One key objective in any scientific field is to learn lessons from past studies in order to improve future work. This talk illustrates this concept by presenting our experience of conducting an ERA on some French former uranium mining sites. A second goal of the workshop is to discuss how better present the assessment findings and associated uncertainties to stakeholders. We will share with the audience our feedback from our participation into a pluralist group including authorities, operators, NGOs, technical experts and scientists. From 2005 to 2011, the GEP mines (group of pluralist expertise on the former uranium sites in the center of France) conducted an exhaustive and rigorous investigation of actual impacts of these legacy sites. The underlying question, at the origin of the group genesis, was to know if the management of these sites complied with their regulatory obligations with regard to protection of both people and the environment. Beginning with a technical analyze of the present situation to clarify the best options for their management and monitoring, the group provided recommendations on option development to reduce the current impacts and to foresee those in the long term according to the preparation of a long term plan. To fulfill the mandate of the GEP, a sub-group of about 25 people representative of all stakeholders focused on the question of the environmental impact, considering both human health and ecological aspects. At the time of the study, dealing in a somewhat regulatory context with radiological impact on wildlife was something new, the international development about radiation protection of the environment just started out. The talk presents the principles of the ERA as applied to the French former uranium mining sites, looking to take into account impacts from both radiological and chemical nature on an aquatic ecosystem. Results from the successive tiers that were necessary to achieve our goals are described; making clear which refinement was applied from one tier to another. By the way, limitations and uncertainties in the applied approach are evocated. To conclude, recommendations are made about the requirements to robustly implement such ERA method, the additional data necessary in complement to direct information on toxic levels in the environment, as well as some perspectives in terms of research and development we identified at this time.