of pollutants is addressed by the use of Kd-based "reactive" transport models. This standard practice has obvious shortcomings, mainly because it can not produce a realistic representation of the system under study. The alternative to overcome these shortages is using more sophisticate models that could represent real complex problems. Reactive transport codes are powerful tools in the evaluation of coupled thermal-hydrological-chemical processes and in the prediction of the long-term performance of remediation strategies. The difference between the predictions from these two approaches can be as high as several orders of magnitude. Generally, conventional approaches produce predicted values higher than the measured ones. On the other hand, the use of reactive transport model requires a good knowledge of the simulated hydrogeochemical system, along with the choice of appropriated algorithms that can represent the most important processes.